Actuator saturation avoidance for fault-tolerant robots

Actuator saturation in robotic systems is a problem area that needs attention. This problem occurs more frequently and at the same time becomes more difficult to address when robots operate under faulty conditions. This work investigates two different approaches, the time regulation, and the torque redistribution methods, in an attempt to avoid saturation in fault-tolerant robots. Fault-tolerant robots considered include general architecture, spatial robots with serial or parallel structures. The control algorithm consists of a model reference algorithm and the computed torque method in the feedforward process, and a PID controller in the feedback process. Upon failure, these methods recover the robot from faulty conditions and complete the task avoiding actuator saturations. Initial simulation results indicate that the time regulation method is more suitable for serial robots, and the torque redistribution method for parallel robots. These methods are expected to be applied not only to a case where a joint is fully failed but also so cases where some joints experience partial failure