Research Of Data Perception And Acquisition Of Underwater Sensor Network Based On Edge Prediction

Exploring the ocean with precious resources has always been the focus of national development and an important support for maintaining national sustainable development.With the rapid development of underwater sensor networks(UWSNs)technology and equipment,underwater IOT equipment has been widely used in energy survey,environmental index detection,military monitoring and disaster event monitoring.Massive underwater data transmission to the cloud for processing and analysis has become the mainstream processing mode,cloud computing has become a mainstream computing paradigm.However,with the rapid growth of marine terminal equipment,UWSNs has also entered the era of big data with information explosion.When the data collected by the underlying sensor needs to be transmitted to the cloud,due to the sensitive and complex underwater environment and the assistance of a variety of heterogeneous devices for data transmission,resulting in large transmission power consumption,high delay and failure to respond to the needs in a timely manner,the data acquisition of the underwater sensor node faces a significant challenge:1)The energy consumption of underwater acoustic communication is always a key problem,and it will lead to multipath effect,frequency selective fading,data delay and other problems,resulting in the failure to obtain underwater data successfully.So how to effectively reduce the underwater acoustic communication is a very important problem.2)Many IOT devices make it difficult for centralized cloud computing to meet the needs of complex underwater environment.The data transmission volume of heterogeneous devices increases the pressure of bandwidth,which results in the increase of the delay of service / access request.As a result,centralized cloud computing has gradually developed into distributed edge computing,which needs to be closer to the underlying device support.3)Traditional underwater data collection mostly uses multi hop routing or AUV to assist in traversal acquisition,which results in large energy consumption delay.However,the existing prediction transmission scheme can not provide differentiated scheme support for different devices,resulting in large prediction error and energy loss when the node data is retransmitted.Based on the above problems,this paper introduces distributed edge computing to bear part of the pressure of cloud computing,and proposes a scheme of data acquisition and perception based on edge prediction for underwater sensor network,which can transform the underwater data acoustic communication transmission part into data prediction transmission,to reduce the energy consumption brought by acoustic communication.The main research work and innovation of this paper are as follows:1.In view of the data acquisition between the underlying sensor node and AUV,a first level bidirectional data perception prediction model based on adaptive exponential smoothing algorithm is proposed.The bidirectional prediction is carried out by both sides of the transmission at the same time,and the transmission party is responsible for proofreading.The purpose is to achieve the prediction result when the storage capacity of the underlying node is low and the computing capacity is weak,and to ensure the simple implementation of the algorithm and the results have high accuracy.2.For the data acquisition between AUV and edge aggregation node,because of the strong computing and storage capacity of edge devices,the second level bidirectional data perception prediction model based on the autoregressive moving average(ARMA)algorithm of extended Kalman filter is constructed,aiming to maximize the utilization of AUV nodes and improve the accuracy of data prediction.3.Aiming at the problem of water quality data prediction in the edge layer,a prediction framework based on the combination of LSTM and BP neural network is constructed,which aims to realize the strong calculation ability based on the edge equipment,further improve the accuracy of water quality data,respond to the needs of the bottom layer in time,and reduce the time delay of cloud layer transmission.In this paper,the data set provided by NOAA is used to verify the experiment.The experiments show that the prediction and data acquisition mechanism based on cloud edge end equipment can greatly reduce the bandwidth and delay consumed in the process of data acoustic communication transmission.At the same time,it shows that the scheme can effectively reduce the energy consumption of the sensor on the premise of ensuring the accuracy of data transmission It can respond to the underlying needs in time,which is obviously better than the existing scheme.