Deep Echo State Q-Network Aided Trust Sharing Provisioning for Internet of Vehicle

Internet of Vehicles is an intelligent transportation network system with complex and highly distributed characteristics. The ability to share information between vehicles, pedestrians, infrastructure, and the cloud has created rich opportunities for deploying various transportation applications and services for Intelligent Transportation System. However, the problem of sharing the results of node trust evaluation has been seriously neglected in existing studies. Given this situation, this paper proposes a distributed trust sharing mechanism based on Reservoir Computing. Considering the limited storage and computing capacity of vehicles, we set up the model as as Partially Observable Markov Decision Process. Besides we novelly combined Echo state network and Reinforcement Learning to support vehicles to share trust information by predicting future driving directions. The current vehicle trust evaluation results are incorporated into a new parameter besides traffic accidents and traffic flow, which protects the security of vehicles to the greatest extent. Finally, the performance of the proposed algorithm is compared with those of Multilayer Perceptron-based DQN and LSTM-based DQN through simulation, which proves that the proposed algorithm can converge to a better solution and has a faster convergence speed, which verifies the feasibility and effectiveness of the method.