摘要
为了使机器人在完成对接(Rendezvous)任务的同时让能耗达到最优,提出一种通过直接配置非线性规划法(DCNLP)来实现五连杆机器手臂能耗最优化的控制方法.使用Euler-Lagrange公式描述机器手臂的动力学方程,并借助符号运算推导相应公式;利用最优化控制的必要条件(Necessary Conditions)来规范最优指标,因是最优控制,求解过程必定遭遇"两端点边界值问题(Two-point Boundary-valued Problem)",本文引入DCNLP法将最优控制问题离散化为非线性规划问题;最后利用非线性规划法求得问题的最优解;仿真结果表明DCNLP对初值估计精度要求不高,鲁棒性好,寻优能力强.
The application of direct collocation with nonlinear programming （ DCNLP ） method applied to a five-linked manipulator arm rendezvous with numerical analysis optimization problem is introduced. Firstly, this study takes advantage of Symbolic Language Programming and uses it to derive the Euler-Lagrange { E-L ） equation symbolically. Since it is an optimal control problem,＂ two-point boundary-valued problem（ TPBVP ）＂ is inevitable. Thus DCNLP is indroduced to seek the numerical solutions of the problem. Then, the optimal control problern/TPBVP is transformed into nonlinear programming problem using DCNLP method. The simulation result demonstrates that DCNLP method is neither acute nor discriminating to the initial guess of state variables; it also shows that DCNLP is robust in obtaining solutions for a large-scaled problem. At last, DCNLP requires large memory space and fast CPU. However these two concerns have not been an issue for now-a-days microcomputer. The major theme of this paper is to use DCNLP to facilitate the use of intelligent robotic manipulator and apply it to enhance the quality of medical care and senior citizen nursing. However, to take up a task such as this kind, the robot has to come up with some algorithm which intelligently gets the maneuvering done with ele- gance. Optimal control and optimization guarantee that the result found is the best by compairing to other methods. It may encounters difficulties in solving the system but the result is unbeatable.
出处
《小型微型计算机系统》
CSCD
北大核心
2012年第12期2749-2752,共4页
Journal of Chinese Computer Systems
基金
浙江省教育厅科研项目(Y201224238)资助
关键词
机器手臂
路径规划
直接定位
最优化控制
非线性规划
robot manipulator
path planning
direct collocation
optimal control
nonlinear programming
