Active control to flexible manipulators

Active control law for flexible structures using nonlinear conventional motor with gear actuators has been a problem without an accepted solution with experimental evidence until now. This is due to the existence of a dead zone in torque caused by the nonlinear frictions inside the actuator. The torques needed to attenuate the vibrations, although calculated by the control law, are consumed by the frictions and do not arrive until the flexible structure. This work proposes a control strategy with friction compensation using neural networks to solve this problem. Experimental results obtained with a flexible manipulator attested to the excellent performance of the proposed control law.