Torque sensor design for light weight arm joint

It outlines a new joint torque sensor's mechanical structure and signal conditioning electronics. The static calibration experiment has been studied. To improve dynamic response, according to the philosophy of inverse model, the dynamic compensation for the sensor is designed based on the least square method, functional link artificial neural network (FLANN) theory and instrumental variable (IV) theory. The experimental results show that the IV theory is more effective and the torque sensor can match the requirement of the light weight arm system.