RFID-Based Vehicle Localization Using Virtual Wideband Multi-Frequency Continuous Wave

Vehicular location information plays a crucial role in intelligent transportation systems. A particular focus is on developing an Internet of Things (IoT)-based sensing system for smart roads to enable high-precision vehicle localization that does not rely on global navigation satellite system (GNSS) and improve traffic safety and efficiency. As a passive low-power IoT technology, radio frequency identification (RFID) has been proposed as an alternative localization approach in GNSS-denied scenarios. However, limited by the licensed narrow bandwidth, the localization accuracy of commercial off-the-shelf (COTS) RFID readers is quite low. To achieve the RFID-based high-precision localization, we develop an RFID localization reader on a universal software radio peripheral (USRP) platform, which realizes virtual broadband multi-frequency continuous-wave (MFCW) transmission and phase extraction of the received backscattered signals. The genetic algorithm (GA) is applied to choose the optimal frequency set to minimize the range difference estimating error. By calculating the phase difference of arrival (PDOA), the localization problem can be modeled as hyperbolic equations and solved. Experimental results validate that our developed RFID-based vehicle localization system could achieve an accuracy less than 5 cm.