Unitary Authentication in the THz and Radiofrequencies domains – AUSTRALE

The main objective of AUSTRALE is to propose a technological solution where a part of the object to be traced will present a unique digital fingerprint obtained from electromagnetic signatures in the MW and THz ranges for the purpose of unitary authentication. It is declined in three axes:
• Identify, design, manufacture and test structures with a random character, therefore difficult to reproduce
• Evaluate the performances and limits of these authentication zones and optimize them without degrading the functional properties of the materials used
• Define a global specification (tag + code + reader) for a possible industrial transfer
In terms of performance, the project aims to bring great improvements compared to what is done in optical. Thus, the authentication part is invisible and integrated into the product, making the solution even more robust and effective against falsification. From 1) the concept of double coding on the surface and in volume and 2) the technologies brought into play, in particular for the THz solution, a new technology of delicate appropriation. This project presents a breakthrough potential that will allow a significant progress in the fight against counterfeiting. This solution allows objects to be identified in a unique way, without requiring the addition of a label or a specific chip that would affect the integrity of the product and especially the packaging, which is an essential communication vector. Here, it is the random characters that are intrinsic and internal to the object, and therefore non-reproducible, that are the basis of its unique signature, which greatly reduces the cost of the solution and makes it robust, with no constraints for recycling.
The richness of the signature of the prints will be amplified by using mainly three ways, which can be combined, in order to control the risks of success of the project:
- Addition of elements (fibers, glitter, pigments...) in one or more layers of cellulosic materials.
- Modification of one or more layers by the creation of periodic patterns, by micro-structuring the dielectric materials of the layers.
- Patterns will also be deposited by printing conductive inks, to create resonant structures at selected frequencies. The random aspect will be induced by an uncontrolled variability of the patterns' dimensions.
Several scientific and technological barriers have been identified and will be studied. They are related to 1) the losses in the targeted materials which could be prohibitive for the thickest devices, 2) the interview applications which may require specific constrained characteristics in terms of tag manufacturing, 3) the impact of the configuration and the environment of the tag during its reading, 4) the reliability of the signature interpretation (probability of recognition or not) and finally 5) the availability of readers.

The AUSTRALE project has allowed the development of several types of structures for authentication in the THz and MW domains, based mainly on cellulosic biosourced materials and therefore recyclable.
In the MW domain, the tags use metallic resonant structures presenting discriminating frequency signatures, the random character being brought by the non-reproducibility of the manufacturing techniques. It has been shown that the use of MW frequencies brings an undeniable advantage over lower frequencies for authentication applications by significantly reducing the probability of error from more than 20% for RF studies to about 0.5% in the worst configurations tested. This project has also shown that measurement (especially for several angular incidence) but also post processing can significantly improve the results. Also an authentication system based on the randomness of the EM response must be addressed both in the design of the tag and its manufacture as well as in the characterization method and post-processing. Finally we have fabricated for the first time resonant chipless tags, with ground plane, using a consumer printer where the substrate is just the printing medium. This is possible in MW where the substrate of a tag can have a thickness of the order of 100µm which corresponds to the films classically used for printing.
In the THz domain, we have finally opted for structures dedicated to imaging and using the natural randomness of materials constituting the patterns to be imaged. These structures can include electrically conductive elements to maximize the contrast of the image obtained, but other totally dielectric solutions have also been studied for optimal discretion. Thus we have designed, manufactured and characterized structures using flocked, calendered, watermarked papers, the insertion of conductive fibers or patterns deposited by printing with conductive inks.
We have also proposed a new technique to extract discriminating features from the obtained THz images in order to efficiently classify and thus authenticate them. This method uses a translation and rotation invariant decomposition of the images into wavelet packets. We have also demonstrated a new entropy-based feature extraction technique. The computed features are then used to train a graphical neural network adapted to a four-tree decomposition which has the huge advantage of taking into account the structural information of the rotation-invariant decomposition.
Finally, a prototype demonstrator integrating concomitantly solutions from both spectral domains has been built.

