The ROAD-AI challenge is a collaborative project between INRIA and Cerema, started in 2021. The objective is to study some innovative solutions that may ease the management of transport infrastructures (roads, structures).
Within this context, the following PhD thesis is proposed to develop a specific solution for the maintenance of secondary road assets, i.e., low-traffic roads, from video-based surface monitoring data only, namely, pavement images.
Keywords: Statistical modeling, image segmentation, machine learning, non-destructive testing
Thesis objectives and contents
To manage and maintain the performance of road assets, stakeholders have been using statistical models to predict road degradations with time and trafic, namely, roadway evolution and/or survival law models. These models are based on measurements of different specific physical quantities, which are related to pavement management and carried out in situ by specialized operators. This management approach prevails on the high-traffic national network and might be adapted to the secondary road network. Compared to the national network, the secondary road network experiences lower trafic and is less instrumented and documented.
The objective of the thesis will be, on the one hand, to assess the feasibility of updating the statistical degradation model laws of roadway from video-based imaging the pavement surface conditions; and on the other hand, to adapt the latter approach to the secondary road network. The expectations of the thesis require a large enough data base of labelled pavement images along with the corresponding detailed knowledge of the underlying pavement structure. Images labelling (with annotation, record of disorders, etc.) will be provided for the study and may cover a 5 years period at least.
To meet the objectives of the thesis, the following issues are considered:
o the existing survival/evolution laws which are based on both the structural and the surface conditions of the roadway.
o the existing automatic image processing techniques, e.g., segmentation and feature engineering, to estimate surface condition indicators.
o the simplifying hypotheses that could be applied on the secondary road network
- The selection and the computation of the surface condition indicators within pavement images which then serve as input parameters to the statistical degradation laws.
- The determination of the relevant image analysis scale being adapted to the estimation of the roadway degradation laws; it will be determined in relation with the diagnostic needs, the gridding scale at which the pavement structure and data on road surface conditions can be provided, etc.
The development of the model versatility for further adaptation to various roadways context and specific management issues, e.g., climate changes, mid-moutain vs . sea-shore climate.
The performance assessment of the statistical model laws should take account of the accuracy on the estimated surface indicators. The performance of the algorithms will be compared to some benchmark method which is routinely used at the operational level.
Host teams: joint Inria/UGE project team (I4S) and Cerema research team (ENDSUM)
Working location: Strasbourg (Cerema)
Start: fall 2024
Duration: 3 years
Application deadline: 30th of June 2024
Management contact
Director: laurent.mevel@inria.fr
Supervisors: philippe.foucher@cerema.fr; vincent.baltazart@univ-eiffel.fr
Associated experts: alain.hebting@cerema.fr; fabien.menant@univ-eiffel.fr
Required Skills and scientific background
- Statistics, applied mathematics, pattern recognition and/or machine learning.
- Computing skills in Python and/or Matlab and/or R and/or Julia languages.
- Background on physics, mechanics and non destructive testing would be appreciated.
- Good writing skills and communication reporting in english. Mastery of Latex & Overleaf are a plus.
Benefits package
Restauration subventionnée
Transports publics remboursés partiellement
Congés: 7 semaines de congés annuels + 10 jours de RTT (base temps plein) + possibilité d'autorisations d'absence exceptionnelle (ex : enfants malades, déménagement)
Possibilité de télétravail (après 6 mois d'ancienneté) et aménagement du temps de travail
Équipements professionnels à disposition (visioconférence, prêts de matériels informatiques, etc.)
Prestations sociales, culturelles et sportives (Association de gestion des oeuvres sociales d'Inria)
Accès à la formation professionnelle
Sécurité sociale
Application
Send a CV, cover letter (in french or english) and academic transcripts to the three following emails: laurent.mevel@inria.fr, philippe.foucher@cerema.fr, vincent.baltazart@univ-eiffel.fr