RBF神经网络算法在汽油机空燃比控制中的比较

Algorithm Comparison of RBF Neural Networks for Air-Fuel Ratio Control of Gasolene Engine

选择具有多干扰和复杂非线性的汽油机空燃比为控制对象,对径向基函数(RBF)神经网络的RLS和LMS在线算法,从训练速度、抗干扰能力、控制精度三个方面,在MATLAB下进行了仿真比较。结果表明:在用离线训练的网络参数作为初始值、且不调整RBF中心值和宽度向量时,RLS算法优于LMS算法。

RLS algorithm is compared with LMS algorithm on the speed,inti-interfere capability and control precision for Gasdene engine of AFR that is used in mul-interfere and complex offline in similnked contrlling.The simulations show that RLS algorithm is superior to LMS algorithm in case of real-time learning used nontime learning of datum and without changing center value and width vector during realtime learning in the anrticle.