Path Planning for Mobile Robot Using Improved Adaptive Rapidly-exploring Random Tree

Path planning based on RRT has achieved great achievements in a complex and high dimensional environment. RRT algorithms require the user to choose an appropriate stepsize and bias probability before path planning, however, selecting algorithm parameters suitable for various environments is not only difficult but also time consuming. An improved adaptive RRT algorithm is proposed in this paper and has two notable features. Firstly, it can automatically determine the initial range of stepsize and bias probability according to the relative complexity of the robots working environment, which is ever highly problem dependent and time-consuming process. Secondly, it can automatically adjust these two parameters on the basis of the collision detection result as the iteration continue. Various numerical experiments are demonstrated and validated to illustrate the effectiveness and obvious advantages of the proposed algorithm over the basic RRT algorithm under different environmental conditions.