Mobile target tracking in non-overlapping wireless visual sensor Networks using Neural Networks

Target tracking using Wireless Visual Sensors Networks (WVSN), is an interesting research area, especially if visual sensors have non-overlapping Field-Of-Views (FOV). In this paper, we propose a new prediction based method to efficient sensor selection. These method uses a Neural Network (NN) to predict the next target movement. We implemented and tested this tracking approach in a flat environment, simulation shows this approach is efficient non overlapping tracking with acceptable accuracy, configuration of WVSN cost and energy conservation.