Optimization of Radio Frequency Identification Reference Tag Location Algorithm Based on Back Propagation Neural Network

Sing-Ant: RFID Indoor Positioning System Using Single Antenna with Multiple Beams Based on LANDMARC Algorithm

“…where S mn represents the signal strength of the n-th reference tag measured by the m-th reader [4]. The reader measures the signal strength matrix of the tag under test, where T mu represents the signal strength of the u-th tag to be detected by the m-th reader [18].…”

“…The Euclidian distance of the vector is calculated as E ij . This is the distance between the i-th reference label and also the j-th label to be located [18].…”

“…In order to find the nearest reference tag to the tracking tag, we can use the coordinate of the reference tag with the smallest E value as the unknown tag's coordinate. It can be called the 1-nearest neighbor algorithm [18]. Unlike conventional LANDMARC approaches, when using antenna beams, the accuracy can vary because of differences when switching between antenna beams, and its positioning is an independent value; hence, finding the equilibrium point between the number of reference points and also the accuracy and range relationship is the key problem.…”

“…However, in the practical test, as a result of the complex environmental factors and multipath effect, the RSSI value from the device could not directly reflect the distance between the tag and the antenna [16]. Some RSSIs of reference tags and tags to be tested can be grouped into theoretical values and measured values according to different formulas [18]. The results can be logged to allow for further study, such as for deep learning models of estimation.…”