2024 – 2025
DELIVERWHO A-LIVR8A
MIA Referent — Model, Intelligence, Autonomy
MATLAB SimulinkPID ControlKiCad PCBSolidWorksMPU-6050
Mechatronics capstone: 2-wheeled self-balancing campus delivery robot. Defined behavior laws, built a MATLAB Simulink digital twin, co-led PID stabilization and MPU-6050 sensor fusion.
Context
ICAM Toulouse mechatronics capstone project. The brief: design and build an agile 2-wheeled delivery robot for the ICAM campus, capable of carrying 5 kg packages (50×30×30 cm), with 2 hours of autonomy, PS5 DualSense controller interface, and a roadmap toward autonomous LIDAR/GPS navigation.
Team
- Clara Vialard — Project Lead
- Erwan De Saint Germain — Quality
- Carla Ferrand — Procurement & Manufacturing
- Paul Chamussy — CAD
- Nathan Pascual — Structural Engineering
- Léo Cirouant — Electronics
- Ori Dhumez-Biagi — MIA (Model, Intelligence, Autonomy)
My Role
- Defined the mathematical behavior laws governing balance and locomotion
- Developed the MATLAB Simulink model for system simulation (PID controllers, motor dynamics) — used as a digital twin for pre-build validation
- Co-led programming with Erwan: PID stabilization loop, MPU-6050 IMU sensor fusion, motor control firmware
- Participated in stability tests and direction-control validation sessions
Technical Highlights
- Mechanics: SolidWorks CAD + topology optimization + finite element analysis (Von Mises stress, displacement, deformation)
- Electronics: Custom PCB designed in KiCad with PlatformIO firmware. System architecture in SysML.
- Control: PID balance control, MPU-6050 IMU sensor fusion, PS5 DualSense Bluetooth interface
- Manufacturing: Fab-lab 3D printing (PLA), €500 budget ceiling verified with SolidWorks Costing
- Sustainability: SolidWorks Sustainability assessment (component supply chains: Asia → Europe and Europe → Europe)
- Project management: OpenProject (Gantt + Kanban), GitHub version control
Test Results
- Motor control: validated
- Stability (2-wheel balance): partially validated — minor oscillation at steady state
- Direction + balance combined: partially validated
Next Steps Identified
- Custom PCB integration (replacing breadboard prototype)
- LIDAR + ultrasonic sensors for autonomous navigation
- GPS module for outdoor routing
- Bumpers and mudguards
- Full CAD completion with electrical integration