A Search Strategy for Motion Planning of Unmanned Crawler Crane

Abstract

The motion planning for unmanned crawler crane (UCC) whose initial and goal described accurately in workspace was challenging. The difficulty dues to the fact that the initial and goal produce low-dimensional self-motion manifolds, which rendered the search of planners bypassing the self-motion manifolds. In this study, a new concept of the neighbor-hoods of self-motion manifold was introduced, and the corresponding search strategy of bias was developed towards extending the nodes within the neighborhoods of self-motion manifolds so as to decrease the probability of the tree nodes walking by the self-motion manifolds. Then this strategy was used to improve the performance of BiMRRTs proposed in the previous study. Finally, several simulation experiments were implemented to demonstrate the effectiveness of the proposed search strategy of the neighborhood of self-motion manifold. The results showed that the proposed search strategy was able to dramatically decrease the planning time and the path length simultaneously.