An Adaptive RRT Algorithm Based on Narrow Passage Recognition for Assembly Path Planning

Rapidly-exploring Random Trees (RRT) algorithm is classic sample-based algorithm for path planning. However, complex products often have compact structure, leading to narrow passage in assembly path planning, which makes the success rate and efficiency of RRT deteriorate significantly. Thus, an adaptive RRT algorithm based on narrow passage recognition is proposed. Firstly, the algorithm divides random tree nodes into neighborhoods with hyper-spheres and then dynamically identifies narrow passage with spatial indicators and temporal indicators. Spatial indicators characterize the narrowness by analyzing nodes within the hyper-sphere; temporal indicators characterize the planning state by analyzing the changes of spatial indicators within different growth cycles. Secondly, to improve performance in narrow passage, a Gaussian sampling method based on spatial indicators are proposed to modify the sampling probability density distribution. The proposed methods are verified by a test scenario in simulation environment and a practical application of assembling smart phone parts, and the results show that the proposed algorithm is feasible and effective.