A FAULT-TOLERANT CONTROL ALGORITHM HAVING A DECENTRALIZED AUTONOMOUS ARCHITECTURE FOR SPACE HYPER-REDUNDANT MANIPULATORS

Adaptation to partial failure is one of the most important requirement s for space robotics, since space robots cannot be repaired after they have been launched. Ne propose a decentralized autonomous control alg orithm for hyper-redundant manipulators that uses parallel processing with low-performance processors to achieve this adaptation. In this pa per, a number of manipulator joints are locked at a certain angle in a computer simulation and the adaptability of the control algorithm to these failures is assessed. The control algorithm successfully continu es its positioning task at a rate of more than 90%, even after half of its joints have failed. The control algorithm is also compared with b ehavior-based control architecture.