基于INSGA-Ⅱ算法的机械手多目标轨迹规划

INSGA-Ⅱ based multi-objective trajectory planning for manipulators

针对机械手时间最优、能量最优、平滑性最优等多目标下的轨迹优化问题,设计新的多目标轨迹优化方法.采用高次B样条曲线插值方法,构造机械手高阶连续且起始和终止的运动参数均可指定的关节轨迹,保证了机械手运动性能.采用改进非支配排序遗传算法(INSGA-Ⅱ)对机械手轨迹进行优化,得到一组Pareto最优解集,该算法采用一维Logistic映射产生初始种群并利用不可行度选择操作处理约束条件.在6自由度串联机械手上的计算结果表明,采用高次B样条轨迹规划方法可以得到高阶连续的机械手分段轨迹,采用INSGA-II方法可以对B样条轨迹实现有效的多目标寻优,得到理想的Pareto分布,为用户提供较多的选择.

A new multi-objective optimization algorithm was proposed to plan the trajectory of manipulators under more objectives,including time optimal,energy optimal,and smoothness optimal.The high-degree B-spline was adopted to construct a continuous path with controllable start-stop kinematic parameters which guaranteed the motion performance of manipulators.The improved non-dominated sorting genetic algorithm-Ⅱ（INSGA-Ⅱ）was applied to optimize the trajectory of manipulators in order to get a set of Pareto optimal solution aggregate.The algorithm used one-dimensional Logistic mapping to generate the initial populations and infeasibility degree selection to handle the constraints.The results on a six-degree of freedom serial robot manipulator show that high-degree B-spline can get high-degree continuous trajectories.INSGA-II provides an effective approach to do a multi-objective optimal for B-spline and can obtain good distributed Pareto solutions providing more choices for users.