OBSTACLE AVOIDANCE PATH PLANNING FOR INTELLIGENT VEHICLES BASED ON SPARROW POTENTIAL FIELD IN MULTI-TYPE SCENARIOS

To overcome the issue of unreachable targets and local optima in traditional artificial potential fields in multi-type scenarios, this paper introduces a method for obstacle avoidance path planning for intelligent vehicles based on the sparrow potential field (SPF). First, by integrating gravity and repulsion adjustment factors into the traditional artificial potential field, we propose a new intermediate potential field and target repulsive potential field. The resulting potential field is then optimized through the vehicle’s heading angle to resolve issues present in structured scenes. Second, we propose an adaptive velocity function and consider dynamic constraints in path planning. Next, we combine the improved artificial potential field with the sparrow search algorithm to resolve local path optimization problems in unstructured scenarios. Finally, simulation experiments are conducted using Simulink and Carsim co-simulation platform. The results show that in the unstructured scenario, the evaluation function score of SPF algorithm is the best, and the number of algorithm iterations is reduced by about half on average. In a structured scenario, the maximum lateral acceleration of the path planned by the SPF algorithm is generally reduced by about 0.1 g, and the average front wheel angle is reduced by about 2.3%.