Objectives and deliverables
The RenovAIte project develops: :
– software based on artificial intelligence to optimize the renovation of housing and roads, on a large scale
-an online platform to bring together and connect data and AI tools for the renovation sector.
Its objectives are multiple:
-improve the overall performance (economic, environmental and social) of renovation projects (savings in energy, materials, social diversity, suitability for local needs, resilience to climate risks, etc.);
– support the acceleration of the pace of renovation;
– integrate environmental footprint data, climate risks, and social data into calculations for renovation; – democratize scientific results and the experience of the most successful projects by integrating them as much as possible into AI models and software rules.
Initially, this software will be applied within its consortium. As an indication, The major French social landlord renovates tens of thousands housing units per year and Eurovia, a subsidiary of the VINCI group, maintains a network of more than 70,000 km of roads (including a large part in France and Germany).
In a second step, they will be offered and adapted more widely through the platform under development..
Work packages
Coordination and platform
Coordination
In addition to the overall management of the project, Leonard ensures:
– support for the strategic and technical management of AI technologies;
– a dynamic of sharing knowledge and scientific results within the project;
– coordination of relations with the renovation and tech network, with a view to scaling up the project.
Plateform & data
ALEIA uses its collaborative and sovereign platform dedicated to AI, and develops:
– a dedicated version of the RenovAIte platform of its marketplace technology, for a collaborative one-stop shop for the entire renovation sector
– an environment dedicated to R&D and experimentation for project developments;
– its functionalities for putting AI models into production;
– monitoring and dialogue for the identification and integration of all data sets potentially useful for renovation.
Housing renovation
AI technologies for decision support, inspection, diagnosis and solution design support can pool their approaches regardless of the infrastructure studied. RenovAIte, in order to pave the way for the development of these technologies for all types of infrastructure, proposes two use cases: housing and the road. 5 work packages concern the renovation of housing.
Urban renewal strategies and testing grounds
The major French social landlord, through its activity of developing social and intermediate housing but also through projects supported by the National Agency for Urban Renovation:
– offers four experimental sites to the project, in mainland France (Valentin, Garge-Lès-Gonnesses) and overseas (Reunion, Mayotte);
– uses its expertise in urban renewal, rehabilitation and digitization of heritage (BIM in particular);
– leverages its possible structured datasets to explore the potential of the different AI models developed in the other work packages.
Application of Adversarial Resilience Learning to renovation
OFFIS has developed advanced AI technology aimed at solving complex problems with resilient solutions. Their solution is composed of an AI model that offers solutions that train against an AI model that generates crises:
– applied to renovation, OFFIS is developing an AI model that offers decision support regarding the renovation strategy for housing assets;
– to bring a dimension of resilience to climate risks, OFFIS is developing, supported by Resallience, an AI model that simulates climate crises;
– once the experimentation phase is over, OFFIS will put the model into production and work on how to roll out this approach to all housing asset managers.
Data acquisition (AI) for renovation
Accompanied by ALEIA, The major French social landlord creates through this work package:
-a proposal for a minimum base of data to be acquired at the scale of the building complex to allow software assistance for renovation assistance, called “renovation passport” (concept taken from past experiments);
– a method for acquiring this data at the scale of the residential building, in order to allow managers of non-digital assets to access the technologies developed within RenovAIte;
– a collection of experiences in data capture (sensors, measurement tools) and data processing (software, algorithms with or without AI).
Performance assessment for resilience and climate change
Resallience provides RenovAIte with its expertise in studying the impact of global warming on infrastructures to:
– create a diagnostic tool for the environmental performance of a renovation project (including both mitigation and adaptation to climate risks);
– bring its experience in data processing to anticipate changes (climate, satellite, geographic data);
– integrate scientific results and global warming scenarios produced by the Intergovernmental Panel on Climate Change (IPCC) into all work packages.
Optimized solutions for energy efficiency in housing renovation
Following previous work to help with financing strategies and the diagnosis of housing assets, the renovation company Qivy Habitat and the Lab Research Environment (partnership between VINCI and ParisTech for environmental performance) wish to explore the potential of AI to propose work recommendations.
Road renovation
Long-term planning of road renovation
VINCI Construction’s Mérignac research center and its AI team:
– develop a physical data acquisition sensor on the roads;
– create a road renovation planning model from this data;
– will make it possible to supplement the theoretical planning of road maintenance with the actual data captured by their technologies.
AI for road condition inspection
In addition to long-term data, the Mérignac research center works within VINCI Construction on visual data:
– by training detection models allowing the automated inspection of roads from smartphones and vehicle fleets;
– using these results, supported by VINCI’s road infrastructure expertise, to produce renovation recommendations for the appropriate authorities (local authorities, private operators).
Real-time route management
VIA IMC, a German subsidiary of VINCI specializing in digital innovation for road infrastructure, is developing different AI models:
– to estimate the safety risks presented by different road infrastructure configurations and develop preventive renovations;
– to improve the environmental performance of roads (adjustments to climate risks, carbon footprint of projects);
– to reduce the cost of inventory and maintenance operations.
This work works specifically on data from modern vehicle driving assistance sensors, as well as open data from cities.
