VIABLE : R&I project to imagine the future of air mobility
The VIABLE project encompasses a wide range of focus areas, all aimed at shaping the mobility of the future. Key domains include:
- Developing sustainable solutions to reduce energy impact
- Advancing the capabilities of autonomous systems
- Innovating methods for modelling, simulation, and intelligent maintenance
Internship objectives
- State of the art in energy management systems (EMS)
- Implementing a reinforcement learning algorithm for optimal eVTOL energy distribution
- Write technical documentation & use the GIT versioning tool
- Participate in an international conference
Year
2023
Skills
Reinforcement Learning, Energy Management System, Python, PyTorch
Client
Capgemini Engineering
Team:
VIABLE R&I project
Results
- State of the art on existing energy management systems and reinforcement learning for hybrid vehicles
- Mathematical formulation of the hybridization model between a battery and a fuel cell in Python
- Implementation of a Q-Learning reinforcement algorithm for optimal power distribution according to a power profile
- Test and validation of results in comparison with the rule-based method (SMC: State Machine Control). The criteria analysed comply with constraints : hydrogen consumption and maintenance of the battery SOC (State of charge)
- Implementation and validation of a deep learning reinforcement algorithm (DQN: Deep Q-Network)
- Use of the GIT version management tool to ensure development traceability
- Detailed documentation of the developed code
- Participation in a paper submited in MEA'24 - More electric aircraft conference in Toulouse
Previous & Next
