Unfalsified direct adaptive control of a two-link robot arm

Abstract

This paper describes an application of unfalsified control theory to the design of an adaptive controller for a nonlinear robot manipulator. A nonlinear "computed torque" control structure is employed. Four parameters representing unknown masses, inertias and other dynamical coefficients are adaptively adjusted in real-time using an linear programming technique to optimally satisfy control-law unfalsification condition. Simulations show that the method yields significantly more precise and rapid parameter adjustments than conventional continuous parameter update rules, especially when the manipulator arm is subject to sudden random changes in mass or load properties.