Dynamic multi-obstacles avoidance of a robot manipulator based on repulsive vector summation for human-robot co-works

In this paper, a whole-arm collision avoidance from multi-obstacles dynamically with repulsive vectors summation method for human-robot collaboration and co-works application is presented. Procedures of the collision avoidance method are described in detail. First, the environment is monitored by perception sensor and the process of collision detection is divided into six parts to effectively perform detection algorithm. Second, an effective algorithm with the concept of repulsive vectors summation from multi-obstacles is presented. Also, the vectors summation problem in three types are solved. Furthermore, based on multi-obstacles algorithm, we calculate the collision reaction of arm on three different types. Finally, the command is sent to Vector-Based Online Trajectory Generator (vOTG) to perform the control motion of a robot manipulator. As a result, the robot manipulator can achieve the whole-arm collision avoidance with vectors summation for multi-obstacles dynamically. We have successfully demonstrated the proof of concept of the system with a 7-DoF NTU-iCeiRA arm designed and developed by ourselves in iCeiRA Lab.