INTELLIGENT PREDICTIVE MAINTENANCE FOR UAVs: A DIGITAL TWIN-BASED APPROACH

Unmanned Aerial Vehicles (UAVs) have experienced rapid growth across a wide range of application domains, including infrastructure inspection, precision agriculture, logistics, civil security, and defense. This widespread adoption is accompanied by increasingly stringent requirements in terms of safety, availability, and reliability, in compliance with European regulations such as the Specific Operations Risk Assessment (SORA) framework. In this context, maintenance has become a major challenge, as critical UAV components—particularly propulsion systems—are subject to aging and degradation processes that may lead to costly failures or even mission loss.

Current maintenance strategies primarily rely on scheduled inspections or post-flight analysis of recorded flight data. These approaches do not enable effective anticipation of failures nor optimization of maintenance operations.

The objective of this PhD project is to develop an innovative predictive maintenance framework based on the integration of a Digital Twin continuously updated in real time using data collected from onboard sensors and the flight controller, combined with advanced Artificial Intelligence (AI) and Deep Learning techniques. This framework will enable continuous health monitoring of UAV propulsion systems, early anomaly detection, estimation of the Remaining Useful Life (RUL) of critical components, and optimization of maintenance decision-making.

The proposed methods will be experimentally validated using instrumented propulsion test benches and experimental campaigns on real UAV platforms in order to demonstrate their robustness, accuracy, and potential for industrial deployment.

ESTACA, an engineering school and member of the ISAE Group, offers a five-year engineering program dedicated to training engineers passionate about technologies addressing the challenges of future mobility. The school also conducts applied research serving stakeholders in the transportation sectors, including aerospace, automotive, space, naval, and rail transportation.

ESTACA offers engineering degrees and specialized master's programs accredited by the French Commission for Engineering Degrees (CTI). It currently welcomes more than 2,500 students across three campuses located in Montigny-le-Bretonneux (Paris-Saclay), Laval, and Bordeaux.

ESTACA'Lab, ESTACA's research laboratory, brings together around thirty faculty members and researchers, as well as approximately forty PhD candidates. The laboratory conducts collaborative applied research focused on sustainable, intelligent, and safe mobility.

MAIN RESPONSIBILITIES :

The PhD candidate will join ESTACA'Lab and will be responsible for :

Réaliser un état de l'art sur les jumeaux numériques, la maintenance prédictive et l'intelligence artificielle appliqués aux UAVs.

• Conduct a comprehensive literature review on Digital Twins, predictive maintenance, and Artificial Intelligence applied to UAVs.
• Develop a real-time Digital Twin of UAV propulsion systems.
• Design Deep Learning models for anomaly detection, fault diagnosis, and Remaining Useful Life (RUL) prediction.
• Develop multi-sensor data fusion methods using IMUs, ESCs, motors, batteries, flight controllers, and onboard sensors.
• Design an embedded health monitoring architecture for online health state estimation.
• Conduct experimental campaigns using instrumented test benches and UAV platforms.
• Evaluate the proposed methods in terms of robustness, accuracy, computational efficiency, and generalization capability.
• Disseminate research results through publications in leading international journals and conferences, and through technological demonstrators developed with industrial partners.