Research On Spatial Correlation Clustering Algorithm For Wireless Sensor Networks

Wireless sensor networks (WSNs) is a self-organization, multi-hop network consists of plenty of sensor nodes, routing algorithm is a hot spot in the research of WSNs, considering the energy limit of sensor nodes whose battery can't be recharged, it is important to put emphasis on the design of energy efficient routing. Clustering routing is regarded as a good strategy, compared with the flat routing, clustering algorithm has the simple topological structure, great chance of data aggregation, making the member nodes dormant, etc. Hence it becomes the key issue in the routing research of WSNs.This paper firstly introduces some basic knowledge of WSNs, including the concept, character, application and key research technology. Then it analyzes the routing algorithm in WSNs, particularly introduces the character and principle in clustering algorithm and also discusses some classical clustering mechanisms including LEACH, pointing the advantages and disadvantages which sets the stage for the proposed routing algorithm in this paper.Connecting the existed clustering mechanisms and considering the spatial data correlation in the monitored area, this paper firstly proposes a grid based data spatial correlation clustering mechanism (GSCC), by dividing grids to make the cluster heads well-distributed, GSCC mainly clusters the nodes which have the high similarity. On the basis of GSCC, this paper then discusses about two different data collecting strategies. Firstly it analyzes how to apply the Slepian-Wolf coding in cluster, including rate allocation in member nodes and how to decode in cluster head, then it proposes a data spatial correlation clustering based approximation framework (SCCAF). SCCAF permits a part of member nodes to sleep and only the rest of the member nodes need to transmit data to cluster head since the same cluster has high data redundancy according to GSCC. In this way it can prove that SCCAF can largely save energy in order to prolong the lifetime of WSNs while ensuring the accuracy. This paper also deduces the optimum number of cluster and transmitting nodes in the same cluster according to GSCC and SCCAF by math proof.Finally, this paper tests the GSCC and SCCAF by experiments. Firstly it uses Matlab to prove that SCCAF based on GSCC can save more energy than LEACH which says it has a longer lifetime, at the same time it proves the optimum number of clusters in SCCAF which proves theorem 1 in chapter 4. And then it sets the experiment scene using Jennic sensor nodes to sense the temperature in the room, from the experiment results, it can be seen that SCCAF has higher accuracy than CASA.