芒果采摘机械臂运动轨迹规划与运动参数优化研究

Motion Trajectory Planning and Motion Parameter Optimization for A Mango-picking Robotic Arm

摘要:

为了提高芒果采摘机械臂的采摘效率，在建立机械臂的正、逆运动学模型的基础上进行轨迹规划，并通过智能算法优化其运动参数。采用 Denavit-Hartenberg (D-H) 法对芒果采摘机械臂进行建模，利用 3-5-3 混合多项式插值方法进行轨迹规划，针对芒果采摘机械臂的特点，提出改进的蜘蛛蜂算法 (ISWO), 主要通过拉丁超立方体采样初始化种群，采用线性递减随机性权重提升搜索稳定性，设计动态切换机制平衡全局探索与局部开发以兼顾搜索效率与收敛速度。实验结果表明，采用 ISWO 的芒果采摘机械臂各个关节的迭代速度和精准度均优于粒子群优化算法 (PSO) 和标准蜘蛛蜂算法 (SWO)。在 4 次实际采摘实验中，ISWO 的总采摘时间分别为 62.94、32.22、39.52、46.11s, 缩短了 18.25%~42.98%。研究表明，ISWO 显著提升了芒果采摘机械臂的采摘效率，为后续相关的应用研究提供了可靠的数据支持。

Abstract:

In order to improve the picking efficieney of the mango-picking robotic arm, trajectory planning was carried out based on the establishment of both positive and negative kinematice models of the robotic arm, with its motion parameters were optimized through the application of intelligent algorithms. The Denavit - Hartenberg (D -H) method was employed to model the mango-picking robotic amm, while the 3 - 5 - 3 hyrid polynomial interpolation method was utilized for effective trajectory planning. Additionally, an improved spider wasp algorithm (ISWO) was proposed to address the specific characteristics of the mango-picking robotic armn. This algorithm primarily focused on initializing the population through Latin hypercube sampling, which enhanced the search stability by incorporating linearly decreasing stochasticity weights. Furthermore, a dynamic switching mechanism was designed to balance global exploration with local exploitation, thereby optimizing both search efficiency and convergence speed. Experimental results indicated that each joint of the mango-picking robotic arm utilizing ISWO significantly outperformed the particle swarm optimization algorithm (PS0) and the standard spider wasp algorithm (SWO) in terms o iteration speed and accuracy. In four actual picking experiments, the total picking time recorded for ISW0 were 62.94 s, 32. 22 s, 39.52 s, and 46. 11 s, demonstrating a reduction of 18.25% to 42.98% in picking time. The research result conclusively showed that the ISWO algorithm significantly enhanced the picking efficiency of the mango-picking robotic arm, providing reliable data support for subsequent related application studies.