Narrow passage identification based robot path planning

Probabilistic Sampling based Path Planning Method presents much lower performance for the case of narrow passage existing in robot configuration space. This paper uses Randomized Star Builder to identify the narrow passage in working space, based on which, to configure the extension point in the narrow passage. We propose two types of new Triple-RRT algorithms to assure rapidly explore open configuration space and effectively acquire the connection in the narrow passage. Simulations of rigid robots with multiple DoFs in 2D and 3D working space verify the validation of the proposed algorithms.