Home // VEHICULAR 2014, The Third International Conference on Advances in Vehicular Systems, Technologies and Applications // View article

Nonlinear Error Modeling of Reduced GPS/INS Vehicular Tracking Systems Using Fast Orthogonal Search

Authors:
Mohamed Maher Atia
Aboelmagd Noureldin
Michael Korenberg

Keywords: Land vehicles Tracking; Reduced IMU; GPS; INS/GPS integration.

Abstract:
Land Vehicle Tracking systems depend mainly on Global Navigation Satellite Systems (GNSS), such as Global Positioning System (GPS). However, GNSS suffer from signal blockage and degradation in urban areas. At the same time, most land vehicles, nowadays, come with low-cost low-power Inertial Measurement Units (IMU). Although these IMU can be used an accurate short-term tracking system using Inertial Navigation Systems (INS) technology, they are currently mostly used only for safety applications. This paper proposes an enhanced land-vehicles tracking system by integrating a reduced IMU system with GPS to enhance the tracking accuracy of land vehicles in downtown and urban areas. Commonly, GPS/INS integration is based on Kalman Filter (KF), where a linearized dynamic models for INS errors is utilized. If Low-Cost MEMS-based inertial sensors with complex stochastic error nonlinearity are used, performance degrades significantly during short periods of GPS-outages. This paper presents a nonlinear INS-errors modelling using a fast nonlinear identification technique called fast orthogonal search (FOS). During reliable GPS coverage, the corrected vehicle state and sensors measurements are input to FOS and the FOS models outputs are trained to predict the INS deviations from GPS. During GPS-outages in urban areas, the trained FOS models along with the most recent vehicle state are used to predict INS deviations from GPS. The predicted INS deviations are then feedback to the system Kalman Filter, as updates to estimate all INS errors. The experimental setup of this work used a very low-cost IMU from Crossbow Inc. (USA based), the vehicle odometer measurements along with a GPS receiver from Novatel, Inc. (Canada based). Experiments were performed in Kingston, Ontario, Canada. Initial results show promising improvement of tracking accuracy in challenging GNSS-denied areas.

Pages: 16 to 21

Copyright: Copyright (c) IARIA, 2014

Publication date: June 22, 2014

Published in: conference

ISSN: 2327-2058

ISBN: 978-1-61208-348-3

Location: Seville, Spain

Dates: from June 22, 2014 to June 26, 2014