Adaptation to partial failure is one of the most important requirements in space robotics, since space robots cannot be repaired after they have been launched. We proposed a decentralized autonomous control algorithm of hyper-redundant manipulators that use parallel processing with very low-performance processors to achieve this adaptation. In this paper, some of manipulator joints are locked at a certain angle by computer simulation and the adaptability for the failure of the control algorithm is assessed. The control algorithm successfully continued the positioning task at the rate of more than 90 percent, even after half of its joints had failed. The control algorithm is also compared with behavior-based control architecture.
