In a suboptimally controlled manipulator arm, friction torque variations cause trajectory deviations and performance deterioration. On the other hand, the on-line computation of nonlinear terms of the suboptimal controller is time consuming or requires large memory space when table lookup techniques are used. In order to obtain robustness against variations in open-loop dynamics and decrease the effect of nonlinear terms due to couplings in suboptimal control systems with quadratic performance indexes, in addition to state proportional feedback, it is necessary to introduce feedback associated with the derivatives of the state variables. This additional feedback may be implemented as minor compensating loops built around the individual joints with analog devices. As an example, torque compensation and acceleration compensation for the M.I.T. Scheinman arm are considered. They yield a simple linear controller with improved robustness.