John Cockerill, enablers of opportunitiesDriven since 1817 by the entrepreneurial spirit and thirst for innovation of its founder, the John Cockerill Group develops large-scale technological solutions to meet the needs of its time: facilitating access to low carbon energies, enabling sustainable industrial production, preserving natural resources, contributing to greener mobility, enhancing security and installing essential infrastructures.Its offer to businesses, governments and communities consists of services and associated equipment for the sectors of energy, defence, industry, the environment, transports, and infrastructures. With more than 8,000 employees, John Cockerill achieved a turnover of € 1.417 billion in 2024 in 28 countries, on 5 continents.John Cockerill has been providing its expertise to the defense industry for 200 years. Today, John Cockerill Defense offers Cockerill® modular gun turrets from 25 to 120 mm caliber, Arquus® high-mobility vehicles, technical and tactical training on these systems, and Agueris simulators.Your Mission:The goal of this internship is to develop a robust and precise control algorithm based on machine-learning principles, simulate its control performances and calibrate it against real turret motorization measurements. The control algorithm should be developed in Matlab/Simulink environment and should be able to interface different type of commands (velocity, position and error commands).By leveraging ML techniques-such as reinforcement learning or supervised learning-the goal is to develop a more adaptive and intelligent control strategy that improves precision, responsiveness, and robustness of turret movements.1. Exploration and need definitionYou will study the motor control algorithm principles and define the place where the machine learning algorithm can take place. You will define the desired control behaviour, as well as the performance to achieve (accuracy, response time,…), based on the hardware constrains.2. Data collection and preprocessingYou will define which data will be used by the model, collect them and put them in form to train its machine learning algorithm.3. Machine learning model designYou will define which model to use, and train its model using the data from the data collection. You will also define a method to train its model against specific turret system.4. MIL and HIL testsThe controller model will be tested using only software simulating tools (Matlab/Simulink) to validate its behaviour and performance. In a second step, the hardware constrains will be added and the model will be validated in a laboratory environment.Subject discussed:
Machine Learning
Motion control algorithm
Matlab / Simulink
Your Profile:
Student in a master's degree education level in electronic
Minimum level B2 in French and English
Available for at least of 2-3 months
Your Internal Support:During this internship, you will be supported by the R&D department and the prototyping department of John Cockerill Defense. The internal supervisors will be:
Matthieu Close (R&D engineer) - project definition
Fabrice Trotta (Test engineer) - day-to-day support - tester support
Quentin Gaspart - FW support
What we offer you:
Immersion in a stimulating, technology-driven environment
Concrete and rewarding responsibilities
A caring and passionate team
We look forward to receiving your application and meeting you.Discover our job opportunities in details on
