异类粒子群算法的机械臂轨迹多目标规划

为了减少机械臂工作过程中的耗时、耗能和冲击,提出了基于异类粒子群算法的机械臂轨迹多目标规划方法。建立了7自由冗余机械臂运动学模型和优化模型;以传统粒子群算法为基础,根据学习信息的不同来源,设计了4种粒子进化行为,根据粒子当前状态计算不同进化行为的当前价值和未来价值,根据当前价值和未来价值的综合价值计算不同进化行为的选择概率,从而完成了异类粒子群算法的构造。以耗时最少单目标优化为例,与传统粒子群算法相比,异类粒子群算法不仅收敛速度快,而且优化程度提高了27.48%;使用异类粒子群算法对机械臂轨迹进行多目标综合优化,给出了Pareto最优前沿面,可根据不同需求和优化重心从中选择优化结果。

Robotic Arm Trajectory Multi-Objective Programming Based on Heterogeneous Particle Swarm Algorithm

To reduce time-cost,energy dissipation and shock under the robotic arm working process,robotic arm trajectory multi-object programming method based on heterogeneous particle swarm algorithm is proposed.Kinematic model and optimization model of 7 degree of freedom redundancy robotic arm are built.On basis of traditional particle swarm algorithm,depend on different source of learning information,evolutionary behavior of four kinds of particles are designed.present value and feature value of different evolutionary behavior relying on particle present state are calculated.Selective probability of different evolutionary behavior is given by its comprehensive value,so that construction of heterogeneous particle swarm algorithm is over.In the case of mainimum time-cost,compared with traditional particle swarm algorithm,heterogeneous particle swarm algorithm possesses faster convergence,and optimization degree improves by 27.48%.robotic arm trajectory multi-object optimization is executed by heterogeneous particle swarm algorithm,and Pareto leading surface is given,so that optimization result can be chosen depending on different requirement and optimization core.