Tracking Acoustic Target in Wireless Sensor Network Using Adaptive Particle Filter

Abstract

Since acoustic signal is ubiquitous in nature and can be measured easily by node sensors, the moving object equipped with an acoustic generator is usually regarded as the tracked target in surveillance region. And tracking acoustic target is equal to processing the data of acoustic power obtained from each node in Wireless Sensor Networks (WSNs). It is well known that the Particle Filter (PF) is well suited to be used in this kind of target-tracking application. However, the common particle filter cannot keep a high quality performance in tracking moving target, especially high maneuvering target, in practice. The main factor contributing to this situation is that it is hard to choose a suitable state model for the unpredictable moving trajectory. So we present an Adaptive Particle Filter (APF) algorithm in this paper. This algorithm employs a Based on Applied Force (BAF) model to capture the complicated trajectory, and it can adaptively adjust the sampling period and model parameters by analyzing the relationships between particles and measurements. According to our computer simulation by MATLAB, the algorithm can obviously improve the track accuracy, even as the target trajectory is highly maneuverable. The results can prove that adaptive particle filter algorithm is effective and practical.