A Clustering Algorithm Based On Software-Defined Wireless Sensor Networks And Further Optimation

With the rise of the Internet of things(IoT),Wireless Sensor Network(WSN)has become more and more widely used.Traditional WSN adopts the distributed control mode,and the network management is not flexible enough,the sensor nodes have limited energy,and the network life is short.The emergence of Software Defined Network(SDN)solves most of the challenges faced by WSN.Based on the idea of central control and separating logic control from data plane,the function of sensor nodes is simplified and many energy-intensive tasks will be performed in the central controller.Applying SDN to WSN,and Software Defined Wireless Sensor Network(SDWSN)comes into being.In SDWSN,we can programmatically implement various algorithms.Energy consumption seriously restricts the development of WSN,so energy conservation is the primary task of WSN.Among many WSN energy-efficient algorithms,energy-efficient routing algorithm is the most convenient method to implement.In energy-efficient routing algorithm,clustering routing algorithm is the most widely used.Since clustering is the core of clustering routing algorithm,this paper,based on SDWSN framework,proposes a clustering algorithm based on dynamic subnet expansion,and introduces a fuzzy system for cluster head selection,making cluster head distribution uniform and node energy consumption balanced.In the clustering algorithm proposed in this paper,the optimization of fuzzy system is very important.In the face of different network scenarios,fuzzy systems need to adjust parameters to better adapt to different network environments.The whole network is complicated and can be regarded as a black box,so black box optimization algorithm could be used to search the global optimal parameters of fuzzy system.In this paper,a bayesian fuzzy system optimization method is proposed to find better system parameters with less search steps.The optimization of fuzzy system can further exert the energy saving effect of clustering algorithm and extend the network life.