Repository hosted by TU Delft Library

Home · Contact · About · Disclaimer ·
 

Vehicle state estimation using GPS/IMU integration

Publication files not online:

Author: Wang, Y. · Mangnus, J. · Kostić, D. · Nijmeijer, H. · Jansen, S.T.H.
Type:article
Date:2011
Source:10th IEEE SENSORS Conference 2011, SENSORS 2011, 28 October 2011 through 31 October 2011, Limerick, Ireland, 1815-1818
series:
Proceedings of IEEE Sensors
Identifier: 462209
ISBN: 9781424492886
Article number: No.: 6127142
Keywords: Electronics · Adaptive kalman filter · Automated tuning · Automotive applications · Driver support system · Driving test · GPS/IMU · Ground stations · High quality · Instrumented vehicle · Kalman-filtering · Motion sensors · Recorded signals · Satellite navigation · Signal disturbances · State Estimators · System specification · Vehicle model · Vehicle motion · Vehicle position · Automobiles · Global positioning system · Kalman filters · Sensors · Specifications · State estimation · Vehicles · Mobility · Mechatronics, Mechanics & Materials · IVS - Integrated Vehicle Safety · TS - Technical Sciences

Abstract

New driver support systems require knowledge of the vehicle position with great accuracy and reliability. Satellite navigation (GNSS) is generally insufficiently accurate for positioning and as an alternative to using a ground station, combinations with high quality motion sensors are used in so-called Inertial Navigation Systems. However the system specifications and related cost are not suitable for Automotive applications. In this article a Vehicle model based concept is presented in a state estimator setup that will use signals that are available on modern vehicles. An extension of a commonly used Bicycle representation of the vehicle is applied with an automated tuning for signal disturbances. For coping with different update frequencies from GNSS and motion sensors a Bezier extrapolation is used. The resulting Adaptive Kalman Filter approach is compared to recorded signals from driving tests with an instrumented vehicle. The comparison shows that with the new setup a clear improvement is achieved for the vehicle motions compared to more commonly used Kalman filtering. This verifies that sensor disturbances can better be compensated with the presented concept, and also better results for positioning can be expected. © 2011 IEEE.