FC-RRT*: A modified RRT* with rapid convergence in complex environments

The Rapidly-exploring Random Tree algorithm (RRT) is currently the preferred algorithm for solving motion planning problems. It enables fast path generation on a large scale with high-latitude spatial species. RRT* as the optimal variant provides an asymptotically optimal solution and inspires the F-RRT* algorithm, which significantly reduces the path cost but performs poorly in complex environments. A modified RRT* algorithm is proposed in this article, FC-RRT*, utilizing the prior knowledge of the mission to expand the path tree at the start point and goal point bidirectionally. Besides, based on F-RRT*, an obstacle proximity node is created to reduce the path cost while modifying its Rewire procedure by including this node as a potential parent node. In this paper, a numerical simulation is performed to compare ARA*, RRT*, F-RRT*, and FC-RRT*, and the advantages of the FC-RRT* algorithm in complex environments is demonstrated.