基于改进人工蜂群算法的并联机器人正运动学解

Forward Kinematics Solution for Parallel Manipulators Based on Improved Artificial Bee Colony Algorithm

并联机器人的正运动学问题可转化为等效的最小化问题,并采用数值优化方法求解。人工蜂群算法是模拟采蜜过程中蜂群行为机制的一种启发式搜索方法,目前还尚未在并联机器人正运动学问题中得到应用。提出一种数值求解并联机器人正运动学问题的改进人工蜂群算法(Improved artificial bee colony algorithm,IABC),该方法通过多维度改变蜜源位置,对算法中雇佣蜂采用以DE/rand/1差分操作为主,并兼以遗传算术交叉为辅两种策略相结合搜索方式增强群体多样性,而对旁观蜂则采用DE/best/2差分操作方式在其邻域内搜索到性能更优蜜源位置;此外在旁观蜂选择蜜源阶段采用一种非线性排序选择操作,减弱群体中超级蜜源(个体)引起的早熟问题;基准函数测试表明该方法较大改善了寻优解质量、求解可靠性。在此基础上以6-SPS并联机器人为例,采用IABC算法数值求解其正运动学问题,数值仿真结果表明了该方法是求解并联机器人正运动学问题的一种有效方法。

The forward kinematics problem of 6-SPS parallel manipulator can be transformed into an equivalent optimization problem and solved through numerical optimization methods. The artificial bee colony algorithm（ABC） is a meta-heuristic optimization method which mimics the foraging behavior of honey bee swarm. To the best of our knowledge, there is very little research work on ABC for forward kinematics problem of parallel manipulator. An improved artificial bee colony algorithm（IABC） is presented for computing numerical solutions to the forward kinematics problem of parallel manipulator. By changing nectar position multi-dimensionally in IABC algorithm, two strategies are combined for the employed bees, that is, DE/rand/1 differential operator is adopted as the main opt^nizing scheme, and genetic arithmetic crossover operator is used as auxiliary manner to enhance the diversity of the bee colony; while for the onlookers, the DE/best/2 differential operator is employed to find better solution in its vicinity. In additional, during the onlookers selection nectar sources phase, non-linear ranking selection is adopted to alleviate premature problem that caused by the super nectar sources in the colony. The experimental results of benchmark functions show that this IABC algorithm greatly improves both the optimal solution quality and the search solution reliability. Based on these, a 6-SPS parallel manipulator is used as an example, and then the proposed IABC is used to solve the forward kinematics problem. Numerical simulation results demonstrate the effecfivenegs afthig aloarithm ta gnlv~, the. fnrward kinamatic~ nrnhlam nfnnrnllal ranninnlatar~