基于MPSO算法的横动伺服控制系统黑箱模型辨识

Identification of Black box Model of Traverse Servo Control System Based on MPSO

针对标准粒子群算法(PSO)在横动伺服控制系统黑像模型辨识过程中出现的局部收敛问题,提出了一种引入多粒子共享策略(multi-particle information share)的改进粒子群算法(MPSO)辨识方法。首先,建立了系统的五阶传递函数模型,其次,在PSO算法的基础上,引入多粒子信息共享和综合判断来修正各粒子的下一次行动策略,避免粒子趋同陷入局部最优,实现了系统模型的优化。最后,为了验证辨识模型的正确性,进行仿真与实测对比实验,结果表明:该算法辨识出的模型准确度较高,具有较好的控制品质,对于同一速度输入信号,仿真与实测的输出曲线跟随性好,误差在-0.2~0.2rad范围内,误差小。

Aiming at the local convergence problem of standard particle swarm optimization（PSO）algorithm in the identification of the black image model of the servo control system,an improved particle swarm optimization（MPSO）algorithm based on multi-particle information share is proposed.First of all,five-order transfer function model of the system was set up.Secondly,based on the PSO algorithm,multi-particle information share and comprehensive judgment were introduced to correct the next action strategy of each particle so as to avoid particle convergence into local optimum and realize the optimization of system model.Finally,in order to validate the correctness of identification model,simulation and actual measurement contrast experiment was done.The experimental results show that：the model identified by this algorithm is accurate and reliable,with good control quality.For the same speed input signal,the output curve of simulation and actual measurement has good tracking property,with a small error（within-0.2～0.2rad）.