CAJ | 학술논문

针对采样运动规划算法效率低，尤其在处理高维空间和复杂障碍环境等问题时，严重依赖于所选采样参数和碰撞检测距离等，提出了一种自适应双向快速密集树(ABiRDT)避碰运动规划方法。首先，深入研究了ABiRDT算法的基础理论和实现方法，可适应调整碰撞检测距离参数和随机采样扩展步长；其次，重点研究了本算法所采用的C-空间加权均匀采样、最近邻位形查找和基于混合包围盒的并行离散碰撞检测等关键自适应策略；最后，通过三维可视化计算机仿真验证了本文提出算法的有效性。
침대채양운동규화산법효솔저，우기재처리고유공간화복잡장애배경등문제시，엄중의뢰우소선채양삼수화팽당검측거리등，제출료일충자괄응쌍향쾌속밀집수(ABiRDT)피팽운동규화방법。수선，심입연구료ABiRDT산법적기출이론화실현방법，가괄응조정팽당검측거리삼수화수궤채양확전보장；기차，중점연구료본산법소채용적C-공간가권균균채양、최근린위형사조화기우혼합포위합적병행리산팽당검측등관건자괄응책략；최후，통과삼유가시화계산궤방진험증료본문제출산법적유효성。
Due to algorithm efficiency of sampling-based motion planning is low, especially when dealing with high-dimensional space and complex environmental obstacles, it is heavily dependent on the selected sampling parameters and collision detection distance, adaptive bidirectional balance rapidly exploring dense tree(ABiRDT) motion planning method has been proposed. First, basic theory and implementation method of the ABiRDT algorithm have been researched deeply, its including adaptive strategies include how to automatically adjust the collision detection distance parameter, random sampling expansion step;Secondly, key adaptive key adaptive strategies strategy used in the ABiRDT about C-space weighted uniform sampling, nearest neighbor configuration searching and hybrid bounding box based parallel discrete collision detection have been investigated strongly;Finally, effectiveness of the proposed algorithms have been verified by three-dimensional visualization computer simulation.