A Cooperative Positioning Enhancement for Self-Powered Vehicular Networks With V2I Trustworthy Data Sharing

The emergence of connected vehicles (CVs) and cooperative vehicle-infrastructure systems (CVISs) have paved the way for new innovative prospects for improving the safety and efficiency of transportation systems. Unleashing the communication and positioning capabilities of CVs and roadside units' (RSUs) storage and computing capabilities, the CVIS can enhance the positioning signal coverage and improve the positioning accuracy of the vehicles with low positioning accuracy by sharing the positioning error correction model. In particular, based on the particle swarm optimization (PSO) algorithm, we first design an improved deep neural network (DNN) to train the positioning error model. Second, the selection crossover factors of the genetic algorithm are used to optimize the network. Then, based on the consensus algorithms, the positioning error model is shared and stored in the blockchain network to ensure the security of vehicles and RSUs that provide positioning correction information. Finally, the trained network is used to predict and correct the positioning errors of other vehicles with low positioning accuracy. Additionally, the proposed method's performance and effectiveness in terms of accuracy, timeliness, and security are verified in different scenarios.