多机器人作业系统交互式碰撞检测与仿真实现

发布时间：2018-06-30 02:53

本文选题：多机器人 + 交互式　；参考：《北方工业大学》2016年硕士论文

【摘要】：随着多机器人技术的不断发展,针对多机器人的仿真技术得到了学术界和工业界越来越高程度的重视。然而随着多机器人仿真场景的不断扩大和复杂度的增加,对仿真系统中的碰撞检测算法的精度和效率都提出了更高的要求。为了提高多机器人协作交互式仿真系统的仿真度,提升用户的使用体验,检验具体工序方案的正确性以及提高机器人仿真系统的运行效率,传统碰撞检测算法已经难以完成针对超高精密机械部件的精确碰撞检测以及大规模仿真场景中的快速碰撞检测。当前常用的仿真系统碰撞检测技术多是基于空间分割方法或者基于包围盒的碰撞检测方法。但是基于空间分割的碰撞检测方法容易出现漏检的情况,而基于包围盒的碰撞检测方法又十分容易出现过检,所以碰撞检测精度比较低。此外,目前的碰撞检测方法在大规模复杂场景中的运行效率都比较低。基于此,本课题提出一种基于物体特征图的碰撞检测方法,通过提取物体模型的表面特征点,然后使用图论连接这些点形成特征图,再用紧密包围物体的特征图的边来做碰撞检测。同时,提出一种自适应的碰撞检测加速方法,使用碰撞检测掩码来快速剔除不需要检测的对象,从而加快碰撞检测的运行效率。通过对多机器人协作交互式碰撞检测方法的研究,在实现多机器人仿真系统的过程中,完成了对多机器人协作的交互式仿真,并且检验了基于物体特征图的碰撞检测算法和自适应的碰撞检测加速方法,使得在复杂大规模复杂场景中的碰撞检测算法能够更加高效的运行,同时更加精确的检测到物体之间的碰撞,减少了过检和漏检率。同时结合实际的多机器人系统,使用实现的多机器人仿真系统加快了实际多机器人系统的开发速度。
[Abstract]:With the development of multi-robot technology, more and more attention has been paid to the simulation technology of multi-robot. However, with the continuous expansion and complexity of multi-robot simulation scene, the accuracy and efficiency of collision detection algorithm in simulation system are required to be higher. In order to improve the simulation degree of the multi-robot cooperative interactive simulation system, enhance the user's experience, verify the correctness of the concrete working procedure scheme and improve the running efficiency of the robot simulation system, Traditional collision detection algorithms have been difficult to complete accurate collision detection for ultra-precision mechanical components and fast collision detection in large-scale simulation scenarios. Most of the commonly used collision detection techniques in simulation systems are based on space segmentation or bounding box based collision detection. However, collision detection based on space segmentation is prone to miss detection, and collision detection based on bounding box is easy to overdetect, so the accuracy of collision detection is relatively low. In addition, the current collision detection methods are less efficient in large scale complex scenarios. Based on this, a collision detection method based on object feature graph is proposed, which is based on the surface feature points of the extract body model, and then connects these points with graph theory to form the feature map. Then the edge of the feature graph is used to detect the collision. At the same time, an adaptive collision detection acceleration method is proposed, in which collision detection masks are used to quickly eliminate objects that do not need to be detected, so as to speed up the efficiency of collision detection. Through the research on the interactive collision detection method of multi-robot cooperation, the interactive simulation of multi-robot cooperation is completed in the process of realizing the multi-robot simulation system. The collision detection algorithm based on the object feature graph and the adaptive collision detection acceleration method are tested to make the collision detection algorithm run more efficiently in the complex large-scale complex scene. At the same time, the collision between objects is detected more accurately, and the rate of overdetection and missed detection is reduced. At the same time, combining with the actual multi-robot system, the development speed of the actual multi-robot system is accelerated by using the realized multi-robot simulation system.
【学位授予单位】：北方工业大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP391.41;TP242

【参考文献】