Commercial indoor spaces such as hospitals, hotels, offices offer great potential for commercial exploitation of logistic robotics. Also, offer advantages for their deployment, since they are required by law to meet stringent building codes, and therefore the navigation space exhibits some structure. In addition, they offer reliable communications infrastructure, since this is required for normal business operation. Thus, commercial spaces are rightfully considered the next great field of logistic robotics deployment. Despite these advantages, today, few solutions exist, and these solutions do not trigger widespread acceptance by the market. This is because existing systems require costly infrastructure installation (arrays of peripheral sensors, mapping, etc.); they do not easily integrate to corporate IT solutions and as a result, they do not fully automate procedures and traceability; they are limited to a single type of service, i.e. transfer of goods. Through transfer of knowledge, multidisciplinary research and cross-fertilization between academia and industry, ENDORSE will address the aforementioned technical hurdles. Four innovation pillars will be pursued: (i) infrastructure-less multi-robot navigation, i.e. minimum (if any) installation of sensors and communications buses inside the building for the localization of robots, targets and docking stations; (ii) advanced HRI for resolving deadlocks and achieving efficient sharing of space resources in crowded spaces; (iii) deployment of the ENDORSE software as a cloud-based service facilitating its integration with corporate software solutions such as ERP, CRM, etc.; (iv) reconfigurable and modular hardware architectures so that diverse modules can be easily swapped. The latter will be demonstrated and validated by the integration of an e-diagnostic support module (equipped with non-invasive sensors/devices) and the Electronic Health Records (EHR) interfacing, which will serve as an e-diagnostic mobile station

Electrification of vehicles and improved efficiency of internal combustion engines (ICE) are the main levers to reduce greenhouse gas emissions. Recent studies indicate that in 2040 thermal cars sales will still remain an important part of the market and the spark-ignition engine (SIE) is seen as the most interesting ICE technology. However, technological challenges must be tackled before meeting real driving emissions expectation due to the diversification and complexity of hybrid applications. For flow aerodynamics, mixing and combustion down to the individual engine cycle, challenges are for example associated to robustness of concepts on a cycle basis, rapid variations of engine loads observed in hybrid technologies during transients, the occurrence of extreme cycles for a wider range of operating conditions. Numerical, experimental and analyzing tools have made significant progress in recent years for the analysis of spatial and temporal scales of the unsteady in-cylinder flows. Large-Eddy Simulation (LES) is an essential tool for the design of robust concepts. While LES has been validated against well-defined experiments, the prediction of internal turbulent dynamics and combustion during a cycle is affected by epistemic uncertainties. Therefore, progress is still needed to obtain optimal and robust design. The main objective of ALEKCIA is to develop game-changing tools for augmented prediction and analysis of turbulent reactive flows with a focus on real SIE operations to better capture time-resolved events and increase understanding and control of the origins of undesired behaviors. The key hypothesis is that future progress and success is tied to the synergistic, strong combination of experimental and numerical tools at every stage of the project, which will provide advancement in the analysis of physical scales and boundary conditions (BCs). The major scientific challenges addressed by ALEKCIA are to 1/ quantify and reduce uncertainties (UQ) due to model parameters and BCs, 2/ develop new Data Assimilation (DA) approaches for coupling LES with experimental measurements, 3/ develop new decomposition methods to analyse big data generated by LES and high-speed PIV, 4/ combine them with UQ and DA methods for detailed analysis of individual SIE cycles during steady operations and fast transients. We stress that this methodology could also be used more widely for industry and energy applications. To achieve its ambitious objectives, work in ALEKCIA is structured into one management task (T0) and three technical tasks (T1 to T3). We will address non-cyclic phenomena under transient and fired operations and develop novel analysis from the acquired experimental and LES databases of a SIE performed respectively at PRISME (T1) and IFPEN (T3) laboratories. The partners of the project will also collaborate on the development of crank-angle resolved spatio-temporal EMD decomposition (T1 and T3) for engine flows to obtain an unprecedented detailed understanding of the mechanisms involved in the generation of in-cylinder flow, turbulent dynamics and their impact on combustion. The development of UQ tools to quantify and reduce uncertainties in complex LES of SIE flows is also targeted (T3). Finally, the capabilities of DA methods to calibrate realistic BCs on-the-fly is investigated by PPRIME (T2 and T3). This task is particularly relevant when assimilating experimental data (in the form of BC and in-cylinder large-scale flow patterns from EMD) obtained in extreme cycles. EMD obtained from a selected number of measured cycles presenting very slow or fast combustion rates will be coupled with UQ and DA tools for their inclusion in LES (T3). In this scenario, LES will be able to properly follow the assimilated aerodynamic behaviour of these cycles while turbulent dynamic will be modelled. Finally, the application of the developed tools will allow to identify the main key parameters controlling internal aerodynamics.

