The Fast Collision Detection Algorithm For Multi-robot And System Experiments

Collision detection is an efficient way to make sure that the coordinate motion of multi-robots working safely with each other and operate orderly.Collision detection can improve the operating beat of multi-robots and the ability to utilize the working space more effectively.Futhermore,study the fast collision detection algorithm is an effective approach to improve the efficiency of multi-robots' trajectory planning and it has a significant technical and economic efficiency.In order to realize the fast collision detection among multi-robots,some detection algorithms and virtual simulations of fast collision are studied in-depth,the main contents are as follows:(1)Create a virtual simulation environment for the collision detection among the multi-robots.Beacse of the existing commercial software cannot be established a collision detection experimental environment for the multi-robots.Therefore,the OpenGL-based self-developed simulation software was proposed for the virtual simulation of multi-robots.And then various collision detection algorithms were embedded in this simulation software to do collision detection experiments among multi-robots,afterward the algorithms' detection efficiency were compared and analysis,and finally,the visualization of collision was realized during detection process.(2)Propose a hybrid bounding volume hierarchy tree(HBVHT)collision detection algorithm.Because of the robot was composed by series multiple rigid slender rod joints,the HBVHT with three layers was proposed based on the acceleration technology as bounding volume hierarchy tree for the collision detection among multi-robots.Simulation results were obtained from six kinds of HBVHT's collision detection,in which the top layer is an AABB MLHT based on LSS has the shortest cost of 23.205ms and highest efficiency among the six structures of HBVHTs.It is 5 times higher than the RAPID algorithm.(3)Achieve parallel computing for the HBVHT collision detection algorithm.Explore the high computational performance of multi-core processors,parallel structure was designed based on the structure of the original HBVHT and achieved its parallel computing by using OpenMP.Experimental results show that the HBVHT' middle layer and bottom layer with parallel computing compared to the HBVHT without parallel computing,collision detection efficiency is about 2 times higher in an Inter i5 quad-core processor hardware platform.Consequently,it means that the parallel structure is effective and it can effectively improve the computational efficiency of the collision detection algorithm.(4)Build a real robot experiment platform to verify the collision detection algorithm.The host computer,controller,servo drives and servo motors were used to design an electronic control for the robot,meanwhile build an experiment platform with the robot to verify the collision detection algorithm.Before this,the upper computer control software was developed for the robot based on OpenGL and parker development library.Then,it realizes an openness motion controller for a 6 DOF robot.The collision detection algorithms were embedded in the upper computer control software for experimental verification.When using collision detection function,the collision warning was obtained before the robot collided with the environment.Furthermore,it stops robot's movement before collide with obstacles in the environment.The experimental results show that the HBVHT collision detection algorithm is effective and suitable for collision detection for robot,and it can provide security for the robots.