28. Aachener Kolloquium Fahrzeug- und Motorentechnik 2019
Sound.AI – Teaching Vehicles How to Hear
Autoren
Markus Strobel, ZF Friedrichshafen AG, Friedrichshafen;
Thomas Keutgens, Christoph Klas, Dr.-Ing. Adrian Zlocki, fka GmbH, Aachen;
Michael Kwade, Thomas Böttcher, Prof. Dr.-Ing. Lutz Eckstein, Institut für Kraftfahrzeuge (ika), RWTH Aachen University
Zusammenfassung
Self-driving cars are by law required to grant the right of way to emergency vehicles. There exist several possibilities to detect emergency vehicles like optical sensors, vehicle-to-vehicle communication and acoustic sensors. Since there is currently and in the near future no sufficient penetration of vehicle-to-vehicle communication for all emergency vehicles, this option is not available for current self-driving vehicles. In order to enable self-driving and automated cars to detect emergency vehicles, ZF and fka tried to estimate the potential of acoustic sensors in comparison to optical vehicle sensors to do a faster classification for intersection scenarios. Because of the lack of commercially available automotive grade acoustic sensors, who are able to detect emergency vehicle sirens, the answering of this question leads to the development of such an acoustic sensor. Within this paper, a method is presented on how to develop an acoustic sensing system, capable of the detection and localization of emergency vehicle sirens. The method presented here includes the requirements for the acoustic sensors, the determination of the sensor positions, the development of a “Martinshorn” detection and the direction detection where the siren is coming from.
Mitglieder des Österreichischen Vereins für Kraftfahrzeugtechnik haben Zugriff auf alle Vorträge der Internationalen Wiener Motorensymposien.