Development of model free longitudinal controllers for autonomous vehicles - Internship Toulouse

EasyMile SAS; Toulouse; Hybride; EUR 10 000 - 12 000; Plein temps

Avantages offerts par ce poste

1000€ gross monthly

Tickets restaurant

CSE

Résumé du poste

EasyMile SAS in Toulouse invites a Master’s/Engineering student to join the EZDrive-Control team for a 6-month internship focused on developing model-free longitudinal controllers for autonomous vehicles.

You will design and test MFC and ADRC, implement algorithms in C++ within ROS, and benchmark against model-based controllers in simulation and on real prototypes, with English and French language skills valued.

Qualifications

• Master’s/Engineering degree in Automatic Control, Robotics.
• Solid knowledge of control theory, state estimation, and observer design.
• Experience with MATLAB or Python and C++.
• Familiarity with ROS, Linux development, Model-Free Control or ADRC is desirable.
• Proficient in English and French.

Responsabilités

• Review literature on MFC and ADRC for autonomous vehicles and ADAS.
• Design MFC (e.g., iPID) and ADRC architectures and derive tuning rules.
• Benchmark performance against model-based controllers in simulation across mass, friction, and slope variations.
• Implement control algorithms in C++ within EasyMile’s ROS-based software stack.
• Validate controller performance on a HIL setup or a real prototype vehicle.

Connaissances

Control theory

State estimation

Observers and controllers design

Autonomy

Team spirit

Curiosity

English

French

Formation

Master’s/Engineering degree in Automatic Control or Robotics

Outils

MATLAB

Python

C++

ROS

Linux

Description du poste

Internship details

• Duration: 6 months (starting date is flexible)

• Location: Toulouse

• Team: EZDrive-Control

• Internship subject: Development of model free longitudinal controllers for autonomous vehicles

• Compensation: 1000€ gross, tickets restaurant, CSE

Internship Context

The EZDrive-Control team is responsible for vehicle dynamics and lateral/longitudinal control across EasyMile’s driverless platforms. Accurate longitudinal control (speed and acceleration) is critical for operational safety, stopping precision, and load stability in industrial transport and logistics vehicles. EasyMile currently relies on model-based controllers. However, massive variations in vehicle payload (ranging from empty to fully loaded platforms) create severe dynamic non-linearities, making model calibration complex across diverse operational scenarios.

The objective of this internship is to investigate and design Model-Free Control (MFC) and Active Disturbance Rejection Control (ADRC) to handle large payload changes. The intern will focus on controller synthesis, disturbance estimation, and formulating tuning methods that reduce calibration effort while maintaining high dynamic performance.

Missions / Responsabilities

• Review literature on MFC and ADRC for autonomous vehicles and ADAS.

• Design MFC (e.g., iPID) and ADRC architectures and derive systematic tuning rules.

• Benchmark performance against model-based controllers in simulation across mass, friction, and slope variations.

• Implement control algorithms in C++ within EasyMile’s ROS-based software stack.

• Validate controller performance on a HIL setup or a real prototype vehicle.

There is no typical profile at EasyMile, we all come from different backgrounds and that is what makes us strong! We are looking for:Student in Master’s/Engineering degree in Automatic Control, Robotics.Technical Skills: Solid knowledge of control theory (e.g., state estimation, observers and controllers design). Proficiency in MATLAB (or Python) and C++.Desirable: Experience with ROS, Linux development, and Model-Free Control / ADRC or control skillsSoft skills: Team spirit, autonomy, and curiosityLanguage skills : English and French