PMT2: A Predictive Mobile Target Tracking Algorithm in Wireless Multimedia Sensor Networks

In this work, we propose a new Predictive-based Mobile Target Tracking Algorithm for Wireless Multimedia Sensor Networks called PMT . Resource management being a critical feature of this kind of networks, the main aim of PMT 2 is to handle the trade-off between the accuracy of the tracking and the energy conservation. Prediction approach seems to be the best candidate to reach this objective. For this purpose, we introduce an enhanced version of the Extended Kalman Filter combined with a change detection mechanism named CuSum for Cumulative Summary. We also propose a deployment strategy to improve the efficiency of the tracking algorithm. Using simulations, we show the performances of the proposed coupled mechanism in the trajectory prediction and in the reactivity to abrupt direction changes. Moreover, we perform a comparative study between PMT 2 and existing works: 1) BASIC where all the Cameras Sensors are always in active mode; 2) OCNS for Optimal Camera Node Selection, a cluster-based solution with a probabilistic sensor selection; 3) PTA, another predictive solution based on standard Kalman Filter. The obtained results illustrate that PMT 2 improves the quality of tracking by up to 35% compared to existing works, while reducing energy consumption by up to 55%.