Initial Evaluation of Vehicle Type Identification using Roadside Stereo Microphones

A key feature of Intelligent Transport Systems (ITS) is the ability to detect and identify vehicles. In this paper, we put forward a stereo microphone-based system capable of detecting and identifying the type of individually, sequentially, and simultaneously passing vehicles in multi-lane environments based on their sound. We find that our proposed system shows improved performance over single-microphone systems thanks to its improved sequential and successive vehicle detection performance. Initial evaluation results using sound data collected from roads on a university campus show a classification accuracy of 95.01 %.