UHF RFID Indoor Localization Algorithm Based on BP-SVR

With the widespread of Internet of Things (IoT), Radio Frequency Identification (RFID) technology is used in various fields. In the complex indoor environment, the traditional RFID indoor localization algorithm cannot give superior performance. Because a large number of antennas cannot be deployed in practical applications, the dimension of the signal strength feature vector obtained is relatively low, and it is not easy to fully describe the information in the environmental field. Thus, must adopt the appropriate antenna placement structure, the phase difference information can be effectively used as the feature. In this paper, a method of indoor localization algorithm based on Back Propagation–Support Vector Regression (BP-SVR) is proposed, which takes the signal strength and phase difference of the RFID tag as the feature inputs and uses the hidden layer of the neural network to enhance the dimensionality of the data. Moreover, the Sequential Minimal Optimization (SMO) algorithm is used to increase the rate of model learning. The method was valid in a 2D scene. Experimental results show that the method has an average positioning error of 9.5cm in the indoor area of 6m
$\times $
8m.