ISLAB/CAISR - User contributions [en]

Detection Unit Imperfections and CPM Reliability in Interoperable RSUs

2026-02-06T10:01:31Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=Collaboration with Mittlogik: The project investigates how data and analysis quality affects reliability of collective perception services. |..."

{{StudentProjectTemplate
|Summary=Collaboration with Mittlogik: The project investigates how data and analysis quality affects reliability of collective perception services.
|Keywords=RSU, CPM, reliability analysis, connectivity
|TimeFrame=Spring-Summer 2026
|Prerequisites=Strong skills in C/C++, Python, and data analysis. Knowledge of real-time systems and a basic understanding of V2X (CPM) and sensor fusion concepts is required
|Supervisor=Mittlogik, Elena Haller, Oscar Molina
|Author=Mittlogik
|Level=Master
|Status=Open
}}
<h2>Background</h2>
<p>
Standardization bodies have defined the structure of Vehicle-to-Everything (V2X) messages such as the
<strong>Collective Perception Message (CPM)</strong>. However, the actual safety utility of these messages
depends entirely on the quality of the data provided by the <strong>Detection Unit (DU)</strong>.
In real-world deployments, DUs (e.g., computer vision sensors) suffer from variable latency,
measurement noise, and intermittent detection gaps.
</p>
<p>
This thesis investigates how these <em>sensor imperfections</em> propagate through the
Roadside Unit (RSU) interworking protocol and affect the reliability and timeliness of
safety warnings sent to vehicles.
</p>

<h2>Objectives</h2>
<ol>
<li>
<strong>Characterize DU Latency and Noise:</strong>
Benchmark the latency and spatial accuracy of the selected DU under various conditions,
such as varying numbers of detected objects and different lighting environments.
</li>
<li>
<strong>Model Detection Imperfections:</strong>
Develop a fault-injection framework capable of introducing artificial delays, noise,
and dropped detections into the data stream between the DU and the RSU.
</li>
<li>
<strong>Analyze CPM Reliability:</strong>
Quantify the impact of detection imperfections on generated CPMs using metrics such as
Age of Information (AoI), positional error, and message frequency stability.
</li>
<li>
<strong>Evaluate System Robustness:</strong>
Identify the system break-point, defined as the level of sensor latency or noise at which
CPMs no longer provide a reliable safety benefit for vulnerable road user (VRU) protection.
</li>
</ol>

<h2>Methodology</h2>
<ul>
<li>
<strong>Benchmarking:</strong>
Conduct physical experiments to measure the baseline latency of the DU-to-RSU communication link.
</li>
<li>
<strong>Fault Injection Framework:</strong>
Implement a Python/C++ tool positioned between the DU and the RSU interworking module to simulate
real-world errors, including noise, jitter, and packet loss.
</li>
<li>
<strong>Stress Testing:</strong>
Execute safety-critical scenarios, such as a pedestrian emerging from behind a vehicle,
while injecting controlled levels of sensor imperfection.
</li>
<li>
<strong>Data Analysis:</strong>
Compare ground-truth pedestrian positions with RSU-reported positions contained in CPMs
to calculate error margins and safety-critical delays.
</li>
</ul>

<h2>Expected Deliverables</h2>
<ul>
<li>
<strong>Robustness Analysis Report:</strong>
A comprehensive assessment of how sensor imperfections influence CPM quality and reliability.
</li>
<li>
<strong>DU Performance Profile:</strong>
A technical benchmark documenting the performance characteristics of the selected detection hardware.
</li>
<li>
<strong>Recommendations for RSU Logic:</strong>
Proposed software-level mitigation strategies, such as data smoothing or dead reckoning,
to reduce the impact of sensor limitations.
</li>
</ul>

Comparative Study on Data Abstraction Methodologies for Interoperable V2X Roadside Units (RSU)

2025-12-19T10:59:17Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=It is a collaboration with MittLogik on development of Interoperable RSU Prototype focused on Vulnerable Road User (VRU) safety. |Keywords=pe..."

{{StudentProjectTemplate
|Summary=It is a collaboration with MittLogik on development of Interoperable RSU Prototype focused on Vulnerable Road User (VRU) safety.
|Keywords=pedestrian safety, computer vision, V2X communication
|TimeFrame=february 26 - june 26
|Prerequisites=Strong skills in embedded C/C++, data structures, computer networks, and a foundational understanding of V2X standards.
|Supervisor=Elena Haller, MittLogik
|Level=Master
|Status=Open
}}

<h1 id="comparative-study-v2x-rsu">
Comparative Study on Data Abstraction Methodologies for Interoperable V2X Roadside Units (RSU)
</h1>

<section id="project-context">
<h2>Project Context</h2>
<p><strong>Interoperable RSU Prototype</strong> focused on <em>Vulnerable Road User (VRU)</em> safety.</p>
</section>

<section id="summary">
<h2>Summary</h2>
<p>
This thesis is a pre-study to inform the architecture of a proprietary interworking protocol.
Theme: Embedded Systems, V2X Data Efficiency, Comparative Protocol Analysis.
</p>
</section>

<section id="metadata">
<h2>Project Metadata</h2>
<dl>
<dt>Theme</dt>
<dd>Embedded Systems, V2X Data Efficiency, Comparative Protocol Analysis</dd>
<dt>Location</dt>
<dd>Lund</dd>
<dt>Timeline</dt>
<dd>2026-02-01 to 2026-10-31</dd>
</dl>
</section>

<section id="background">
<h2>Background</h2>
<p>
To achieve true RSU interoperability and scalability, the core RSU application must be decoupled
from vendor-specific detection unit (DU) sensor data and diverse communication methods (e.g., C‑V2X/DSRC).
The design of a lean, efficient <strong>Minimal Dataset Schema</strong> is critical for low-latency
<em>VAM (Vulnerable Road User Awareness Message)</em> generation. This research will compare various approaches
for defining this minimal schema, but will not disclose the final proprietary implementation.
</p>
</section>

<section id="objectives">
<h2>Objectives</h2>
<ol>
<li>
<strong>Analyze DU Data Complexity:</strong>
Analyze and characterize the data payload complexity and velocity from representative Computer Vision (CV)
detection units (DU), identifying challenges in converting raw data (e.g., bounding boxes, tracking IDs)
into V2X data elements.
</li>
<li>
<strong>Compare Methodologies:</strong>
Conduct a comparative study of different encoding methodologies for V2X safety applications, evaluating each
based on latency, CPU overhead on embedded platforms, and message size efficiency.
</li>
<li>
<strong>Define Abstraction Rules:</strong>
Propose a set of general best-practice rules and conversion algorithms for mapping high-frequency CV tracking data
onto the necessary minimal dataset fields required for standardized VAM generation
(e.g., position, dynamics, object type).
</li>
</ol>
</section>

<section id="methodology">
<h2>Methodology</h2>

<h3>Comparative Review</h3>
<p>
Detailed review of three candidate serialization/encoding techniques (Objective 2) suitable for the RSU's embedded platform,
including profiling their processing speed in a laboratory environment (simulation/modelling).
</p>

<h3>Algorithm Design</h3>
<p>
Design generalized algorithms and flowcharts for data aggregation and abstraction (Objective 3).
</p>
</section>

<section id="deliverables">
<h2>Expected Deliverables</h2>
<ul>
<li>
<strong>Comparative Analysis Report:</strong>
A detailed technical report comparing the performance and suitability of the investigated
data serialization methodologies for the RSU application.
</li>
<li>
<strong>Generalized Abstraction Rules:</strong>
A documented set of algorithms and principles for mapping complex detection data to V2X minimal datasets.
</li>
<li>
<strong>Thesis Report:</strong>
A thesis submitted for publication focusing on the methodologies, analysis, and generic findings.
The final proprietary RSU protocol schema will not be documented or published.
</li>
</ul>
</section>

In-Vehicle Data Acquisition for Driving Behavior Profiling and Automotive AI Applications

2025-11-25T14:38:49Z

Cclab:

{{StudentProjectTemplate
|Summary=Research focus in-vehicle data acquisition techniques, building upon existing work and addressing key limitations in collecting comprehensive, real-world automotive datasets.
|Keywords=In-vehicle networks;CAN protocol;data acquisition;datasets generation
|References=A. Dos Santos Roque, L. M. Da Silva Alves and E. P. de Freitas, "CAN-Modes: In-vehicle datasets generation and analysis in different driving situations," 2024 Workshop on Communication Networks and Power Systems (WCNPS), Brasilia, Brazil, 2024, pp. 1-7, doi: 10.1109/WCNPS65035.2024.10814379.
|Prerequisites=Communication, programming and hardware skills.
|Supervisor=ALEXANDRE DOS SANTOS ROQUE
|Level=Master/Bachelor
|Status=Open
}}
General Description:
To design, implement, and validate an in-vehicle data acquisition system that contributes in dataset generation, and to develop methodologies for profiling diverse driving behaviors from this data. This research will establish a foundation for future data-driven applications.

Goals:
- Investigating and addressing the limitations of current data acquisition methods, specifically those highlighted in works like CAN-Modes, regarding data diversity (e.g., more car brands, model variations), modern protocols (e.g., CAN FD), and advanced systems (e.g., ADAS, authenticated OBD access).
- Developing an enhanced in-vehicle data acquisition system capable of collecting more extensive and varied data types, including those from ADAS sensors and potentially beyond standard CAN.
- Creating methodologies for generating novel, diverse, and well-labeled datasets under a broader range of driving conditions and vehicle types.
- Data Preprocessing and Feature Engineering: Implement robust pipelines for data cleaning, synchronization, and feature extraction.
- Study of standards for data generation, as the MDF4 file format, and programming libraries.
- Study of methods of autentication to access in-vehicle networks (Gateway access problem).
- Investigating and implementing non-invasive data access methods (e.g., CAN crocodile)

In-Vehicle Data Acquisition for Driving Behavior Profiling and Automotive AI Applications

2025-11-25T14:35:39Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=Research focus in-vehicle data acquisition techniques, building upon existing work and addressing key limitations in collecting comprehensive..."

{{StudentProjectTemplate
|Summary=Research focus in-vehicle data acquisition techniques, building upon existing work and addressing key limitations in collecting comprehensive, real-world automotive datasets.
|Keywords=In-vehicle networks;CAN protocol;data acquisition;datasets generation
|References=A. Dos Santos Roque, L. M. Da Silva Alves and E. P. de Freitas, "CAN-Modes: In-vehicle datasets generation and analysis in different driving situations," 2024 Workshop on Communication Networks and Power Systems (WCNPS), Brasilia, Brazil, 2024, pp. 1-7, doi: 10.1109/WCNPS65035.2024.10814379.
|Prerequisites=Communication, programming and hardware skills.
|Supervisor=ALEXANDRE DOS SANTOS ROQUE
|Level=Master/Degree
|Status=Open
}}
General Description:
To design, implement, and validate an in-vehicle data acquisition system that contributes in dataset generation, and to develop methodologies for profiling diverse driving behaviors from this data. This research will establish a foundation for future data-driven applications.

Goals:
- Investigating and addressing the limitations of current data acquisition methods, specifically those highlighted in works like CAN-Modes, regarding data diversity (e.g., more car brands, model variations), modern protocols (e.g., CAN FD), and advanced systems (e.g., ADAS, authenticated OBD access).
- Developing an enhanced in-vehicle data acquisition system capable of collecting more extensive and varied data types, including those from ADAS sensors and potentially beyond standard CAN.
- Creating methodologies for generating novel, diverse, and well-labeled datasets under a broader range of driving conditions and vehicle types.
- Data Preprocessing and Feature Engineering: Implement robust pipelines for data cleaning, synchronization, and feature extraction.
- Study of standards for data generation, as the MDF4 file format, and programming libraries.
- Study of methods of autentication to access in-vehicle networks (Gateway access problem).
- Investigating and implementing non-invasive data access methods (e.g., CAN crocodile)

Explainable GNNs for Security Verification of RISC-V Cores

2025-10-22T04:43:01Z

Cclab:

{{StudentProjectTemplate
|Summary=develop an explainable graph-neural-network (GNN) workflow that localises security-relevant weaknesses in open-source RISC-V cores at RTL.
|References=Reimann, Lennart M., et al. ”Qtflow: Quantitative timing-sensitive information flow for security-aware hardware design on rtl.” 2024 International VLSI Symposium on Technology, Systems and Applications (VLSI TSA). IEEE, 2024.

Gosch, Lukas, et al. ”Provable Robustness of (Graph) Neural Networks Against Data Poisoning and Backdoor Attacks.” Transactions on Machine Learning Research.
|Supervisor=Mahdi Fazeli
|Level=Master
|Status=Open
}}
In this thesis, you will develop an explainable graph-neural-network (GNN) workflow that localises security-relevant weaknesses in open-source RISC-V cores at RTL. You will (i) extract circuit graphs from RTL, (ii) train/finetune a GNN with Jumping-Knowledge to avoid over-smoothing, and (iii) integrate XAI (e.g., GNNExplainer) to produce “vulnerability heatmaps.” As a validation case, you will introduce a small, data-dependent prefetcher-like feature into a RISC-V design and evaluate whether your pipeline flags and localises the risky structure.

Explainable GNNs for Security Verification of RISC-V Cores

2025-10-22T04:39:13Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=develop an explainable graph-neural-network (GNN) workflow that localises security-relevant weaknesses in open-source RISC-V cores at RTL. |R..."

{{StudentProjectTemplate
|Summary=develop an explainable graph-neural-network (GNN) workflow that localises security-relevant weaknesses in open-source RISC-V cores at RTL.
|References=Reimann, Lennart M., et al. ”Qtflow: Quantitative timing-sensitive information flow for security-aware hardware design on rtl.” 2024 International VLSI Symposium on Technology, Systems and Applications (VLSI TSA). IEEE, 2024.

Gosch, Lukas, et al. ”Provable Robustness of (Graph) Neural Networks Against Data Poisoning and Backdoor Attacks.” Transactions on Machine Learning Research.
|Supervisor=Mahdi Fazeli
|Level=Master
}}
n this thesis, you will develop an explainable graph-neural-network (GNN) workflow that localises security-relevant weaknesses in open-source RISC-V cores at RTL. You will (i) extract circuit graphs from RTL, (ii) train/finetune a GNN with Jumping-Knowledge to avoid over-smoothing, and (iii) integrate XAI (e.g., GNNExplainer) to produce “vulnerability heatmaps.” As a validation case, you will introduce a small, data-dependent prefetcher-like feature into a RISC-V design and evaluate whether your pipeline flags and localises the risky structure.

Sensitivity‑Aware Hardening and Run‑Time Detection of Stealthy Weight‑Drift Trojans in ML Accelerators

2025-10-22T04:32:36Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=design and evaluate sensitivity-aware defences that detect and mitigate stealthy, gradual weight-drift Trojans in FPGA-based ML accelerators ..."

{{StudentProjectTemplate
|Summary=design and evaluate sensitivity-aware defences that detect and mitigate stealthy, gradual weight-drift Trojans in FPGA-based ML accelerators
|References=Grimsholm, Filip, and Cassandra Westergren. "Resilience of Machine Learning Hardware Accelerators Against Accuracy Degrading Trojans." (2025).
|Supervisor=Mahdi Fazeli
|Level=Master
|Status=Open
}}
In this thesis, you will design and evaluate sensitivity-aware defences that detect and mitigate stealthy, gradual weight-drift Trojans in FPGA-based ML accelerators. You’ll combine PD-guided “canary” monitors with selective integrity checks on the most critical weights, then measure detection speed, false alarms, and hardware cost. The work includes reproducing a baseline attack setup and delivering a practical, low-overhead defence recipe with code.

