测绘学报 ›› 2018, Vol. 47 ›› Issue (11): 1446-1456.doi: 10.11947/j.AGCS.2018.20170649

• 大地测量学与导航 • 上一篇    下一篇

移动机器人视觉动态定位的稳健高斯混合模型

程传奇1, 郝向阳2, 李建胜2, 胡鹏2, 张旭2   

  1. 1. 武警工程大学, 新疆 乌鲁木齐 830000;
    2. 信息工程大学地理空间信息学院, 河南 郑州 450001
  • 收稿日期:2017-11-15 修回日期:2018-05-29 出版日期:2018-11-20 发布日期:2018-11-29
  • 通讯作者: 郝向阳 E-mail:xiangyanghao2004@163.com
  • 作者简介:程传奇(1989-),男,博士,讲师,研究方向为计算机视觉、导航定位与位置服务。E-mail:legend3q@163.com
  • 基金资助:
    国家863计划(2015AA7034057A)

Robust Gaussian Mixture Model for Mobile Robots' Vision-based Kinematical Localization

CHENG Chuanqi1, HAO Xiangyang2, LI Jiansheng2, HU Peng2, ZHANG Xu2   

  1. 1. Engineering University of PAP, Urumqi 830000, China;
    2. Institute of Geographical Spatial Information, Information Engineering University, Zhengzhou 450001, China
  • Received:2017-11-15 Revised:2018-05-29 Online:2018-11-20 Published:2018-11-29
  • Supported by:
    The National High-tech Research and Development Program of China (No. 2015AA7034057A)

摘要: 针对动态场景中运动路标点严重影响传统视觉自主定位算法精度,甚至产生定位失效的问题,提出一种顾及动态路标点的稳健高斯混合模型。在传统图优化视觉定位模型的基础上,增加“运动指数”描述图优化模型中路标点的运动概率,把传统图优化高斯模型增强为高斯混合模型,以约束运动路标点对图优化结果的影响;为增强模型对噪声的稳健性,采用方差膨胀模型约束残差方程;详细推导了该高斯混合模型的期望-最大化求解方法,把该问题转化为经典迭代最小二乘问题进行解算。仿真试验和真实数据试验表明:强动态场景中,提出的算法绝对精度指标和相对精度指标均优于传统优化算法;静态或弱动态场景中,提出的算法仍与传统优化算法定位性能相当。本文方法可有效减小场景中运动路标点对优化结果的影响,更适用于移动机器人的自主定位。

关键词: 视觉定位, 图优化, 动态路标, 方差膨胀, EM算法

Abstract: In dynamic environments,the moving landmarks can make the accuracy of traditional vision-based localization worse or even failure.To solve this problem,a robust Gaussian mixture model for vision-based localization with dynamic landmarks is proposed.The motion index is added to the traditional graph-based vision-based localization model to describe landmarks' moving probability,changing the classic Gaussian model to Gaussian mixture model,which can reduce the influence of moving landmarks for optimization results.To improve the algorithm's robustness to noise,the covariance inflation model is employed in residual equations.The expectation maximization method for solving the Gaussian mixture problem is derived in detail,transforming the problem into classic iterative least square problem.Experimental results demonstrate that in dynamic environments,the proposed algorithm outperforms the traditional method both in absolute accuracy and relative accuracy,while maintains high accuracy in static environments.The proposed method can effectively reduce the influence of the moving landmarks in dynamic environments,which is more suitable for the autonomous localization of mobile robots.

Key words: vision-based localization, graph optimization, dynamic landmarks, covariance inflation, expectation maximization

中图分类号: