Methods for collision-free arm teleoperation in clutter using constraints from 3D sensor data

Abstract

We introduce CAT, a constraint-aware teleoperation method that can track continuously updating 6-DOF end-effector goals while avoiding environment collisions, self-collisions, and joint limits. Our method uses sequential quadratic programming to generate motion trajectories that obey kinematic constraints while attempting to reduce the distance to the goal with each step. Environment models are created and updated at run-time using a commodity depth camera. We compare our method to three additional teleoperation strategies, based on global motion planning, inverse kinematics, and Jacobian-transpose control. Our analysis, using a real robot in a variety of scenes, highlights the strengths of each method, and shows that the CAT method we introduce performs well over a wide range of scenarios.