Application Study of Passive Multi-Sensor Data Fusion in Flying Robot State Estimation

• Main Authors: Guang-Fong jhong, 鍾光峯
• Other Authors: Li-Wei Fong
• Format: Others
• Language: zh-TW
• Published: 2018
• Online Access: http://ndltd.ncl.edu.tw/handle/xb22a3

碩士 === 育達科技大學 === 資訊管理所 === 106 === In the development of flying robots for shipping services, navigation is one of the most important key technique issues. In order to realize the navigation, the robot must be capable of accurate position estimate. Consider a particular scenario where GPS, ultrasound and laser rangefinder are not used as the positioning devices. By using a passive sensor network, the present research proposes a method for the robot location estimation. The inside of passive sensor network mainly includes multiple landmarks and an angle-measuring sensor installed on the robot. Each landmark knows its position and radiates electromagnetic wave or infrared ray. The onboard sensor receives angular signals from multiple landmarks for producing measurements namely azimuth and elevation. Based on a group of hybrid coordinate filters, a recursive fusion algorithm developed and used to estimate position and velocity parameters of the robot. By utilizing only two landmarks, simulation results of proposed algorithm show that convergence of estimation errors accelerated greatly. The algorithm markedly improves the estimation accuracy as well.