An Automotive-grade INS/HD Maps/Odometer/GNSS Integration Scheme for Lane-level Navigation Application in Urban Area

Abstract

Advanced Driver Assistance System (ADAS) technology is increasingly expanding recently which primary core is to get accurate location information of vehicles Positioning and Orientation System (POS) which integrates Global Navigation Satellite System (GNSS) and Inertial Navigation System (INS) is widely used to complete outdoor navigation mission However GNSS positioning is sensitively affected by satellite shielding and signal reflection by buildings or other obstacles in urban canyon which cause multipath interference and Non-Line-Of-Sight (NLOS) reception The degraded GNSS would affect the performance of POS after integration process In addition the error model of those influences is significantly varied with surroundings and cannot be reduced entirely by physical or mathematical methods This research utilizes the sophisticated Mobile Mapping System (MMS) to establish the High-Definition Maps (HD Maps) which is the new component in self-driving technology as reference information for vehicular navigation process The high precision maps provide information such as the horizontal position of centerline of lane height value road network connection and road curvature etc First the proposed HD maps searching mechanism for selecting the information from Autoware map format Next the HD maps aided GNSS optimization scheme can remove or mitigate the error pseudorange measurements and positions due to multipath and NLOS reception then feedback to navigation process by height constraints system which can significantly improve the accuracy in both horizontal and vertical direction Last the proposed INS/HD maps/Odometer/GNSS integration scheme applies the Extended Kalman filter (EKF) in Loosely-Coupled (LC) architecture to obtain seamless navigation states which the optimized GNSS can provide reliable update measurements Moreover the horizontal position height and heading update process derived from HD maps are applied to improve the overall performance The experiment utilizes automotive-grade GPS/Galileo/QZSS L1/L5 receiver in GNSS-hostile urban environments Besides the code-based differential GNSS (DGNSS) positioning is implemented with HD maps aiding information Then the tactical-grade Inertial Measurement Unit (IMU) which most HAD vehicular navigation applies recently is applied in the proposed integration scheme The results in dense urban area indicate that the INS/HD maps/Odometer/GNSS integration scheme can achieve 1 5 meters in the horizontal direction and less than half a meter in the vertical direction in terms of Root Mean Square Error (RMSE) The lane-level maps searching correct rate is more than 93% which can provide credible reference most of the time On account of this research using HD maps as reference information in vehicular navigation process can provide a better performance which makes ADAS driving more robust and feasible in urban area

Date of Award	2019
Original language	English
Supervisor	Kai-Wei Chiang (Supervisor)