摘要
以经典控制的弹体闭环回路作为受控对象，研究了基于神经元的导弹自适应控制器。以整体式冲压发动机超音速空地导弹为算例，进行了末制导段弹道仿真。仿真结果表明所设计的控制规律实时性强、抑制干扰性能好，且结构简单，易于工程实现。
In the terminal guidance stage of longrange airtoground missile, its parameters vary widely and abruptly, causing the missile to miss the target. We designed a neurone controller to try to lower the miss distance greatly. Fig.2 shows schematically the neurone adaptive controller we designed. The lower portion of Fig.2 shows the conventional controller and the upper portion corresponds to the neurone part of the controller. For the controller in Fig.2, we derived eqs.(7) through (17). Table 1 uses simulation results to compare the miss distances for neurone controller with those for conventional controller. For example, for ground object moving only in one direction at 14 m/s and for 22.5 s flying time of terminal guidance stage of airtoground missile, the miss distance for neurone controller is only 0.817 m whereas that of conventional controller is 6.926 m.
出处
《西北工业大学学报》
EI
CAS
CSCD
北大核心
1999年第1期113-117,共5页
Journal of Northwestern Polytechnical University
关键词
神经元
自适应控制器
中远程空地导弹
制导律
neurone adaptive controller, longrange airtoground missile, miss distance
