作者: L. Kavraki , J.-C. Latombe
DOI: 10.1109/ROBOT.1994.350966
关键词:
摘要: This paper presents a new approach to path planning for robots with many degrees of freedom (DOF) operating in known static environments. The consists preprocessing and stage. Preprocessing, which is done only once given environment, generates network randomly, but properly selected, collision-free configurations (nodes). Planning then connects any initial final the robot two nodes computes through between these nodes. Experiments show that after paying cost (on order hundreds seconds), extremely fast fraction second difficult examples involving 10-DOF robot). particularly attractive many-DOF have perform successive point-to-point motions same environment. >