College of Mechanical and Electrical Engineering,Northeast Forestry University,Harbin 150040,China
Show less
文章历史+
Received
Published
2025-10-14
2026-05-20
Issue Date
2026-06-11
PDF (5601K)
摘要
当下林果采摘大量采用人工作业的方式,其采摘效率低、生产成本高,且具有一定危险性,林果采摘机械臂的发展和应用势在必行,避障路径规划是机械臂作业的关键。快速随机扩展树(rapidly-exploring random tree,RRT)算法作为常用的路径规划方法,存在搜索效率低、规划成本高和规划路径曲折冗长等问题。为此提出一种基于多策略融合的改进APF-RRT算法,APF(artificial potential field)为人工势场法,在RRT算法中引入动态目标偏置策略,融合改进后的人工势场法,引导随机树朝着目标点方向扩展,结合步长自适应调整策略,提升算法的性能;对初始路径进行剪枝和平滑优化,剔除冗余节点,提升路径质量。利用Matlab进行对比仿真试验的三维空间仿真结果显示,相较于RRT算法、RRT*算法和文献[8]算法,改进APF-RRT算法的搜索时间分别下降90.91%、96.92%、77.78%,路径长度分别缩短33.22%、22.52%、25.94%,随机树中节点分别减少92.58%、96.23%、83.33%。机械臂采摘试验结果表明,改进APF-RRT算法下机械臂采摘时间为18.9 s,采摘成功率为92%。改进的APF-RRT算法可应用于林果采摘作业,提高林果采摘效率。
Abstract
Forest fruit harvesting still relies heavily on manual labor, which suffers from low efficiency, high production costs, and certain safety risks. The development and application of forest fruit-harvesting robotic arms are therefore imperative. Obstacle avoidance path planning is a critical aspect of robotic arm operations. The RRT algorithm, as a commonly used path planning method, suffers from issues such as low search efficiency, high planning costs, and excessively convoluted and lengthy planned paths. This paper proposes an improved APF-RRT algorithm based on a multi-strategy fusion approach. The approach introduces a dynamic target bias strategy into the RRT algorithm, integrates an improved artificial potential field method to guide the random tree expansion toward the target point, and combines an adaptive step-size adjustment strategy to improve algorithm performance. Additionally, the initial path is optimized through pruning and smoothing to eliminate redundant nodes and enhance path quality. Using Matlab for comparative simulation experiments,the three-dimensional space simulation results demonstrate that the improved APF-RRT algorithm reduces search time by 90.91%, 96.92%, and 77.78% compared to RRT, RRT*, and the algorithm in reference [8] respectively, while shortening path lengths by 33.22%, 22.52%, and 25.94%, and decreasing nodes in the random tree by 92.58%, 96.23%, and 83.33%. The experimental results of robotic arm harvesting demonstrate that under the improved APF-RRT algorithm, the harvesting time is 18.9 seconds with a success rate of 92%. This shows that the improved APF-RRT algorithm can be applied to forest fruit harvesting operations, improving harvesting efficiency.
当下常用的路径规划方法包括基于图搜索的A*算法[5]和基于采样的快速随机扩展树(rapidly-exploring random tree,RRT)算法。由Lavalle[6]提出的RRT算法搜索快速、概率完备,但在复杂环境中搜索效率低,生成的路径并不是最优解,国内外学者对RRT算法进行了诸多改进[7-9]。Karaman等[10]提出了RRT*算法,该算法通过重选父节点获取渐进最优路径,但计算成本较高。Kuffner等[11]提出双向RRT(RRT-connect)算法,该算法极大地提高了搜索速度,但随机性强,规划的路径不一定是最优解。Tahir等[12]提出了双向RRT*算法(RRT*-connect),提高了算法的收敛速度,减少了算法的内存占用。黄杰等[13]提出了预置树的概念,将探索区域均匀地分布到整个空间,提高了搜索效率。李子建等[14]在基本RRT基础上提出目标偏置和剪枝优化策略,降低采样点随机性,减少路径长度。栾庆磊等[15]通过引入全局自适应步长策略对基本RRT算法地图探索能力进行提升,基于节点附近障碍物信息动态调整搜索步长。王杲等[16]和万程龙等[17]使用人工势场法与基本RRT算法相融合的方法,通过人工势场引导随机树进行采样,提高了避障性能。
YUG R, GUOK X, HAOT X.Preliminary exploration on the construction of forestry ecological economic system in the new era[J].Scientia Silvae Sinicae,2025,61(7):8-22.
YANGS, HONGT, ZHUL K.Forest fire prevention mobile robot path planning based on improved ant colony algorithm[J].Forest Engineering,2024,40(1):152-159.
LUD J, MENGQ G.A crane path planning algorithm based on improved RRT[J].Ordnance Industry Automation,2025,44(7):1-4.
[17]
KARAMANS, FRAZZOLIE.Sampling-based algorithms for optimal motion planning[J].The International Journal of Robotics Research,2011,30(7):846-894.
[18]
KUFFNERJ J, LAVALLES M.RRT-connect:an efficient approach to single-query path planning[C]//Proceedings 2000 ICRA.Millennium Conference.IEEE International Conference on Robotics and Automation.Symposia Proceedings.April 24-28,2000,San Francisco,CA,USA.IEEE,2000:995-1001.
[19]
TAHIRZ, QURESHIA H, AYAZY,et al.Potentially guided bidirectionalized RRT* for fast optimal path planning in cluttered environments[J].Robotics and Autonomous Systems,2018,108:13-27.