激光通信快速反射镜系统辨识与平衡截断

System Identification and Balanced Truncation of Fast Steering Mirror for Laser Communication

中文关键词  激光通信  音圈电机快速反射镜  系统辨识  平衡截断

英文关键词  laser communication  voice coil actuator fast steering mirror  system identification  balanced truncation

中文摘要

      音圈电机快速反射镜是激光通信精跟踪的执行机构, 其柔性结构与双轴耦合特性会导致系统谐振模态增多、模型阶次高等复杂控制问题, 基于递推最小二乘等传统辨识方法建立精确模型是十分困难的, 难以满足高精度控制器的设计要求. 对于该问题, 本文提出了基于Hankel矩阵模型解算与平衡截断建立高精度模型的方法. 首先, 根据系统激励序列与响应序列建立Hankel矩阵, 其奇异值分解确定模型阶次, 可观、可控矩阵解算确定模型参数; 其次, 建立系统的平衡实现, 并基于平衡截断原理完成模型降阶. 本文从频率特性角度评价了系统模型的辨识精度,结果表明, 相比于经典的模型辨识方法, 本文建立的系统模型阶次更低且辨识精度更高, 充分满足现代理论高性能 控制系统的设计与应用要求.

英文摘要

      The voice coil actuator fast steering mirror is an actuator for precise tracking of laser communication, its flexible structure and dual-axis coupling characteristics must lead to complex control problems with increasing resonant modes and higher model order. Traditional identification methods, such as recursive least squares, struggle to establish accurate models that meet the requirements for high-precision controller design. For this problem, this paper proposed a method to build a high-precision model for fast steering mirrors based on the Hankel matrix model solution and balanced truncation. Firstly, the Hankel matrix can be established according to the system excitation sequence and response sequence, the model order can be determined according to the singular value decomposition, and the model parameters are determined based on the observable and controllable matrix solution. Secondly, the system’s balanced realization can be established, and the model order can be reduced based on the principle of balanced truncation; the identification accuracy of the system model was compared from the frequency characteristics. The results showed that compared to the classical identification method, the system model established in this paper had lower order and higher identification accuracy. This approach fully satisfies the design and application requirements of modern theoretical high-performance control systems.