北京航空航天大学学报

• •    

视觉导航下基于H2&H∞的航迹跟踪

吴成良1,曹云峰2,庄丽葵1,谢也1,丁萌1   

  1. 1. 南京航空航天大学
    2. 南京航空航天大学航天学院
  • 收稿日期:2015-06-05 修回日期:2015-09-18 发布日期:2015-10-14
  • 通讯作者: 曹云峰

Path Tracking for Visual Navigation Based on H2&H∞

  • Received:2015-06-05 Revised:2015-09-18 Published:2015-10-14

摘要: 为了从理论上研究视觉导航系统的特性,首先在物理上将视觉导航系统划分为传感器(摄像头、激光测距仪)和视觉处理计算机。针对视觉处理计算机,建立视觉测量模型。将视觉位置测量模型的不确定性分为参数不确定性和输入干扰两种,通过仿真详细分析了视觉位置估计针对不同模型不确定性的敏感程度。最后在引入视觉测量模型不确定性的基础上研究H2/H∞混合鲁棒控制在航迹跟踪上的应用,并给出了保证航迹控制系统稳定性和跟踪性能情况下的视觉测量模型摄动范围,对计算机视觉算法的研究具有一定的指导意义。

关键词: 视觉导航, 视觉测量模型, 模型不确定性, H2&H∞混合控制, 鲁棒性能

Abstract: Visual navigation system was physically divided into sensors (camera and laser rangefinder) and computer vision processing, so as to research the characteristics of visual navigation system theoretically. And then visual measurement model was established, which was based on the above computer vision processing. The uncertainties of visual position measurement model were divided into parameters uncertainties and input interference. The sensitivity degrees that triggered by different kinds of model uncertainties in visual position estimation were analyzed in detail. Finally, based on the introduction of visual measurement model uncertainty, visual measurement model perturbation ranges, under the constraint conditions of both track control system stability and tracking performance, were given in the study of the hybrid robust control application on trajectory tracking; which was of certain guiding significance to computer vision algorithm。

Key words: visual navigation, visual measurement model, model uncertainty, H2&H∞ hybrid control, robust performance

中图分类号: 


版权所有 © 《北京航空航天大学学报》编辑部
通讯地址:北京市海淀区学院路37号 北京航空航天大学学报编辑部 邮编:100191 E-mail:jbuaa@buaa.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发