基于改进粒子群优化六自由度机械臂时间轨迹规划

Improved Particle Swarm Optimization Based Six Degree of Freedom Robotic Arm Time Trajectory Planning

摘要: 在运动学限制的框架下，六自由度工业机器人采用标准的粒子群优化算法进行时间优化时，往往容易陷入局部最优解的困境。针对这一挑战，本研究提出了一种新型的多生境自适应变异粒子群优化算法(MNAM-PSO)以实现时间最优的轨迹规划。为了降低粒子陷入局部最优的风险，本研究从粒子的初始随机分布入手，通过采用多生境策略来分散粒子的起始位置，并结合扰动机制对每次迭代后的最优全局粒子进行变异处理。每一次变异的界限都与粒子各维度中的最小值相关联，从而形成了一种自适应的变异策略。在此基础上，本研究还采用了“3-5-3”混合多项式插值方法进行轨迹规划，并在MATLAB仿真环境中实现了机器人关节轨迹的拟合。实验结果显示，MNAM-PSO算法能够确保机器人在运动中的速度和加速度平滑无突变，相较于传统粒子群算法，规划时间缩短了大约10.9%，从而证明了该改进算法的优越性、有效性和实用性。

Abstract: In the framework of kinematic constraints, six-degree-of-freedom industrial robots often tend to fall into the dilemma of locally optimal solutions when using standard particle swarm optimization algorithms for time optimization. To address this challenge, this study proposes a novel multi-habitat adaptive variational particle swarm optimization algorithm (MNAM-PSO) for time-optimal trajectory planning. In order to reduce the risk of particles falling into local optimality, this study starts from the initial random distribution of particles, disperses the starting positions of particles by employing a multi-habitat strategy, and combines a perturbation mechanism to mutate the optimal global particles after each iteration. The bounds of each mutation are associated with the minimum value in each dimension of the particle, thus forming an adaptive mutation strategy. On this basis, this study also adopts the “3-5-3” hybrid polynomial interpolation method for trajectory planning, and realizes the fitting of robot joint trajectories in the MATLAB simulation environment. The experimental results show that the MNAM-PSO algorithm is able to ensure that the velocity and acceleration of the robot in motion are smooth and free of sudden changes, and the planning time is shortened by about 10.9% compared with that of the traditional particle swarm algorithm, which proves the superiority, effectiveness and practicability of this improved algorithm.