Research On RFID Network Planning Based On Multi-objective Optimization

As a key supporting technology in the Internet of Things,RFID has developed rapidly in recent years.This technology uses electromagnetic waves to identify targets,and the identification process does not require the assistance of manpower,nor does it need to be achieved through contacting.Readers and tags are an important part of the RFID system,and large-scale deploying readers will be bounded to cause conflict,which lead to the decrease of reading efficiency of the reader and the increase of economic costs.In order to solve the conflict between readers,there have two commonly methods,including signal processing and reader management,among which the method of reader management can fundamentally solve the conflict problem.The key to RFID network planning is to determine the location and number of readers to achieve the maximum benefit of the system,which is essentially a multi-objective optimization problem.Based on the actual application scenarios,this thesis adds a global archive to save elite individuals on the basis of the NSGA2 algorithm,replaces the original tournament selection with the tournament selection method with elite retention,and replaces the polynomial mutation with Gaussian mutation,which is called G＿NSGA2algorithm.Compare G＿NSGA2 with NSGA2,AW＿GA,MOEA/D three different multi-objective optimization algorithms to verify the superiority of the G＿NSGA2 algorithm.The innovative work of this thesis are as follows:Firstly,mathematically modeling the RFID network planning problem,taking coverage,collisions between readers,and load balance as important factors to measure RFID network planning,and combining point coverage and multi-objective optimization algorithms to lay out the readers.The entire RFID network planning problem is simple and feasible,which improves the efficiency of the reader.Secondly,improving the NSGA2 algorithm and proposing the G＿NSGA2 algorithm,we examine the performance difference between the G＿NSGA2 algorithm and the other three algorithms on the WFG,BNH,and UF multi-objective test function sets,simulation results show that the G＿NSGA2 algorithm has advantages in the distribution of solution sets and its convergence is better than the other three algorithms.Finally,in order to verify the superiority of the G＿NSGA2 algorithm in solving the RFID network planning problem,combining point coverage and 0-1 antenna coverage model for simulation experiment.The results show that the solution set produced by G＿NSGA2algorithm are better than others,which verifies the effectiveness of the algorithm in solving RFID problems.