Profiling ML Side-Channel on CiM for Input Reconstruction

2025-10-21T14:52:46Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=investigate whether supervised models (e.g., U-Net/pix2pix) can reconstruct pri- vate inputs from CiM-generated “power-feature matrices” ..."

{{StudentProjectTemplate
|Summary=investigate whether supervised models (e.g., U-Net/pix2pix) can reconstruct pri- vate inputs from CiM-generated “power-feature matrices” and how noise/sampling constrain feasibility.
|TimeFrame=Spring 2026 (Jan–Jun)
|References=Wang, Ziyu, et al. "PowerGAN: a machine learning approach for power side‐channel attack on compute‐in‐memory accelerators." Advanced Intelligent Systems 5.12 (2023): 2300313.
|Supervisor=Mahdi Fazeli
|Level=Master
|Status=Open
}}
We study profiling side channels in which an adversary trains a supervised model to map CiM power/timing features to user inputs. The setting is an RRAM crossbar CiM performing vector–matrix multiplication with ADC conversion. The attacker records or simulates peroperation features (e.g., tile start/stop markers, accumulate/ADC phase activity, coarse power samples) on a set of known inputs to learn a feature-input mapping. At attack time, the trained model reconstructs private inputs from features obtained on unknown inputs. Prior work demonstrates that a conditional generative model can recover meaningful images from CiM leakage and retains salient structures while remaining effective under substantial measurement noise, indicating practical privacy risk when profiling data and device characteristics are available.

Security Vulnerabilities in Multi-Model Computing-in-Memory Systems

2025-10-21T14:41:27Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=Discover and quantify timing side-channels and model-fingerprinting risks in multi-tenant CiM accelerators using open simulators and a custom..."

{{StudentProjectTemplate
|Summary=Discover and quantify timing side-channels and model-fingerprinting risks in multi-tenant CiM accelerators using open simulators and a custom runtime
|TimeFrame=Spring 2026 (Jan–Jun)
|References=Kim, Seah, et al. "Moca: Memory-centric, adaptive execution for multi-tenant deep neural networks." 2023 IEEE International Symposium on High-Performance Computer Architecture (HPCA). IEEE, 2023.

|Prerequisites=1. Programming in Python; experience with PyTorch and Linux.
2. Background in computer architecture and basic ML (CNNs/transformers).
3. Interest in hardware security or systems security (timing/throughput analysis).
|Supervisor=Mahdi Fazeli
|Level=Master
|Status=Open
}}
Computing-in-Memory (CiM) accelerators increasingly host multiple neural networks on shared analog arrays to raise utilization. Throughput-oriented mechanisms such as tenant-level area allocation, operator splitting or duplication, and fine-grained inter-layer pipelining also increase interactions among co-resident models. This thesis evaluates whether these mechanisms create exploitable timing side-channels and model-specific execution signatures in realistic edge settings without physical access. We design two primary experiments and one goal: (i) cross-model timing leakage, where an unprivileged co-tenant infers binary properties of a victim’s inputs using only its own per-inference latencies and queuing behavior; (ii) model fingerprinting, which identifies the victim’s architecture family from contention-driven timing patterns; and (iii) exploratory parameter-structure inference on small fully connected layers.

Increase Data Rate over Error-prone networks

2025-10-17T12:17:46Z

Cclab:

{{StudentProjectTemplate
|Summary=Increase Data Rate over Error-prone networks (AIRBUS)
|References=1) D. E. Lucani, M. V. Pedersen, D. Ruano, C. W. Sørensen, F. H. P. Fitzek, J. Heide, O. Geil, V. Nguyen, M. Reisslein — “Fulcrum: Flexible Network Coding for Heterogeneous Devices.” IEEE Access, 2018.

2) V. Nguyen, J. A. Cabrera, D. You, H. Salah, G. T. Nguyen, F. H. P. Fitzek — “Advanced Adaptive Decoder Using Fulcrum Network Codes.” IEEE Access, 2019.

3) A. Shahzad, R. Ali, A. Haider, H. S. Kim — “RS-RLNC: A Reinforcement Learning-Based Selective Random Linear Network Coding Framework for Tactile Internet.” IEEE Access, 2023.
|Supervisor=EDISON PIGNATON DE FREITAS
|Level=Master
|Status=Open
}}
This work aims to devise a mechanism to reduce latency by sending redundant data together with the original one, allowing for packet reconstruction at the destination. In particular, FEC through Random Linear Network Coding (RLNC) is able to reduce the number of distinct packet transmissions in a network and minimizes packet transmissions due to poor network conditions. In particular Fulcrum Network Coding (FNC), is a variation of RLNC that aims to reduce the computational complexity and decoding delay. However, Fulcrum Codes parameters are statically chosen before data transmission, while using feedback or retransmission is impractical in rapidly changing network conditions. Hence, this work aims to devise a mechanism to allow FNC to adapt to the available capabilities of nodes in a dynamic network.

Work in partnership with AIRBUS Central Research and Technology, Germany.

Anomaly Detection in Time Series Data Using Generative Models

2025-10-16T19:40:23Z

Cclab:

{{StudentProjectTemplate
|Summary=Anomaly Detection in Time Series Data Using Generative Models
|Keywords=Anomaly detection, time series, generative model
|TimeFrame= Fall 2025
|References=https://github.com/exathlonbenchmark/exathlon
|Prerequisites=Deep learning
|Supervisor=Guojun Liang
|Level=Master
|Status=Open
}}
Research Question:
How effectively can generative models detect anomalies in time series data compared to traditional statistical or supervised methods?

Type:
This research is software-oriented, focusing on algorithm design, model training, and data analysis rather than hardware implementation.

Brief Description of 3–4 Related Works:

Variational Autoencoders (VAE) for Time Series:
VAEs learn the normal patterns of time series data by reconstructing input sequences. Anomalies are detected when the reconstruction error exceeds a certain threshold.

Diffusion Models for Probabilistic Detection:
Diffusion-based generative models capture complex temporal distributions and allow for likelihood-based anomaly detection by modeling the data generation process step by step.

Transformer-based Generative Models:
Attention mechanisms in transformer architectures can capture long-term dependencies in time series, improving the ability to model and identify subtle temporal anomalies.

Hybrid Generative–Predictive Models:
Combining generative models with forecasting networks (e.g., VAE + LSTM) enables learning both the underlying data distribution and predictive patterns, enhancing anomaly detection robustness.

Expected Outcome:
Students will develop a working prototype of a generative-model-based anomaly detection system for time series data. They will:

Gain hands-on experience with deep learning frameworks (e.g., PyTorch or TensorFlow).

Learn how to design, train, and evaluate generative models on real or simulated time series datasets.

Analyze model performance compared to traditional methods.

Produce a short research thesis summarizing the findings, methodology, and potential applications (e.g., fault detection, health monitoring, or finance).

Identification and Classification of Automotive Radar Interference using Data-driven Methods

2025-10-16T07:30:10Z

Cclab:

{{StudentProjectTemplate
|Summary=This is a collaboration with Radar Reticence. The project investigates how different detection methods handle radar interference, a crucial factor for ensuring accurate perception in automated driving.
|Keywords=Automotive radar, ADAS, interference, anomaly detection,
|TimeFrame=Spring 2026
|References=Radar interference: https://www.analog.com/en/resources/analog-dialogue/articles/automotive-radar-sensors-and-congested-radio-spectrum-an-urban-electronic-warfare.html

Equipment: https://www.ti.com/tool/AWR2944EVM, and https://www.ti.com/tool/DCA1000EVM,
|Prerequisites=Programming for data analysis, experience with Matlab/Python (with ML-libraries), signal processing foundations
|Supervisor=Emil Nilsson, Sławomir Nowaczyk, Elena Haller
|Level=Master
|Status=Open
}}
Background

Transportation systems are currently becoming automated for improved efficiency and safety. The automation requires various types of sensors, and for vehicle automation, cameras and radars are common. The cameras and radars are integrated together with computing devices into Advanced Driving and Assisting Systems (ADAS). Drivers become increasingly dependent on ADAS, and eventually the vehicle will require no human operator at all.

Transportation systems are becoming increasingly automated to improve efficiency and safety. These systems rely on integrated sensor suites — cameras and radars — feeding Advanced Driving and Assistance Systems (ADAS). Radar sensors are particularly valuable because they function well in adverse weather and low-light conditions. Unlike passive cameras, radars are active sensors: they transmit radio-frequency signals and receive weak echoes reflected by the environment. Because received echoes are weak, radars are vulnerable to other transmitters operating in the same frequency band; interference can mask echoes or create false detections. As the number of radar-equipped vehicles (and radars per vehicle) grows, interference will become more frequent and varied.
A better empirical understanding of how automotive radar responds to different interference types is required. Radar hardware is complex and hard to model in full fidelity, therefore controlled experiments and data-driven analysis are preferred approaches for a master thesis.

Project Goal

Develop and evaluate a data-driven pipeline that (1) detects the presence of interference in automotive radar data and (2) classifies the interference into meaningful types (e.g., FMCW, CW, GMSK, pulsed, wideband noise). The project will produce an experimental dataset, baseline algorithms, and an evaluation of accuracy and robustness in controled environment.

Research Questions

1. Which radar data representations (ADC/IQ, range profiles, range-Doppler maps, range-Doppler-angle cubes) provide the “best” information for interference detection and classification?

2. How well can an autoencoder (can be smth else) trained only on “good” data detect interference through reconstruction-error–based anomaly scores, and what are suitable thresholds or statistical criteria to separate “clean” from interfered data?

3. Can residual-based features (e.g., reconstruction error, spectral characteristics of the residual, anomaly scores over time) support accurate supervised classification of interference types?

4. How do different interference types (FMCW overlap, CW tone, pulsed bursts, wideband noise) look like in autoencoder residuals, and which features are most important for classification?

Anomaly Detection in Time Series Data Using Generative Models

2025-10-15T19:36:35Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=Anomaly Detection in Time Series Data Using Generative Models |Keywords=Anomaly detection, time series, generative model |TimeFrame= Fall 202..."

{{StudentProjectTemplate
|Summary=Anomaly Detection in Time Series Data Using Generative Models
|Keywords=Anomaly detection, time series, generative model
|TimeFrame= Fall 2025
|References=https://github.com/EQTPartners/TSDE/tree/main
https://github.com/exathlonbenchmark/exathlon
|Prerequisites=Deep learning
|Supervisor=Guojun Liang
|Level=Master
|Status=Open
}}
Research Question:
How effectively can generative models detect anomalies in time series data compared to traditional statistical or supervised methods?

Type:
This research is software-oriented, focusing on algorithm design, model training, and data analysis rather than hardware implementation.

Brief Description of 3–4 Related Works:

Variational Autoencoders (VAE) for Time Series:
VAEs learn the normal patterns of time series data by reconstructing input sequences. Anomalies are detected when the reconstruction error exceeds a certain threshold.

Diffusion Models for Probabilistic Detection:
Diffusion-based generative models capture complex temporal distributions and allow for likelihood-based anomaly detection by modeling the data generation process step by step.

Transformer-based Generative Models:
Attention mechanisms in transformer architectures can capture long-term dependencies in time series, improving the ability to model and identify subtle temporal anomalies.

Hybrid Generative–Predictive Models:
Combining generative models with forecasting networks (e.g., VAE + LSTM) enables learning both the underlying data distribution and predictive patterns, enhancing anomaly detection robustness.

Expected Outcome:
Students will develop a working prototype of a generative-model-based anomaly detection system for time series data. They will:

Gain hands-on experience with deep learning frameworks (e.g., PyTorch or TensorFlow).

Learn how to design, train, and evaluate generative models on real or simulated time series datasets.

Analyze model performance compared to traditional methods.

Produce a short research thesis summarizing the findings, methodology, and potential applications (e.g., fault detection, health monitoring, or finance).

Identification and Classification of Automotive Radar Interference using Data-driven Methods

2025-10-14T10:17:19Z

Cclab: This master thesis project will investigate how automotive radar systems are affected by interference and explore machine learning approaches to detect and characterize interference. The work will be

{{StudentProjectTemplate
|Summary=In this project, we will investigate how different detection methods handle radar interference, a crucial factor for ensuring accurate perception in automated driving.
|Keywords=Automotive radar, ADAS, interference, anomaly detection,
|TimeFrame=Spring 2026
|References=Radar interference: https://www.analog.com/en/resources/analog-dialogue/articles/automotive-radar-sensors-and-congested-radio-spectrum-an-urban-electronic-warfare.html

Equipment: https://www.ti.com/tool/AWR2944EVM, and https://www.ti.com/tool/DCA1000EVM,
|Prerequisites=Programming for data analysis, experience with Matlab/Python (with ML-libraries), signal processing foundations
|Supervisor=Emil Nilsson, Sławomir Nowaczyk, Elena Haller
|Level=Master
|Status=Open
}}
Background

Transportation systems are currently becoming automated for improved efficiency and safety. The automation requires various types of sensors, and for vehicle automation, cameras and radars are common. The cameras and radars are integrated together with computing devices into Advanced Driving and Assisting Systems (ADAS). Drivers become increasingly dependent on ADAS, and eventually the vehicle will require no human operator at all.

Transportation systems are becoming increasingly automated to improve efficiency and safety. These systems rely on integrated sensor suites — cameras and radars — feeding Advanced Driving and Assistance Systems (ADAS). Radar sensors are particularly valuable because they function well in adverse weather and low-light conditions. Unlike passive cameras, radars are active sensors: they transmit radio-frequency signals and receive weak echoes reflected by the environment. Because received echoes are weak, radars are vulnerable to other transmitters operating in the same frequency band; interference can mask echoes or create false detections. As the number of radar-equipped vehicles (and radars per vehicle) grows, interference will become more frequent and varied.
A better empirical understanding of how automotive radar responds to different interference types is required. Radar hardware is complex and hard to model in full fidelity, therefore controlled experiments and data-driven analysis are preferred approaches for a master thesis.

Project Goal

Develop and evaluate a data-driven pipeline that (1) detects the presence of interference in automotive radar data and (2) classifies the interference into meaningful types (e.g., FMCW, CW, GMSK, pulsed, wideband noise). The project will produce an experimental dataset, baseline algorithms, and an evaluation of accuracy and robustness in controled environment.

Research Questions

1. Which radar data representations (ADC/IQ, range profiles, range-Doppler maps, range-Doppler-angle cubes) provide the “best” information for interference detection and classification?

2. How well can an autoencoder (can be smth else) trained only on “good” data detect interference through reconstruction-error–based anomaly scores, and what are suitable thresholds or statistical criteria to separate “clean” from interfered data?

3. Can residual-based features (e.g., reconstruction error, spectral characteristics of the residual, anomaly scores over time) support accurate supervised classification of interference types?

4. How do different interference types (FMCW overlap, CW tone, pulsed bursts, wideband noise) look like in autoencoder residuals, and which features are most important for classification?

Coordinated Multi-Drone Pattern Formation with Crazyflie & Lighthouse

2025-10-13T20:52:32Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=Design and implement synchronized multi-pattern formation (circle, square, hexagon, spiral) for 2–6 Crazyflie drones using Lighthouse posit..."

{{StudentProjectTemplate
|Summary=Design and implement synchronized multi-pattern formation (circle, square, hexagon, spiral) for 2–6 Crazyflie drones using Lighthouse positioning and ROS 2/Crazyswarm2, with smooth formation transitions and quantified accuracy/synchronization.
|References=Crazyswarm/Crazyswarm2 (multi-robot micro-quadrotor control) https://www.bitcraze.io/tag/swarm/

Bitcraze Lighthouse positioning (setup & calibration)

https://www.bitcraze.io/documentation/tutorials/getting-started-with-lighthouse/

Olfati-Saber, “Flocking for multi-agent dynamic systems” (IEEE, 2006)

Ren & Beard, Distributed Consensus in Multi-vehicle Cooperative Control
|Prerequisites=Basic Python, ROS 2, linear algebra & coordinate frames; Git/Linux
|Supervisor=EDISON PIGNATON DE FREITAS
|Level=Master
|Status=Open
}}
Problem: Achieving precise, synchronized formations with multiple micro-UAVs indoors is challenged by timing jitter, radio bandwidth, and pose noise, yet a generic, reproducible formation framework would enable reliable group behaviors.

