Force Control of a 1-DoF Cable Robot Using Deep Reinforcement Learning

This paper presents the force control of a single module of a cable robot based on a Deep Reinforcement Learning approach. This can be used for an underlying force control of a cable robot. Parameter studies of the actor and critic networks and the reward function are also presented. The controller is tested in a simulation as well as on the physical system. For the experiments on the physical system, the controller is compared to an expert tuned PID controller.