非线性系统模糊神经网络控制的改进策略

Improving strategies on fuzzy neural network control for nonlinear object

中文关键词  非线性系统  PID型模糊神经网络  最小二乘支持向量机  混沌优化  量子粒子群优化算法

英文关键词  nonlinear system  PID--type fuzzy neural network  least squares support--vector--machine  chaos optimization  quantum-behaved particle swarm optimization algorithm

中文摘要

      针对以模糊神经网络自适应方法为核心的不确定非线性系统控制问题, 以常规静态模糊神经网络控制结构为基础, 分别就控制器、辨识器及优化算法3个方面展开改进研究. 以一种改进结构的动态PID型模糊神经网络为控制器, 最小二乘支持向量机为辨识器构成控制系统. 利用带混沌搜索的量子粒子群算法离线优化结合在线误差反传微调的寻优策略优化控制器参数, 带混沌扰动的粒子群离线优化支持向量机的核参数, 并通过对系统稳定性的讨 论将改进的控制系统逐步完善. 对某热交换对象模型的数值仿真验证了该改进方法的可行性和有效性.

英文摘要

      Based on the conventional control methods, we study and improve the fuzzy-neural-network-adaptive control for a system with unknown nonlinearities. The controller, identifier and optimization algorithm of the scheme are designed respectively by the improved methods. A structure-improved PID--type fuzzy-neural-network is used as the controller, and the least squares support-vector-machine(LS--SVM) is employed as the identifier. The parameters of the controller are optimized by the offline quantum-behaved particle-swarm-optimization(QPSO) with chaos strategy combined with the online-error-back-propagation tuning. The kernel parameters of the LS--SVM are optimized by PSO with chaos optimization. The stability of the improved scheme is discussed in the conclusion section to complete the presentation of the whole design method. Finally, simulation results on a heat exchanger show the feasibility and validity of the designed control system.