Ultra-large bandwidth (BW) wireless and the seamless connectivity to the cloud and core networks are driving the access to sub-THz bands and the massive deployment of small cells. To compensate sub-THz path loss, high-gain antennas must be used. Current solutions rely on conservative quasi-optical systems (mostly reflectors or lenses) and, generally, do not offer reconfiguration. Moreover, the bulkiness of such systems precludes their efficient integration on mobile platforms or urban furniture. To overcome these issues, a change of paradigm must be adopted: RF front-ends must not only satisfy the link budget over broad BWs, but be also amenable for integration on the chassis of vehicles or smart urban furniture. AROMA’s ambition is to leverage metasurface (MTS) antennas to develop ultra-thin smart skins that meet these needs. MTS antennas consist of modulated impedance surfaces that gradually radiate the power carried by a surface wave launched by one port. Unfortunately, high-gain MTS antennas exhibit relatively narrow gain BWs. To overcome the physical bounds in the gain-BW product of single-port MTS antennas, we will explore aperiodically-tiled MTS apertures, with a limited number of input ports, through which we can sense the electromagnetic environment. AROMA builds on two research hypothesis, validated in the proposal by extremely promising preliminary results: 1. Since the BW of a MTS aperture is inversely proportional to its radius, we can increase the BW without impacting the gain (G) by dividing the total radiating aperture into sub-apertures with smaller radius, each one fed by its own port. 2. The grating lobes (related to the spacing between ports) can be avoided by tiling aperiodically the aperture and appropriately tailoring the shape of the embedded tile patterns. Starting from these assumptions AROMA will pursue 4 main scientific objectives. 1) First, we will theoretically study the physical bounds of aperiodically tiled MTS apertures. By increasing the number N of sub-apertures, the gain G will not change and the BW will increase proportionally to N. Larger values of N will also densify the array, precluding grating lobes, at the expense of more complex feeding networks. Hence, one must find the smaller N that enables the objective BW for each tiling. 2) Second, based on the developed theoretical framework, we will select the most appropriate surface partition and impedance modulation. The ideal surface impedance will then be implemented by changing the size and orientation of sub-wavelength metallic elements arranged in a periodic lattice. The antenna design will be completed with the appropriate feeding structures. 3) The third objective deals with the fabrication of proofs of concept (PoC) for aperiodic MTSs at J-band. To that end, we will use photolithography on low-loss substrates or Deep Reactive Ion Etching to micro-machine Si wafers subsequently metallized by sputtering. System compactness will be privileged for volume and mass reduction. Intermediate PoC will be also demonstrated at Ka-band. 4) Finally, the manufactured PoC antennas will be tested at M2ARS platform of IETR, using far-field/near-field measurements with ultra-accurate positioning and J-band extension modules. AROMA targets breakthrough architectures to surpass the state-of-the-art with ultra-thin, broad-band (>20%) and directive (>40dB) antennas at J-band. AROMA will have a duration of 42 months and it is a fundamental research project that gathers around the young researcher IETR’s expertise on electromagnetic theory, antenna design, microfabrication and antenna metrology. AROMA will be also advised by an external scientific board involving high-level experts in MTSs and THz technology along with one telecom operator (Orange Labs) and one equipment provider (RFS) with whom the above-mentioned target performance values will be refined based on the selected use cases (point-to-point communications).

Wireless communication systems constantly evolve towards more sophistication in order to meet ever-growing efficiency requirements. This evolution entails more complex data processing, which can be handled via classical signal processing techniques or the more recent machine learning approaches. Signal processing tends to be computationally efficient but can rely on simplistic analytic models, while machine learning is data-adaptive by nature but requires heavy computations to be trained. MoBAIWL aims to take the best of both worlds by designing computationally efficient AI methods built on models usually used in signal processing.

Modern multiprocessor Systems-on-Chip (SoCs) widely rely on Trusted Execution Environments (TEEs) in order to address data confidentiality and integrity. However, composed of heterogeneous components and several processors, the real surface of attacks of modern SoCs is very vast and still underestimated. In particular, more and more features for power and clock management from software greatly extend the surface of attacks. They provide a remote attacker with significant information on the system (e.g., through embedded temperature, power, etc sensors) at runtime, as well as with the means to maliciously exploit dynamic voltage and frequency scaling. This can therefore lead to covert channels and side-channel attacks (SCA), as well as to timing fault injection (FI). CoPhyTEE will investigate these recent attack vectors that can circumvent TEE security. We state that these attacks can be combined in order to perform powerful FI assisted SCA, and SCA assisted FI. Finally, CoPhyTEE will propose practical countermeasures by benefiting and exploiting modern SoCs capabilities as power and clock management for increasing the security of the system, instead of removing or limiting these features. The objective is to benefit from open-hardware to build a reliable secure SoC by integrating dedicated security blocks able to first, monitor the state of the system and detect at runtime vulnerable execution scenarios, and second, to dynamically react by hiding and masking the (timing, power, temperature) signature of the system.

