基于改进增益规划法的导弹稳定控制回路设计

定点设计和多点拟合是利用传统增益规划法进行控制回路设计的基本环节。针对传统增益规划方法中存在的问题 ,本文将基于遗传算法的特征点处控制系统参数设计方法和基于神经网络的稳定控制律拟合方法组合在一起 ,形成一种改进增益规划方法用于导弹稳定控制回路设计 ,以克服原有设计方法中人工试凑较多的弊端 ,降低人为因素对设计工作的影响 ,实现设计工作程序化、自动化和通用化。

A New Design Approach for Air Defense Missile Autopilot Based on Genetic Algorithms and Artificial Neural networks

Following analysis of the traditional air defense missile autopilot design approach, a new gain scheduling method based on Genetic Algorithms (GA) and Artificial Neural networks (ANN) is presented, in which GA is used to automatically adjust the controller gains while ANN is used for the fitting of the control law over the missile flight envelop based on the prophase optimization done by GA Design methods and control schemes are tested by numerical simulations Both theoretical research and simulation results lead to the conclusion that the new approach is more designer-independent, automatic and general in use compared with the traditional ways