This article presents a new class of adaptive schemes for the motion c
ontrol of robot manipulators. The proposed controllers are very genera
l and computationally efficient because they do not require knowledge
of either the mathematical model or the parameter values of the manipu
lator dynamics, and are implemented without calculation of the robot i
nverse dynamics or inverse kinematic transformations. It is shown that
the control strategies are globally uniformly bounded in the presence
of bounded disturbances, and that in the absence of disturbances the
ultimate bound on the size of the tracking errors can be made arbitrar
ily small. Computer simulation results are given for a PUMA 560 manipu
lator, and demonstrate that accurate and robust trajectory tracking ca
n be achieved by using the proposed controllers. (C) 1994 John Wiley &
Sons, Inc.