摘要
研究非线性系统辨识问题。针对非线性系统中单输入单输出Hammerstein模型,由于传统辨识方法对Hammerstein模型中非线性部分具有不易辨识的缺陷,造成辨识精度低、辨识效果差等问题。为此,在基本粒子群算法的基础上,提出了一种带有收缩因子的改进的粒子群算法对非线性系统进行辨识的方法,可将参数辨识问题转换为参数空间上的函数优化问题,然后利用粒子群算法的并行搜索能力进行参数寻优。通过MATLAB软件进行仿真,并与基本粒子群算法进行比较,结果表明,利用改进算法不仅提高了辨识精度而且获得了良好的辨识效果,从而验证了算法的有效性和可行性。
Research nonlinear system identification problems. In the basis of the particle swarm algorithm, a new method with a shrinkage factor of the particle swarm algorithm for nonlinear system identification was presented in the paper. First, we changed the parameter identification problem into the function optimization problem on the parameter space. Then we used the particle swarm optimization algorithm' s parallel search capability for the parameter optimiza- tion. The MATLAB simulation software was used to carry out the experiments. The result was compared with that of the particle swarm algorithm, which shows that using the improved algorithm can improve the identification accuracy, obtain good recognition results, and verify the effectiveness and feasibility of the algorithm.
出处
《计算机仿真》
CSCD
北大核心
2013年第3期269-272,共4页
Computer Simulation
关键词
系统辨识
粒子群优化算法
非线性系统模型
System identification
Particle swarm optimization algorithm
Nonlinear system model