Detection Algorithm for Multi-Vehicular Target Tracking in Wireless Acoustic Sensor Networks

Abstract

In this paper, we introduce Algorithm for Detection of Multi-targets in Wireless Acoustic Sensor Networks (ADMAN). Wireless acoustic sensors are popular in networked data fusion systems and those also can be applied to vehicular target tracking systems. Even though data association is not possible when we use acoustic sensors (because measurement signal is superimposed signals of multiple sources), we can track vehicular-targets identifiably that are detected. We detach detection part from the whole tracking procedure (usually `tracking' means by both detection and estimation with data association) and provide with detection algorithm for the variable number of vehicular-targets in wireless acoustic sensor networks. ADMAN attaches every target to specific sensor exclusively so that detected vehicular target's location can be known approximately depending on the range of the sensor during the detection procedure. Estimation step can be applied after that and completes the tracking procedure. We focus on only detection part in this paper.