Integrating IMU with OBD to design DR model to improve accuracy of GPS positioning

• Main Authors: Chih-Min Lan, 藍志明
• Other Authors: 黃有評
• Format: Others
• Language: zh-TW
• Online Access: http://ndltd.ncl.edu.tw/handle/7u2m93

碩士 === 國立臺北科技大學 === 電機工程系研究所 === 105 === The Vehicle Navigation System is the most commonly used equipment in the Global Positioning System (GPS). In general condition the Vehicle Navigation System could give the receiver correct position at open locations. However, in actual operation of GPS in downtown areas, underpass roadway or tunnels the accuracy of GPS locations could be influenced by the surroundings area environments. The purpose of this thesis focused on how to overcome the GPS location accuracy by using the OBD (On-Board Diagnostics) that measures the car speed and IMU (Inertial Measurement Unit) to ensure the moving vehicle status. It allows users to trace the moving vehicle data by applying the neural network to establish the DR (Dead Reckoning) model. To solve the GPS shelter problem a model is created to estimate vehicle’s moving distance and direction. Based on the Kalman filter amendment it can provide more accurate and stable positioning of electric vehicle system. Experimental results verify that the combination of OBD, neural network and Kalman Filter had better improvement over the efficiency, accuracy and smoothness on vehicle’s tracking