The Optimal Strategy of an Autonomous Vehicle by Analyzing Not Only the External Factors, But Also the Internal Factors of the Vehicle

Abstract

Self-driving cars will need algorithms that can cope with dynamic and complicated urban junctions. For self-driving cars to be dependable, several challenges must be solved, including as correctly recognising other drivers' routes and balancing safety and efficiency while interacting with other vehicles. Based on conflict resolution theory, we created a vehicle tactical decision-making model to predict the paths of oncoming vehicles. The method presented here may help self-driving cars negotiate junctions safely. Data collected at intersections using subgrade sensors and a retrofit autonomous car were used to build Gaussian process regression models to predict incoming vehicle trajectories. Next, a decision-making approach for MOPs based on the efficient intersection conflict resolution theory was put forward. After that, a nondominated sorting genetic algorithm (NSGA-II) and a deep deterministic policy gradient are used to compare and choose the most effective steps (DDPG). Finally, Matlab/Simulink and PreScan were used to create a simulation and verification platform. Simulations were used to confirm the tactical decision-making model's accuracy and dependability. As a result of its superior performance in solving the MOP model, DDPG lays the theoretical groundwork for future studies on decision-making in a complex and uncertain junction environment.