Research On Energy-efficient Cluster-based Routing Protocol For Wireless Sensor Networks

With the development of wireless communication technology, embedded computer technology and MEMS technology, wireless sensor networks has increasingly penetrated into people's lives. Wireless sensor networks is a new type of non-infrastructure self-organizing network, in which sensor nodes are usually battery powered and are often deployed in unattended areas, so that failure is easy to occur on account of drained energy. Therefore, to improve energy efficiency and reliability of data transmission is the core issue of wireless sensor networks. In large-scale network with nodes randomly deployed, how to prolong the network's lifetime and improve the effectiveness in data transmission so as to enhance the overall performance of the network is a significant research topic. Aimed at routing and topology control and other key issues in the network layer during data transmission in wireless sensor networks, in-depth investigations have been implemented in this thesis. Clustering mechanism, cluster-based multi-hop routing protocol, hierarchical clustering protocol and passive cluster-based multipath routing protocol are in turn proposed, the corresponding theoretical analysis and simulation on these protocols are implemented as well. The main contents include:(1) A regional-centralized clustering hierarchical routing protocol (RCCHR) is proposed. The cluster head is treated as a regional center in this protocol, while the virtual center algorithm is used to generate the cluster head of next round. At the same time, the inter-cluster routing is also established through drawing lessons from the virtual grid technology in the geographical routing of wireless sensor networks. The results of the research show that the energy efficiency of wireless sensor networks is effectively improved in this protocol, as well as the network life cycle is extended.(2) Started from optimizing cluster heads, balancing the network load in order to distribute energy consumption to all nodes, a hierarchical clustering routing protocol EERMA is proposed based on OWA operator and improved AntNet algorithm. In EERMA, sensor node calculates the time that it broadcasts its own cluster head information in terms of the OWA operator after collecting neighborhood information, based on which the optimal cluster constructions are produced by using density method. An improved AntNet algorithm is adoppted to search the optimal path in-inter-cluster, the affiliations of cluster heads are determined as well. EERMA protocol can improve the energy efficiency of wireless sensor networks, and it can also be suitable to large-scale wireless sensor networks. (3) Taken into account that there are many actual event-driven sensor network applications, a practical passive clustering node-disjoint many to one multipath routing protocol (PPCMP) is advanced in this thesis. In PPCMP protocol, passive clustering approach is firstly used to divide the entire network into several clusters, at the same time, a two phases node disjoint many to one multipath routing protocol is implemented among cluster heads, based on which the network coding technique is applied. The simulation results indicate PPCMP increases the energy efficiency, prolong the working life of the network, and improves the reliability of data transmission.(4) A new cluster head selection mechanism FAHP is proposed after applying the fuzzy multiple criteria decision making methods to the cluster head selection. In FAHP algorithm, multiple factors that can influence cluster head selection are taken into account synthetically, meanwhile, the multiple criteria decision making approach that is based on trapezoidal fuzzy AHP and hierarchical fuzzy integral is used to organically combine these factors together to get an integrated value which can represent the state of the node, then this integrated value is regard as a criterion for further filtration. The fuzzy multiple criteria decision making approach not only effectively prevents the inter-relationship of these factors, but also has been superb in all aspects of the program. The simulations show that using multiple criteria decision making to select cluster heads achieved good results. The new proposed cluster head selection mechanism not only enhances energy efficiency, but also improves the QoS performance.