水下机器人动力学系统的小波辨识算法  

System Identification of Underwater Vehicle Dynamic Model Based on Wavelet Approximation

作　　者：邓志刚[1,2] 朱大奇[2] 方建安[1] 

机构地区：[1]东华大学信息科学与技术学院,上海201620 [2]上海海事大学水下机器人与智能系统实验室,上海201306

出　　处：《系统仿真学报》2014年第11期2546-2551,共6页Journal of System Simulation

基　　金：国家自然科学基金(51279098);上海市优秀学术带头人计划(11XD1402500);上海海事大学基金项目(20120106)

摘　　要：动力学模型辨识是机器人控制设计的基础,是水下机器人研究的核心内容之一。以Falcon水下机器人的动力学模型为研究对象,在模型适当简化的基础上,提出基于小波级数模型的水下机器人动力学在线辨识方法,选取DOG(Derivative of Gaussian)小波作为小波函数,对Falcon水下机器人纵向自由度动力学模型进行自适应辨识。针对水下机器人动力学模型的时变特性,研究负载特性变化情况下的系统辨识,得出了基于DOG小波级数模型的水下机器人动力学在线算法更加有效。The dynamic model identification was the basis of development of control system and one of the most important research areas. The dynamic model of “Falcon” underwater vehicle was taken as the research object. An online identification algorithm for underwater vehicle dynamic model based on wavelet series model was proposed, and it was based on the simplification of the dynamical model The longitudinal dynamic model of ＂Falcon＂ underwater vehicle was identified, using the DOG （Derivative of Gaussian） wavelet as the wavelet function. According to the time-varying character of the dynamic model of underwater vehicle, the identification experiments were presented when the payload was changed. The results show that the identification algorithm based on the DOG wavelet series is more valuable and feasible.

关 键 词：水下机器人 系统辨识 动力学模型 非参数辨识 正交小波级数 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]