We are witnessing unprecedented Cyber-Physical Systems deployments to monitor and control the environment, including critical infrastructures. The Internet of Things federates these edge computing systems with the cloud, all of which make increasing use of Artificial Intelligence to provide high inference quality and autonomous decision-making capabilities. As a result, there is an enormous network composed of billions of highly heterogeneous computing systems with high exposure to cyber attacks which require unprecedented amounts of energy to operate. We have never witnessed before as many and as varied attacks at all levels of computer systems. And neither have we ever before required as much energy as several times the lifetime emissions of one car to train an AI model, clearly posing severe sustainability challenges. After decades focused on performance, this has awakened the community to seriously address the security and energy efficiency of computing systems, which largely drives research today. There is a large body of work focused on improving the performance and energy efficiency of AI systems. Approximate computing (AxC) is an emerging paradigm that proposes to relax the accuracy requirements of computing systems, tolerating errors in computations to trade-off quality of results with a reduced usage of computational resources. The reduced computational complexity allows building faster, simpler and less power-hungry computing systems. In parallel, with the accessibility and connectivity of CPS unveiling an enormous attack surface, cybersecurity is a major concern and security is embraced as a key design goal. Renowned cyber attacks to critical infrastructures in recent years have shown large potential to cause major impacts on security, safety and privacy of personal and corporate data. Research on hardware security has largely focused on providing side-channel resistant cryptographic systems. However, we are seeing how increasing numbers of attacks are now directed to AI systems, compromising private data and secret corporate IP models. In this scenario, where ubiquitous deployments of AI-enabled CPS are increasingly making autonomous decisions and operating without human intervention, it is of paramount importance to secure our cyberinfrastructures to protect the upcoming smart society we are heading into. This is the ultimate objective ATTILA aims to contribute to: to secure AI. ATTILA addresses this challenge by focusing on the security of hardware implementations of AI-enabled edge CPS that use AxC, present in increasingly large shares of systems to reduce energy consumption. With the interplay between AxC and security only starting to be considered, we are largely unaware of its resulting impact. To the best of our knowledge, ATTILA is the first work studying side-channel vulnerabilities associated with AxC techniques like extreme voltage overscaling and quantization in approximate Deep Neural Networks (DNN) accelerators running in reconfigurable devices, uniquely suited for AxC. We focus initially on power side-channels and build an experimental SCA set-up to study the impact of approximations on leakage behavior and SCA resistance. To contribute to building more secure implementations, we perform a design space exploration of DNN implementations with configurable approximation levels, building pareto fronts that facilitate trading-off SCA resistance with inference quality. ATTILA also addresses countermeasures through an intelligent run-time manager that leverages on AxC to render attacks more difficult, enabling self-adaptation of the system approximation level to modify side-channels leakage behavior at run time. We initially consider standard SCA techniques to shed light on the impact of AxC in SCA resistance of approximate DNNs, and then move to study ML-based techniques and consider EM side-channels to study the applicability and generalizability of our previous findings.

1) Wider Research Context / Theoretical Framework: Navigating the challenges posed by wave scattering in complex systems is crucial for numerous daily applications, spanning wireless communication, sound transmission, environmental sensing and imaging. In the META-INCOME project, we aim to make significant progress in these areas by employing novel inverse design techniques that don’t just optimize individual wavefronts, but many of them in parallel. 2) Hypotheses/Research Questions/Objectives: Our first goal is to demonstrate the possibility of achieving full transparency of scattering media by placing custom-made anti-reflection metastructures in front of them. Specifically, we target the development of large-scale anti-reflection metastructures in the centimeter wavelength range to explore possibilities for enhancing wireless connectivity in indoor environments. Our second objective involves designing electromagnetic reverberant cavities capable of perfectly absorbing multiple radiation modes simultaneously, broadening the applications of coherent perfect absorption. Our third objective is to route the Fisher information to a desired observer in view of sensing and imaging tasks. 3) Approach/Methods: Our study is based on inverse design approaches, integrating concepts from wave physics such as generalized impedance matching, coherent perfect absorption, exceptional points and Fisher information. The employed methods include the development of large-scale anti-reflection metastructures, electromagnetic reverberant cavities, and the enhancement of sensing capabilities through inverse design. 4) Level of Originality / Innovation: The original and innovative aspect of project META-INCOME lies in the fact that our approach goes beyond established wavefront shaping concepts in which individual incoming wavefronts are optimized for specific tasks. Instead, we design meta-structures for which many radiation modes satisfy the desired property (such as transparency or full absorption), simultaneously. Building on an established and successful collaboration between the two involved teams, we will address the stated goals both from a theoretical perspective (covered by the Austrian group) and from the experimental side (covered by the French team). 5) Primary Researchers Involved: Project META-INCOME will be implemented collaboratively, involving Stefan Rotter from TU Wien (Austria), an expert in the theory of complex scattering phenomena, and Matthieu Davy, an expert in microwave experiments in disordered and resonant media. The team already has a proven track-record of collaborating successfully. Through the funding from this grant, the team aims to hire a graduate student and a postdoc.