Machine Learning Approach with Receiver Autonomous Integrity Monitoring Algorithm for Position Estimation in Urban Scenario
Integrity is one benchmark to assess Global Navigation Satellite System(GNSS)performance,and the aviation industry first initiated it.GNSS is widely used around the world for numerous purposes.Many GNSS-based applications appear in the urban environment,so the integrity algorithm is highlighted by civil users.Several countermeasures,such as ω – testing,multiple-hypothesis RAIM,and advanced RAIM(ARAIM)algorithm,have been implemented to secure the GNSS.These algorithms are not so effective in the applications of urban environments because they normally consider only one or a few number of failures.However,in urban scenarios,limited satellite visibility,multipath effect,and other interferences are the main reasons for GNSS performance degradation.This study proposes a compound strategy based on the RAIM algorithm and a machine learning method to detect numerous failures and determine the user receiver's correct positions in the urban area for land vehicles.RAIM is a powerful tool to detect failures.Still,it remains challenging to detect and identify multiple faulty signals in urban scenarios.Thus,an unsupervised learning method named density-based spatial clustering of applications with noise(DBSCAN)along with RAIM algorithm is associated and finally comes across as correct as possible position estimation.We find that the proposed strategy can work effectively without assumptions on the number of faulty signals and is computationally efficient.
GNSS positioning machine learning failure detection and isolation urban scenario
Nawshin Mannan Proma Xin Chen
Shanghai Jiao Tong University,Shanghai.200240,China
国内会议
北京
英文
1-18
2022-12-01（万方平台首次上网日期，不代表论文的发表时间）