IPSO—BP算法在半主动悬架控制中的应用

为了改善半主动悬架的性能,提出采用改进的粒子群优化(improved particle swarm optimization,IPSO)-向后传播(back propagation,BP)算法作为半主动悬架自适应控制,该算法将标准粒子群算法进行改进,用以改善粒子群全局收敛性和收敛速度,并将改进后的IPSO算法作为BP神经网络的学习算法,用于半主动悬架的自适应控制.自适应控制器采用了双神经网络单元结构,一个作为输入端的控制器,根据路面输入调节半主动悬架阻尼值,另一个作为半主动悬架的辨识器,并进行在线识别.通过该控制器进行半主动悬架自适应控制数值仿真,结果表明,基于该算法的控制器明显改善了汽车的舒适性和平顺性,使得车身的垂向加速度比粒子群优化(particle swarm optimization,PSO)-BP半主动悬架的降低了21.73%,提高了汽车悬架的性能.

Application of IPSO-BP Network Algorithm in Semi-active Suspension Control

To improve the performance of semi-active suspension,the paper proposes an IPSO-BP algorithm as an adaptive semi-active suspension control algorithm.The IPSO algorithm improves the standard PSO algorithm to perfect the convergence rate and the capability of global convergence and is used for the BP neural network learning algorithm for adaptive semi-active suspension control.Adaptive controller has a two-unit structure neural network.One is an input controller,which adjusts the value of semi-active suspension damping in accordance with the road input;and the other is a semi-active suspension identifier which is used for online identification.Through the adaptive control test of the semi-active suspension based on the semi-active suspension controller,results show that the controller based on the IPSO-BP algorithm obviously improves the comfort and ride quality of the car.The vertical acceleration of the car body reduces 21.73% compared with the PSO-BP semi-active suspension.This algorithm improves the vehicle suspension performance.