论文部分内容阅读
文中提出了基于人工神经网络和遗传算法的防空导弹控制系统设计方法。借鉴模型参考自适控制和预测控制思想 ;通过建立一个以小种群遗传算法 ( u GA)作为在线寻优工具 ,以人工神经网络作为控制器和被控对象在线模型的控制回路 ,来补偿被控对象的非线性、模型不确定性和在线扰动。理论分析和数学仿真表明 ,本方法相对传统控制策略具有较少依赖人工试凑和通用程度高的特点 ,同时可以在一定程度上解决传统增益规划方法缺少在线自适应能力、传统自适应控制技术难以有效利用离线设计结果和缺少并行处理硬件支持而导致在线实现性差的不足。
The paper presents a design method of air defense missile control system based on artificial neural network and genetic algorithm. Reference to model reference adaptive control and predictive control theory; through the establishment of a small population genetic algorithm (u GA) as an online optimization tool, artificial neural network as the controller and the controlled object online model control loop to compensate for the accused Non-linearity of objects, model uncertainty and online disturbance. Theoretical analysis and mathematical simulation show that compared with the traditional control strategy, the proposed method has less dependence on artificial test-taking and high versatility. At the same time, this method can solve the problem that the traditional gain planning method lacks the on-line adaptive ability and the traditional adaptive control technique is difficult Making good use of off-line design results and lack of parallel processing hardware support leads to poor online implementation.