Collision Detection for CoSCARA Considering Mechanical Structure and Friction Compensation

The measurement of collision detection performance is a crucial factor in evaluating the effectiveness of collaborative robots. The accuracy of sensorless force control in collaborative robots is limited by the presence of nonlinear friction at the joints, which affects collision detection performance. This analysis focuses on the influence of friction components and mechanical structure design on friction in collaborative robots, with a specific emphasis on the nonlinear friction of the joints in these robots. The DLinear-based neural network model is utilized to simulate joint friction, enhancing the precision of friction identification. The feedforward torque is employed to compensate for the friction force, hence enhancing the sensitivity of collision detection. The experimental investigation demonstrates that utilizing a synchronous belt with a reduced transmission ratio resulted in decreased friction and increased collision detection sensitivity for the robot, hence enhancing its safety level.