摘要
针对大多数机器人最优轨迹规划控制算法成本较高的问题,提出了一种两阶段分析-进化算法,分析算法基于间接开环最优控制问题,进化算法基于遗传算法。首先,利用遗传算法产生最优控制;然后,新的次优轨迹通过最优控制生成,计算每个最优解的成本函数,并为下一个步骤选择最优解做准备;最后,遗传算法使用获得的轨迹产生新一代起始点,迭代直到实现最小成本值。算法的验证由两个仿真系统组成:具有完整约束的两关节机械手和具有非完整约束的移动机械手。运用本文算法后,第一个仿真系统的成本降低了92.3%,第二个仿真系统获得的最优成本明显低于只使用最优控制时的成本,表明其算法可应用于实际机器人轨迹规划控制系统。
As the problem of high cost in optimal manipulator trajectory planning algorithm,a two-stage analysis-evolutionary algorithm is proposed.The analysis algorithm is based on indirect open-loop optimal control problem.The evolutionary algorithm is based on Genetic algorithm（GA）.Firstly,the starting point of optimal control problems results from the optimal solution by GA.Then,a new sub-optimal path is generated by the optimal control,and the cost function of every optimal solution is calculated,preparing for the next step to choose the optimal solution.Finally,the new starting point is obtained by the new generation of the path using GA.The iteration is ended until the minimum cost value can be obtained.Verification algorithm simulation system consists of two components,a system is a two-joint mobile manipulator with a complete constraint,and the other is a mobile manipulator with a non-complete constraint.By using the proposed algorithm,the cost of the first simulation system reduces 92.3%,and the optimal cost of a second simulation system is clearly less than that of only using optimal controlling.Simulation results show that the algorithm can be applied to any robot path planning.
出处
《微型电脑应用》
2016年第4期18-22,共5页
Microcomputer Applications
基金
江苏省高校自然科学研究项目(14KJB520036)
关键词
机器人
移动机械手
最优轨迹控制
分析-进化
两阶段遗传算法
Robot
Moving Manipulator
Optimal Trajectory Control
Analysis-evolutionary
Two-stage Genetic Algorithm
