Rouen smart mobility for everyone

• Partners: Lead partner: Métropole Rouen Normandie
• Call for projects: call for projects Territoire d’Innovation Grande Ambition (TIGA)
• CESI budget for the project: €443k
• Total PIA funding: €5.2m
• Project launch: Phase 2 – January 2020
• Project duration: 48 months for phase 2 of implementation

The Rouen Metropolitan Area and its 36 partners want to develop a large-scale integrated mobility system that will enable the development of chosen modes of transport without constraints, while reducing their environmental and physical impact. The project will profoundly transform the way people travel thanks to innovative solutions, whether in terms of autonomous, carbon-free, digital, shared, and connected mobility, control of public space usage, or support for behavioral change.

CESI LINEACT’s involvement in the project:
The CESI LINEACT research team at the Rouen Campus is involved in three main areas:

• Participation in the project’s scientific and technical steering committee,
• Development of R&D work on big data analysis, simulation, and the development of predictive models that represent significant challenges for the provision of user services and the development of decision-making tools for city stakeholders in the context of smart cities and intelligent transportation systems,
• Support for ideation and innovation, including benchmarking existing systems and facilitating ideation workshops as part of the project’s Living Lab.

Achievements as of March 31, 2025:

Hypervisor Action.

• The scientific paper submitted to the IEEE ITSC (Intelligent Transportation Systems Conference) was accepted and presented in Thailand. This paper introduces an innovative methodology for generating a realistic synthetic population adapted to autonomous vehicle on-demand (AMoD) services.
• The deep learning model, which aims to integrate the semantic dimension of the data to better understand passengers’ sequential travel habits, is currently being developed and validated. A preliminary prototype was proposed and successfully tested on a small sample of simulated data, showing a promising improvement in the accuracy of predicting users’ travel sequences.
• In addition, the demand forecasting algorithm has been refined to achieve hourly granularity, thereby increasing accuracy and enabling finer supervision of the system.
• Finally, work has shifted more towards dynamic carpooling, specifically addressing issues related to vehicle sharing and optimal passenger sequencing. A model integrating real-time dynamic management of passenger demand and preferences is currently in an advanced stage of validation through simulation.
• These contributions were the subject of an in-depth report and a successful thesis defense in January 2025, confirming the significant benefits of the approach developed for smart and sustainable urban mobility.

In the long term, work will focus on the following actions:

• Extension and generalization of the deep learning model to incorporate more contextual and behavioral dimensions of users, in particular by exploiting advanced architectures such as attention neural networks to better capture the complex dynamics of urban mobility.
• Large-scale validation of the dynamic carpooling management model, integrating real-time consideration of passenger requests, based on experiments with extensive and varied datasets from diverse urban contexts.
• Development of an operational platform enabling the continuous integration of real-time data (vehicle GPS, mobile data, socio-economic events, and public calendars) for dynamic predictive adjustment and optimal system responsiveness.
• Exploration of integrated optimization strategies to dynamically adapt pricing and resource management (vehicle fleet, availability, routes) based on instantaneous variations in demand and operational constraints, with a view to sustainably improving the profitability and environmental efficiency of the service offered.

Living Lab Action:

• Actions taken as organizer:
  o Organized a visit to Le Mix for a delegation from the C-Care project of the INTERREG FMA program
  o Led two “Mobility Fresco” workshops to raise awareness among 120 learners about the carbon challenges of personal mobility
  o Facilitation of a Lego® Serious Play® workshop with 10 learners to identify mobility-related problems in Madrillet and propose possible solutions
  o Organization of a visit to the CESI Rouen Campus for Mix members
• Actions taken as a participant:
  o Participation in co-creation workshops organized by Le Mix
  o Participation in partner meetings organized by Mix
  o Participation in the “Mobility Fresco” workshop organized by Mix
  o Participation in Mix partner visits organized by Transdev, Les Copeaux numériques, and Energytronik
  o Participation in the “Climate Fresco” workshop organized by Mix
  o Participation in the “2 tons” workshop organized by Mix
• Characterization of a LivingLab:
  o Interview with 6 OpenLabs (Le WIP, La French Tech Caen, Normandie Incubation and Espace Public Numérique de Gonneville en Auge, Le Dôme, and Blue LivingLab by Nausicaa)
  o Review of the literature on OpenLabs and LivingLabs
  o Definition and distinctions of OpenLabs
  o Identifying the characteristics of a LivingLab
  o Presentation of a scientific article at a conference: E. Pillon “Exploring stakeholder-defined activities and outcomes in Living Labs: the case of Living Lab Mix.” In proc. WOIC 2023, Bilbao, November 2023
  o Submission of a scientific article at a conference: E. Pillon “Uncovering Living Labs’ Activities and Outcomes: Key Insights for Stakeholder Expectations,” in: Proc. R&D Management 2024, Stockholm, June 2024
• Role of Living Labs in business model innovation:
  o Recruitment of a doctoral student (Gaëtan SAVARIT) since December 2021
  o Scientific article in a conference: G. Savarit, E. Pillon, A. Louis, “Business model development through testing: lessons from both the business model literature and Living Lab literature regarding stakeholders’ engagement and context,” in: Proc. EURAM 2023, Dublin, June 2023.