Improving the Positioning Accuracy of Robotic Manipulators Subject to Base Oscillations

Abstract. Several robotic system applications require manipulators to carry heavy payloads while operating from moving vehicles. These manipulators are mounted on moving/compliant bases, introducing additional degrees of freedom that are not actuated, increasing the complexity of the dynamic modeling and control. In this paper, a control strategy to improve the performance of mobile robotic manipulators is presented. In the proposed architecture, a joint level controller needs to be designed to account for gravity, base motions, and manipulator joint friction. This is achieved by combining two methods: (i) a linear feed-forward dynamic disturbance compensation (DDC) method, which combines a dynamic model of the physical system with sensory feedback of the base oscillations to compensate for inertial effects; and (ii) Base Sensor Control, which compensates for nonlinear joint characteristics such as high joint friction. To evaluate the performance of the control architecture, a 3 DOF planar mobile manipulator is simulated including base oscillations, several payload weights, and Coulomb and viscous joint friction effects.