SiRF’s Federated Filter Architecture for Automotive Dead Reckoning

Keith Brodie

Abstract:	The urban canyon is a challenging environment for automotive GPS because of blockage and multipath. SiRF Technology undertook the SiRFDRiveÔ program to improve automotive navigation in the urban canyon by integrating GPS with a low cost gyroscope and the automobile's onboard speed sensor. The primary objectives were to allow the receiver to navigate through a two minute GPS outage with less than 150 m (2s) of horizontal position error growth and to minimize the effects of multipath in the urban canyon with a minimal increase in throughput and memory requirements. Previous studies demonstrate that the performance of a tightly integrated, single Kalman filter navigation system design can be very nearly matched by a federated architecture. The SiRF federated filter architecture uses separate GPS-navigation, heading (3-state) and speed (2-state) Kalman filters. These instrument filters are updated with the GPS velocity estimate from the primary navigation filter. The gyroscope and vehicle speed sensors are used to propagate the state continuously once calibrated, allowing for a 20-hertz navigation output rate. At start-up SiRFDRive uses a zero-velocity gyro bias update to enable accurate navigation before GPS velocity data is available. A SiRFDRive equipped receiver can navigate immediately after power-up, using only stored initial heading and position, vehicle speed sensor measurements, and the zero-velocity calibrated gyroscope measurements. This mechanism provides the navigation solution in a blocked GPS environment, such as a parking garage, immediately after starting the car and overcomes changes in gyro bias induced by large temperature changes that are common in the automotive environment. The SiRFDRive system has been implemented and tested, and meets are horizontal position accuracy objective. The federated filter architecture requires an increase of less than 15% in throughput, less than 10% in code space, and less than 5% in RAM.
Published in:	Proceedings of the 12th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GPS 1999) September 14 - 17, 1999 Nashville, TN
Pages:	1251 - 1258
Cite this article:	Brodie, Keith, "SiRF’s Federated Filter Architecture for Automotive Dead Reckoning," Proceedings of the 12th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GPS 1999), Nashville, TN, September 1999, pp. 1251-1258.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In