Distributed fault detection for modular and reconfigurable robots with joint torque sensing: A prediction error based approach

• A fault detection method that does not require motion states of other joints is proposed.
• The proposed method is robust to uncertainties in the system parameters.
• Independent of measurements or estimations of joint accelerations.

A distributed fault detection scheme for modular and reconfigurable robots (MRRs) with joint torque sensing is proposed in this paper. With the proposed scheme, the joint torque command is filtered and compared with a filtered torque estimate derived from the nonlinear dynamic model of MRR with joint torque sensing. Common joint actuator faults are considered with fault detection being performed independently for each joint module. The proposed fault detection scheme for each module does not require motion states of any other module making it an ideal modular approach for fault detection of modular robots. Experimental results have confirmed the effectiveness of the proposed fault detection scheme.