Unlike Trust in Social Science, in which interactions between humans are measured, thanks to the integration of numerous smart devices, Trust in IoT security focuses more on interactions between nodes. Moreover, As IoT nodes can somehow benefit from ‘Group’/’Community’ since they form by similar interests or functionalities, the assessment of Group-Individual and Inter-Individual Trust is also important. However, handling limitations brought by potential threats

(e.g., trust-related attack and BAR Model) and inherent vulnerability due to Trust Management (TM) architecture remains challenging. This thesis investigates Trust from three perspectives in the Service-Oriented Internet of Things (SO-IoT): Inter-Group Trust, Group-Individual Trust, and Inter-Individual Trust. Firstly, a role-based dynamic model is developed to assess intra- and inter-community(group), enhancing service-oriented activities and addressing security issues within and between communities. A locally centralized four-phase approach is employed, focusing on countermeasures against attacks on services within the community. Additionally, a three-phase mechanism is devised to measure cooperativeness between communities. An implementation based on the ROS 2 system was implemented to analyze the performance of the proposed model based on the preliminary results. Secondly, to address misbehavior in SO-IoT in terms of Inter-Individual trust, a Stochastic Bayesian Game (SBG) is introduced. This game theoretical model considers the heterogeneity of IoT nodes and complex behavioral schemes of service providers, encouraging cooperation and penalizing malicious strategical actions. Lastly, the work of assessing the Trust of V2X messages in IoV demonstrates the possibility of implementing Trust Management in a concrete IoT environment.

Keywords: Trust Management, IoT security, Trust modeling, misbehavior detection, Service-Oriented IoT

• Ye-Qiong Song, Université de Lorraine
• Riahi Arbia, Cylab, CISS, Royal Military Academy, Belgium