Joint chairs and laboratories

The OpenLab PSA@Paris-Saclay is a structure of excellence in research, created in 2016, which brings together PSA and the laboratories GeePs, L2S et SATIEIts aim is to develop skills and technological innovation in the field of electrical engineering through scientific research.

These include in particular:

• Propose and carry out concept studies to contribute to sustainable mobility,
• Develop disruptive technologies and explore methodologies with high scientific content,
• Ensure, in collaboration with the STELLANTIS expertise sector, the technological transfer of results and deliverables from research work with a view to automotive application within a 10-year horizon.

Lines of research

Openlab's research is structured around 4 axes:

• Components and systems for electrical energy
• Materials for electrical components
• Physics and Engineering of Electromagnetism
• Information systems and environmental perception

SPMS and Pytheas Technology have partnered to create a LabCom supported by the ANR, SPyMS, to design the electroactive materials of tomorrow. Pytheas Technology is an SME based in La Ciotat, specializing in the design and production of innovative piezoelectric devices, with applications in energy harvesting, electro-hydraulics, underwater acoustics, and vibration damping. SPMS has long-established expertise, particularly in electroactive materials. This LabCom is led by Pierre-Eymeric Janolin (SPMS, Director) and Alice Aubry (Pytheas Technology, Deputy Director).

Image

Image

Image

A research institute dedicated to smart electricity networks

The RISEGrid institute (Research Institute for Smarter Electric Grids) was jointly launched by Supélec and EDF (Électricité de France) in December 2012 and is dedicated to the study, modeling and simulation of smart electricity distribution networks and their interactions with the entire electricity system.

Research topics cover both theoretical and more practical and industrial aspects. The RISEGrid Institute aims to be perfectly aligned with the strong and stimulating developments in electricity systems around the world: the opening of electricity markets, the development of decentralized production, the ubiquitous availability of information and communication technologies, etc.

The research conducted within the Institute combines the multidisciplinary aspects of smart grids (electrical systems, control, information systems, telecommunications networks, etc.). In addition, new tools for modeling and simulating complex systems are studied in depth.

Scientific program in four axes

• Evaluation of new solutions for smart electricity networks

The rapid and significant development of decentralized electricity generation is having a significant impact on existing distribution networks, which were designed at a time when dispersed energy resources were marginal. In this context, the objective is to design new solutions to increase network flexibility, not only for the massive integration of decentralized generation, but also for new uses and applications of electricity, while achieving high levels of reliability and security.

• Observability of the electrical system

The development of new intelligent and automated functionalities will allow the optimization of the entire system, taking advantage of various flexibilities (generation, consumer demand management, storage, electricity grid flexibility). However, this requires improving the real-time observability of system components. Current trends are based on the use of intelligent and innovative signal processing methods or data analysis algorithms.

• Information and communication systems

Information systems and communication resources and infrastructure are essential for implementing new features in smart grids. The RISEGrid Institute is interested in modeling these systems, given their strong interactions with the electricity grid. Interdependencies, quality of service, and network reliability are some examples of the topics covered in this research area, as well as new approaches and simulation tools.

• Advanced Modeling and Simulation

Smart grids consist of many interacting subsystems (electricity grids, automated meter management, centralized and decentralized production, demand management (including smart charging of electric vehicles), storage, information systems, telecommunications network, etc.).
All these subsystems must be taken into account to obtain a sufficiently realistic representation of the behavior of the entire system. This requires the development of new multi-simulation tools aimed at combining dedicated subsystem simulators. New subsystems and interactions must also be studied, such as automated meter management systems or smart charging of electric vehicles.

LIAGORA, Generative Artificial Intelligence Laboratory for Agent-based Research Orchestration.

Created by the MICS laboratory and the ILLUIN TECHNOLOGY company with the support of the ANR, LIAGORA aims to enable the deployment and use of Generative AI within companies, while ensuring the use of trusted and frugal systems. Capable of creating original content, formulating answers to complex questions on a knowledge base, acting as a versatile assistant for a wide range of professions, generative AI offers considerable potential for optimizing business processes. ILLUIN is developing a suite of products capable of processing documentary data (search engine in complex documentary corpora, document parser) or conversational data (conversational agents, conversation parser, corpus analyzer), as well as a multimodal orchestration platform that allows the design of tailor-made AI use cases.

Image

However, the limitations of current assistants based on generative AI are still too numerous. They are too prone to hallucinations that result in inaccurate or inconsistent responses. They use a RAG (Retrieval Augmented Generation) system to respond to many queries; yet the RAG approach itself has many limitations, particularly in searching for information for complex queries within large specialized documentary corpora. They generally lack long-term context, which can lead to omissions or contradictions in extended conversations. They too rarely use the company's business rules, which can be formulated in various types of structures (documents, ontologies, databases, etc.). Finally, they require too large an infrastructure for their large-scale use to be acceptable for most companies.

The three major challenges within this LabCom will therefore be to develop assistants based on Generative AI:

• Efficient, by improving their relevance and diversifying possible use cases,
• Trusted, by reducing hallucinations through the optimization of the foundation models used (LLM, VLM, etc.) or information search paradigms (RAG, Agents, use of Tools, etc.),
• Frugal, to reduce the carbon, energy and financial impact of the systems developed (through hybridization, use of small foundation models, model compression, etc.) in order to enable massive use and adoption with reasonable consumption.