In order to solve the local path planning of self-driving car in the structured road environment, an improved path planning algorithm named Regional-Sampling RRT (RS-RRT) algorithm was proposed for obstacle avoidance conditions. Gaussian distribution sampling and local biasing sampling were integrated to improve the search efficiency in the sampling phase. In the expansion phase, considering the actual size of the vehicle and obstacles, combined with the goal of safety and comfort, the separating axis theorem (SAT) method and vehicle dynamics were used to detect the collision among vehicle and surrounding obstacles in real time. In the post-processing stage, the driver’s driving consensus and path smoothing algorithm were combined to correct the planning path. In order to track the generated path, the MPC tracking algorithm was designed based on the Four-Wheel-Independent Electric Vehicle (FWIEV) model. The co-simulation software platform of CarSim and MATLAB/Simulink was employed to verify the effectiveness and feasibility of the path planning and tracking algorithm. The results show that compared with basic RRT and Goal-biasing RRT, the proposed RS-RRT algorithm has advantages in terms of number of nodes, path length and running time. The generated path can meet the FWIEV dynamics and path tracking requirements.
Efficient Local Path Planning Algorithm Using Artificial Potential Field Supported by Augmented Reality
