Application of distributed fault detection in WSN to dangerous chemical products based on Bayesian approach

Wireless Sensor Networks (WSN) are employed in several applications, including military, medical, environmental and household applications. Among the most important problems in WSN are their vulnerability to faults and high energy consumption. Both of these problems require the implementation of fault distributed detection methods, capable of distinguishing faulty nodes from normal nodes while taking into account energy problems. Several approaches of distributed fault detection are proposed; for example, we include the Bayesian approach, which ensures a study oscillating between Theory of Signal Detection (TSD), fault tolerance and energy optimization by determining optimal parameters of the detection. It is in this context that our paper is inscribed, where we try to apply this approach in the industrial security´s field. The Bayesian approach ensures a pre study of WSN´s deployment in the dangerous chemical products stores and guarantees a best performing fault detection. The experimental simulations will testify that the proposed algorithm is able to provide a good detection performance while optimizing energy in WSN.