Continuous precise position estimation is vital for a parking assistance system, which is designed to provide the driver with exact information on how to steer the vehicle into the desired parking spot. The possibility of computing the actual position of a passenger car based on different independent internal signals available in a production car is investigated. A new front wheel based dead reckoning approach is compared to a model based approach as well as to an EKF (Extended Kalman Filter). Depending on the method the input signals are the wheel revolutions of the front wheels, the steering angle and the yaw rate. It was clearly shown, that front axle based dead reckoning, the model based position estimation algorithm and the EKF approach are to a simple rear axle based dead reckoning. Modern passenger cars usually provide all signals, for the advanced algorithms. Experiments have proven that basically all position estimation algorithms yield reasonable results for the parking assistance system. provide position estimates within an accuracy of 5 cm. Only the EKF shows superior precision of about 1 cm.
Data fusion for precise dead-reckoning of passenger cars
Verarbeitung mehrerer Daten für die genaue Positionsbestimmung von Personenkraftfahrzeugen
2002
6 Seiten, 9 Bilder, 1 Tabelle, 10 Quellen
Conference paper
Storage medium
English
TIBKAT | 1970
|Field Trials of a Combined GPS/Dead-Reckoning System for Cars
British Library Conference Proceedings | 1994
|IEEE | 2020
|Dead reckoning mechanization problems
NTRS | 1970
Adaptive dead reckoning algorithm
British Library Online Contents | 1996
|