Compressive Sensing-Based Clustering Joint Annular Routing Data Gathering Scheme for Wireless Sensor Networks
Yicong Yuan,Wei Liu,Tian Wang,Qingyong Deng,Anfeng Liu,Houbing Song
DOI: https://doi.org/10.1109/access.2019.2935462
IF: 3.9
2019-01-01
IEEE Access
Abstract:Compressed sensing technology is one of the effective techniques to effectively reduce the amount of data transmission in wireless sensor networks. Compressed sensing technology can reduce the amount of data that a node undertakes from n to m, where m ≪ n, but we still hope to further reduce the amount of data that the node bears to improve network lifetime. In this paper, a Compressive Sensing based Clustering Joint Annular Routing Data Gathering (CS-CARDG) scheme is proposed to improve the network life. The key technology adopted by CS-CARDG scheme is: data is collected by cluster. The network first forms a cluster, and each node in the cluster sends the data packet to the cluster head. Each cluster forms m-dimensional data according to the requirements of the compressed sensing technology to ensure that the data can be recovered. When the cluster head node routes the m-dimensional data to the sink, the CS-CARDG scheme adopts a two-stage routing scheme with the same ring routing and shortest path that is completely different from the previous scheme. The same ring routing means that the cluster heads with the same number of sink hops are routed around the ring for one week to route the compressed sensing data of the same ring to a node in the ring. In this way, each sub-dimension data in the same ring is routed to the corresponding node of each ring through the same-loop route, and then the shortest-circuit strategy of the second phase is started. That is, from the outermost ring, the same fractal data is sequentially compressed from the outside to the inside, and is routed to the sink by the shortest path. In this round of data collection, the number of data packets that the nodes in the near-sink one-hop range bears is only m, and the nodes in the near-sink region directly send data to the sink node, thereby reducing the amount of data that the node bears to m/k + 1, where k is the number of nodes within the node's broadcast radius.. In this paper, the compressed sensing strategies proposed in the past are compared by detailed theoretical analysis. The theoretical analysis results show that the CS-CARDG strategy proposed in this paper can effectively reduce the amount of data carried by nodes. This scheme reduces the amount of data in the network from (1+r/l2) λ + (1+r/l3) λ + ··· + m/k+d to m/k+1. Reduced the amount of data by at least 20%. In the network with R=480 m, the energy utilization rate can reach more than 90%.