Lane Level Vehicle Positioning Based on Integrated Multi-sensor Fusion

Abstract

A lane level vehicle positioning system is designed based on integrated multisensor fusion which can be provided by a smart phone. Traditional vehicle positioning is always based on a low-precision global position system (GPS) which has gross error. This paper makes full use of a smart phone, using its sensors such as camera, low-precision GPS and accelerometer, to provide lane level positioning for a car when the lane line can be detected. GPS-based lateral deviation is calculated, with the help of the prior knowledge of roads. Then, multiple lanes are detected using visual approach and vision-based lateral deviation is calculated based on the results of lane detection. Both lateral deviations are fused using a Kalman filter, with the help of accelerometer, to estimate the current location. Experimental results using real vehicle experiments indicate the effectiveness of our system. Comparison between our system and high-precision RTK-DGPS is also given in the experimental results.