Vehicle detection is one of the fundamental tasks in the ITS (intelligent transportation system). Although automatic vehicle sensors have already been deployed to retrieve real-time traffic data, these sensors suffer from high deployment costs, resulting in limited deployment at high traffic roads. This paper presents a new approach of vehicle sensing using sidewalk microphones. We install stereo microphones at one side of a road and draw a sound map, i.e., a time-difference map of vehicle sound on the two microphones. A state-machine based algorithm analyzes vehicle passing drawn a sound map to detect vehicles. A simple image processing technique is applied to a sound map prior to analysis to reduce false negative detections. Experimental evaluations demonstrated that our vehicle sensing system successfully detected vehicles with an F-measure of 0.92. This paper also presents big challenges we are facing to extend our system to multi-lane high traffic roads.
|Title of host publication||Smart Sensors and Systems|
|Subtitle of host publication||Technology Advancement and Application Demonstrations|
|Publisher||Springer International Publishing|
|Number of pages||14|
|Publication status||Published - Jan 1 2020|
All Science Journal Classification (ASJC) codes