Research On Energy-Efficient Data Collection Of Wireless Sensor Network Employing Unmanned Aerial Vehicle

Since the 21 st century,the capabilities of sensors in Wireless Sensor Network(WSN)continue to increase,such as sensing,processing,storage,and communication.Data collection is one of the basic functions of WSN.However,the traditional WSN data collection methods have the problem of “energy hole”.In order to solve this problem,some scholars have proposed to introduce an Unmanned Air Vehicle(UAV)into the WSN.Utilizing the mobility of the UAV,the distances between the sensors and the drone can be effectively reduced,the energy consumption can be reduced,and the transmission energy burden can be more uniform.However,the UAV-WSN system also faces many challenges,such as trajectory planning of the drone,resource management of the system,high energy efficiency and high reliability of data transmission.This research focuses on the data acquisition technology based on the UAV relay platform with fixed flight path and flight altitude.The research content of this paper is as follows:Firstly,this paper proposes an energy-saving data acquisition algorithm(JSPTA)that jointly plans the flight speed of the UAV,the transmit power of sensors,and time slots resource allocation of the UAV-WSN system in order to solve the energy optimization problem of the WSN with unknown location and data amount.When the UAV collects data,the transmission energy consumption optimization problem of the sensor network can be expressed as a mixed integer non-convex optimization problem.In this paper,successive convex optimization techniques and differential evolution algorithms are used to solve the problem.In this paper,the performance of the proposed algorithm and the benchmark algorithm are compared by simulating when the average data amount of the sensor,the network scale,and the sensor position distribution are different.The results show that when the average data amount of the sensor is 15 Mbit,the number is 100,and the position is uniformly distributed,the JSPTA algorithm can increase the data acquisition ratio by about 22%,and reduce the energy consumption by about 18%compared with the benchmark algorithm.Secondly,the energy optimization problem of the wireless sensor network in Intelligent Transportation System(ITS)employing UAV is analyzed,and a data acquisition method based on predicted data amounts is given.The method is divided into two stages:the prediction of the data amount and the collection of data.In ITS,the amount of data in the WSN can be predicted.Therefore,three deep learning methods including Long Short-Term Memory,Gated Recurrent Unit and Stacked Auto-Encoders are applied.The data set provided by the California Department of Transportation is used to predict the amount of data collected by the sensors in the future.Then based on the predicted data vamounts,the concept of virtual node is introduced,and the WSN energy optimization problem is re-presented.The differential evolution algorithm is proposed to solve the problem.In this paper,the simulation experiment of data acquisition is carried out on the obtained prediction data set.This paper simulates based on the simulated data.The results show that when the sensors average data amount is 15 Mbit,the number is 100,and the position is uniformly distributed,the proposed whole time slot resource allocation strategy can increase the data acquisition ratio by about 9%,and reduce the energy consumption by about 3% compared with the JSPTA method.In this paper,based on the prediction data set,the simulation experiment of data acquisition is carried out.The simulation results show that after introducing traffic prediction,the full time slot resource allocation strategy based on the Stack Auto-Encoders can increase the data acquisition ratio by about 6%,and reduce the energy consumption by about 5% compared with the JSPTA method,.Finally,although this paper studies the energy-efficient of data acquisition in the UAV-WSN system,there are still several directions that can be studied: The first is the3-D trajectory planning for the UAV,the second is the joint prediction of data amounts and locations of sensors.