Research On Collision Avoidance Method For Tele-Operation Robot End-Effector Based On Virtual Force Compliance

In tele-operation systems, because of the lack of direct perception of the remote real working environment and the time delay of the data transmitting, the operator can't get enough real-time information about the real working environment. It is highly possible for the collision between the robot arm especially the end of the robot arm and obstacles in the working environment to happen. In order to keep the robot arm safe, the research on how to avoid the collisions is necessary.Through the analysis of the advantages and disadvantages of the traditional methods, combined with the characteristics and needs of the tele-operation system, an approach based on virtual force compliance was presented. In order to avoid the collisions, three layers of collision avoidance buffer zone next to the bounding box were built. Meanwhile the distance between the end of the robot arm and the obstacles was acquired by the collision detection technology, to determine whether the end of the robot arm entered the buffer zone. And virtual force calculation model was presented to generate virtual force when the end of the robot arm entered the buffer zone. Combined with the model of three layers of buffer zone, impedance control of the robot arm was used to avoid the collision. In the end, a tele-operation system was set up, and experiments on the system verified the effectiveness of the approach.In the approach, a new kind of bounding box based on center axis was used to simplify the geometrical models and speed up the algorithm. And through an experiment, the time spent in constructing bounding box and collision detection using the traditional bounding box and the new one was analyzed and compared. The superiority of the new bounding box based on center axis was verified.Compared with the traditional methods, this approach will not limit the robot arm in a planned trajectory which will increase the space of the free motion. Meanwhile, it will maximize the advantage of the real-time control of human-computer reaction. And because the approach is based on virtual collision force, the operator can feel the virtual collision directly which is good for the operator to perceive the real working environment.