Mobile Robot Path Planning Based on Improved RRT Algorithm

In complex and narrow environments, mobile robots usually face significant challenges when performing path planning tasks, as evidenced by long search times and low success rates. This paper proposes an improved RRT path planning algorithm. The algorithm is designed with concentric semi-elliptical distribution to divide the sampling region into four sub-regions. It dynamically adjusts regional sampling and distance-target bias strategies based on graded bias probabilities, and combines with the step selection function to guide the random tree to converge to the target region faster. The algorithm utilizes an obstacle avoidance strategy, designs a multi-stage expansion method. By analyzing the positional characteristics of the expansion points, the obstacles are classified into wall obstacles, narrow passages, and concave obstacles, and the sampling direction is determined to enhance obstacle avoidance performance. Simulation results show that compared with other algorithms, the improved RRT algorithm performs better in terms of path length, number of sampling points and success rate.