Goal: Develop a Lighthouse-based multi-drone system that executes parameterized formations and smooth transitions for 2–6 Crazyflies with measured accuracy, synchronization, and scalability.

Proposed Solution & Tasks: Implement parametric pattern generators; use Crazyswarm2 to upload/start time-aligned trajectories with per-drone offsets; add transition scheduling between patterns; manage radio channels and logging; record flight data and analyze formation error, phase/sync, and tracking.

Evaluation Criteria: Formation RMS error, minimum inter-drone separation, trajectory tracking error during transitions, scalability vs. number of drones.

Suggested Tools & Platforms: ROS 2, Crazyswarm2 (control), Bitcraze Lighthouse (positioning), Python scripts for trajectory generation and data analysis.

Project in cooperation with Sivadinesh Ponrajan (sivadinesh.ponrajan@hh.se)

Posture Estimation for Motorcycle Simulator Riders Using Remote Sensing and IMU-Based Systems

2025-10-08T12:26:17Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=Develop and compare posture estimation methods for motorcycle simulator riders using camera-based remote sensing and IMU-based systems, focus..."

{{StudentProjectTemplate
|Summary=Develop and compare posture estimation methods for motorcycle simulator riders using camera-based remote sensing and IMU-based systems, focusing on hip movement, body lean angles, and leg positions.
|Keywords=Posture Estimation, Motorcycle Simulator, IMU, Camera-Based Tracking, ROS, CARLA, BikeSim
|TimeFrame=Fall 2025 - Spring 2027
|Prerequisites=Familiarity with ROS and CARLA simulator
Knowledge of computer vision (e.g., pose estimation frameworks)
Experience with IMU data processing and sensor fusion
Python programming for data analysis and integration
|Supervisor=Oscar Amador Molina
|Examiner=Emil Nilsson
|Level=Master
|Status=Open
}}
<h3>Research Question</h3>
<blockquote>
How can remote sensing (camera-based) and IMU-based systems be used to estimate rider posture in a motorcycle simulator, and how do these approaches compare in terms of accuracy and robustness?
</blockquote>

<h3>HW/SW Focus</h3>
<blockquote>
The project emphasizes algorithm development and system integration, leveraging CARLA and ROS for simulation and data synchronization. Both vision-based and IMU-based methods will be explored and compared.
</blockquote>

<h3>Work Packages</h3>
<ul>
<li><b>WP1: System Setup & Data Collection</b>
<ul>
<li>Integrate CARLA simulator with ROS for data streaming.</li>
<li>Configure camera-based tracking system for rider posture.</li>
<li>Prepare IMU modules for hip, torso, and legs; design enclosures for secure attachment.</li>
<li>Collect synchronized datasets from both systems.</li>
</ul>
</li>
<li><b>WP2: Posture Estimation Algorithms</b>
<ul>
<li>Implement vision-based posture estimation (e.g., OpenPose, MediaPipe).</li>
<li>Develop IMU-based orientation and joint angle estimation.</li>
<li>Extract key posture metrics: hip movement, body lean angles, leg positions.</li>
</ul>
</li>
<li><b>WP3: Validation & Comparative Analysis</b>
<ul>
<li>Define ground truth using simulator kinematics or marker-based system.</li>
<li>Compare accuracy, latency, and robustness of both approaches.</li>
<li>Analyze trade-offs for real-world applicability.</li>
</ul>
</li>
</ul>

<h3>Deliverables / Outcomes</h3>
<ul>
<li>Integrated CARLA-ROS simulation environment with posture tracking.</li>
<li>Two posture estimation pipelines: camera-based and IMU-based.</li>
<li>Comparative analysis report with performance metrics and recommendations.</li>
</ul>

Posture Estimation for Motorcycle Riders Using Multi-IMU Systems and Video-Based Ground Truth

2025-10-08T11:59:40Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=Develop and validate a motorcycle rider posture estimation system using multi-IMU sensors and video-based ground truth for accuracy evaluatio..."

{{StudentProjectTemplate
|Summary=Develop and validate a motorcycle rider posture estimation system using multi-IMU sensors and video-based ground truth for accuracy evaluation.
|Keywords=IMU, Sensor Fusion, Posture Estimation, microROS, Motorcycle Safety, Motion Tracking
|TimeFrame=Fall 2025 - Spring 2027
|Prerequisites=Embedded systems programming (ESP32, microROS)
Sensor fusion techniques (e.g., Extended Kalman Filter)
Basic machine learning (optional, for posture classification)
Familiarity with Python or MATLAB for data analysis
|Supervisor=Oscar Amador Molina
|Examiner=Pererik Andreasson
|Level=Master
|Status=Open
}}
Research Question:
How can multi-IMU systems integrated into a motorcycle jacket and helmet be used to accurately estimate rider posture, and how does the performance compare to a video-based ground truth system?

Hardware/Software Focus:
The project is primarily algorithm-oriented, with some hardware integration for data collection and synchronization. Emphasis will be on sensor fusion, drift compensation, and validation against ground truth.

Work Packages

WP1: Hardware Setup & Data Collection

- Configure jacket with six IMUs and helmet IMU.

- Implement microROS communication over Wi-Fi.

- Synchronize IMU and video data streams.

- Collect datasets under controlled riding scenarios.

WP2: Sensor Fusion & Posture Estimation

- Implement orientation estimation algorithms (Madgwick, Mahony, EKF).

- Investigate magnetometer interference and apply correction strategies (e.g., soft/hard iron calibration, magnetometer-free fusion).

- Compute joint angles and posture metrics from IMU orientations.

WP3: Validation & Analysis

- Use video-based marker system as ground truth.

- Evaluate accuracy (e.g., RMSE of joint angles) and robustness under different conditions.

- Compare performance of different fusion algorithms.

Deliverables / Outcomes

- Synchronized dataset of IMU and video-based posture data.

- Posture estimation algorithm with documented performance metrics.

- Comparative analysis report of IMU-based vs. video-based posture estimation.
- Recommendations for improving robustness in real-world motorcycle environments.

Service Centric Mobility Management in LEO Constellations

2025-10-06T23:38:03Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=Service Centric Mobility Management in LEO Constellations |References=1) User-Centric Flexible Resource Management Framework for LEO Satellit..."

{{StudentProjectTemplate
|Summary=Service Centric Mobility Management in LEO Constellations
|References=1) User-Centric Flexible Resource Management Framework for LEO Satellites with Fully Regenerative Payload — Sovit Bhandari, Thang X. Vu, Symeon Chatzinotas; IEEE Journal on Selected Areas in Communications (JSAC), May 2024.

2) LISP-LEO: Location/Identity Separation-based Mobility Management for LEO Satellite Networks — Jun Hu, Tian Pan, Yujie Chen, Xuebei Zhang, Tao Huang, Yunjie Liu; IEEE Global Communications Conference (GLOBECOM), 2022.

3) SKYCASTLE: Taming LEO Mobility to Facilitate Seamless and Low-latency Satellite Internet Services — Jihao Li et al.; IEEE INFOCOM 2024
|Supervisor=EDISON PIGNATON DE FREITAS
|Level=Master
|Status=Open
}}
This work aims to tackle the problem of mobility management in IP-based LEO constellation networks, based on the notion of service mobility, and not based on the notion of host mobility as done by prior-art.

The goal is to allow customers (e.g. aircraft) to always be well connected to communication services, even in the presence of rapid network topology to change constantly.

Work in partnership with AIRBUS Central Research and Technology, Germany.

Data muling services over a constellation of aircraft

2025-10-06T23:34:30Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=Data muling services over a constellation of aircraft |References=1) “Unmanned Aerial Vehicles as Data Mules: An Experimental Assessment”..."

{{StudentProjectTemplate
|Summary=Data muling services over a constellation of aircraft
|References=1) “Unmanned Aerial Vehicles as Data Mules: An Experimental Assessment” — Palma, Zolich, Jiang, Johansen, IEEE Access, 2017.

2) “GeoSaW: A Location-Aware Waypoint-Based Routing Protocol for Airborne DTNs in Search and Rescue Scenarios” — Bujari, Calafate, Cano, Manzoni, Palazzi, Ronzani, Sensors (Basel), 2018.
|Supervisor=EDISON PIGNATON DE FREITAS
|Level=Master
|Status=Open
}}
A Data Muling service is a method for transferring data between locations that lack a continuous, end-to-end network connection.

The core concept relies on a specialized mobile node, called a data mule, that physically carries data and acts as a mobile relay.

This work aims to analyse the viability of having aircraft as data mules, in a scenario where aircraft may be interconnected by some omnidirectional or directional links.

Work in partnership with AIRBUS Central Research and Technology, Germany.

Segment Routing based on Geographic Checkpoints

2025-10-06T23:31:47Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=Segment Routing based on Geographic Checkpoints: |References=1) M. Moy, R. Kassouf-Short, N. Kortas, J. Cleveland, B. Tomko, D. Conricode, Y...."

{{StudentProjectTemplate
|Summary=Segment Routing based on Geographic Checkpoints:
|References=1) M. Moy, R. Kassouf-Short, N. Kortas, J. Cleveland, B. Tomko, D. Conricode, Y. Kirkpatrick, R. Cardona, B. Heller and J. Curry — “Contact Multigraph Routing: Overview and Implementation.” 2023 IEEE Aerospace Conference.

2) J. Hu, L. Cai, C. Zhao and J. Pan — “Directed Percolation Routing for Ultra-Reliable and Low-Latency Services in Low Earth Orbit (LEO) Satellite Networks.” 2020 IEEE Vehicular Technology Conference (VTC-Fall).

3) G. Stock, J. A. Fraire and H. Hermanns, "Distributed On-Demand Routing for LEO Mega-Constellations: A Starlink Case Study," 2022 11th Advanced Satellite Multimedia Systems Conference and the 17th Signal Processing for Space Communications Workshop (ASMS/SPSC), Graz, Austria, 2022, pp. 1-8, doi: 10.1109/ASMS/SPSC55670.2022.9914716
|Supervisor=EDISON PIGNATON DE FREITAS
|Level=Master
|Status=Open
}}
Segment Routing based on Geographic Checkpoints:

Improve performance in starlink, using other algorithms in Florasat (e.g. CGR), and the "greedy geographic" routing approach of starlink as a benchmark.

Analyse how the routing protocol behaves in other constellations like Oneweb, Iridium, Kuiper or Telesat.

Analyse the behaviour of the routing protocol in an Aircraft constellation.

The student(s) may choose 2 out of 3 to perform the thesis, or one, if a deeper he/she/them want(s) to perform a deeper study.

Work in partnership with AIRBUS Central Research and Technology, Germany.

Real-time flow compression for TCP and QUIC at the network edge

2025-10-06T23:24:09Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=Real-time flow compression for TCP and QUIC at the network edge |References=1) “Dynamic Semantic Compression for CNN Inference in Multi-Acc..."

{{StudentProjectTemplate
|Summary=Real-time flow compression for TCP and QUIC at the network edge
|References=1) “Dynamic Semantic Compression for CNN Inference in Multi-Access Edge Computing: A Graph Reinforcement Learning-Based Autoencoder”
Nan Li, Alexandros Iosifidis, Qi Zhang — IEEE Transactions on Wireless Communications, Vol. 24, Issue 3, Mar 2025.

2) “CStream: Parallel Data Stream Compression on Multicore Edge Devices” Xianzhi Zeng, Shuhao Zhang — published in 2023 (preprint / arXiv)

3) “To Talk or to Work: Flexible Communication Compression for Energy Efficient Federated Learning over Heterogeneous Mobile Edge Devices”
Liang Li, Dian Shi, Ronghui Hou, Hui Li, Miao Pan, Zhu Han — FLIP-type / IEEE / arXiv etc., 2020.
|Supervisor=EDISON PIGNATON DE FREITAS
|Level=Master
|Status=Open
}}
This work aims to minimize redundant data and overhead using features inherent to the protocols or their upper layers (like HTTP/2 and HTTP/3). For instance using header compression by replacing common headers with a single-byte index or Store custom or previously seen header-value pairs during the connection, replacing them with a much smaller index on subsequent requests. Other approaches may include having edge devices compressing the payload data stream in real time before transmission using algorithms like LZ4 or LZO. The work may include using a system able to monitor local resources (CPU utilization, battery level) and network conditions (latency, bandwidth) to dynamically adjust the compression level or even the algorithm itself. An advanced setup may also include using Deep Reinforcement Learning agents to learn the optimal trade-off between compression ratio, compression time, and power consumption based on real-time feedback from the network and device, ensuring that real-time latency requirements are always met.

Work in partnership with AIRBUS Central Research and Technology, Germany.

Increase Data Rate over Error-prone networks

2025-10-06T23:21:39Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=Increase Data Rate over Error-prone networks |References=1) D. E. Lucani, M. V. Pedersen, D. Ruano, C. W. Sørensen, F. H. P. Fitzek, J. Heid..."

{{StudentProjectTemplate
|Summary=Increase Data Rate over Error-prone networks
|References=1) D. E. Lucani, M. V. Pedersen, D. Ruano, C. W. Sørensen, F. H. P. Fitzek, J. Heide, O. Geil, V. Nguyen, M. Reisslein — “Fulcrum: Flexible Network Coding for Heterogeneous Devices.” IEEE Access, 2018.

2) V. Nguyen, J. A. Cabrera, D. You, H. Salah, G. T. Nguyen, F. H. P. Fitzek — “Advanced Adaptive Decoder Using Fulcrum Network Codes.” IEEE Access, 2019.

3) A. Shahzad, R. Ali, A. Haider, H. S. Kim — “RS-RLNC: A Reinforcement Learning-Based Selective Random Linear Network Coding Framework for Tactile Internet.” IEEE Access, 2023.
|Supervisor=EDISON PIGNATON DE FREITAS
|Level=Master
|Status=Open
}}
This work aims to devise a mechanism to reduce latency by sending redundant data together with the original one, allowing for packet reconstruction at the destination. In particular, FEC through Random Linear Network Coding (RLNC) is able to reduce the number of distinct packet transmissions in a network and minimizes packet transmissions due to poor network conditions. In particular Fulcrum Network Coding (FNC), is a variation of RLNC that aims to reduce the computational complexity and decoding delay. However, Fulcrum Codes parameters are statically chosen before data transmission, while using feedback or retransmission is impractical in rapidly changing network conditions. Hence, this work aims to devise a mechanism to allow FNC to adapt to the available capabilities of nodes in a dynamic network.

Work in partnership with AIRBUS Central Research and Technology, Germany.

Segment-based coordination of Congestion Control

2025-10-06T23:17:26Z

Cclab: Development of a decentralized congestion control mechanism among multiple transport-layer connections (e.g., TCP or QUIC) between the source and destination

{{StudentProjectTemplate
|Summary=Segment-based coordination of Congestion Control
|References=1 - S. N. S. Hashemi & A. Bohlooli — “3CP: Coordinated Congestion Control Protocol for Named-Data Networking”, IEEE Transactions on Network and Service Management (TNSM), 2021.

2 - W. Yang, L. Cai, S. Shu, J. Pan — “Mobility-Aware Congestion Control for Multipath QUIC in Integrated Terrestrial and LEO Satellite Networks”, IEEE Transactions on Mobile Computing (TMC), 2024.

