Résumé

The increasing complexity of urban traffic systems has made congestion a significant challenge, leading to severe economic, environmental, and social impacts. Intelligent Transportation Systems (ITSs) have emerged as a promising solution to mitigate these challenges by enabling dynamic traffic management. However, the reliability of data within ITSs represents an increasingly significant challenge. The introduction of erroneous data by defective or malicious sensors can lead to malfunctions or intentional disruptions of the system. In this context, trust management models assume a crucial importance. Most of the existing trust models propose vehicle-centric approaches. How ever, the high mobility and dynamic nature of the ITS environments affects the stability and scalabity of such systems. Accordingly, proposing novel trust models designed specifically for ITSs to enhance the accuracy, security, scalability and stabilty of traffic information dissemination constitutes the overall goal of this thesis. First, we propose a generic architecture for a trust framework, leveraging Smart Road Signs (SRSs). The conception of this architecture is built upon the output of a deep investigation of the state of the art. This framework is, then, devel oped to propose two novel trust models. The first model, considers the contex tual information and multi-source data aggregation to assess the trustworthiness of reported traffic events and the different nodes of the network. Additionally, the model applys a bi-level trust evaluating combining Bayesian Inference and a dynamic weighted sum approach. Furthermore, we propose a predictive-based Baysian Inference to enhance the accuracy of trust evaluation. Thereafter, we propose a communication trust model to complement the previ ous contribution. This model introduces a self-organizing trust model to track the SRSs’ behaviours and establishes stable environments usinsg a fuzzy-based Dempster Shafer Theory (DST). The main objective of this model is to ensures that the systems remains operational even in hostile environments, by mitigating the inherent single point of failure vulnerability characteristic of centralized net work architectures. The proposed models were validated through simulations, showing their effec tiveness in identifying malicious nodes and mitigating erroneous traffic reports. The results demonstrate that the integration of trust management systems in ITSs significantly improves data reliability, enabling more effective congestion man agement and increasing the overall robustness of transportation infrastructures. Keywords: ITS,trustmanagementmodels, requirements, context-awareness, multi-source aggregation, self-organizing trust