Path planning for manipulators based on the planar constraint RRT* algorithm

Path planning is crucial for automatic measurement to ensure a collision-free process for manipulators. However, the more complex the measurement environment, the more complex the path planning scheme is often required in order to meet the above requirements. To overcome this problem, a planar constraint RRT* method (PC-RRT*) is proposed to limit nodes to a specific plane, reduce the blindness of RRT*, and smooth paths. The measurement space model is built according to the position of the manipulator, the obstacle environment, the starting point, and the ending point. A set of equiangular interval collinear planes is built by utilizing the starting and ending line as the central axis. A local path on each plane is planned by extended rapidly exploring random tree strategy. By optimizing the planned local path set, the high-quality measurement path will be solved. Both numerical simulation and experimental analysis are carried out to verify the effectiveness of the PC-RRT* method. The experimental results show that the average path lengths (APL) are 216.03 mm for PC-RRT*, 346.83 mm for RRT, 241.93 mm for RRT*, and 383.6 mm for Q-learning (QL), the average number of nodes is 9.5 for PC-RRT*, 11.12 for RRT, 11.6 for RRT*, and 18.92 for QL. The suggested PC-RRT* has an 37.71% improvement for APL and 14.57% improvement for path nodes compared to the RRT algorithm. PC-RRT* has a 10.71% improvement for APL and 18.1% improvement for path nodes compared to the RRT* algorithm. Additionally, PC-RRT* achieves a 43.68% improvement for APL and a 49.79% improvement for path nodes compared to the QL algorithm. In all, the PC-RRT* method is superior to other traditional methods.