The REFINE project focuses on the experimental investigation and numerical simulation of real-fluid injection and mixing processes under sub-, trans- and super-critical conditions. The domain of interest of the present proposal concerns the propulsion with application to the automotive and aerospace science and technology where supercritical fluids may be considered as propellants. Indeed, the need for higher efficiency and lower emission levels leads to increase pressure and temperature levels, i.e. to reach supercritical properties of fluids. The objective of REFINE is to build a simple well-controlled test-bench able to study a fluid injection under sub-, trans- and super-critical conditions and to associate experimental and numerical diagnostics to deliver the finest information. An ethane injection occurs in a 5-liter high-pressure experimental test-bench. The X-ray diagnostics setting-up will be the project keystone, as it allows for delivering a non-polluted density measurement. Indeed, such diagnostics are not disrupted by the index gradient observed in corrugated flows, contrary to laser techniques. Colored background oriented Schlieren visualization is used for backup as well as a more classical shadowgraphy technique. Numerical simulations will be realized in parallel to consolidate physics understanding and for model validation.

In the last few years, neural models have allowed spectacular progress in natural language processing (NLP). The DeepTypo project proposes to use multilingual models of speech to design methods for automatically extracting, from audio recordings, typological information useful for language documentation and research (phonological and morphosyntactic complexity indices, similarities between languages…). Based on a collaboration between linguists and NLP researchers, the DeepTypo project sits squarely in the space of digital humanities by addressing fundamental questions of both communities. It will help linguists in their work of documenting and analyzing languages, especially “rare” or “poorly endowed” languages, by providing them with new tools and methods that will allow them, for example, to bring out new information on similarities between languages. Beyond the “tool development” aspect, the DeepTypo project aims, above all, at showing that the representations at the heart of neural networks can be used to answer fundamental questions in linguistic, by taking, as an example, current issues in creolistics (the study of creoles) and dialectology of Sino-Tibetan languages. Extracting typological information, the core of the DeepTypo project, will also contribute to the identification of the limits of fine-tuning. This approach has made it possible to develop, at low cost, NLP systems for several languages and many tasks and is often presented today as "THE" solution to all NLP problems. The identification of linguistic features captured by neural networks will allow us to verify if this is indeed the case: if a model is, for example, not able to detect and represent the tones of a language, it is more than likely that it cannot be used to develop a system for tonal languages. To achieve this ambitious goal, we will use neural representation analysis methods to interpret and understand the decisions of neural networks and will develop them along four original axes: 1. Based on the collaboration with the different partners of the project, we will try to identify richer features than those considered in the state of the art: if the existing works have focused on “simple” features (speaker gender, language of the utterance, ...), we will also consider information related to the diversity of the languages and to the linguistic characteristics of these languages (phonemic inventory, identification of tonal languages, ...). 2. In addition to existing analysis methods (e.g. linguistic probes), we will develop new methods to measure similarity between languages. Again, close collaboration between linguists and NLP researchers will be essential to define a linguistically relevant similarity (or similarities). 3. We will apply our methods to the 230 languages of the Pangloss collection (an archive of rare languages managed by LACITO) and to 15 creoles (collected mainly by LLL). These large-scale experiments will allow us to test state-of-the-art pre-trained models on languages with a wide variety of linguistic features rarely considered in NLP work. 4. We will apply these methods to language documentation support tasks, an application that has, until now, never been considered.