3 - W. Li, H. Zhang, S. Gao, C. Xue, X. Wang, S. Lu — “SmartCC: A Reinforcement-Learning Approach for Multipath TCP Congestion Control in Heterogeneous Networks”, IEEE Journal on Selected Areas in Communications (JSAC), 2019.
|Supervisor=EDISON PIGNATON DE FREITAS
|Level=Master
|Status=Open
}}
This work aims to develop a decentralized congestion control mechanism among multiple transport-layer connections (e.g., TCP or QUIC) between the source and destination set up as a chain in different segments.

Transport-layer connections in each segment may be set up between proxy servers, middleboxes, or even relays of new transport standards like Media over QUIC.

Non-Invasive Safety Monitoring via Integrated Sensing and Communication

2025-10-06T11:41:19Z

Cclab:

{{StudentProjectTemplate
|Summary=Smart sensing for motion, presence, and accident detection—without cameras
|Keywords=ISAC, IoT, Networking, Sensing
|TimeFrame=Fall 2025 - Spring 2026
|Supervisor=Oscar Amador Molina
|Level=Master
|Status=Open
}}
Research question:

How can Integrated Sensing and Communication (ISAC) using Wi-Fi (802.11 b/g/n), Bluetooth RSSI, and sensor-equipped ESP32-C3 devices be used to detect motion, presence, and accidents in a non-invasive way to improve safety for elderly individuals living alone?

Focus:

The project is primarily software-oriented, with some hardware integration involving ESP32-C3 microcontrollers and basic sensors.

Work packages:

WP1: Wi-Fi Sensing – Analyze 802.11 access-layer data (e.g., RSSI) to infer motion and presence.

WP2: Bluetooth RSSI Communication – Use BLE RSSI for proximity detection and device-to-device signaling.

WP3: Sensor Integration – Connect motion and environmental sensors to ESP32-C3 devices and develop logic for accident detection.

WP4: System Coordination – Enable communication between devices to share sensing data and trigger alerts.

Deliverables/outcomes/results:

A set of prototype systems demonstrating motion, presence, and accident detection.
Evaluation of detection accuracy and system responsiveness in a simulated home environment.

Non-Invasive Safety Monitoring via Integrated Sensing and Communication

2025-10-06T11:39:49Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=Smart sensing for motion, presence, and accident detection—without cameras |Keywords=ISAC, IoT, Networking, Sensing |Supervisor=Oscar Amado..."

{{StudentProjectTemplate
|Summary=Smart sensing for motion, presence, and accident detection—without cameras
|Keywords=ISAC, IoT, Networking, Sensing
|Supervisor=Oscar Amador Molina
|Level=Master
|Status=Open
}}
Research question:

How can Integrated Sensing and Communication (ISAC) using Wi-Fi (802.11 b/g/n), Bluetooth RSSI, and sensor-equipped ESP32-C3 devices be used to detect motion, presence, and accidents in a non-invasive way to improve safety for elderly individuals living alone?

Focus:

The project is primarily software-oriented, with some hardware integration involving ESP32-C3 microcontrollers and basic sensors.

Work packages:

WP1: Wi-Fi Sensing – Analyze 802.11 access-layer data (e.g., RSSI) to infer motion and presence.

WP2: Bluetooth RSSI Communication – Use BLE RSSI for proximity detection and device-to-device signaling.

WP3: Sensor Integration – Connect motion and environmental sensors to ESP32-C3 devices and develop logic for accident detection.

WP4: System Coordination – Enable communication between devices to share sensing data and trigger alerts.

Deliverables/outcomes/results:

A set of prototype systems demonstrating motion, presence, and accident detection.
Evaluation of detection accuracy and system responsiveness in a simulated home environment.

AI-Driven Semantic Encoding for Efficient Communication

2025-10-03T22:34:38Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=To develop and evaluate an AI-based semantic encoding model capable of transforming raw data into compact, structured representations. |Refe..."

{{StudentProjectTemplate
|Summary=To develop and evaluate an AI-based semantic encoding model capable of transforming raw data into compact, structured representations.
|References=1) Pezone, Francesco. "Semantic communication based on generative AI: a new approach to image compression and edge optimization." arXiv preprint arXiv:2502.01675 (2025).

2) Qiao, Li, Mahdi Boloursaz Mashhadi, Zhen Gao, Chuan Heng Foh, Pei Xiao, and Mehdi Bennis. "Latency-aware generative semantic communications with pre-trained diffusion models." IEEE wireless communications letters (2024).

3) Islam, Azharul, and KyungHi Chang. "Navigating the future of wireless networks: A multidimensional survey on semantic communications." ICT Express 10, no. 4 (2024): 747-773.

4) Li, Nan, Alexandros Iosifidis, and Qi Zhang. "Dynamic semantic compression for cnn inference in multi-access edge computing: A graph reinforcement learning-based autoencoder." IEEE Transactions on Wireless Communications (2024).
|Supervisor=EDISON PIGNATON DE FREITAS
|Level=Master
|Status=Open
}}
Contextualization:
Semantic Communication Networks represent a fundamental shift in how we think about and design communication systems. Unlike traditional communication, which focuses on the accurate and reliable transmission of bits, SCNs prioritize the meaning or intent behind the data.

One of the key element in SCN is the semantic encoder. Instead of compressing data using a general-purpose algorithm (like JPEG for images), a sophisticated AI model analyzes the raw data and extracts only the most important semantic features. The encoder compresses this semantic information into a low-dimensional vector that captures the properties of the original data.

Goals:
To develop and evaluate an AI-based semantic encoding model capable of transforming raw data into compact, structured representations. The objective is to reduce transmission payload while preserving the essential meaning of the information. The proposed encoder will focus on text sentences and/or images as primary data modalities.

Testing;
Datasets recommendation: European Parliament for text and MNIST for images.

Evaluation Criteria:
Compare the proposed solution with traditional communication (the baseline) in terms of compression ratio; semantic similarity; CPU/GPU consumption and encoding/decoding time.

Extra (if you have time): behavior/evaluation when a normal noise is inserted in both baseline and proposed solution.

Main Tasks:
Task 1: Study semantic encoders concepts, especially NLPs, LLMs, ANNs Knowledge Graphs and Topos.
Task 2: Study the addressed dataset.
Task 3: Develop, train and improve the model.
Task 4: Test the model, capturing the evaluation criteria.
Task 5: Analisys the results and compare with baseline.
Task 6: Write thesis and prepare presentation.

Deliverables (Besides the final thesis document):
- AI model specification to encode and decode text/image.
- Tests description with visual result representation (graphs).
- Performance analysis report.
- Thesis documentation.

Resilient Recovery: A Byzantine Fault Tolerance (BFT) Framework for Tactical Mesh Networks

2025-10-02T10:50:36Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=This proposal addresses the critical state-poisoning vulnerability in tactical networks during post-partition recovery. The research proposes..."

{{StudentProjectTemplate
|Summary=This proposal addresses the critical state-poisoning vulnerability in tactical networks during post-partition recovery. The research proposes the development of a recovery protocol centered on Byzantine Fault Tolerance (BFT)
|Keywords=Byzantine Fault Tolerance (BFT), Tactical Mesh Networks, Distributed Consensus, Network Recovery, Critical Systems
|Supervisor=ALEXANDRE DOS SANTOS ROQUE
|Level=Master
|Status=Open
}}
Summary:
This proposal addresses the critical state-poisoning vulnerability in tactical networks during post-partition recovery. The research proposes the development of a recovery protocol centered on Byzantine Fault Tolerance (BFT) that forces a distributed consensus among nodes before accepting state updates. The goal is to prevent compromised nodes from corrupting mission-critical state information, such as targeting data or command logs, ensuring that network re-synchronization occurs securely and reliably, even in the presence of malicious actors.

General Information:
Goals: To design, implement, and validate a recovery protocol for tactical mesh networks that utilizes BFT principles to guarantee network state integrity during the re-integration of partitioned segments, thereby preventing data poisoning attacks.

Research Focus and Objectives:
The research will focus on adapting classic BFT algorithms to the dynamic and resource-constrained environment of tactical mobile ad-hoc networks (MANETs).

The objectives include:
- Analyzing the feasibility and overhead of different BFT protocols in tactical network scenarios.
- Designing a lightweight consensus mechanism that allows rejoining network segments to agree upon a unified and trustworthy state.
- Developing a state synchronization process that is resilient to Byzantine (compromised) nodes attempting to provide falsified state histories.
- Evaluating the trade-off between the level of security provided by BFT and the impact on recovery performance (latency and bandwidth consumption).

Robotics for Urban Traffic Support: Design, Prototyping, and Intelligent Interaction in Smart Cities

2025-10-01T12:09:05Z

Cclab:

{{StudentProjectTemplate
|Summary=This proposal focuses on the application of robotics to enhance traffic safety and efficiency in smart cities. The research considers the interaction with VRU
|Keywords=Robotics, Smart Cities, Traffic Support, Human-Robot Interaction, Vulnerable Road Users (VRU), Urban Safety
|Supervisor=ALEXANDRE DOS SANTOS ROQUE
|Level=Master
|Status=Open
}}
Summary: "This proposal focuses on the application of robotics to enhance traffic safety and efficiency in smart cities. The research will encompass the design and prototyping (potentially using 3D printing) of robots capable of interacting with people, vulnerable road users (VRU), and transit agents. It will explore robot communication with smart city infrastructure, regulatory considerations, and case studies highlighting their benefits in intense traffic, accident response, and VRU guidance, aiming to create safer urban environments."

Goals: To design, prototype, and evaluate robotic systems for smart city traffic support, enhancing safety for all road users, particularly in challenging urban scenarios, through intelligent interaction and integration with smart infrastructure.

Research Focus and Objectives: This thesis will focus on:

- Conceptual design and prototyping (e.g., 3D printing) of traffic support robots.
- Mechanisms for effective robot communication with smart city systems.
- Exploration of regulatory frameworks and ethical considerations for urban robot deployment.
- Analysis of case studies where robotic intervention improves safety and flow (e.g., congestion, accidents, VRU guidance).
- Investigating human-robot interaction principles for intuitive and safe engagement.

Methodology and Implementation:
The research will combine conceptual design for various traffic robot types with rapid prototyping (e.g., 3D printing).
Evaluation and study of communication protocols for robot interaction with smart city infrastructure, or simulation-based. An option is Traffic simulation softwares to model robotic impact in high-density traffic, accident response, and VRU guidance scenarios. Human-robot interaction strategies will be developed, and existing regulations for urban robot deployment will be reviewed. Communication protocols and the reliability dependency need to be considered.

Project Deliverables:
Key deliverables include conceptual robot designs, a physical or simulated prototype, communication architecture designs, simulation results demonstrating positive impacts on traffic and safety, and a report on regulatory and ethical considerations.

Main Tasks:
Task 1: Conduct a literature review on urban robotics, smart city traffic management, human-robot interaction (HRI), and 3D printing applications in robotics.
Task 2: Develop conceptual designs for traffic support robots targeting specific urban challenges (e.g., traffic intensity, accidents, VRU guidance).
Task 3: Investigate and apply 3D printing or other rapid prototyping techniques to create a physical or detailed virtual prototype of a selected robot design.
Task 4: Design and simulate communication protocols for seamless integration of robots with existing smart city infrastructure and traffic management systems.
Task 5: Model and simulate various urban scenarios (e.g., peak hour traffic, accident response, pedestrian crossing assistance) to evaluate the impact of robotic intervention on safety and efficiency.
Task 6: Analyze current regulations and propose guidelines or recommendations for the safe and effective deployment of urban traffic support robots.
Task 7: Evaluate the effectiveness of the robot's interactions with people, VRUs, and transit agents in simulated or controlled environments.
Task 8: Finalize thesis document and prepare for academic dissemination.

Deliverables (Besides the final thesis document):

Detailed conceptual designs and technical specifications for urban traffic support robots.
A physical or simulated prototype of a key robotic component or full system.
Design of communication architecture for robot integration with smart city systems.
Simulation results demonstrating improvements in traffic flow, safety, and VRU protection due to robotic intervention.
A comprehensive report on regulatory, ethical, and societal considerations for urban robotics deployment.

Evaluation Criteria:

Innovation and practicality of the robot designs and prototyping approach.
Measurable improvements in traffic safety and efficiency demonstrated through simulations.
The robustness and scalability of the robot's communication and integration capabilities within smart city ecosystems.
The depth of analysis of regulatory challenges and the utility of proposed guidelines for urban robotics.

Robotics for Urban Traffic Support: Design, Prototyping, and Intelligent Interaction in Smart Cities

2025-10-01T10:48:03Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=This proposal focuses on the application of robotics to enhance traffic safety and efficiency in smart cities. The research considers the int..."

{{StudentProjectTemplate
|Summary=This proposal focuses on the application of robotics to enhance traffic safety and efficiency in smart cities. The research considers the interaction with VRU
|Keywords=Robotics, Smart Cities, Traffic Support, 3D Printing, Human-Robot Interaction, Vulnerable Road Users (VRU), Urban Safety
|Supervisor=ALEXANDRE DOS SANTOS ROQUE
|Level=Master
|Status=Open
}}
Summary: "This proposal focuses on the application of robotics to enhance traffic safety and efficiency in smart cities. The research will encompass the design and prototyping (potentially using 3D printing) of robots capable of interacting with people, vulnerable road users (VRU), and transit agents. It will explore robot communication with smart city infrastructure, regulatory considerations, and case studies highlighting their benefits in intense traffic, accident response, and VRU guidance, aiming to create safer urban environments."

Goals: To design, prototype, and evaluate robotic systems for smart city traffic support, enhancing safety for all road users, particularly in challenging urban scenarios, through intelligent interaction and integration with smart infrastructure.

Research Focus and Objectives: This thesis will focus on:

- Conceptual design and prototyping (e.g., 3D printing) of traffic support robots.
- Mechanisms for effective robot communication with smart city systems.
- Exploration of regulatory frameworks and ethical considerations for urban robot deployment.
- Analysis of case studies where robotic intervention improves safety and flow (e.g., congestion, accidents, VRU guidance).
- Investigating human-robot interaction principles for intuitive and safe engagement.

Methodology and Implementation:
The research will combine conceptual design for various traffic robot types with rapid prototyping (e.g., 3D printing).
Evaluation and study of communication protocols for robot interaction with smart city infrastructure, or simulation-based. An option is Traffic simulation softwares to model robotic impact in high-density traffic, accident response, and VRU guidance scenarios. Human-robot interaction strategies will be developed, and existing regulations for urban robot deployment will be reviewed. Communication protocols and the reliability dependency need to be considered.

Project Deliverables:
Key deliverables include conceptual robot designs, a physical or simulated prototype, communication architecture designs, simulation results demonstrating positive impacts on traffic and safety, and a report on regulatory and ethical considerations.

Main Tasks:
Task 1: Conduct a literature review on urban robotics, smart city traffic management, human-robot interaction (HRI), and 3D printing applications in robotics.
Task 2: Develop conceptual designs for traffic support robots targeting specific urban challenges (e.g., traffic intensity, accidents, VRU guidance).
Task 3: Investigate and apply 3D printing or other rapid prototyping techniques to create a physical or detailed virtual prototype of a selected robot design.
Task 4: Design and simulate communication protocols for seamless integration of robots with existing smart city infrastructure and traffic management systems.
Task 5: Model and simulate various urban scenarios (e.g., peak hour traffic, accident response, pedestrian crossing assistance) to evaluate the impact of robotic intervention on safety and efficiency.
Task 6: Analyze current regulations and propose guidelines or recommendations for the safe and effective deployment of urban traffic support robots.
Task 7: Evaluate the effectiveness of the robot's interactions with people, VRUs, and transit agents in simulated or controlled environments.
Task 8: Finalize thesis document and prepare for academic dissemination.

