Power efficiency degradation of machines often provides intrinsic indication of problems associated with their operation conditions. Inspired by this observation, in this thesis work, a simple yet effective power efficiency estimation base health monitoring and fault detection technique is proposed for modular and reconfigurable robot with joint torque sensor.
The design of the Ryerson modular and reconfigurable robot system is first introduced, which aims to achieve modularity and compactness of the robot modules. Critical components, such as the joint motor, motor driver, harmonic drive, sensors, and joint brake, have been selected according to the requirement. Power efficiency coefficients of each joint module are obtained using sensor measurements and used directly for health monitoring and fault detection.
The proposed method has been experimentally tested on the developed modular and reconfigurable robot with joint torque sensing and a distributed control system. Experimental results have demonstrated the effectiveness of the proposed method.
Deep Learning Based Pavement Inspection Using Self-Reconfigurable Robot