Efficient Particle Filter Localization Algorithm in Dense Passive RFID Tag Environment

Abstract

The means of distributing dense passive radio-frequency identification (RFID) tags has been widely utilized for accurate indoor localization. However, they suffer a disadvantage on low localization precision due to the increasing interference of RFID tag collisions and the variation of behavior of tags. Current localization algorithms used in passive RFID location systems are mostly deterministic and have a limited capability on improving localization precision in a dynamic environment with uncertain sensor measurement. This paper investigates the feasibility of using particle filter technique as an efficient localization approach to deliver both relatively good accuracy and precision in dense passive RFID tag distribution applications. A position feature-based system model is first built to apply the typical particle filter technique in passive RFID location applications. Then, a new particle filter algorithm by using a moving direction estimation-based feature improvement scheme is proposed to enhance localization precision in a dense passive RFID tag environment. Experimental results show that the proposed method can provide relatively good accuracy and precision for passive RFID location applications, with an improved performance over the typical particle filter algorithm and a state-of-the-art deterministic method.