Deliverables (Besides the final thesis document):

Detailed conceptual designs and technical specifications for urban traffic support robots.
A physical or simulated prototype of a key robotic component or full system.
Design of communication architecture for robot integration with smart city systems.
Simulation results demonstrating improvements in traffic flow, safety, and VRU protection due to robotic intervention.
A comprehensive report on regulatory, ethical, and societal considerations for urban robotics deployment.

Evaluation Criteria:

Innovation and practicality of the robot designs and prototyping approach.
Measurable improvements in traffic safety and efficiency demonstrated through simulations.
The robustness and scalability of the robot's communication and integration capabilities within smart city ecosystems.
The depth of analysis of regulatory challenges and the utility of proposed guidelines for urban robotics.

Leveraging LLM for Proactive Fault Analysis and Prediction in V2X Communication Systems

2025-10-01T10:35:18Z

Cclab:

{{StudentProjectTemplate
|Summary=This proposal outlines research into applying LLMs for advanced fault analysis and prediction within V2X communication systems. The study will involve analyzing public datasets, developing novel datasets for specific V2X scenarios.
|Keywords=V2X Communication, Fault Analysis, Fault Prediction, Automotive Systems, Data collection
|Supervisor=ALEXANDRE DOS SANTOS ROQUE
|Level=Master
|Status=Open
}}
Goals: Design and implement an LLM-based framework to analyze and predict faults in V2X communication systems, enhancing system reliability and safety. The research aims to establish a robust methodology and knowledge base applicable to critical automotive scenarios.

Research Focus and Objectives: This thesis will explore the capabilities of LLMs in processing V2X data for identifying anomalous patterns and predicting potential failures. It will focus on:

- Investigating the applicability of various LLMs on existing public V2X datasets.
- Developing new V2X datasets that simulate critical failure modes and challenging communication environments.
- Simulating specific case studies where V2X communication integrity is paramount for automotive safety.
- Researching and evaluating existing simulation software, frameworks, and other AI approaches within this scope.

Methodology and Implementation:

- Data Collection and Generation: Sourcing public V2X datasets and generating synthetic datasets through V2X simulation tools (e.g., SUMO, Veins, OMNeT++).
- LLM Application and Fine-tuning: Applying pre-trained LLMs and potentially fine-tuning them for V2X-specific fault recognition and prediction tasks.
- Case Study Simulation: Developing and simulating critical V2X scenarios (e.g., platooning, emergency vehicle warning, intersection assistance) to test fault detection and prediction capabilities.
- Evaluation and application of specific metrics (Prediction Accuracy, False Positive Rate, Scalability, Robustness, Methodology Utility)

Project Deliverables:
Deliverables include a comprehensive literature review, study or development of V2X datasets (synthetic and annotated), LLM models analysis, simulation results for one critical case study.

Main Tasks:
Task 1: Conduct a thorough literature review on LLM applications in fault diagnosis/prediction and V2X communication systems.
Task 2: Identify suitable public V2X datasets for initial LLM training and evaluation.
Task 3: Design and implement a methodology for generating novel V2X datasets simulating critical fault scenarios.
Task 4: Select and adapt appropriate LLMs for V2X fault analysis and prediction, potentially involving transfer learning or fine-tuning.
Task 5: Develop simulation environments and case studies focusing on V2X communication criticality for automotive safety.
Task 6: Evaluate the LLM-based fault analysis and prediction system using both public and custom datasets, as well as simulated case studies.
Task 7: Propose and document a methodological framework or knowledge base for V2X fault management using LLMs.
Task 8: Finalize thesis document and prepare for dissemination (e.g., journal and conference papers).

Deliverables (Besides the final thesis document):

Curated and documented public V2X datasets.
Novel synthetic V2X datasets simulating critical fault scenarios.
Implemented and evaluated LLM models for V2X fault analysis and prediction.
Detailed simulation results and analysis for critical V2X case studies.

Safety Assurance in Critical Automotive Control Systems: A Human-Centric and Standards-Compliant Approach

2025-10-01T10:34:30Z

Cclab:

{{StudentProjectTemplate
|Summary=This research proposes an in-depth study into the safety aspects of critical automotive control systems, such as those found in autonomous vehicles. It will investigate the interplay between communication dependency in specific scenarios.
|Keywords=Automotive Safety, ISO 26262, ISO 21448, Communication Dependency, Human Factors, Cultural Aspects, Critical Control Systems.
|Supervisor=ALEXANDRE DOS SANTOS ROQUE
|Level=Master
|Status=Open
}}
Goals: To comprehensively study vehicle safety, for example, autonomous cars, but not limited to, by integrating technical standards (ISO 26262, ISO 21448), communication dependency, and socio-cultural factors, aiming for safer and more acceptable autonomous mobility solutions.

Research Focus and Objectives: This thesis will focus on:

- Analyzing safety implications of communication dependency in critical autonomous driving scenarios.
- Investigating challenges and applications of ISO 26262 and ISO 21448 for advanced autonomous features.
- Exploring user perceptions, trust, and cultural influences on autonomous vehicle safety and acceptance.
- Identifying and addressing gaps between technical safety implementations and human expectations/behavior.

Methodology and Implementation:
The study will involve an extensive literature review on safety standards, communication protocols, and human factors. Critical driving scenarios will be analyzed for communication-related safety risks. Current autonomous vehicle discussions need to be reviewed, also considering the recent applications of standards as ISO 26262 and ISO 21448, and others. User perspectives on safety and acceptance could be gathered via surveys or interviews. Evaluate the application of a conceptual framework to integrate technical safety with human-centric design.

Project Deliverables:
Key deliverables include a detailed analysis report on communication-dependent safety aspects, a comparative study of safety standards application, a report on user perception and cultural challenges, and a conceptual framework for integrated automotive safety design.

Main Tasks:
Task 1: Conduct a comprehensive literature review on autonomous vehicle safety, communication dependency, ISO 26262, ISO 21448, and human factors in automotive contexts.
Task 2: Analyze specific critical autonomous driving scenarios to identify safety risks stemming from communication reliance and potential failures.
Task 3: Examine how existing and emerging autonomous vehicle technologies align with, or deviate from, the requirements of ISO 26262 and ISO 21448.
Task 4: Design and execute a qualitative and/or quantitative study (surveys, interviews, focus groups) to capture user views, cultural aspects, and challenges related to autonomous vehicle safety and trust.
Task 5: Synthesize findings from technical analysis, standards compliance, and human factors to identify key safety gaps and potential mitigation strategies.
Task 6: Develop a conceptual framework or set of guidelines for integrating technical safety standards with user expectations and cultural considerations in autonomous vehicle design.
Task 7: Finalize thesis document and prepare for academic dissemination.

Deliverables (Besides the final thesis document):

Detailed report on the safety implications of communication dependency in critical autonomous driving scenarios.
Analysis of compliance and challenges of ISO 26262 and ISO 21448 in autonomous systems.
Report summarizing user perceptions, trust, and cultural factors affecting autonomous vehicle safety.
A conceptual framework for holistic safety design in critical automotive control systems.

Evaluation Criteria:

The depth of analysis of communication-dependent safety risks.
The thoroughness in assessing standards compliance and identifying practical challenges.
The robustness and insights derived from the user studies on human factors and cultural aspects.
The practical relevance and novelty of the proposed conceptual framework.

Safety Assurance in Critical Automotive Control Systems: A Human-Centric and Standards-Compliant Approach

2025-10-01T10:33:29Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=This research proposes an in-depth study into the safety aspects of critical automotive control systems, such as those found in autonomous ve..."

{{StudentProjectTemplate
|Summary=This research proposes an in-depth study into the safety aspects of critical automotive control systems, such as those found in autonomous vehicles. It will investigate the interplay between communication dependency in specific scenarios, adherence to international safety standards.
|Keywords=Automotive Safety, ISO 26262, ISO 21448, Communication Dependency, Human Factors, Cultural Aspects, Critical Control Systems.
|Supervisor=ALEXANDRE SANTOS ROQUE
|Level=Master
|Status=Open
}}
Goals: To comprehensively study vehicle safety, for example, autonomous cars, but not limited to, by integrating technical standards (ISO 26262, ISO 21448), communication dependency, and socio-cultural factors, aiming for safer and more acceptable autonomous mobility solutions.

Research Focus and Objectives: This thesis will focus on:

- Analyzing safety implications of communication dependency in critical autonomous driving scenarios.
- Investigating challenges and applications of ISO 26262 and ISO 21448 for advanced autonomous features.
- Exploring user perceptions, trust, and cultural influences on autonomous vehicle safety and acceptance.
- Identifying and addressing gaps between technical safety implementations and human expectations/behavior.

Methodology and Implementation:
The study will involve an extensive literature review on safety standards, communication protocols, and human factors. Critical driving scenarios will be analyzed for communication-related safety risks. Current autonomous vehicle discussions need to be reviewed, also considering the recent applications of standards as ISO 26262 and ISO 21448, and others. User perspectives on safety and acceptance could be gathered via surveys or interviews. Evaluate the application of a conceptual framework to integrate technical safety with human-centric design.

Project Deliverables:
Key deliverables include a detailed analysis report on communication-dependent safety aspects, a comparative study of safety standards application, a report on user perception and cultural challenges, and a conceptual framework for integrated automotive safety design.

Main Tasks:
Task 1: Conduct a comprehensive literature review on autonomous vehicle safety, communication dependency, ISO 26262, ISO 21448, and human factors in automotive contexts.
Task 2: Analyze specific critical autonomous driving scenarios to identify safety risks stemming from communication reliance and potential failures.
Task 3: Examine how existing and emerging autonomous vehicle technologies align with, or deviate from, the requirements of ISO 26262 and ISO 21448.
Task 4: Design and execute a qualitative and/or quantitative study (surveys, interviews, focus groups) to capture user views, cultural aspects, and challenges related to autonomous vehicle safety and trust.
Task 5: Synthesize findings from technical analysis, standards compliance, and human factors to identify key safety gaps and potential mitigation strategies.
Task 6: Develop a conceptual framework or set of guidelines for integrating technical safety standards with user expectations and cultural considerations in autonomous vehicle design.
Task 7: Finalize thesis document and prepare for academic dissemination.

Deliverables (Besides the final thesis document):

Detailed report on the safety implications of communication dependency in critical autonomous driving scenarios.
Analysis of compliance and challenges of ISO 26262 and ISO 21448 in autonomous systems.
Report summarizing user perceptions, trust, and cultural factors affecting autonomous vehicle safety.
A conceptual framework for holistic safety design in critical automotive control systems.

Evaluation Criteria:

The depth of analysis of communication-dependent safety risks.
The thoroughness in assessing standards compliance and identifying practical challenges.
The robustness and insights derived from the user studies on human factors and cultural aspects.
The practical relevance and novelty of the proposed conceptual framework.

Leveraging LLM for Proactive Fault Analysis and Prediction in V2X Communication Systems

2025-10-01T10:20:03Z

Cclab:

{{StudentProjectTemplate
|Summary=This proposal outlines research into applying LLMs for advanced fault analysis and prediction within V2X communication systems. The study will involve analyzing public datasets, developing novel datasets for specific V2X scenarios.
|Keywords=V2X Communication, Fault Analysis, Fault Prediction, Automotive Systems, Data collection
|Supervisor=ALEXANDRE SANTOS ROQUE
|Level=Master
|Status=Open
}}
Goals: Design and implement an LLM-based framework to analyze and predict faults in V2X communication systems, enhancing system reliability and safety. The research aims to establish a robust methodology and knowledge base applicable to critical automotive scenarios.

Research Focus and Objectives: This thesis will explore the capabilities of LLMs in processing V2X data for identifying anomalous patterns and predicting potential failures. It will focus on:

- Investigating the applicability of various LLMs on existing public V2X datasets.
- Developing new V2X datasets that simulate critical failure modes and challenging communication environments.
- Simulating specific case studies where V2X communication integrity is paramount for automotive safety.
- Researching and evaluating existing simulation software, frameworks, and other AI approaches within this scope.

Methodology and Implementation:

- Data Collection and Generation: Sourcing public V2X datasets and generating synthetic datasets through V2X simulation tools (e.g., SUMO, Veins, OMNeT++).
- LLM Application and Fine-tuning: Applying pre-trained LLMs and potentially fine-tuning them for V2X-specific fault recognition and prediction tasks.
- Case Study Simulation: Developing and simulating critical V2X scenarios (e.g., platooning, emergency vehicle warning, intersection assistance) to test fault detection and prediction capabilities.
- Evaluation and application of specific metrics (Prediction Accuracy, False Positive Rate, Scalability, Robustness, Methodology Utility)

Project Deliverables:
Deliverables include a comprehensive literature review, study or development of V2X datasets (synthetic and annotated), LLM models analysis, simulation results for one critical case study.

Main Tasks:
Task 1: Conduct a thorough literature review on LLM applications in fault diagnosis/prediction and V2X communication systems.
Task 2: Identify suitable public V2X datasets for initial LLM training and evaluation.
Task 3: Design and implement a methodology for generating novel V2X datasets simulating critical fault scenarios.
Task 4: Select and adapt appropriate LLMs for V2X fault analysis and prediction, potentially involving transfer learning or fine-tuning.
Task 5: Develop simulation environments and case studies focusing on V2X communication criticality for automotive safety.
Task 6: Evaluate the LLM-based fault analysis and prediction system using both public and custom datasets, as well as simulated case studies.
Task 7: Propose and document a methodological framework or knowledge base for V2X fault management using LLMs.
Task 8: Finalize thesis document and prepare for dissemination (e.g., journal and conference papers).

Deliverables (Besides the final thesis document):

Curated and documented public V2X datasets.
Novel synthetic V2X datasets simulating critical fault scenarios.
Implemented and evaluated LLM models for V2X fault analysis and prediction.
Detailed simulation results and analysis for critical V2X case studies.

Leveraging LLM for Proactive Fault Analysis and Prediction in V2X Communication Systems

2025-10-01T10:19:26Z

Cclab:

{{StudentProjectTemplate
|Summary=This proposal outlines research into applying LLMs for advanced fault analysis and prediction within V2X communication systems. The study will involve analyzing public datasets, developing novel datasets for specific V2X scenarios, and simulating critical automotive case studies.
|Keywords=V2X Communication, Fault Analysis, Fault Prediction, Automotive Systems, Data collection
|Supervisor=ALEXANDRE SANTOS ROQUE
|Level=Master
|Status=Open
}}
Goals: Design and implement an LLM-based framework to analyze and predict faults in V2X communication systems, enhancing system reliability and safety. The research aims to establish a robust methodology and knowledge base applicable to critical automotive scenarios.

Research Focus and Objectives: This thesis will explore the capabilities of LLMs in processing V2X data for identifying anomalous patterns and predicting potential failures. It will focus on:

- Investigating the applicability of various LLMs on existing public V2X datasets.
- Developing new V2X datasets that simulate critical failure modes and challenging communication environments.
- Simulating specific case studies where V2X communication integrity is paramount for automotive safety.
- Researching and evaluating existing simulation software, frameworks, and other AI approaches within this scope.

Methodology and Implementation:

- Data Collection and Generation: Sourcing public V2X datasets and generating synthetic datasets through V2X simulation tools (e.g., SUMO, Veins, OMNeT++).
- LLM Application and Fine-tuning: Applying pre-trained LLMs and potentially fine-tuning them for V2X-specific fault recognition and prediction tasks.
- Case Study Simulation: Developing and simulating critical V2X scenarios (e.g., platooning, emergency vehicle warning, intersection assistance) to test fault detection and prediction capabilities.
- Evaluation and application of specific metrics (Prediction Accuracy, False Positive Rate, Scalability, Robustness, Methodology Utility)

Project Deliverables:
Deliverables include a comprehensive literature review, study or development of V2X datasets (synthetic and annotated), LLM models analysis, simulation results for one critical case study.

