基于改进变异蚁群算法神经网络的空间大型可展开天线动态响应辨识  

Identification of Large-space Deployable Antenna Dynamic Response Based on Improved Mutated Ant Colony Algorithm

作　　者：王娟[1] 刘明治[1] 杨要恩[1] 

机构地区：[1]西安电子科技大学,西安710071

出　　处：《中国机械工程》2009年第1期86-89,共4页China Mechanical Engineering

基　　金：国家自然科学基金资助项目(59775052)

摘　　要：针对空间大型可展开天线柔性大、展开过程中弹性变形与刚体运动相互耦合、机构运动参数时变的特点,提出了基于改进变异蚁群算法神经网络的辨识模型用于可展开天线动态响应辨识的方法。该方法采用改进变异蚁群算法优化神经网络权值,将变异机制引入蚁群算法,解决了蚁群算法收敛慢的缺点,对变异蚁群算法进行改进,提高了算法跳出局部最优的能力,进一步加快了收敛速度。仿真结果表明,该辨识模型兼具神经网络和蚁群算法的优点,不仅具有优异的非线性逼近能力,还具有高的运算效率。该辨识模型能够准确地辨识天线的动态响应,辨识的收敛速度快且精度高。Large--space deployable antenna is tlexible structure, tlexible distortion couples with rigid movement and the movement parameter is time--varying during the deploy process. System i- dentification model based on mutated ant--colony algorithm neural networks was present to utilize the identification of dynamic response of deployable antenna. Improved mutated ant--colony algorithm was adopted to optimize the weights of neural networks in this method, mutated evolutionary was added into ant colony algorithm, the disadvantage of slow convergence rate was settled, mutated ant colony algorithm was improved to quicken the convergence rate. The simulation results show that the model colligates the advantage of both neural networks and ant--colony algorithm, has excellent ability to non--linear approach, and enhance the operation efficiency. The identification sensor can exactly identify the dynamic response of the antenna, and has quick convergence rate and high pr＇ecision.

关 键 词：神经网络 系统辨识 柔性结构 天线 动态响应 变异蚁群算法 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]