A Game Theory approach for communication security and safety assurance in cyber-physical systems with Reputation and Trust-based mechanisms

Cyber-physical systems' security and safety assurance is a challenging research problem for Smart City concept development. Technical faults or malicious attacks over communication between its elements can jeopardize the whole system and its users. Reputation systems implementation is an effective measure to detect such malicious agents. Each agent in the group has its indicator, which reflects how trustworthy it is to the other agents. However, in the scenario when it is not possible to calculate the Reputation indicator based on objective characteristics, malicious or defective agents can negatively affect the system's performance. In this paper, we propose an approach based on Game Theory to address the Reputation and Trust initial values calculation challenge. We introduced a mixed strategies game concept and a probability indicator. The possible outcomes of using different strategies by the system agents are represented with a payoff matrix. To evaluate the approach effectiveness, an empirical study using a software simulation environment was conducted. As a Cyber-physical system implementation scenario, we considered an intersection management system with a group of unmanned autonomous vehicles, the aim of which is to perform conflict-free optimal intersection traversal. To simulate the attack scenario, some vehicles were able to transmit incorrect data to other traffic participants. The obtained results showed that the Game Theory approach allowed us to increase the number of detected intruders compared to the conventional Reputation and Trust model.