基于参数变化粒子群算法的MAF传感器建模

Modeling of the MAF sensor using PSO with changeable parameters

在用Hammerstein模型描述热膜式空气质量流量(MAF)传感器时,应用多项式回归分析建立其静态非线性环节的模型,应用参数线性变化的粒子群优化(PSO)算法建立其动态线性环节的模型。文章给出PSO算法的适应度函数及算法流程,并说明了参数设置的方法。研究表明,与基本粒子群算法相比,参数线性变化粒子群算法的建模精度及收敛速度有很大提高。应用参数变化粒子群算法进行传感器动态建模是非常有效的。

The hot-film mass air flow （MAF） sensor is described by the Hamlnerstein model. The polynomial regression analysis is used to fit the expression of the static non-linear part, and the particle swarm optimization （PSO） with changeable parameters is utilized to build the model of the dynamic linear part. The fitness function of PSO is presented, the computation procedure is introduced, and the method of setting parameters is explained in this paper. The modeling results show that both the modeling precision and the convergence speed of PSO with changeable parameters are better than those of basic PSO. It is very effective using PSO with changeable parameters to build the dynamic models of sensors.