Performance Optimization Of Heterogeneous Wireless Sensor Networks Based On Extreme Learning Machine

Internet of Things(Io T)is the third revolution of information technology after computers and the Internet.As the key technology of the underlying perception of the Internet of Things,wireless sensor networks integrate information perception,wireless communication and information processing technologies to realize the interconnection of the physical world,the computing world and the ternary world of human society.Among them,Heterogeneous Wireless Sensor Networks(HWSNs)have the advantages of universality,high efficiency and scalability,and have becoming a new trend in the current evolution of wireless sensor network technology.This paper takes the complex industrial application environment as the research background,focusing on the research on the key issues of data processing and performance optimization of HWSNs that are suitable for large-scale,simple information and time-delay tolerance.Extreme learning machines and swarm intelligence optimization algorithms are used to study key issues of data processing and performance optimization of HWSNs to meet the stringent requirements of industrial applications for network performance,and to ensure the energy-saving,efficient,stable and reliable operation of HWSNs in complex industrial application environment.The research content includes data processing and performance optimization issues such as node deployment and optimal coverage in an industrial application environment,energy-saving and reliable clustering data collection,stable and efficient data fusion,and node fault diagnosis.The main research contents are as follows:1.Aiming at the coverage blind area and redundancy problems caused by the random deployment of HWSNs sensor nodes,With the goal of reducing node deployment costs,reducing coverage redundancy and voids,and increasing the coverage of the monitoring area,a method for node deployment and coverage optimization for HWSNs based on monarch butterfly optimized by particle swarm algorithm and extreme learning machine is proposed.Firstly,the coverage optimization model is established,and the extreme learning machine is used to optimize the coverage of the network,the work efficiency of the network,and the weight coefficient of the energy balance coefficient.In the process of calculating coverage,a particle swarm optimization monarch butterfly algorithm is proposed to optimize the coverage of HWSNs.The particle swarm algorithm is mainly used to optimize the butterfly migration rate and adjustment ratio of the monarch butterfly algorithm,which can improve the convergence speed of the algorithm and find the global optimal solution.Simulation results show that the proposed algorithm improves the coverage of the network and effectively avoids the coverage blind spots and coverage redundancy in the network.2.The problem of the cluster head selection and the optimal number of clusters in the process of clustering data collection of HWSNs is studied.In order to effectively reduce the energy consumption,improve the efficiency of the data collection and prolong the lifetime of the network,a clustering data collection method for HWSNs based on online sequence extreme learning machine and gray wolf algorithm is proposed.The best cluster heads are selected through adaptive learning of the online sequence extreme learning machine to avoid improper and frequent selection of cluster heads,accelerating the energy consumption of sensor nodes and shorten the life cycle of the network,and undermining the stability of the entire HWSNs.At the same time,a method for the optimal number of clusters of HWSNs based on gray wolf optimization algorithm is designed to avoid most or least cluster heads to optimize the energy consumption of communication between the cluster heads and the base station and balance the energy consumption of all the nodes in the network.The simulation results show that the proposed algorithm can reduce the energy consumption,improve the efficiency and reliability of the data collection,and prolong the lifetime of the network under the condition of ensuring the data delay requirements.3.Aiming at the problems of large data communication volume,high energy consumption and low fusion rate in the data fusion process of HWSNs,the bat algorithm is used to optimize the weights and thresholds of the extreme learning machine.A data fusion method of HWSNs based on extreme learning machine optimized by bat algorithm is proposed.Combining the spatio-temporal correlation between the data of the sensor nodes,an improved extreme learning machine is used to process the data collected by the nodes in the cluster routing structure of HWSNs to reduce the amount of data transmitted to Sink.The simulation results show that the proposed data fusion algorithm not only improves the efficiency of data fusion,balances the energy load of the network,but also reduces the energy consumption of the network and prolongs the network's lifetime.4.The problem of low fault diagnosis accuracy of the nodes of HWSNs is studied.In view of the problems of low accuracy and complex calculation of diagnosis process of the nodes,the regularization coefficient and the kernel parameter in kernel extreme learning machine are taken as the parameters of the fault diagnosis model of HWSNs.A node fault diagnosis method for HWSNs based on the kernel extreme learning machine optimized by the improved artificial bee colony algorithm is proposed.The artificial bee colony algorithm is used to optimize the regularization factor and nuclear parameters of the nuclear extreme learning machine,and the Cauchy mutation operation is used to encode the bee colony algorithm to fall into premature maturity.The method has fewer parameter settings,simpler implementation and faster network training speed.The simulation experiment results show that the proposed algorithm improves the accuracy of the hardware fault diagnosis of the sensor nodes,and can be better applied to the fault diagnosis for HWSNs.The research in this paper not only improves the data processing and performance optimization theory of HWSNs,but also provides strong support for the development of the Internet of Things and HWSNs technology,and broadens the application range of heterogeneous wireless sensor networks.It also provides an effective way to construct heterogeneous wireless sensor networks suitable for complex industrial application environments.