Main Tasks:
Task 1: Conduct a thorough literature review on LLM applications in fault diagnosis/prediction and V2X communication systems.
Task 2: Identify suitable public V2X datasets for initial LLM training and evaluation.
Task 3: Design and implement a methodology for generating novel V2X datasets simulating critical fault scenarios.
Task 4: Select and adapt appropriate LLMs for V2X fault analysis and prediction, potentially involving transfer learning or fine-tuning.
Task 5: Develop simulation environments and case studies focusing on V2X communication criticality for automotive safety.
Task 6: Evaluate the LLM-based fault analysis and prediction system using both public and custom datasets, as well as simulated case studies.
Task 7: Propose and document a methodological framework or knowledge base for V2X fault management using LLMs.
Task 8: Finalize thesis document and prepare for dissemination (e.g., journal and conference papers).

Deliverables (Besides the final thesis document):

Curated and documented public V2X datasets.
Novel synthetic V2X datasets simulating critical fault scenarios.
Implemented and evaluated LLM models for V2X fault analysis and prediction.
Detailed simulation results and analysis for critical V2X case studies.

Leveraging LLM for Proactive Fault Analysis and Prediction in V2X Communication Systems

2025-10-01T10:18:37Z

Cclab:

{{StudentProjectTemplate
|Summary=This proposal outlines research into applying LLMs for advanced fault analysis and prediction within V2X communication systems.
|Keywords=V2X Communication, Fault Analysis, Fault Prediction, Automotive Systems, Data collection
|Supervisor=ALEXANDRE SANTOS ROQUE
|Level=Master
|Status=Open
}}
Goals: Design and implement an LLM-based framework to analyze and predict faults in V2X communication systems, enhancing system reliability and safety. The research aims to establish a robust methodology and knowledge base applicable to critical automotive scenarios.

Research Focus and Objectives: This thesis will explore the capabilities of LLMs in processing V2X data for identifying anomalous patterns and predicting potential failures. It will focus on:

- Investigating the applicability of various LLMs on existing public V2X datasets.
- Developing new V2X datasets that simulate critical failure modes and challenging communication environments.
- Simulating specific case studies where V2X communication integrity is paramount for automotive safety.
- Researching and evaluating existing simulation software, frameworks, and other AI approaches within this scope.

Methodology and Implementation:

- Data Collection and Generation: Sourcing public V2X datasets and generating synthetic datasets through V2X simulation tools (e.g., SUMO, Veins, OMNeT++).
- LLM Application and Fine-tuning: Applying pre-trained LLMs and potentially fine-tuning them for V2X-specific fault recognition and prediction tasks.
- Case Study Simulation: Developing and simulating critical V2X scenarios (e.g., platooning, emergency vehicle warning, intersection assistance) to test fault detection and prediction capabilities.
- Evaluation and application of specific metrics (Prediction Accuracy, False Positive Rate, Scalability, Robustness, Methodology Utility)

Project Deliverables:
Deliverables include a comprehensive literature review, study or development of V2X datasets (synthetic and annotated), LLM models analysis, simulation results for one critical case study.

Main Tasks:
Task 1: Conduct a thorough literature review on LLM applications in fault diagnosis/prediction and V2X communication systems.
Task 2: Identify suitable public V2X datasets for initial LLM training and evaluation.
Task 3: Design and implement a methodology for generating novel V2X datasets simulating critical fault scenarios.
Task 4: Select and adapt appropriate LLMs for V2X fault analysis and prediction, potentially involving transfer learning or fine-tuning.
Task 5: Develop simulation environments and case studies focusing on V2X communication criticality for automotive safety.
Task 6: Evaluate the LLM-based fault analysis and prediction system using both public and custom datasets, as well as simulated case studies.
Task 7: Propose and document a methodological framework or knowledge base for V2X fault management using LLMs.
Task 8: Finalize thesis document and prepare for dissemination (e.g., journal and conference papers).

Deliverables (Besides the final thesis document):

Curated and documented public V2X datasets.
Novel synthetic V2X datasets simulating critical fault scenarios.
Implemented and evaluated LLM models for V2X fault analysis and prediction.
Detailed simulation results and analysis for critical V2X case studies.

Leveraging LLM for Proactive Fault Analysis and Prediction in V2X Communication Systems

2025-10-01T10:12:11Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=This proposal outlines research into applying Large Language Models (LLMs) for advanced fault analysis and prediction within Vehicle-to-Every..."

{{StudentProjectTemplate
|Summary=This proposal outlines research into applying Large Language Models (LLMs) for advanced fault analysis and prediction within Vehicle-to-Everything (V2X) communication systems. The study will involve analyzing public datasets, developing novel datasets for specific V2X scenarios, and simulating critical automotive case studies. The aim is to establish a methodological framework or knowledge base for enhancing V2X system reliability and safety through AI-driven fault management.
|Keywords=V2X Communication, Fault Analysis, Fault Prediction, Automotive Systems, Data collection
|Supervisor=ALEXANDRE SANTOS ROQUE
|Level=Master
|Status=Open
}}
Goals: Design and implement an LLM-based framework to analyze and predict faults in V2X communication systems, enhancing system reliability and safety. The research aims to establish a robust methodology and knowledge base applicable to critical automotive scenarios.

Research Focus and Objectives: This thesis will explore the capabilities of LLMs in processing V2X data for identifying anomalous patterns and predicting potential failures. It will focus on:

- Investigating the applicability of various LLMs on existing public V2X datasets.
- Developing new V2X datasets that simulate critical failure modes and challenging communication environments.
- Simulating specific case studies where V2X communication integrity is paramount for automotive safety.
- Researching and evaluating existing simulation software, frameworks, and other AI approaches within this scope.

Methodology and Implementation:

- Data Collection and Generation: Sourcing public V2X datasets and generating synthetic datasets through V2X simulation tools (e.g., SUMO, Veins, OMNeT++).
- LLM Application and Fine-tuning: Applying pre-trained LLMs and potentially fine-tuning them for V2X-specific fault recognition and prediction tasks.
- Case Study Simulation: Developing and simulating critical V2X scenarios (e.g., platooning, emergency vehicle warning, intersection assistance) to test fault detection and prediction capabilities.
- Evaluation and application of specific metrics (Prediction Accuracy, False Positive Rate, Scalability, Robustness, Methodology Utility)

Project Deliverables:
Deliverables include a comprehensive literature review, study or development of V2X datasets (synthetic and annotated), LLM models analysis, simulation results for one critical case study.

Main Tasks:
Task 1: Conduct a thorough literature review on LLM applications in fault diagnosis/prediction and V2X communication systems.
Task 2: Identify suitable public V2X datasets for initial LLM training and evaluation.
Task 3: Design and implement a methodology for generating novel V2X datasets simulating critical fault scenarios.
Task 4: Select and adapt appropriate LLMs for V2X fault analysis and prediction, potentially involving transfer learning or fine-tuning.
Task 5: Develop simulation environments and case studies focusing on V2X communication criticality for automotive safety.
Task 6: Evaluate the LLM-based fault analysis and prediction system using both public and custom datasets, as well as simulated case studies.
Task 7: Propose and document a methodological framework or knowledge base for V2X fault management using LLMs.
Task 8: Finalize thesis document and prepare for dissemination (e.g., journal and conference papers).

Deliverables (Besides the final thesis document):

Curated and documented public V2X datasets.
Novel synthetic V2X datasets simulating critical fault scenarios.
Implemented and evaluated LLM models for V2X fault analysis and prediction.
Detailed simulation results and analysis for critical V2X case studies.

Predicting Unit Behavior in Tactical Scenarios Using Deep Learning

2025-09-25T14:29:34Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=Develop and evaluate deep learning models capable of predicting future unit behavior and movement trajectories in tactical scenarios, using a..."

{{StudentProjectTemplate
|Summary=Develop and evaluate deep learning models capable of predicting future unit behavior and movement trajectories in tactical scenarios, using a combination of historical trajectory data, environmental context, and mission objectives.
|References=1. Fernandes, R., Hieb, M. R., & Costa, P. C. “Levels of Autonomy: Command and Control of Hybrid Forces”, 21st ICCRTS, 2016.

2. Dunin-Keplicz, B., & Verbrugge, R. “Teamwork in Multi-agent Systems: A formal approach”, John Wiley & Sons, 2011.

3. Alberts, D. S. “The Agility Advantage: A Survival Guide for Complex Enterprises and Endeavors”, CCRP Publication Series, 2011.
|Supervisor=EDISON PIGNATON DE FREITAS
|Level=Master
|Status=Open
}}
Problem Statement
In tactical military operations, situational awareness and timely decision-making are critical for mission success. Commanders often rely on historical movement patterns, environmental conditions, and mission objectives to anticipate the future actions of friendly and adversarial units. However, manual or rule-based predictive models fail to capture the complexity and non-linear dynamics of real-world tactical behavior.
Recent advances in deep learning — particularly Recurrent Neural Networks (RNNs), Long Short-Term Memory (LSTM) networks, and Transformers — have shown strong performance in sequence modeling, trajectory prediction, and spatiotemporal reasoning. Applying these techniques to predict the behavior and movement of units could enhance proactive decision-making in dynamic environments, improve coordination in coalition networks, and anticipate adversary maneuvers.

Project Goal
Develop and evaluate deep learning models capable of predicting future unit behavior and movement trajectories in tactical scenarios, using a combination of historical trajectory data, environmental context, and mission objectives.

Proposed Solution & Specific Tasks
1. Literature Review
- Survey existing work on trajectory prediction, spatiotemporal deep learning, and military/tactical movement modeling.
- Identify key datasets, modeling techniques, and challenges (e.g., uncertainty, adversarial behaviors, incomplete data).

2. Dataset Preparation
- Collect or simulate trajectory datasets relevant to tactical scenarios:
- Public datasets (e.g., CAIDA military mobility traces, CRAWDAD mobility datasets, OpenSky flight data).
- Synthetic datasets generated using NS-3, OMNeT++ or SUMO (for vehicle/unit movement).
- Include contextual information: terrain features, mission objectives, coalition/adversary interactions.

3. Model Development
- Implement baseline trajectory prediction models (e.g., RNN, LSTM).
- Explore transformer-based models (e.g., Temporal Fusion Transformers, Trajectory Transformers).
- Integrate environmental context embeddings (terrain, weather, obstacles) and mission constraints into the model.

4. Scenario Simulation
- Develop tactical scenarios (e.g., convoy movements, UAV swarms, adversarial pursuit/avoidance).
- Train models on historical/simulated data, then predict future positions and actions.

5. Evaluation & Analysis
- Compare performance of RNN, LSTM, and transformer-based approaches.
- Analyze predictive accuracy under different tactical conditions (dense vs. sparse units, contested vs. uncontested environments).
- Investigate uncertainty estimation (e.g., Monte Carlo dropout, Bayesian DL).

Evaluation Criteria:
- Prediction Accuracy:
Mean Squared Error (MSE), Average Displacement Error (ADE), Final Displacement Error (FDE).
Accuracy of predicting mission-relevant behaviors (e.g., retreat, advance, encirclement).

- Robustness:
Model performance with noisy or incomplete trajectory data.
Ability to generalize across different terrains/missions.

- Operational Relevance:
Improvement in situational awareness (quantified as lead-time advantage in decision-making).
Computational efficiency for deployment in constrained tactical systems.

Tools & Frameworks:
- Deep Learning Libraries: PyTorch, TensorFlow/Keras.
- Simulation Tools (for synthetic data): NS-3, OMNeT++, SUMO (for vehicle/convoy scenarios), UAVSim.

Datasets:
- CRAWDAD (military mobility traces, wireless connectivity).
- OpenSky Network (UAV/aircraft movements).
- Synthetic tactical datasets generated via simulators.

Visualization: Matplotlib, Seaborn, Plotly (trajectory plots, heatmaps).

Coalition Interoperability in Multi-Domain Trust Frameworks

2025-09-25T14:22:08Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=Develop and test mechanisms for trust interoperability across coalition forces using different trust frameworks, ensuring secure and effectiv..."

{{StudentProjectTemplate
|Summary=Develop and test mechanisms for trust interoperability across coalition forces using different trust frameworks, ensuring secure and effective collaboration in joint missions. Trust Framework Analysis
|References=1. Fernandes, R., Hieb, M. R., & Costa, P. C. “Levels of Autonomy: Command and Control of Hybrid Forces”, 21st ICCRTS, 2016.

2. Dunin-Keplicz, B., & Verbrugge, R. “Teamwork in Multi-agent Systems: A formal approach”, John Wiley & Sons, 2011.

3. Alberts, D. S. “The Agility Advantage: A Survival Guide for Complex Enterprises and Endeavors”, CCRP Publication Series, 2011.
|Supervisor=EDISON PIGNATON DE FREITAS
|Level=Master
|Status=Open
}}
Goals
Develop and test mechanisms for trust interoperability across coalition forces using different trust frameworks, ensuring secure and effective collaboration in joint missions.

Trust Framework Analysis
The study analyzes existing trust frameworks in allied military systems to identify interoperability challenges and security constraints.

Protocol Design
Translation and reconciliation protocols for trust metrics are designed to enable interoperability across diverse trust frameworks.

Simulated Coalition Testing
An interoperability layer is implemented in a simulation environment to test effectiveness in joint operations with mixed trust models.

Evaluation Criteria
The system is evaluated based on trust translation success, security preservation, and operational effectiveness in missions.

Main Tasks:
- Task 1: Study existing trust frameworks used in allied military systems.
- Task 2: Identify interoperability challenges and security constraints.
- Task 3: Design translation and reconciliation protocols for trust metrics.
- Task 4: Implement interoperability layer in a simulated coalition environment.
- Task 5: Test effectiveness in joint operations with mixed trust models.
- Task 6: Document findings and prepare thesis defense.

Deliverables (Besides the final thesis document)
- Trust interoperability protocol
- Simulation of coalition operations with heterogeneous trust systems
- Evaluation report on collaboration effectiveness and security

Evaluation Criteria
- Success rate of trust translation across domains
- Security preservation during trust reconciliation
- Operational effectiveness in coalition missions

Dynamic Trust Recalibration Based on Mission Phase and Autonomy Shifts

2025-09-25T14:19:41Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=Design and implement a mechanism for dynamic trust recalibration that adapts to evolving mission phases and changes in agent autonomy levels,..."

{{StudentProjectTemplate
|Summary=Design and implement a mechanism for dynamic trust recalibration that adapts to evolving mission phases and changes in agent autonomy levels, incorporating plan recognition and collective intention modeling.
|References=1. Fernandes, R., Hieb, M. R., & Costa, P. C. “Levels of Autonomy: Command and Control of Hybrid Forces”, 21st ICCRTS, 2016.

2. Dunin-Keplicz, B., & Verbrugge, R. “Teamwork in Multi-agent Systems: A formal approach”, John Wiley & Sons, 2011.

3. Alberts, D. S. “The Agility Advantage: A Survival Guide for Complex Enterprises and Endeavors”, CCRP Publication Series, 2011.
|Supervisor=EDISON PIGNATON DE FREITAS
|Level=Master
|Status=Open
}}
Goals
Design and implement a mechanism for dynamic trust recalibration that adapts to evolving mission phases and changes in agent autonomy levels, incorporating plan recognition and collective intention modeling.

Research Focus and Objectives
The thesis explores dynamic trust recalibration adapting to mission phases and changing agent autonomy levels.

