Acoustic Detection of Wrong-Way Drivers using Microphone Arrays

Reliable traffic surveillance is of upcoming interest to meet safety and security standards on roads and highways. Especially wrong-way drivers must be detected carefully to avoid severe damages to health and property. Automatic traffic surveillance methods are therefore an appropriate method to get information about the traffic situation on highways A few 'non-intrusive' systems became more prominent in the last years whereas in most cases video image detection technologies or microwave techniques were applied. Especially at fog or poor sight the video systems have no chance to offer reliable results. Therefore an alternative acoustic system based on a microphone array was developed to obtain more stable and reliable results. The microphone array was mounted overhead on a gantry and covered the whole lane of the highway. This paper is about the development of an acoustic system for detecting wrong-way drivers. The system is based on a microphone array (beamformer) with a consecutive vehicle and driving-direction detector. The first step in the work was the development of a beamformer simulation, to evaluate the theoretical performance of different microphone-array designs. After the designing process of the array a lot of measurements were done on freeways in Austria to get enough data for testing the algorithms. The evaluation results show that the performance of the algorithms is really excellent. Nevertheless the number of false alarms has to be reduced in future.