Sensorless Force Admittance Control Considering Robot Actuator Deadzone

External force detection is a necessary capability of cooperative robots to ensure the safety and convenience of human-machine interaction. Sensorless external torque estimation method has a cost advantage in many low-precision force control applications. The difficulties of sensorless external force detection lie in the joint torque detection with actuator deadzone. In this paper, a LuGre model-based static friction identification scheme is proposed, which is based on the identification of sliding friction and link dynamics. In the scheme, the static friction model is applied to estimate the external torque of joint with actuator deadzone, and an improved sensorless force admittance control algorithm is proposed. The practicability of the static friction identification and sensorless force admittance control schemes are verified in experiments. The method proposed can effectively improve the external torque estimation accuracy of robot joints considering the actuator deadzone without a force (moment) sensor.