CONTROL OF A CLIMBING ROBOT USING REAL-TIME CONVEX OPTIMIZATION

This paper presents a controller for a free-climbing robot. Given a pre-planned path, a loop is closed around desired robot chassis position to generate cartesian feedback forces. A convex optimization problem is then solved in real-time to find a torque input to the robot that will achieve these feedback forces while not exceeding torque limits or violating friction constraints. The effectiveness of this controller is demonstrated both in simulation and with experimental hardware.