A fused Global Navigation Satellite System with an Inertial Navigation System (GNSS/INS System) for automotive applications may operate with reduced degrees of freedom due to car dynamics. Approaches to use these constraints as pseudo measurements (non holonomic constraints) of a GNSS/INS system have shown an improvement of the navigation solution (position, velocity, and attitude). Applied constraints reduce the drift and error in several states and at the same time reduce the remaining information content of these states. This paper explores robustness improvements of GNSS/INS navigation filters analyzing simulative and real sensor data. For this purpose state depending constraints and GNSS raw data weighting are investigated. The paper compares a common Tightly Coupled filter with a state reduced Tightly Coupled GNSS/INS filter. The filters use onboard sensor information including wheel odometer measurements and a reduced inertial sensor set (RISS). In automotive applications costs of sensors and their integration are essential. Models using already integrated infrastructure are preferable and their usability is investigated. Model constraints depending on the state of the car (curve, straight road) are applied to both filters. Especially during reduced GPS availability the car dependent constraints and odometer information significantly aid the navigation solution. The evaluation includes at the one hand the processing of simulated sensor data. The artificial inertial data is provided by a professional car dynamic simulation tool in order to model accurately automotive trajectories and its sensor data. On the other hand real time automotive sensor data recorded in suburban scenarios are evaluated. The simulated sensor data corresponds to these scenarios concerning the dynamics. Real time GPS raw data (pseudorange measurements) vary in their quality depending on the surrounding. Common raw data weighting and gating are applied to obtain robust filter performance.
Robustness improvements and comparisons of automotive GPS-INS navigation filters in suburban scenarios fusing automotive on-board sensors
2012
14 Seiten, 7 Bilder, 7 Quellen
Conference paper
English
ENHANCED LAND VEHICLES NAVIGATION BY FUSING AUTOMOTIVE RADAR AND SPEEDOMETER DATA
British Library Conference Proceedings | 2020
|Tema Archive | 2006
|Tema Archive | 1994
|Springer Verlag | 2014
|TIBKAT | 1979
|