Human-like motion planning for robotic arm system

We explore the problem of human-like motion planning for robotic arms based on researches in biophysics, robotics and motion planning. Two common manipulation tasks, reaching point movement in obstacle free environment and reaching point movement through a narrow passage, are separately studied. First, we proposed a hypothesis named "Target Arm Pose" (TAP) to interpret the natural motion of human arm. Second, based on TAPs and the minimum jerk model, we develop a new robotic inverse kinematic algorithm in joint jerk level to conduct the reaching point movement in obstacle free environment. Furthermore, based on the new algorithm we utilize the redundancy of the robot to avoid joint velocity limits. Next, we use existing Bi-RRT algorithm combined with TAPs to implement the reaching point movement through a narrow passage. Finally, comparison between these two tasks and two algorithms is made to give a systematic analysis on the human-like motion planning problem.