Path planning in dynamic environments based on the improved Rapidly-exploring random tree and dynamic window approach

Path planning in dynamic environments holds paramount significance within the realm of mobile robotics. The Rapidly-Exploring Random Tree (RRT) algorithm stands as one of the most extensively employed path planning algorithms, distinguished by its probabilistic completeness. However, the algorithm suffers from the long time to obtain the initial solution and the blindness of the expansion. To address these shortcomings, an improved RRT algorithm based on dynamic goal-biased sampling method (Dyn-RRT) is proposed to enhance the algorithm’s orientation and reduce the time obtaining the initial solution. To bolster the algorithm’s utility in dynamic settings, Dynamic Window Approach (DWA) is integrated into the Dyn-RRT algorithm to enhance adaptability to the dynamic milieu. In addition, simulation experiments show that proposed dynamic goal-biased sampling method is also applicable to other RRT series algorithms and can greatly reduce the time obtaining the initial solution.