Analysis of the variability of RSSI values for active RFID-based indoor applications

Authors: GÜLBEN ÇALIŞ, BURÇİN BECERİK GERBER, AHMET BURAK GÖKTEPE, SHUAI LI, NAN LI

Abstract: Radio frequency identification (RFID) technology-based indoor localisation solutions have been widely adapted by many industries. An important factor affecting the performance of RFID applications is the received signal strength indication (RSSI) values. Indoor setting and layout may have direct effects on the RSSI variability, which, in turn, affect the robustness of RFID applications' outcomes. Effects of different environmental factors on RSSI values for RFID tags have been observed and reported in previous literature; however, there is a lack of research that addresses the effects of such parameters on RSSI values in a holistic and quantified manner. In this study, 2 different test scenarios are used to compare and assess the relationships between the RSSI values and system parameters for active RFID tags in indoor applications. These parameters include: 1) type of materials that tags are attached to, 2) obstructions between tags and an antenna, 3) relative elevation between tags and an antenna, and 4) relative orientation of tags and an antenna. The effect of each parameter was evaluated by statistical analyses. The Shapiro--Wilk, Bartlett's, Levene's, Kruskal--Wallis, Mann--Whitney, and Games--Howell tests were conducted. The results show that the data groups for all of the system parameter tests were nonnormally distributed and heterogeneous. Based on the Mann--Whitney test results, U values for comparisons of RSSI values in both test beds were found to be 0.00 for all of the system parameters. The results indicate that different system parameters caused variations in RSSI values as well as the detectability rates. Nevertheless, different test beds changed the influence of each parameter on the RSSI values. It is concluded that testing environment is a prominent parameter that affects the results and fluctuations in the RSSI values, and that the detectability rates cannot be attributed to any specific parameter that has been included in this study.

Keywords: Received signal strength indication, radio frequency identification, indoor asset tracking, indoor localisation, hypothesis testing, attached material, obstruction, elevation, orientation, test environment

Full Text: PDF