FACILITY RFID LOCALIZATION SYSTEM BASED ON ARTIFICIAL NEURAL NETWORKS

作者：Holland William S; Young William A II^*; Weckman Gary R

来源：International Journal of Industrial Engineering-Theory Applications and Practice, 2011, 18(1): 16-24.

• Radio frequency identification
• Inventory location systems
• Artificial neural networks
• Interference

Radio frequency identification (RFID) technology is used for asset tracking due to its accuracy and speed. RFID tracking systems are being used to locate tagged objects in indoor environments, however; reliability is low due to interferences. To overcome this limitation, artificial neural networks (ANNs) can be used to determining a device's location in the proximity of interference. This research presents a proof of concept to an industrial application of using ANNs as an RFID localization algorithm when objects are subjected to metallic and human interference. To prove this concept, random samples are collected using the received signal strength indication (RSSI) values from passive RFID readers and antennas. The test results show that ANNs can determine the location of a passive RFID tag accurately in the presence of noise and shows that data preprocessing techniques can improve the predictive capabilities of the ANN-RFID localization algorithm.
Significance: Research shows that passive tags are better suited for tracking items that are high volume, low cost, or have a short shelf life. A passive RFID tag system is explored in this research due to the inexpensive nature of the tags. Thus, there is a great potential for many industrial applications if advancements are made to increase the reliability of determining a tag's location. Constructing an ANN-RFID localization algorithm is significant because ANNs are capable of predicting non-linear, noisy, or incomplete readings that are obtain from RFID antennas. These models can ultimately decrease the setup time needed to implement and increase the accuracy of a location system in the presence of noise.

• 出版日期2011