Node Scheduling and Compressed Sampling for Event Reporting in WSNs

This work focuses on developing a node scheduling algorithm for detecting events in a sensor field such that few random samples from a set of the active sensor nodes are transmitted to the cluster head and are further used for almost complete reconstruction of the cluster data. A node scheduling algorithm is proposed to achieve maximum coverage of the physical sensor field with correlated sensor readings. Random samples of the correlated data, obtained from the active nodes, are collected at the cluster head using the compressed sensing principle. Targeting the importance of minimum in-network communication, the node scheduling algorithm and the compressed sensing based data gathering, aim at generating random yet correlated sampling matrices for accurate data recovery. A pseudo probabilistic model is proposed to perceive the essential understanding of the monitoring region, ensuring that the joint sensing probability of the event is always more than the predefined threshold $\varepsilon$. Experimental analysis on different sized networks of TelosB motes and extensive simulation analysis demonstrate that the proposed scheme outperforms the existing schemes in terms of average coverage ratio, in-network transmissions and network lifetime.