Decentralized tracking an event using optimal coverage sensor networks

A variety of distributed cooperative strategies of mobiles agents are discovered to perform covering and tracking tasks. However, energy is a challenge for them. Energy-out may cause the failure of the mission. In this paper, a decentralized tracking problem of some events is considered by using the optimal coverage sensor networks. We emphasize that the occurrence of the events do not have a uniform probability distribution, so we can design a preferred fixed sensor configuration which can help to maximize the probability of detecting the event. That is to say, we deploy the sensors to the positions which can achieve the optimal coverage corresponding to the probability density function obtained from the statistical data in history, which may minimize the energy consumption. Then the tracking method by using the standard Luenberger observer will be shown under the condition that the events really happen in the considered environment. The novelty of our algorithm is that we use the properties of Voronoi partition to achieve decentralized estimation of the position of the events. In fact, the sensor technical reflected on the quantitative effect of the signal, and the quantitative results of position error indirectly affect the performance of the tracking system. So, what is more significant is that we take the sensor quantitative effect on the system into consideration. Simulation results are provided to demonstrate the validity of the algorithm.