Fuzzy Decentralized Cooperative Global Localization for a Multi-Ballbot System

• Main Authors: Ye-Min Tan, 陳奕名
• Other Authors: Ching-Chih Tsai
• Format: Others
• Language: en_US
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/92627527418107356806

碩士 === 國立中興大學 === 電機工程學系所 === 104 === This thesis presents two cooperative global localizationmethods usingfuzzy decentralized extended information filtering (FDEIF) and fuzzy decentralized extended Kalman filtering (FDEKF) for a multi-ballbot system composed of several ballbots. When any known landmarks or QR codes are detect by a RGB-D sensor in a ballbot, a pose initialization algorithm is proposed to estimate both start-up positions and orientations of the ballbot, and acooperative localization method is proposed to find initial poses of other ballbots that can be directly detected by a laser scanner mounted at thelocalized ballbot. Once the initial global posesof all ballbots have been roughly determined,both FDEIF and FDEKF dynamic cooperative localization methodsare presented to fuse external measurements from oneRGB-D sensor, or one laser scanner and the four encoders mounted on each ballbot, in order to keep track of the moving poses of all the ballbotscontinuously and precisely. Several simulations and experiments are conducted to show that the proposed methodsare effective in finding accurate estimation of both unknown initial and continuous moving poses of a multi-ballbot system with threefollower ballbots.