Adaptive Control of A Servo System Based on Multiple Models
参考中译:基于多模型的伺服系统自适应控制


          

刊名：Asian Journal of Control: Affiliated with ACPA, the Asian Control Professors' Association
作者：Gan, Ming-Gang（Beijing Inst Technol, Sch Automat, State Key Lab Intelligent Control & Decis Complex, Beijing 100081, Peoples R China）
Zhang, Meng（Beijing Inst Technol, Sch Automat, State Key Lab Intelligent Control & Decis Complex, Beijing 100081, Peoples R China）
Ma, Hui-Xia（Beijing Inst Technol, Sch Automat, State Key Lab Intelligent Control & Decis Complex, Beijing 100081, Peoples R China）
Chen, Jie（Beijing Inst Technol, Sch Automat, State Key Lab Intelligent Control & Decis Complex, Beijing 100081, Peoples R China）
刊号：737C0022
ISSN：1561-8625
出版年：2016
年卷期：2016, vol.18, no.2
页码：652-663
总页数：12
分类号：TP13
关键词：Parameter estimation；adaptive control；servo systems；filtering；observers
参考中译：参数估计；自适应控制；伺服系统；滤波；观测器
语种：eng
文摘：This paper presents an alternative technology for adaptive control of a DC motor servo system based on multiple models. A dynamic mechanical model of the controlled plant is built, where the unmeasurable variables can be estimated by a filter observer. According to the mechanical model, an adaptive controller is designed. Specific attention is given to the jumping parameters in the control process, which motivate the proposition of multiple models, including fixed models, identified model, and adaptive model, to approximate the global dynamic characteristics of the plant model. A model switching rule is proposed to select the optimal model matching the plant, and the identified and adaptive models are reset when switching occurs, minimizing the effect caused by jumping parameters. Simulation results demonstrate that the introduced scheme is superior to the conventional adaptive control in that it yields a significant improvement of transient stability and response speed as well as steady accuracy, guaranteeing better low-speed performance.
参考中译：本文提出了一种基于多模型的直流电机伺服系统自适应控制的替代技术。建立了被控对象的动态力学模型，不可测变量可由过滤观测器估计。根据建立的力学模型，设计了自适应控制器。特别关注了控制过程中的跳跃参数，这些跳跃参数激发了多个模型的提出，包括固定模型、辨识模型和自适应模型，以逼近对象模型的全局动态特性。提出了一种模型切换规则来选择与被控对象匹配的最优模型，并在发生切换时对辨识模型和自适应模型进行重置，使跳变参数的影响降到最低。仿真结果表明，与传统的自适应控制相比，该方案能显著提高系统的暂态稳定性、响应速度和稳态精度，保证了较好的低速性能。



©2016机械工业出版社（机械工业信息研究院） 京ICP备05055788号-35