Artificial Neural Networks in WSNs design: Mobility prediction for barrier coverage

Abstract

Barrier coverage provides intrusion detection for various national security applications. If the network is randomly deployed, in moderately dense networks, the full end-to-end barrier line might not be provided. To fill the breaks and to assure the intrusion detection, additional nodes have to be introduced. The network should be designed in a way that enables the good (cost/benefit) balance between the number of initially deployed static nodes and the (added) mobile nodes. This research, for the first time introduces the artificial neural networks (ANNs) in predicting the number of the additionally supplied static nodes or simultaneously deployed mobile nodes for barrier coverage setup after the network's initial installation. The results show a high degree of predictability, with the R-factor of over 0.99 regarding the test data. Besides its primary results, the importance of the research relies also in fact that the approach can be extended to the prediction of k-barrier coverage, the mobility range, and to the other network design objectives.