关键词: 兵器科学与技术, 灰预测, 粒子群优化算法, 小波神经网络, 变结构, Lyapunov稳定, 半实物仿真

Abstract: A new type of wavelet neural network and grey prediction control strategy is proposed for the complex nonlinearity of some artillery servo system of electric load simulator and the influence of extra torque on the system. The control strategy is mainly composed of a variable structure wavelet neural network controller with particle swarm optimization and a grey prediction compensator(GPC). The variable structure wavelet neural network controller optimizes the parameters of wavelet neural network with particle swarm optimization(PSO) to speed up the convergence of the system, and changes the number of hiden neurons using the self-learning algorithm dynamically to reduce the calculation complexity and improve the dynamic and static performances of the system. The grey prediction compensator is constructed based on the stability of the system in the sense of Lyapunov, which predicts the input torque deviation and further improves the stability and accuracy of the system. The hardware-in-the-loop simulation results show that the hybrid control strategy has strong robustness and high control precision and ensures the stability and anti-interference ability of the system under dynamic load.

Key words: ordnance science and technology, grey prediction, particle swarm optimization algorithm, wavelet neural network, variable structure, Lyapunov stability, hardware-in-the-loop simulation