Um algoritmo anticolisão para RFID de larga escala em ambientes ruidosos

Abstract

The Industrial Internet of Things (IIoT) is often presented as a concept that is significantly changing the technological landscape of industries, through automation procedures and identification of relevant objects. Therefore, reliability and performance problems must be considered when providing the communication services provided. By using Radio Frequency Identification (RFID) in the context of IIoT, several previous researches have worked to improve the efficiency of their communication systems, generally defining mathematical models for planning and quality assessment. However, such models are designed based on error-free communications, which is in fact unrealistic when considering the fault-prone nature of wireless communications in industrial plants. Therefore, this thesis proposes a new anti-collision algorithm for RFID, together with a formal model based on Generalized Stochastic Petri Nets (GSPN) to evaluate RFID communications, modeling different possibilities of errors between readers and tags. Since this proposal uses the EPCGlobal UHF Class 1 Gen2 parameters as a reference, which are already adopted by the Dynamic Frame Slotted Aloha anti-collision protocol for passive RFID systems, this model can be explored to assess the performance and reliability of different access protocols to the RFID means by assuming noisy channels, supporting better comparisons between different algorithms and protocols. The results showed that the proposed algorithm is able to present a better result in relation to the other evaluated protocols, mainly in the presence of noisy channels and a large number of tags to be read. The simulation scenarios are defined to present results of reliability and performance, when evaluating RFID tag readings, which are essential when designing and maintaining IIoT applications.