Regressor-Free Adaptive Control of Flexible Joint Robot Manipulators with Reduced Number of Estimators

In this paper, an adaptive controller is proposed for controlling flexible joint robots containing time-varying uncertainties. The controller is free from the computation of the regressor matrix which largely simplifies its implementation. The feedback of link accelerations and their higher derivations are avoided without the support of Slotine and Li's modification. The number of update laws in the estimation of uncertainties by using the function approximation technique is significantly reduced in this new design which further improves its applicability. The closed loop stability and boundedness of internal signals are proved by the Lyapunov-like theory. Computer simulations are performed to verify its effectiveness.