Biologically inspired, self organizing communication networks.

Abstract

The problem of energy-efficient, reliable, accurate and self-organized target tracking in Wireless Sensor Networks (WSNs) is considered for sensor nodes with limited physical resources and abrupt manoeuvring mobile targets. A biologically inspired, adaptive multi-sensor scheme is proposed for collaborative Single Target Tracking (STT) and Multi-Target Tracking (MTT). Behavioural data obtained while tracking the targets including the targets’ previous locations is recorded as metadata to compute the target sampling interval, target importance and local monitoring interval so that tracking continuity and energy-efficiency are improved. The subsequent sensor groups that track the targets are selected proactively according to the information associated with the predicted target location probability such that the overall tracking performance is optimized or nearly-optimized. One sensor node from each of the selected groups is elected as a main node for management operations so that energy efficiency and load balancing are improved. A decision algorithm is proposed to allow the “conflict” nodes that are located in the sensing areas of more than one target at the same time to decide their preferred target according to the target importance and the distance to the target. A tracking recovery mechanism is developed to provide the tracking reliability in the event of target loss. The problem of task mapping and scheduling in WSNs is also considered. A Biological Independent Task Allocation (BITA) algorithm and a Biological Task Mapping and Scheduling (BTMS) algorithm are developed to execute an application using a group of sensor nodes. BITA, BTMS and the functional specialization of the sensor groups in target tracking are all inspired from biological behaviours of differentiation in zygote formation. Simulation results show that compared with other well-known schemes, the proposed tracking, task mapping and scheduling schemes can provide a significant improvement in energy-efficiency and computational time, whilst maintaining acceptable accuracy and seamless tracking, even with abrupt manoeuvring targets.