No-signal Area Recognition of Mobile Wireless SensorNetworks Based on Cohomology

In recent years, wireless sensor networks data is taken into consideration as an ideal source, in terms of speed, accuracy and cost, in order to study the Earth's surface. One of the most important challenges in this area, is the signaling network coverage and finding holes. For this purdpose, previous studies used random methods or complex assumptions. Their heavy calculations an time-consuming approaches, prevent algorithms to be implemented to dynamic sensors. This Article has provided a new cost-effective and efficient method of coverage hole detection in wireless sensor networks with no location information, using algebraic topology. The proposed method is a new and unique method that it's tools use no coordinates and assumes. Also in order to reduce costs, the nodes are lacking capabilities such as localizing or navigation. In the proposed method, first, the null space of the first Laplacian is investigated. This space is equal to the first (co)homology group. If the amount is trivial, the network coverage is complete. But when the (co)homology is non-trivial, it is a sign of the existence of holes in network coverage. In other words, the rank of harmonics matrix shows the number of no-coverage areas. In the following, hole location extraction method is discussed using divide and conquer algorithms combined with the use of features of elements in the laplacian matrix null space. Since the method has not been used to diagnose areas without coverage in mobile networks yet, in order to evaluate the results, we estimated the performance of the algorithm in detection of areas without network coverage in a moving sensor network. The final results indicate the powerful performance and accuracy of the proposed algorithm.