Mehdi Sookhak

Abdolazim Rezaei, Mehdi Sookhak, Ahmad Patooghy

Connected and Autonomous Vehicles (CAVs) rely on a range of devices that often process privacy-sensitive data. Among these, roadside units play a critical role particularly through the use of AI-equipped (AIE) cameras for applications such as violation detection. However, the privacy risks associated with captured imagery remain a major concern, as such data can be misused for identity theft, profiling, or unauthorized commercial purposes. While traditional techniques such as face blurring and obfuscation have been applied to mitigate privacy risks, individual privacy remains at risk, as individuals can still be tracked using other features such as their clothing. This paper introduces a novel privacy-preserving framework that leverages feedback-based reinforcement learning (RL) and vision-language models (VLMs) to protect sensitive visual information captured by AIE cameras. The main idea is to convert images into semantically equivalent textual descriptions, ensuring that scene-relevant information is retained while visual privacy is preserved. A hierarchical RL strategy is employed to iteratively refine the generated text, enhancing both semantic accuracy and privacy. Evaluation results demonstrate significant improvements in both privacy protection and textual quality, with the Unique Word Count increasing by approximately 77\% and Detail Density by around 50\% compared to existing approaches.

Abdolazim Rezaei, Mehdi Sookhak

This study introduces PEFT-DML, a parameter-efficient deep metric learning framework for robust multi-modal 3D object detection in autonomous driving. Unlike conventional models that assume fixed sensor availability, PEFT-DML maps diverse modalities (LiDAR, radar, camera, IMU, GNSS) into a shared latent space, enabling reliable detection even under sensor dropout or unseen modality class combinations. By integrating Low-Rank Adaptation (LoRA) and adapter layers, PEFT-DML achieves significant training efficiency while enhancing robustness to fast motion, weather variability, and domain shifts. Experiments on benchmarks nuScenes demonstrate superior accuracy.

Ndagijimana Cyprien, Mehdi Sookhak, Hosein Zarini et al.

Joint deployment of unmanned aerial vehicles (UAVs) and unmanned ground vehicles (UGVs) has been shown to be an effective method to establish communications in areas affected by disasters. However, ensuring good Quality of Services (QoS) while using as few UAVs as possible also requires optimal positioning and trajectory planning for UAVs and UGVs. This paper proposes a joint UAV-UGV-based positioning and trajectory planning framework for UAVs and UGVs deployment that guarantees optimal QoS for ground users. To model the UGVs' mobility, we introduce a road graph, which directs their movement along valid road segments and adheres to the road network constraints. To solve the sum rate optimization problem, we reformulate the problem as a Markov Decision Process (MDP) and propose a novel asynchronous Advantage Actor Critic (A3C) incorporated with meta-learning for rapid adaptation to new environments and dynamic conditions. Numerical results demonstrate that our proposed Meta-A3C approach outperforms A3C and DDPG, delivering 13.1\% higher throughput and 49\% faster execution while meeting the QoS requirements.

Abdolazim Rezaei, Mehdi Sookhak, Mahboobeh Haghparast

The proliferation of AI-powered cameras in Intelligent Transportation Systems (ITS) creates a severe conflict between the need for rich visual data and the right to privacy. Existing privacy-preserving methods, such as blurring or encryption, are often insufficient due to creating an undesirable trade-off where either privacy is compromised against advanced reconstruction attacks or data utility is critically degraded. To resolve this challenge, we propose RL-MoE, a novel framework that transforms sensitive visual data into privacy-preserving textual descriptions, eliminating the need for direct image transmission. RL-MoE uniquely combines a Mixture-of-Experts (MoE) architecture for nuanced, multi-aspect scene decomposition with a Reinforcement Learning (RL) agent that optimizes the generated text for a dual objective of semantic accuracy and privacy preservation. Extensive experiments demonstrate that RL-MoE provides superior privacy protection, reducing the success rate of replay attacks to just 9.4\% on the CFP-FP dataset, while simultaneously generating richer textual content than baseline methods. Our work provides a practical and scalable solution for building trustworthy AI systems in privacy-sensitive domains, paving the way for more secure smart city and autonomous vehicle networks.

Nabil Abdelaziz Ferhat Taleb, Abdolazim Rezaei, Raj Atulkumar Patel et al.

Large Language Models (LLMs) offer significant promise for intelligent traffic management; however, current chain-based systems like TrafficGPT are hindered by sequential task execution, high token usage, and poor scalability, making them inefficient for complex, real-world scenarios. To address these limitations, we propose GraphTrafficGPT, a novel graph-based architecture, which fundamentally redesigns the task coordination process for LLM-driven traffic applications. GraphTrafficGPT represents tasks and their dependencies as nodes and edges in a directed graph, enabling efficient parallel execution and dynamic resource allocation. The main idea behind the proposed model is a Brain Agent that decomposes user queries, constructs optimized dependency graphs, and coordinates a network of specialized agents for data retrieval, analysis, visualization, and simulation. By introducing advanced context-aware token management and supporting concurrent multi-query processing, the proposed architecture handles interdependent tasks typical of modern urban mobility environments. Experimental results demonstrate that GraphTrafficGPT reduces token consumption by 50.2% and average response latency by 19.0% compared to TrafficGPT, while supporting simultaneous multi-query execution with up to 23.0% improvement in efficiency.

Nader Sohrabi Safa, Carsten Maple, Steve Furnell et al.

Previous studies show that information security breaches and privacy violations are important issues for organisations and people. It is acknowledged that decreasing the risk in this domain requires consideration of the technological aspects of information security alongside human aspects. Employees intentionally or unintentionally account for a significant portion of the threats to information assets in organisations. This research presents a novel conceptual framework to mitigate the risk of insiders using deterrence and prevention approaches. Deterrence factors discourage employees from engaging in information security misbehaviour in organisations, and situational crime prevention factors encourage them to prevent information security misconduct. Our findings show that perceived sanctions certainty and severity significantly influence individuals' attitudes and deter them from information security misconduct. In addition, the output revealed that increasing the effort, risk and reducing the reward (benefits of crime) influence the employees' attitudes towards prevent information security misbehaviour. However, removing excuses and reducing provocations do not significantly influence individuals' attitudes towards prevent information security misconduct. Finally, the output of the data analysis also showed that subjective norms, perceived behavioural control and attitude influence individuals' intentions, and, ultimately, their behaviour towards avoiding information security misbehaviour.