Real time fault tolerant control of robot manipulators

The complexity of computing the generalized forces or torques acting at the joints of a robot manipulator with n degrees of freedom directly effects the speed and accuracy of control. For real-time robust control, the large number of mathematical computations performed must satisfy the speed requirements and must also be error-free. In this paper, a combination of parallel and distributed processing techniques are proposed to compute the generalized forces using the Lagrangian formulation. The complexity of computation is reduced from the order theta(n(3)) to theta(n). It is shown that with minor modification in the architecture, all single errors that occur during computation can be detected, located, and corrected for robust control. (C) 2003 Elsevier Science Ltd. All rights reserved.