Sensor data fusion can greatly improve the performance of driver assistance and safety systems. One such application is pre-crash, which takes protective action before an imminent collision. The imminent impact is detected through a multi-target multi-sensor tracking system, from which a time-to-collision (TTC) is determined. However, due to different sensor characteristics, the measurements from the sensors can arrive out-of-sequence. In time critical applications, it is vital to efficiently process incoming measurements, despite this out-of-sequence problem. This paper presents the application of advanced algorithms for out-of-sequence measurement (OOSM) processing in an automotive safety system. The different algorithms are implemented on a test vehicle, where real collision tests were carried out with an impact sensor to evaluate the TTC. Through the advanced OOSM algorithms, the TTC estimation is more accurate and increases the overall performance of the pre-crash system.
Time-to-collision estimation for pre-crash with out-of-sequence measurements
Berechnung der Zeit bis zur Kollision für das Pre-Crash-System mit nichtsequentiellen Messungen
2010
6 Seiten, 10 Bilder, 2 Tabellen, 7 Quellen
(nicht paginiert)
Aufsatz (Konferenz)
Datenträger
Englisch
Time-to-collision estimation for pre-crash with out-of-sequence measurements
Kraftfahrwesen | 2010
|VEHICLE COLLISION/CRASH POSITION DISCRIMINATION DEVICE
Europäisches Patentamt | 2015
|Collision speed estimation using a single normalised crash depth-impact speed characteristic
Kraftfahrwesen | 1992
|Estimation of Real-Time Crash Risk
Transportation Research Record | 2011
|