基于动量观测器和双编码器的协作机器人碰撞检测

为提高协作机器人高速运行时碰撞检测灵敏性，提出一种基于动量观测器和双编码器的碰撞检测方法。基于广义动量的动力学模型构建机器人外力矩观测器；利用双编码器的实时位置偏差和机器人关节传动结构等效柔性特征估计外力矩；设计反应策略实现碰撞后机器人快速停车及能被人工推动。仿真和实验结果表明：在机器人高速运行时，相比传统的基于动量偏差观测器算法，此算法响应时间更短，最大碰撞力更小，提升了高速下碰撞的检测灵敏度，增强了安全防护效果。

Cobot Collision Detection Based on Momentum Observer and Dual Encoders

To improve the sensitivity of collision detection when cobot runs at high speed, a collision detection method based on momentum observer and dual encoders was proposed. An external torque observer was constructed according to the dynamics model of the robot based on generalized momentum. The external torque was estimated by the real-time position deviation of dual encoders and the equivalent flexibility characteristics of the robot joint transmission structure. A reaction strategy was designed to realize the robot to stop quickly and could be pushed manually after collision. The simulation and experiment results show the proposed algorithm has shorter response time and smaller maximum collision force than the traditional algorithm based on momentum deviation observer when the robot runs at high speed, which improves the sensitivity of collision detection and the safety protection effect.