Research On Load Balancing And Anomaly Detection Of Wireless Sensor Network Based On SDN

Wireless sensor network is a self-organizing wireless network system composed of multiple sensor nodes.It collects information from the surrounding environment and transmits the collected information to the aggregation node or base station in a multi-hop relay manner.Sensor nodes consisting of sensing,data processing,communicating,and power subsystems are key components of wireless sensor networks.Because sensor nodes have limited energy,once energy is exhausted,it is difficult to replenish them in time.Therefore,it is necessary to focus on energy consumption issues when conducting wireless sensor network routing protocol research or sensor network deployment.So that the node can maintain normal working ability for a long time.In addition,as the main equipment for data collection,the node collects and sends incorrect data once it fails.As a device with limited energy consumption,how to effectively detect abnormal data while saving energy consumption also needs in-depth study.Traditional network architectures are tightly coupled and cannot be flexibly updated with routing protocols.The software-defined network decouples the traditional tightly coupled architecture,separating the control plane from the data plane,and better enabling the flexibility and control of the entire network.This new network architecture is introduced in wireless sensor networks.The limited energy consumption of wireless sensor network nodes makes the life cycle of the whole network shorten,and the fault acquisition information of sensor nodes is wrong,so that decision makers make inaccurate decisions.The thesis carries on the theory analysis and the algorithm research based on the software definition network.The main research contents are as follows:Aiming at the problem of energy limitation and load imbalance of wireless sensor network nodes,a multi-path load equalization routing algorithm for neighborhood subspace cooperation is proposed.In the wireless sensor network architecture based on software defined network,there are common sensor nodes and software-defined sensor nodes.Firstly,the perceptual node adopts the unit circle model,and the algorithm divides the nodes in the node one-hop neighborhood near the base station area into different regions,according to the experiment,the multi-path routing via three paths for forwarding is the best.To this end,the nodes are divided into three regions,then using the improved particle swarm optimization algorithm,with the shortest distance and minimum energy consumption as the optimization goal,a fitness function is designed to find the optimal node in each region as the relay node,find three forwarding paths,distribute the load proportionally to each path,and forward the load in parallel through three different paths.The load of a single node is reduced,the energy consumption of the node itself decreases,the performance of the whole network is improved,and the life cycle of the network is prolonged.The data collected by the wireless sensor network is abnormal due to node failure,which makes the reliability of the wireless sensor network application system low.Therefore,a cluster-based distributed anomaly data detection algorithm is proposed.The algorithm is based on a distributed architecture with two layers above and below.The bottom layer adopts the K-means algorithm and combines the improved particle swarm optimization algorithm to design the fitness function of Euclidean distance as an index of intra-class clustering.The KNN algorithm performs centralized exception class detection,and then sends the exception class information to all nodes.The node detects the abnormal data by comparing its own data vector with the exception class data vector.The distributed clustering algorithm performs two steps of intra-class and inter-class clustering,the detection precision is very high,the false detection rate is very low,and in the process of the underlying clustering,the upper-layer parent node sends the information one by one when sending the information.Information,so the communication overhead of the entire network is reduced overall.Finally,through the MATLAB,the algorithm proposed in this paper is compared and analyzed by simulation experiment.The experimental results show that the wireless sensor network equilibrium load algorithm based on software-defined network shows better performance compared with the existing algorithms in terms of network energy consumption and equilibrium load.The distributed anomaly detection algorithm of wireless sensor network based on software-defined network,which detects abnormal data under distributed architecture,has high detection accuracy and low error detection rate,and low network communication overhead.