A dual-arm self-collision detection method and computer-readable storage medium

Abstract

The invention provides a dual-arm self-collision detection method and a computer-readable storage medium. The present invention uses the bounding box for rough collision detection and then uses the grid for fine collision detection. For the dual-arm collaborative robot, given task-level motion planning, real-time detection of the left arm, left arm, right arm, and body of the dual-arm during the motion planning process Or when the right arm collides with the right arm or the body, the robot controller responds in time according to the collision feedback signal. The invention has the beneficial technical effect of more optimized and more accurate collision detection.

Description

technical field [0001] The invention relates to the field of robots, in particular to a dual-arm self-collision detection method and a computer-readable storage medium. Background technique [0002] There are currently two main methods for self-collision detection: one is based on torque sensor detection. The second is the collision detection based on the geometric model, and the collision detection of the actual model is surrounded by the basic geometry. However, the existing self-collision detection methods have the following deficiencies: 1) The detection method of the torque sensor is too expensive, and the installation of the torque sensor needs to be considered in the robot design stage, that is to say, this method is not necessarily applicable In addition, the simple torque detection will cause many misjudgments during the robot debugging process; third, based on the detection of the torque sensor, the collision has already occurred at the time of detection, and it c...

AI Technical Summary

Problems solved by technology

However, the existing self-collision detection methods have the following disadvantages: 1) The detection method of the torque sensor is too expensive, and the installation of the torque sensor needs to be considered in the robot design stage, that is to say, this method is not necessarily applicable In addition, the simple torque detection will cause many misjudgments during the robot debugging process; third, based on the detection of the torque sensor, the collision has already occurred at the time of detection, and it cannot play a preventive role.

2) The geometric model collision detection method envelops the actual model through the basic geometry, but the disadvantage is that its collision accuracy depends on the closeness between the enveloped model and the actual model

When performing dual-arm task planning, considering different application requirements, the method based on the bounding box will reduce the working space of the left and right arm cooperation areas due to the expansion of the actual model volume, which may trigger collision detection and fail to complete the task.

Images