Study on real time autonomous collision avoidance in telerobotic systems

In conventional telerobotic systems, the robot controller fulfills the motion control and leaves the trajectory producing to the human operator. This makes the teleoperation a weighty burden to human operators, especially in cases of unknown environment, large communication delay and limited bandwidth. In this paper a simple real time collision avoidance (CA) approach is presented to provide CA function to the robot motion controller by adding a collision detection module and a trajectory adjustment module to a RMRC (resolved motion rate control) controller. This approach is unique threefold. First, CA can be easily implemented based on any motion controller with joint velocity servo loop. Second, CA action is taken place in joint space without mapping obstacles into joint space from Cartesian space. Third, biology-inspired skills of CA, slowdown/speedup action and bypassing action are adopted in joint velocity adjustment algorithm that can effectively realize stable CA function while tracking the predefined trajectory.