Vehicle Counting and Lane Estimation with ad-hoc Microphone Array in Real Road Environments

In this paper, we present a new traffic lane estimation method that involves analyzing the difference between the power peaks of channels in an acoustic traffic monitoring technique based on an ad-hoc microphone array. Then, we apply traffic monitoring systems involving two traffic lane estimation methods to data recorded in various real road environments. With our proposed traffic monitoring systems, we perform channel synchronization based on single source activities, estimate the number of vehicles by analyzing the power envelope of each channel and detecting the peaks, and estimate the traffic lane in which a vehicle is moving by analyzing the time differences or power differences between the power peaks of the channel. We demonstrate the effectiveness of our proposed method by comparing the experimental results obtained with our proposed traffic monitoring systems in various real road environments.