An Effective Tag Estimation Method Based upon Artificial Neural Networks and Signal Strength for Anticollision in Radio Frequency Identification Systems

Abstract Radio frequency identification (RFID) technology has been widely used in applications such as access control, inventory management, spatial positioning, and object identification. Accurate tag estimation is one of the major challenges in RFID reader systems particularly in areas where large tag populations are to be identified such as shopping carts, warehouse inventory monitoring, and small ruminant farms. This paper proposes a new tag estimation technique employing artificial neural networks (ANNs) and signal strength to read large tag populations. The technique estimates the number of tags through the signal strength of the backscatter channel for efficient implementation of dynamic framed slotted Aloha (DFSA) protocol by analyzing the RN16 and the received signal strength indicator (RSSI). The ANN model is trained using the signal strength of various tag populations and can identify the number of tags with minimal errors. The proposed technique does not require any modification in the tags and is implemented as a minimal software script to be added to the tag estimation module of the reader. The proposed signal strength-ANN model is able to estimate the accurate number of tags thereby improving the performance of the employed DFSA model.