A Cross Layer Design for Energy Aware Target Tracking in Wireless Sensor Networks

摘要

In this paper we consider a cross layer approach to design a target tracking wireless sensor network (WSN), and show how unifying network layers can improve the energy efficiency of tracking in WSN. The tracking problem is formulated as a Markov Decision Process (MDP) and solved through learning target trajectories. Though MDPs have been used before for target tracking, we show how is that possible to control physical layer parameters, such as transmit power, to use the inherent communication errors of a wireless channel to our benefit and hence further improve the performance of the tracking network.