An Optimal Two-Stage Path Planne

RM) planner applied to the robot with multiple degrees of freedom moving th static obstacles is proved to be efficient. But it is clumsy to environment ed path is usually not optimal and not smooth enough. In order to make the re flexible, this paper integrates the PRM planner with simulated annealing -stage path planner: at first a path is generated by PRM planner, then this ed by SA. If the original path becomes infeasible because of environment can also be obtained from the original path by SA. The simulations are efficiency of the proposed planner. g; Probabilistic Roadmap; Simulated Annealing basic and long standing problem of robotics, which is required to seek for a brings the robot from its initial configuration to the specified goal ce, a robot may have many degrees of freedom (DOF), which is a great th planning. Many deterministic path planners especially the complete pplied for robot with many DOF are short of efficiency if not infeasible at tation resource. Therefore the researchers turn to the probabilistic complete such successful planners are the randomized path planner (RPP) and r (PPP) [21]. The former is actually a potential field planner, which escapes rforming random walks [2]. The latter finds the required path according to a that is constructed by randomly selecting configurations from the free space) and interconnecting certain pairs by simple local powerful planners. e roadmap can be a graph [10] or a tree [15]. Two popular PPPs are PRM lanner [13,15]. The theoretic analysis of PRM is present in [11,14], and it is nner is probabilistic complete. by NSFC Projects (No. 60334020, 60440420130, 60475030, and 60575047), MOST Projects (No. 100), and the Outstanding Overseas Chinese Scholars Fund of Chinese Academy of Sciences (No.