Path planning algorithm based on Gb informed RRT* with heuristic bias
Siyu Luo, Shirong Liu, Botao Zhang, Chaoliang Zhong · 2017
Recently, rapidly-exploring random trees(RRT) is widely used in path planning for its nature of single-query. The optimized algorithm RRT∗ extends RRT algorithm to find the optimal path, but it needs to search every state from the initial state to the global scope asymptotically. This method is not only inefficient, but also contrary to the single-query of RRT. In this paper, a new variant of RRT∗-Gb informed RRT∗ is presented. The goal biasing is used in the algorithm to guide the search to the goal. Once an initial path is found, and an ellipsoidal subset in the work space can be constructed by the path for the refine planning. The simulations in 2D and 3D environments show that this algorithm can find an optimal path efficiently.