基于混合遗传算法优化的制导炸弹BP网络PID控制器

经典的PID控制方法虽然其物理意义明确、参数调整简便,但是面对现代的过程控制对象所具有的非线性、时变和多种不确定干扰等特性,使得基于精确线性数学模型的设计思路很难直接获得满意的动静态控制效果.在BP神经网络控制的基础上,采用混合遗传算法优化BP反向传播权系数值,推导了算法过程,设计了基于混合遗传算法优化的制导炸弹BP网络PID控制器.最后,以某制导炸弹纵向通道设计为例,仿真结果表明控制器能使得制导炸弹准确命中目标,并且具有较好的鲁棒性能.

The BP Networks PID Controller Optimized for a Guided Bomb based on Hybrid Genetic Algorithm

With traditional PID method the physical signification is explicit and parameters can be briefly adjusted, while its need for precise linear model can hardly acquire satisfying control result for modern process control object is nonlinear, time-variant and interfere-uncertain. The PID controller is designed for a guided bomb(GB) with hybrid genetic algorithm, which is to optimize the back propagation weigh coefficient of the BP networks. After the algorithm is produced, the example of a GB's portrait channel controller is designed and simulated. The result shows that the target is attacked accurately and the controller is robust.