Research On SVM-based RFID Location

The extensive use of electronic tags provides great convenience for the operation and management of intelligent transportation,industrial automation,cargo warehousing,animal tracking management,and access security.For example,in a storage system,an electronic tag can receive and return a signal as long as it is within the coverage of the signal read/write antenna,and is thus perceived by the system.However,this method can only determine whether the cargos are in the warehouse and cannot obtain the cargos'location information.In order to improve the efficiency of finding a specific cargo in the entire storage area,if the cargo can be properly positioned and the possible scope of the cargo can be decreased,the search time can be reduced.In this paper,the radio frequency identification(RFID)indoor location system is taken as the research object,and a support vector machine(SVM)based RFID location method is proposed.We also studied the optimization method to improve the accuracy of SVM classification.Using the Received Signal Strength Indication(RSSI)of RFID equipment to obtain the location features,regional location recognition system is designed to verify the effectiveness of the proposed method.The main research contents of the thesis are as follows:First,the entire storage area is divided into a number of small areas,and the RFID object location problem is transformed into a problem of determining which of the small areas the RFID object belongs to.The relationship between the ID of each small area and the number of reading times is established,and the small area where the RFID object is located is determined according to the number of reading times.Secondly,the fuzzy SVM method is used to solve the classification problem.Based on the analysis of the influence of the misclassified samples on the performance of SVM,a membership degree assignment method is proposed,and the fuzzy iterative SVM algorithm is used to process the misclassified samples.The experimental results of the standard SVM algorithm and the fuzzy iterative SVM algorithm are compared by multiple data sets to verify the validity of the membership degree assignment method.Then,for the indivisible phenomenon of the one-versus-one multi-classification algorithm(OVO),the minimum distance from the inseparable samples to each region is calculated based on the minimum operator membership,which achieve a secondary classification of non-separable samples.Experiments with multiple data sets show that the method can effectively improve the classification accuracy.According to the actual needs of warehouse storage management of a project,a set of regional location identification system was designed by using SVM-based RFID location method.By using the number of reading times as the classification feature and the area ID as the classification mark,the iterative OVO-FSVM algorithm is used to locate the cargos,which can improve the efficiency of searching for cargos.The experimental results show that the location and identification system has a location accuracy of 77.8%,which can meet the needs of system location.