论文部分内容阅读
针对单入单出离散时间非线性动态系统提出一种辨识方法.该方法采用带误差修正的改进泛模型作为非线性系统的结构模型,模型中的时变特征参量及误差修正系数采用粒子群(PSO)算法优化,优化后的模型可以逼近非线性系统.该方法简单、易于实现.通过对Box-Jenkins煤气炉数据等非线性被控对象的仿真研究及对模型的分析,表明了所提出算法的有效性.
Aiming at the single-in-single-out discrete-time nonlinear dynamic system, an identification method is proposed, which adopts an improved pan-model with error correction as the structural model of the nonlinear system. The time-varying characteristic parameters and error correction coefficients in the model adopt the particle swarm optimization PSO) algorithm is optimized, and the optimized model can be approximated to a nonlinear system.The method is simple and easy to implement.Based on the simulation of nonlinear controlled objects such as Box-Jenkins gas furnace data and the analysis of the model, the proposed algorithm Effectiveness.