Methodology and Implementation
Plan recognition, intention modeling, and trust logic are developed and tested in simulated multi-agent environments.

Evaluation and Success Metrics
Trust update speed, accuracy, team cohesion, and reduced misalignment define the mechanism's effectiveness.

Project Deliverables
Deliverables include recalibration algorithms, mission simulations, evaluation reports, and the thesis document.

Main Tasks:
- Task 1: Research plan recognition and intention modeling in multi-agent systems.
- Task 2: Define mission phase transitions and autonomy shift scenarios.
- Task 3: Develop trust recalibration logic based on cognitive and operational context.
- Task 4: Implement recalibration mechanism in a simulated mission environment.
- Task 5: Evaluate adaptation speed and trust accuracy during dynamic changes.
- Task 6: Finalize thesis and prepare presentation.

Deliverables (Besides the final thesis document):
- Trust recalibration algorithm
- Simulation of evolving mission scenarios
- Evaluation report on adaptation and trust accuracy

Evaluation Criteria
- Speed and accuracy of trust updates during autonomy shifts
- Effectiveness in maintaining team cohesion across mission phases
- Reduction in trust misalignment due to outdated assumptions

Conflict-free Replicated Data Type (CRDT)-based Distributed Trust Propagation in Partitioned Networks

2025-09-25T14:16:53Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=Implement a CRDT-based mechanism for distributed trust computation and conflict-free updates during network partition recovery. |References=1..."

{{StudentProjectTemplate
|Summary=Implement a CRDT-based mechanism for distributed trust computation and conflict-free updates during network partition recovery.
|References=1. Fernandes, R., Hieb, M. R., & Costa, P. C. “Levels of Autonomy: Command and Control of Hybrid Forces”, 21st ICCRTS, 2016.

2. Dunin-Keplicz, B., & Verbrugge, R. “Teamwork in Multi-agent Systems: A formal approach”, John Wiley & Sons, 2011.

3. Alberts, D. S. “The Agility Advantage: A Survival Guide for Complex Enterprises and Endeavors”, CCRP Publication Series, 2011.
|Supervisor=EDISON PIGNATON DE FREITAS
|Level=Master
|Status=Open
}}
Goals
Implement a CRDT-based mechanism for distributed trust computation and conflict-free updates during network partition recovery.

CRDTs and Trust Systems
The thesis reviews CRDT concepts and distributed trust systemsas the foundation for trust propagation.

Trust Propagation Protocol
A trust propagation protocol using CRDTs is designed forreliable trust updates in partitioned networks.

Simulation and Testing
The protocol is tested in simulated partitioned networkscenarios to measure performance and update accuracy.

Evaluation Metrics
Success rate of conflict-free updates and accuracy of trustreconciliation post-recovery are evaluated.

Main Tasks:
Task 1: Review CRDTs and distributed trust systems.
Task 2: Design trust propagation protocol using CRDTs.
Task 3: Implement protocol in a simulated partitioned network.
Task 4: Test trust updates during and after partitions.
Task 5: Measure performance degradation and recovery accuracy.
Task 6: Finalize thesis and prepare defense materials.

Deliverables (Besides the final thesis document):
- CRDT-based trust propagation protocol
- Simulation of partitioned network scenarios
- Evaluation report on update consistency and recovery

Evaluation Criteria:
- Conflict-free trust update success rate
- Performance under varying partition durations
- Accuracy of trust reconciliation post-recovery

Game-Theoretic Trust Evaluation for Objective Conflict Resolution

2025-09-25T14:04:23Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=Design and implement a trust evaluation mechanism that distinguishes between legitimate goal divergence and malicious behavior using game-the..."

{{StudentProjectTemplate
|Summary=Design and implement a trust evaluation mechanism that distinguishes between legitimate goal divergence and malicious behavior using game-theoretic principles.
|References=1. Fernandes, R., Hieb, M. R., & Costa, P. C. “Levels of Autonomy: Command and Control of Hybrid Forces”, 21st ICCRTS, 2016.

2. Dunin-Keplicz, B., & Verbrugge, R. “Teamwork in Multi-agent Systems: A formal approach”, John Wiley & Sons, 2011.

3. Alberts, D. S. “The Agility Advantage: A Survival Guide for Complex Enterprises and Endeavors”, CCRP Publication Series, 2011.
|Supervisor=EDISON PIGNATON DE FREITAS
|Level=Master
|Status=Open
}}
Goals
Design and implement a trust evaluation mechanism that distinguishes between legitimate goal divergence and malicious behavior using game-theoretic principles.

Trust Evaluation Mechanism
Designs a mechanism to distinguish legitimate goal divergence from malicious behavior using game theory.

Multi-Agent System Modeling
Models agent objectives and conflicts to analyze interactions and trust decisions through simulations.

Performance Evaluation Criteria
Evaluates precision, false trust decision reduction, and scalability across team sizes and missions.

Thesis Deliverables:
Includes algorithm, simulations, performance report, and comprehensive thesis document.

Main Tasks:
- Task 1: Study game theory in multi-agent systems and trust dilemmas.
- Task 2: Model agent objectives and potential conflicts.
- Task 3: Develop game-theoretic trust evaluation algorithm.
- Task 4: Implement and simulate agent interactions with conflicting goals.
- Task 5: Analyze trust decisions and false positive/negative rates.
-Task 6: Write thesis and prepare presentation.

Deliverables (Besides the final thesis document)
- Game-theoretic trust evaluation algorithm
- Simulation scenarios with conflicting objectives
- Performance analysis report

Evaluation Criteria
- Precision in distinguishing malicious vs. conflicting behavior.
- Reduction in false trust decisions.
- Scalability across team sizes and mission types.

Autonomy-Aware Trust Modeling in Heterogeneous IoBT - Teams

2025-09-25T12:09:28Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=Develop and evaluate a trust model that dynamically weights behavioral indicators based on agents' autonomy levels (Response Demand, Response..."

{{StudentProjectTemplate
|Summary=Develop and evaluate a trust model that dynamically weights behavioral indicators based on agents' autonomy levels (Response Demand, Response Production, Response Selection).
|References=1. Fernandes, R., Hieb, M. R., & Costa, P. C. “Levels of Autonomy: Command and Control of Hybrid Forces”, 21st ICCRTS, 2016.

2. Dunin-Keplicz, B., & Verbrugge, R. “Teamwork in Multi-agent Systems: A formal approach”, John Wiley & Sons, 2011.

3. Alberts, D. S. “The Agility Advantage: A Survival Guide for Complex Enterprises and Endeavors”, CCRP Publication Series, 2011.
|Supervisor=EDISON PIGNATON DE FREITAS
|Level=Master
|Status=Open
}}
Goals
Develop and evaluate a trust model that dynamically weights behavioral indicators based on agents' autonomy levels (Response Demand, Response Production, Response Selection).

Dynamic Trust Weighting
Trust model dynamically adjusts behavioral indicators based on agent autonomy levels to improve accuracy.

Simulation and Testing
Simulation environment tests heterogeneous agents under various autonomy configurations for model validation.

Model Evaluation Criteria
Success measured by accuracy in detecting compromised agents and robustness in network partitions.

Project Deliverables
Includes trust model specification, simulation setup, evaluation report, and thesis documentation.

Main Tasks:
Task 1: Literature review on trust modeling and HyCCo autonomy levels; define agent profiles.
Task 2: Design autonomy-aware trust metrics and behavioral indicators.
Task 3: Implement simulation environment with heterogeneous agents.
Task 4: Integrate trust model into simulation; run initial tests.
Task 5: Evaluate trust accuracy under different autonomy configurations.
Task 6: Finalize documentation and prepare thesis defense.

Deliverables (Besides the final thesis document)
- Trust model specification
- Simulation environment with agent profiles
- Evaluation report on trust accuracy

Evaluation Criteria:
- Accuracy in detecting compromised agents across autonomy levels
- Robustness under simulated network partitions
- Clarity and adaptability of trust metrics

Blockchain-Assisted Data Integrity in Eventually Consistent IoBT Systems

2025-09-25T12:04:25Z

Cclab:

{{StudentProjectTemplate
|Summary=Build a lightweight blockchain or distributed ledger tailored for IoBT that ensures post-partition reconciliation and prevents data tampering.
|References=An Efficient Lightweight Blockchain for Decentralized IoT.
https://arxiv.org/abs/2508.19219

A Scalable Blockchain Based Framework for Efficient IoT Data Management Using Lightweight Consensus.
https://www.nature.com/articles/s41598-024-58578-7.pdf

TON_IoT dataset (2021) – has telemetry and cyberattack traces for IoT/IIoT.
https://ieee-dataport.org/documents/toniot-datasets
|Supervisor=Edison Pignaton de Freitas
|Level=Master
|Status=Open
}}
Problem:
In IoBT, eventual consistency creates windows of vulnerability where adversaries can inject false or manipulated data. Ensuring integrity without requiring strong synchronization is critical.

Goal:
Build a lightweight blockchain or distributed ledger tailored for IoBT that ensures post-partition reconciliation and prevents data tampering.

Proposed Solution & Tasks:
Develop an append-only log system with cryptographic hashing and digital signatures.
Implement a lightweight blockchain consensus protocol suitable for low-power IoBT devices (e.g., Proof-of-Authority instead of PoW).
Test against false data injection attacks and partition/reintegration scenarios.
Compare performance with Conflict-Free Replicated Data Types (CRDTs).

Evaluation Criteria:
Integrity verification success rate (% of tampered data detected).
System overhead (memory, CPU, bandwidth).
Recovery time after partition rejoining.
Energy efficiency on constrained devices.

Suggested Tools & Platforms:
Blockchain Frameworks:
Hyperledger Fabric (permissioned blockchain, supports Proof-of-Authority).
Ethereum Private Testnet with Geth (can tune consensus protocols).
IOTA Tangle (DAG-based ledger) → lightweight, designed for IoT devices.

IoT Simulation:
NS-3 with DTN (Delay-Tolerant Networking) extensions → to simulate partitions and reintegration.
Cooja Simulator (in Contiki-NG) → deploy constrained IoBT nodes with blockchain overlays.

Validation Tools:
Wireshark for protocol analysis,
Log comparison for tampering detection.

Federatad Learning (FL) improving security and privacy in tactical networks

2025-09-23T14:36:56Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=Investigate and prototype how FL can improve security and privacy in tactical networks, with a focus on intrusion detection or anomaly detect..."

{{StudentProjectTemplate
|Summary=Investigate and prototype how FL can improve security and privacy in tactical networks, with a focus on intrusion detection or anomaly detection. The project will combine a literature review with a simulation-based implementation to assess feasibility.
|References=Federated Learning in Intrusion Detection: Advancements, Applications, and Future Directions
https://link.springer.com/article/10.1007/s10586-025-05325-w

Federated Learning for Anomaly Detection: A Systematic Review on Scalability, Adaptability, and Benchmarking Framework
https://www.mdpi.com/1999-5903/17/8/375
|Supervisor=Edison Pignaton de Freitas
|Level=Master
|Status=Open
}}
Project Goal:
Investigate and prototype how FL can improve security and privacy in tactical networks, with a focus on intrusion detection or anomaly detection. The project will combine a literature review with a simulation-based implementation to assess feasibility.

Proposed Solution & Specific Tasks:
Literature Review & Gap Analysis:
Survey current research on FL in military/tactical or IoBT networks.
Identify open problems in privacy, security, robustness, and efficiency of FL for tactical applications.

System Design:
Define a tactical network scenario (e.g., coalition IoBT with partitioned nodes).
Choose a security task (e.g., intrusion detection, anomaly detection, malware traffic classification).
Design a federated learning framework for this task.

Implementation:
Simulate tactical nodes using a federated learning framework such as Flower, PySyft, or TensorFlow Federated.
Integrate privacy-preserving mechanisms (e.g., differential privacy, secure aggregation).
Model adversarial attacks on FL (e.g., data poisoning, model poisoning, inference attacks).

Experiments & Evaluation:
Train baseline centralized ML vs. federated ML.
Evaluate under varying tactical conditions (network partitions, node dropouts, adversarial clients).
Compare trade-offs: accuracy, latency, bandwidth, privacy leakage, robustness to attack.

Evaluation Criteria:

Model Accuracy:
Classification performance (e.g., F1-score, ROC-AUC) compared to centralized training.

Security & Privacy:
Resistance to data/model poisoning attacks (attack success rate).
Privacy leakage quantified via inference attacks.

Efficiency:
Bandwidth usage reduction compared to centralized training.
Computational overhead on constrained devices.

Robustness:
Performance under intermittent connectivity and node dropouts.

Tools & Frameworks:

Federated Learning Frameworks:
Flower (flexible, Python-based).
TensorFlow Federated (robust, integrates with TF/Keras).
PySyft (privacy-preserving ML with secure aggregation).

Datasets:
CIC-IDS 2017 / CIC-IDS 2020 (intrusion detection).
TON_IoT dataset (2021) (IoT/IIoT telemetry with attacks — closer to tactical IoBT).

Simulation Tools (optional extension):
NS-3 or CORE to emulate tactical network constraints (bandwidth limits, node dropouts).

Simulating cyber attacks and countermeasures using Cyber Operations Research Gym (CybORG)

2025-09-23T14:32:05Z

Cclab:

{{StudentProjectTemplate
|Summary=Develop a simulation-based study of cyber attacks and countermeasures in CybORG, comparing baseline scripted defenses with reinforcement learning–based adaptive defenders to identify effective strategies for protecting networked systems.
|References=Quantitative Resilience Modeling for Autonomous Cyber Defense
https://rlj.cs.umass.edu/2025/papers/RLJ_RLC_2025_99.pdf

Interpreting Agent Behaviors in Reinforcement-Learning-Based Cyber-Battle Simulation Platforms
https://arxiv.org/html/2506.08192v1

CybORG:
https://github.com/cage-challenge/cyborg?utm_source=chatgpt.com
|Supervisor=Edison Pignaton de Freitas
|Level=Master
|Status=Open
}}
Work in partnership with the Fraunhofer FKIE Institute in Bonn, Germany

Project Goal:
Develop a simulation-based study of cyber attacks and countermeasures in CybORG, comparing baseline scripted defenses with reinforcement learning–based adaptive defenders to identify effective strategies for protecting networked systems.

Proposed Solution & Specific Tasks:

Set up CybORG environment:
Install CybORG and familiarize with its baseline scenarios (e.g., red team exploiting vulnerable hosts, blue team defending services).
Select representative scenarios (e.g., privilege escalation, lateral movement, persistence).

Model Attacks:
Implement or tune red-team agents using scripted attack strategies (e.g., scanning, brute force, exploit chaining).
Explore adaptive attackers using RL (e.g., Deep Q-Networks, PPO).

Design Defenses:
Implement baseline defenses (patching, service monitoring, port blocking).
Develop blue-team agents with reinforcement learning to dynamically choose defense actions (e.g., isolate host, restart service, deception).

Experimentation:
Run controlled experiments where red-team agents attack and blue-team agents defend.
Compare performance across different strategies (static rules vs. adaptive RL).

Analysis & Visualization:
Collect metrics on attacker success rate, defender cost, time-to-compromise, and system availability.
Visualize attack-defense dynamics over simulation episodes.

Evaluation Criteria
Defensive Effectiveness:
Reduction in attacker success rate (% of compromised hosts).
Mean time to compromise (MTTC) improvements.

Efficiency:
Resource overhead of defenses (CPU, memory, action cost in CybORG).

Adaptability:
Ability of RL-based defenders to learn effective countermeasures against novel attack patterns.

Comparative Performance:
Benchmark RL defenders vs. rule-based defenders.
Benchmark scripted attackers vs. adaptive RL attackers.