The AUSTRALE project has shown the potential of THz imaging to meet the challenge of authentication and providing images of devices with hazards and therefore non-clonable, moreover using cellulosic materials with high potential both in terms of the richness of the information contained and the potential in terms of recycling. It is planned to continue to develop this track to develop more efficient structures. In addition, it is planned to implement the processing methods used in the AUSTRALE project in the context of another application area, substance identification, which is also the subject of a research project financed by NATO and which began in mid-2022. Finally, from an experimental point of view, it is planned to develop a solution using a THz camera to carry out «real time« images and thus get closer to the application constraints linked to the duration of an image in the THz domain.
The imaging of reasoning structures in MW has been addressed in this project and the first results show the possibility of increasing the amount of information contained in a tag. It would be interesting to be able to quantify this increase in the form of an error probability in order to position ourselves on authentication applications.
The possibility to produce MW tags, resonant and with a ground plan with a simple consumer printer is a strong vector to democratize significantly this solution. This goes hand in hand with the introduction on the market of new MW radar reader systems which should provide the missing piece to this type of system. Two readers of this type have been purchased within the framework of the project, and the first results concerning the evaluation of their performances should be obtained during 2023.

The results of the AUSTRALE project have been the subject of 22 scientific communications, 4 in papers in journal, 10 in international conferences, 6 in national conferences and 3 in workshops. Most of these papers (74%) involve at least two of the project partners.

Counterfeiting and more generally identity theft, is a phenomenon that affects all the industrial sectors, from luxury to mass distribution and whose losses are colossal for economy, employment, brand image: in France a cost of 7.3 €billion (0.3% of the Gross Domestic Product) and of 25 000 jobs (about 300 €billion worldwide), adding health and technological risks associated with the counterfeit of certain products.
In this context, the ambition of AUSTRAL project is to propose new solutions for authentication applications using two large frequency ranges: millimeter-wave (MW) and Terahertz (THz). These solutions will be low costs, consistent with the techniques of paper industry for mass production, biobased and easily recyclable, these latter two criteria promote a more sustainable development.
Project objectives result from previous developments in the field of identification (ANR VERSO "THID", 2009-2013). This project has demonstrated the ability to associate two coding solutions on a single tag: a surface solution for MW encoding and volume solution for a THz encoding. A further study has since been performed on these structures that showed that the quantity of information contained in the electromagnetic signature of such tags is potentially usable for unitary authentication applications.
This concept is based on the idea that it can be extremely difficult to reproduce exactly some materials that have a random part, in the image of the distribution of cellulose fibers in a sheet of paper. From this finding, the tag principle for authentication is to put this randomness on the EM tag signature. However, unicity does not ensure it is not possible to clone the tag and to fight against counterfeiting, we must ensure the authenticity of the tag by comparing two signatures: the first generally measured of chain outlet manufacturing and the second when a user needs to authenticate. To enhance security solution, we will look into this project to integrate the solution directly into the materials that comprise the objects, the packaging, or the Mariana on the bottles ... to perform a tamper proof fingerprint. Moreover, we will evaluate the quantity of information contained in such fingerprint that is a key point for applications of interest.
More specifically, we propose to design, manufacture, characterize and analyze the quantity of information in several structures using surface and volume encodings that use the MW and THz frequency ranges. Finally, we will choose the most relevant MW and THz structures to achieve an efficient encoding solution and we will define the specifications of a whole encoding system (MW and THz structures, method for information encoding, readers, regulatory aspects).
From its main objective, which is to provide new security solutions to fight against a non-violent form of crime that is fraud and counterfeiting, this project fits naturally and primarily in the challenge B.9 (Freedom and security of Europe, its citizens and its residents) - axis 1 (Fundamental research related to the challenge). It indexes several major areas of this challenge about "the safety of persons", "methods for proof research" and "traceability of consumer goods". The expected results of this project affect secondarily the challenge B.7 (Information and communication society) -axis 7 (Micro and nanotechnology for information processing and communication), as the project targets demonstrating quantifiable performance improvements" and "breaks with existing knowledge", based on RF and THz technologies in electronics and photonics, notably in response to the application challenges related to the fight against counterfeiting.