基于改进DBO算法的非线性系统辨识  

Identification of Nonlinear Systems Based on Improved DBO Algorithm

作　　者：王鑫鑫 刘朝涛[1] 王正杰 瞿蒋江 WANG Xinxin;LIU Chaotao;WANG Zhengjie;QU JiangJiang(School of Mechatromechanical and Vehicle Engineering,Chongqing Jiaotong University,Chongqing 400074,China)

机构地区：[1]重庆交通大学机电与车辆工程学院,重庆400074

出　　处：《传感器世界》2023年第12期10-14,共5页Sensor World

摘　　要：针对传统方法在单输入单输出Hammerstein模型的辨识上存在辨识精度低、辨识效果差等问题,文章提出一种基于蜣螂优化算法(Dung Beetle Optimizer,DBO)的非线性系统辨识方法。为了克服该算法在局部开发和全局探索上能力不平衡,易陷入局部最优的问题,引入改进的正弦余弦优化算法(Sine Cosine Algorithm with Self-learning strategy and Lévy,SCASL)用于平衡局部和全局搜索阶段,提高算法辨识精度,同时引入莱维飞行(Lévy flight)策略,帮助DBO算法在迭代后期跳出局部最优。通过数值仿真,对蜣螂优化算法和改进的蜣螂优化算法辨识结果进行比较,实验结果表明,改进的蜣螂优化算法辨识速度得到显著提升,并且辨识精度也得到了提高。Aiming at the problems of traditional method in identification of single-input single-output Hammerstein model,such as low identification accuracy and poor identification effect,a nonlinear system identification method based on Dung Beetle Optimizer(DBO)is proposed.In order to overcome the unbalanced ability of the algorithm in local development and global exploration,it is easy to fall into the problem of local optimization.The improved Sine Cosine Algorithm with Self-learning strategy and Lévy(SCASL)is introduced to balance the local and global search stages and improve the identification accuracy of the algorithm.At the same time,Levy flight strategy is introduced to help DBO algorithm jump out of local optimal in the late iteration.Through numerical simulation,the identification results of the dung beetle optimization algorithm and the improved dung beetle optimization algorithm are compared.The experimental results show that the identification speed of the improved dung beetle optimization algorithm is significantly improved,and the identification accuracy is also improved.

关 键 词：非线性系统辨识 HAMMERSTEIN模型 蜣螂优化算法 

分 类 号：TP202.7[自动化与计算机技术—检测技术与自动化装置]