Robust Dead Reckoning System for Mobile Robots Based on Particle Filter and Raw Range Scan

Sensors

Zixing Cai¹  Huaqing Min²  Zhuohua Duan³ 

[1] School of Information Science and Engineering, Central South University, No. 204, Minzhu Building,Changsha 410083, China;School of Software, South China University of Technology, Higher Education Mega Center, Guangzhou 510006, China;Zhongshan Institute, University of Electronic Science and Technology of China, No. 713,Mingde Building, Zhongshan 528400, China;

关键词: mobile robots;    fault diagnosis;    robust dead reckoning;    particle filters;    raw scan matching;   

DOI  :  10.3390/s140916532

来源: DOAJ

【 摘 要 】

Robust dead reckoning is a complicated problem for wheeled mobile robots (WMRs), where the robots are faulty, such as the sticking of sensors or the slippage of wheels, for the discrete fault models and the continuous states have to be estimated simultaneously to reach a reliable fault diagnosis and accurate dead reckoning. Particle filters are one of the most promising approaches to handle hybrid system estimation problems, and they have also been widely used in many WMRs applications, such as pose tracking, SLAM, video tracking, fault identification, etc. In this paper, the readings of a laser range finder, which may be also interfered with by noises, are used to reach accurate dead reckoning. The main contribution is that a systematic method to implement fault diagnosis and dead reckoning in a particle filter framework concurrently is proposed. Firstly, the perception model of a laser range finder is given, where the raw scan may be faulty. Secondly, the kinematics of the normal model and different fault models for WMRs are given. Thirdly, the particle filter for fault diagnosis and dead reckoning is discussed. At last, experiments and analyses are reported to show the accuracy and efficiency of the presented method.

【 授权许可】

Unknown