Adaptive sampling for energy-efficient collaborative multi-target tracking in wireless sensor networks

Abstract

The problem of energy-efficient multi-target tracking (MTT) in wireless sensor networks is considered for sensor nodes with limited energy resources and sharp manoeuvring targets of different classes. A distributed multi-sensor multi-target tracking scheme is proposed for energy-efficient MTT with adaptive sampling. Behavioural data obtained while tracking the target including the target's previous locations are recorded as metadata to compute the sampling interval so that the tracking continuity and energy efficiency are improved. Following this, the next tasking sensors group is selected proactively according to the predicted target location probability distribution. A ‘main node’ is elected from the selected tasking sensors so that the energy efficiency is improved. Sensor nodes that detect more than one target at the same time determine their preferred target according to the target importance and the distance to the target. Simulation results show that compared to other well-known target tracking schemes, the proposed scheme can provide a significant improvement in energy efficiency while maintaining acceptable accuracy and seamless tracking, even with sharp manoeuvring targets. Additionally, more important targets experience better tracking accuracy.