ELM Localization Algorithm Introducing Virtual Nodes And Compressed Sensing For Wireless Sensor Network

The wireless sensor network(WSN)is composed of a series of nodes with different work tasks.Real-time monitoring of complex environments is performed by collecting and monitoring data from the nodes.As a prerequisite for other work development and technical exploration of the WSN,node location technology faces the challenge of positioning as ineffective and susceptible to environmental damage,poor scalability and network excessive power consumption.It can be seen that it is difficult and valuable to design a WSN location algorithm with low power consumption,good resistance to external environment,strong endurance and high positioning accuracy to deal with network costs,battery capacity,hardware conditions,and external harsh environments.The research of wireless sensor network positioning technology in this paper is as follows:This paper summarizes the research background and status at home and abroad.Then,it introduces the basic points of wireless sensor network node positioning and the evaluation index of positioning performance.It also analyzes several classical node position calculation methods and Range-based and Range-free algorithm.For most Rang-free algorithms,positioning accuracy can be improved by increasing the anchor node ratio.ELM localization algorithm is proposed which is introducing virtual nodes in wireless sensor network.Through the introduction of virtual nodes and extreme learning machine(ELM)in combination with node positioning to find suitable unknown nodes to upgrade to sub-anchor nodes,thereby increasing the proportion of anchor nodes and improving positioning accuracy.Aiming at the disadvantages of large power consumption in the process of node localization in wireless sensor networks,the WSN positioning algorithm based on compressed sensing not in the grid center node positioning accuracy is not high and have high algorithm complexity,a node localization algorithm of WSN is proposed which combining Compressed Sensing(CS)and ELM.Firstly,Initial positioning by compressed sensing,detect whether unknown nodes in the grid.If unknown nodes in the grid,then use ELM to correct errors.Finally,unknown node position is calculated by trilateration.MATLAB simulations show that the ELM localization algorithm introducing virtual nodes in wireless sensor network generally has less error than the DV-Hop algorithm and BP location algorithm.Therefore,this algorithm is suitable for localization for wireless sensor networks,but there are some certain requirements for the anchor node ratio,communication radius and the total number of nodes;CS and ELM based localization algorithm for WSN reduces the errors generated by using RSSI and reduces the amount of data to be uploaded,thereby reducing the power consumption of the WSN.