Adaptive friction compensation of flexible-joint manipulators with parametric uncertainties

In this paper, we propose an adaptive friction compensation strategy for flexible-joint robotic manipulators, taking into account Coulomb friction for the actuator and the load. For such systems, the exact inverse model is unrealizable and only an approximation can be found. The proposed adaptive control strategy consists of a rigid-based linear in parameter feedforward model that approximates the flexible-joint inverse model and a feedback controller that acts as a friction compensator. A reference model is introduced as a trade off strategy to cope with the joint elasticity errors. Results highlight the performance of the controller in friction compensation, in particular nonlinear Coulomb friction terms. Internal stability, a potential problem with such a system, is also verified. Furthermore, the stability of the proposed control system is guaranteed by Lyapunov stability theory.