Hadi FARHAT

Thesis defence
27/03/2019
Amphithéâtre IRCICA


Summary:

UHF RFID technology, also known as RAIN RFID, is a passive technology for the automatic identification of objects. It bridges the gap between the physical and digital worlds by allowing an object to become "alive" on the Internet of Things using low-cost, battery-free RFID tags. Passive UHF RFID technology is undergoing significant development due to the accelerated growth of sophisticated technological devices. This development is largely driven by the supply chain and mass retail. In these areas, second-generation (Gen2) passive UHF RFID, among other possible tracking and identification solutions, is a logical choice given the low cost of tags when produced in high volumes, their ease of printing and the absence of batteries.
The main concern for retailers is product availability. This can be affected by visibility-related errors (storage discrepancies, theft, loss) and human error. It is therefore necessary to improve visibility and automate the process. Despite its benefits, the deployment of RFID technology in retail still faces many obstacles and challenges. In this thesis, we set out to list all the component and reader technologies available on the market or still in the prototype stage, and then, after identifying and analysing the various challenges associated with this particular field of application, to develop and test new solutions.
Firstly, the maximum reading distances of passive tags are evaluated in different environments in order to identify the factors impacting them. As a result of this study, an alternative characterisation method is proposed to monitor tag performance and identify tagged objects with poor performance. Secondly, we use this method to propose a new solution for improving detection. The algorithms designed and implemented to translate this solution allow readers to negotiate their configurations with the environment and with each other. Finally, a new localisation algorithm is proposed to improve accuracy. It has been developed and implemented by exploiting the response of tags identified as the most reliable in the environment. The proposed solutions are particularly advantageous because they are universal, compatible with readers on the market and do not add any cost to the hardware used for detection.

Abstract:

UHF RFID technology, also known as RAIN RFID, is a passive technology that enables the automatic identification of items. Thus, it bridges the gap between the physical and digital worlds by allowing an item to become alive on the Internet of things thanks to inexpensive and battery-free RFID tags. Passive UHF RFID technology has witnessed a significant development due to the accelerated growth of sophisticated technological devices. This development is largely driven by the supply chain and the retail industries. Passive UHF RFID Gen2, among other tracing and identification solutions, is the logical choice given the low costs of large-volume tags, ease of printing and being battery-free, the need for maintenance is completely eliminated.
The main concern of resellers, the withdrawal of the product, is mainly affected by errors related to visibility (stock gap, theft, loss) and human errors. It is, therefore, necessary to improve visibility and automate the process. Despite its advantages, RFID deployment in supermarkets is still facing many obstacles and challenges. In this thesis, we focus on technological availability by defining and analyzing the different challenges and possibly proposing the appropriate solutions.
First, the maximum reading distances of passive tags are evaluated in different environments in order to identify the factors impacting them. At the end of this study, an alternative characterization method is proposed to control tag performance and identify tagged objects with poor performance. Secondly, we will use this method to propose a new solution to improve detection. The algorithms behind this solution allow readers to negotiate their configurations with the environment and with each other. Finally, a new location algorithm is proposed to improve accuracy. This algorithm is based on the exploitation of the answers of the reliable tags of the medium. The proposed solutions are universal, compatible with market readers and add no cost to the hardware used for detection.

JURY :

- Nathalie ROLLAND-HAESE, University of Lille, Thesis supervisor

- Robert STARAJ, Université Nice Sophia Antipolis, Rapporteur

- Aziz BENLARBI-DELAI, Pierre-et-Marie-Curie University, Rapporteur

- Divitha SEETHARAMDOO, IFSTTAR, Examiner

- Philippe MARIAGE, University of Lille, Examiner

- Plamen ILIEV, Decathlon, Examiner

- Mohamed HIMDI, University of Rennes 1, Examiner

- Claude TéTELIN, GS Global Office, Examiner