Tools & Frameworks

Simulation Environment: CybORG* (Python-based cyber operations simulator).

Reinforcement Learning: Stable Baselines3 (PPO, DQN, A2C) or Ray RLlib for scalability.

Visualization & Analysis: Matplotlib, NetworkX (for network attack graphs), TensorBoard.

Expected Contributions
A reproducible simulation framework for testing cyber attack/defense strategies in CybORG.
Empirical insights on when adaptive defense outperforms static defense.
Recommendations for deploying RL-based defenses in real-world cyber ranges.

Simulating cyber attacks and countermeasures using Cyber Operations Research Gym (CybORG)

2025-09-23T14:27:51Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=Develop a simulation-based study of cyber attacks and countermeasures in CybORG, comparing baseline scripted defenses with reinforcement lear..."

{{StudentProjectTemplate
|Summary=Develop a simulation-based study of cyber attacks and countermeasures in CybORG, comparing baseline scripted defenses with reinforcement learning–based adaptive defenders to identify effective strategies for protecting networked systems.
|References=Quantitative Resilience Modeling for Autonomous Cyber Defense
https://rlj.cs.umass.edu/2025/papers/RLJ_RLC_2025_99.pdf

Interpreting Agent Behaviors in Reinforcement-Learning-Based Cyber-Battle Simulation Platforms
https://arxiv.org/html/2506.08192v1
|Supervisor=Edison Pignaton de Freitas
|Level=Master
|Status=Open
}}
Work in partnership with the Fraunhofer FKIE Institute in Bonn, Germany

Project Goal:
Develop a simulation-based study of cyber attacks and countermeasures in CybORG, comparing baseline scripted defenses with reinforcement learning–based adaptive defenders to identify effective strategies for protecting networked systems.

Proposed Solution & Specific Tasks:

Set up CybORG environment:
Install CybORG and familiarize with its baseline scenarios (e.g., red team exploiting vulnerable hosts, blue team defending services).
Select representative scenarios (e.g., privilege escalation, lateral movement, persistence).

Model Attacks:
Implement or tune red-team agents using scripted attack strategies (e.g., scanning, brute force, exploit chaining).
Explore adaptive attackers using RL (e.g., Deep Q-Networks, PPO).

Design Defenses:
Implement baseline defenses (patching, service monitoring, port blocking).
Develop blue-team agents with reinforcement learning to dynamically choose defense actions (e.g., isolate host, restart service, deception).

Experimentation:
Run controlled experiments where red-team agents attack and blue-team agents defend.
Compare performance across different strategies (static rules vs. adaptive RL).

Analysis & Visualization:
Collect metrics on attacker success rate, defender cost, time-to-compromise, and system availability.
Visualize attack-defense dynamics over simulation episodes.

Evaluation Criteria
Defensive Effectiveness:
Reduction in attacker success rate (% of compromised hosts).
Mean time to compromise (MTTC) improvements.

Efficiency:
Resource overhead of defenses (CPU, memory, action cost in CybORG).

Adaptability:
Ability of RL-based defenders to learn effective countermeasures against novel attack patterns.

Comparative Performance:
Benchmark RL defenders vs. rule-based defenders.
Benchmark scripted attackers vs. adaptive RL attackers.

Autonomous Trust and Access Control in Coalition IoBT Networks

2025-09-23T14:18:55Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=Design a decentralized, behavior-based trust and access control system that adapts autonomously under disconnected, adversarial conditions. |..."

{{StudentProjectTemplate
|Summary=Design a decentralized, behavior-based trust and access control system that adapts autonomously under disconnected, adversarial conditions.
|References=Trust-based Blockchain Authorization for IoT
https://arxiv.org/pdf/2104.00832

Blockchain-based Decentralized Trust Management in IoT: Systems, Requirements and Challenges
https://link.springer.com/article/10.1007/s40747-023-01058-8

UAVouch
https://ieeexplore.ieee.org/document/9448085
|Supervisor=Edison Pignaton de Freitas
|Level=Master
|Status=Open
}}
Problem:
Coalition operations require dynamic trust and access control across multiple organizations. Centralized RBAC/PKI solutions fail in partitioned environments, leading to risks of insider threats and impersonation attacks.

Goal:
Design a decentralized, behavior-based trust and access control system that adapts autonomously under disconnected, adversarial conditions.

Proposed Solution & Tasks:
Implement local ABAC/RBAC policies enforced with ephemeral tokens.
Develop behavioral trust models that score nodes based on mobility patterns, resource use, and anomaly detection.
Incorporate location-validation techniques (e.g., UAVouch-style*) to detect Sybil or impersonation attacks.
Test resilience in coalition networks with different trust domains.

Evaluation Criteria:
Accuracy of malicious/insider node detection (precision, recall, F1-score).
Policy enforcement latency and overhead.
Scalability with increasing coalition size.
Success rate in preventing unauthorized access during adversarial attacks.

Suggested Tools & Platforms:
Simulation Frameworks:
OMNeT++ with INET/Veins modules → simulate coalition networks with adversarial behaviors.
NS-3 with mobility models → test trust policies with moving IoBT nodes (drones, vehicles).

Trust/Access Control Libraries:
Open Policy Agent (OPA) → ABAC/RBAC policies in distributed systems.
Keycloak → identity management (extend with ephemeral tokens).
Machine Learning Tools (for behavioral trust scoring):
Scikit-learn or TensorFlow Lite → train anomaly detectors on node behavior (bandwidth use, mobility patterns).

Datasets:
UNSW-NB15 and CIC-IDS2017 → intrusion detection datasets for anomaly detection.
Mobility traces:
CRAWDAD datasets (e.g., military mobility traces, UAV movement).
Synthetic mobility models (Random Waypoint, Gauss-Markov) in NS-3/OMNeT++.

Blockchain-Assisted Data Integrity in Eventually Consistent IoBT Systems

2025-09-23T14:01:25Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=Build a lightweight blockchain or distributed ledger tailored for IoBT that ensures post-partition reconciliation and prevents data tampering..."

{{StudentProjectTemplate
|Summary=Build a lightweight blockchain or distributed ledger tailored for IoBT that ensures post-partition reconciliation and prevents data tampering.
|References=An Efficient Lightweight Blockchain for Decentralized IoT.
https://arxiv.org/abs/2508.19219

A Scalable Blockchain Based Framework for Efficient IoT Data Management Using Lightweight Consensus.
https://www.nature.com/articles/s41598-024-58578-7.pdf

TON_IoT dataset (2021) – has telemetry and cyberattack traces for IoT/IIoT.
https://ieee-dataport.org/documents/toniot-datasets

|Supervisor=Edison Pignaton de Freitas
|Level=Master
}}
Problem:
In IoBT, eventual consistency creates windows of vulnerability where adversaries can inject false or manipulated data. Ensuring integrity without requiring strong synchronization is critical.

Goal:
Build a lightweight blockchain or distributed ledger tailored for IoBT that ensures post-partition reconciliation and prevents data tampering.

Proposed Solution & Tasks:
Develop an append-only log system with cryptographic hashing and digital signatures.
Implement a lightweight blockchain consensus protocol suitable for low-power IoBT devices (e.g., Proof-of-Authority instead of PoW).
Test against false data injection attacks and partition/reintegration scenarios.
Compare performance with Conflict-Free Replicated Data Types (CRDTs).

Evaluation Criteria:
Integrity verification success rate (% of tampered data detected).
System overhead (memory, CPU, bandwidth).
Recovery time after partition rejoining.
Energy efficiency on constrained devices.

Suggested Tools & Platforms:
Blockchain Frameworks:
Hyperledger Fabric (permissioned blockchain, supports Proof-of-Authority).
Ethereum Private Testnet with Geth (can tune consensus protocols).
IOTA Tangle (DAG-based ledger) → lightweight, designed for IoT devices.

IoT Simulation:
NS-3 with DTN (Delay-Tolerant Networking) extensions → to simulate partitions and reintegration.
Cooja Simulator (in Contiki-NG) → deploy constrained IoBT nodes with blockchain overlays.

Validation Tools:
Wireshark for protocol analysis,
Log comparison for tampering detection.

Secure Key Management for Partitioned IoBT Environments

2025-09-23T13:55:42Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=Design a decentralized and lightweight key management scheme that ensures secure communication even under network partitions. |References=A R..."

{{StudentProjectTemplate
|Summary=Design a decentralized and lightweight key management scheme that ensures secure communication even under network partitions.
|References=A Review of the Authentication Techniques for Internet of Things Devices in Smart Cities: Opportunities, Challenges, and Future Directions. https://www.mdpi.com/1424-8220/25/6/1649

Authentication in Internet of Things, Protocols, Attacks, and Open Issues: A Systematic Literature Review. https://link.springer.com/article/10.1007/s10207-023-00806-8
|Supervisor=Edison Pignaton de Freitas
|Level=Master
|Status=Open
}}
Problem:
IoBT nodes often lose connectivity with centralized Public Key Infrastructure (PKI), leaving them vulnerable to key compromise, replay attacks, and confidentiality breaches when operating offline.

Goal:
Design a decentralized and lightweight key management scheme that ensures secure communication even under network partitions.

Proposed Solution & Tasks:
Implement a RAM-only key storage system for tactical nodes (keys vanish if devices are captured).
Develop a peer-to-peer ephemeral certificate exchange system based on self-issued credentials.
Integrate elliptic-curve lightweight cryptography (e.g., Curve25519, ChaCha20) for constrained devices.
Simulate adversarial scenarios such as node capture and network jamming to test resilience.

Evaluation Criteria:
Key compromise resistance (measured by % of scenarios where captured nodes reveal useful credentials).
Cryptographic overhead (CPU and memory usage).
Communication resilience in partitioned networks (latency, delivery ratio).

Trajectory prediction algorithms for intention sharing in Micromobility

2024-10-01T12:55:22Z

Cclab: Created page with "{{StudentProjectTemplate |Summary=Self-prediction of trajectories by Vulnerable Road Users to share their intentions with vehicles. |Keywords=Prediction, vehicular networks, V..."

{{StudentProjectTemplate
|Summary=Self-prediction of trajectories by Vulnerable Road Users to share their intentions with vehicles.
|Keywords=Prediction, vehicular networks, VRU
|TimeFrame=Fall 2024 - Spring 2025
|References=[1] Elfing, Johan, and Joel Pålsson. "V2X Intention Sharing for E-bikes and E-scooters: Design and implementation of a vehicular network protocol for Vulnerable Road Users intention sharing." (2024).
[2] Croall, Ruben, and Douglas Jonsson Lundqvist. "Here I go: A prediction model for e-bike and e-scooter positioning inside a CCAM environment." (2024).
|Prerequisites=Matlab or Python, ML-methods, polynomial fitting, programming for embedded systems.
|Supervisor=Elena Haller, Amira Soliman, Oscar Amador Molina
|Level=Flexible
|Status=Open
}}
Vehicle-to-everything (V2X) describes wireless communication between a vehicle and any entity that may affect or may be affected by, the vehicle. "Here I go" project aims to connect vulnerable road users (VRUs) to V2X services by making them active participants of Future Transport System. This bachelor thesis will be focused on refining prediction algorithms of VRUs future states in their movement trajectory adopting the pipeline presented in [1,2]. These prediction algorithms should propose the optimal shape and dimensions for the reservation area needed for VRUs on the road. The current implementation uses Least Squares (LS) method and allows for improvements, for example using Deep Neural Networks (DNNs) for timeseries data. The main purpose for this bachelor project would be then to apply various prediction algorithms e.g. weighted LS, Kalman Filter, DNNs (e.g., LSTM and GRU) and choose the optimal one for data available from both simulated scenarios and real-life testing.

This project is more software related and we would use, besides synthetic data, our own gathered data and data from publicly-available datasets.

ISLAB/CAISR - User contributions [en]

Detection Unit Imperfections and CPM Reliability in Interoperable RSUs

Comparative Study on Data Abstraction Methodologies for Interoperable V2X Roadside Units (RSU)

In-Vehicle Data Acquisition for Driving Behavior Profiling and Automotive AI Applications

In-Vehicle Data Acquisition for Driving Behavior Profiling and Automotive AI Applications

Explainable GNNs for Security Verification of RISC-V Cores

Explainable GNNs for Security Verification of RISC-V Cores

Sensitivity‑Aware Hardening and Run‑Time Detection of Stealthy Weight‑Drift Trojans in ML Accelerators

Profiling ML Side-Channel on CiM for Input Reconstruction

Security Vulnerabilities in Multi-Model Computing-in-Memory Systems

Increase Data Rate over Error-prone networks

Anomaly Detection in Time Series Data Using Generative Models

Identification and Classification of Automotive Radar Interference using Data-driven Methods

Anomaly Detection in Time Series Data Using Generative Models

Identification and Classification of Automotive Radar Interference using Data-driven Methods

Coordinated Multi-Drone Pattern Formation with Crazyflie & Lighthouse

Posture Estimation for Motorcycle Simulator Riders Using Remote Sensing and IMU-Based Systems

Posture Estimation for Motorcycle Riders Using Multi-IMU Systems and Video-Based Ground Truth

Service Centric Mobility Management in LEO Constellations

Data muling services over a constellation of aircraft

Segment Routing based on Geographic Checkpoints

Real-time flow compression for TCP and QUIC at the network edge

Increase Data Rate over Error-prone networks

Segment-based coordination of Congestion Control

Non-Invasive Safety Monitoring via Integrated Sensing and Communication

Non-Invasive Safety Monitoring via Integrated Sensing and Communication

AI-Driven Semantic Encoding for Efficient Communication

Resilient Recovery: A Byzantine Fault Tolerance (BFT) Framework for Tactical Mesh Networks

Robotics for Urban Traffic Support: Design, Prototyping, and Intelligent Interaction in Smart Cities

Robotics for Urban Traffic Support: Design, Prototyping, and Intelligent Interaction in Smart Cities

Leveraging LLM for Proactive Fault Analysis and Prediction in V2X Communication Systems

Safety Assurance in Critical Automotive Control Systems: A Human-Centric and Standards-Compliant Approach

Safety Assurance in Critical Automotive Control Systems: A Human-Centric and Standards-Compliant Approach

Leveraging LLM for Proactive Fault Analysis and Prediction in V2X Communication Systems

Leveraging LLM for Proactive Fault Analysis and Prediction in V2X Communication Systems

Leveraging LLM for Proactive Fault Analysis and Prediction in V2X Communication Systems

Leveraging LLM for Proactive Fault Analysis and Prediction in V2X Communication Systems

Predicting Unit Behavior in Tactical Scenarios Using Deep Learning

Coalition Interoperability in Multi-Domain Trust Frameworks​

Dynamic Trust Recalibration Based on Mission Phase and Autonomy Shifts

Conflict-free Replicated Data Type (CRDT)-based Distributed Trust Propagation in Partitioned Networks

Game-Theoretic Trust Evaluation for Objective Conflict Resolution

Autonomy-Aware Trust Modeling in Heterogeneous IoBT - Teams​

Blockchain-Assisted Data Integrity in Eventually Consistent IoBT Systems

Federatad Learning (FL) improving security and privacy in tactical networks

Simulating cyber attacks and countermeasures using Cyber Operations Research Gym (CybORG)

Simulating cyber attacks and countermeasures using Cyber Operations Research Gym (CybORG)

Autonomous Trust and Access Control in Coalition IoBT Networks

Blockchain-Assisted Data Integrity in Eventually Consistent IoBT Systems

Secure Key Management for Partitioned IoBT Environments

Trajectory prediction algorithms for intention sharing in Micromobility

Coalition Interoperability in Multi-Domain Trust Frameworks

Autonomy-Aware Trust Modeling in Heterogeneous IoBT - Teams