Kinematic control for fault-tolerant modular robots based on joint angle increment redistribution

Based on the numerical inverse kinematic algorithm developed for modular robots, this paper presents a new control method, termed joint angle increment redistribution, which makes that the maximum allowable joint rates of a redundant robot can be specifically defined according to the internal and external constrains such as joints, tasks, and environments. For example, if some joints have failures and need to be locked, the corresponding joint angle increments can be redistributed to be zero or very small values. The proposed approach can be readily used for online fault-tolerant control of redundant robots. It also makes the optimal control with multiple performance criteria easy. The effectiveness of the proposed algorithms has been demonstrated by a 7-DOF serial modular robot for the avoidance of joint angle limits.