Tutorial — Picooz Helicopter
A complete MBD project, end to end
This six-part tutorial walks through a real Model-Based Design project from zero to deployment using a cheap toy helicopter rotor (the Picooz) as the plant.
You will acquire motor data on a running dsPIC33A target using External Mode, identify a physics-based model from the measurements (using five different methods — four of which need no MATLAB add-on toolbox), design a controller in Simulink, simulate the closed loop, generate embedded C code in one click, and validate on hardware — all with the same model.
This is the 30-minute demo that opens the “Model-Based Design: From Laboratory to Embedded Deployment” session at École Centrale Nantes (December 2025), expanded into a self-paced, reproducible tutorial with downloadable models, scripts, and data.
The six phases — choose your entry point
| # | Page | What you learn | Duration |
|---|---|---|---|
| 1 | Hardware | What the Picooz is, its electronics, why it makes a great MBD teaching example | ~10 min |
| 2 | Acquisition with External Mode | External Mode concept, chirp excitation, capturing live data from a running MCU | ~20 min |
| 3 | Identification — Five Methods | Grey-box physical model and five ways to fit it — from a closed-form backslash on base MATLAB all the way up to the System Identification Toolbox GUI | ~45 min |
| 4 | Controller Design | Closed-loop control — PI baseline, then Super-Twisting sliding-mode for the non-linear plant | ~30 min |
| 5 | Simulate & Deploy | Closed-loop simulation, one-click code generation to the target | ~15 min |
| 6 | Validate & Beyond | Hardware response vs. simulation — the “zero tuning iteration” payoff, then pointers to go further | ~15 min |
Audience. No prior MBD experience required. Basic Simulink familiarity helps. Every concept is introduced when it is first used.
Why a Picooz?
The Picooz is a ~€15 toy helicopter. Opening one up reveals a small brushed DC motor driving a plastic rotor against significant aerodynamic drag. As an MBD teaching vehicle it hits every sweet spot:
| Property | Why it matters |
|---|---|
| Cheap (~€15) | Buy spares. Break things. |
| Safe | A few grams, a plastic rotor on a bench clamp. |
| Non-trivial dynamics | Non-linear aerodynamic drag — forces grey-box identification, a linear model is not enough. |
| Sensor-less speed | Speed is inferred from back-EMF, the same technique used in industrial brushless controllers. |
| Complete | Every MBD phase (acquisition · identification · control · deployment · validation) is demonstrated in an afternoon. |
Once you have the workflow in your hands, apply it to a PMSM motor, a buck converter, or an industrial axis — the four-phase loop is identical.
The four phases you are about to walk through are the loop of every MBD project — Picooz is simply a concrete vehicle for them:
Acquire → Identify → Control → Validate (then Deploy and Validate again on hardware).
What you need before starting
Hardware — a Picooz (or any small brushed DC motor) on a clamp, plus one of:
- dsPIC33AK Curiosity Platform Development Board (EV74H48A) — used by default in the tutorial models
- Curiosity Nano EV17P63A — also supported
- Any dsPIC33C / PIC32 / SAM board — the tutorial models rebuild for any supported target by changing one field in the Master block
Software
- MATLAB + Simulink + Embedded Coder (any release R2018b or newer)
- MPLAB Device Blocks for Simulink — install from MATLAB File Exchange
- MPLAB X IDE + XC-DSC compiler (free from Microchip )
- No other toolbox is strictly required — methods L0-L2 on the identification page and the Super-Twisting controller all work on base MATLAB
See the Installation page for the complete setup.
Downloadable assets — reproduce the entire tutorial
Every page ships a self-contained asset bundle: the Simulink model, the recorded dataset, and the scripts. Reproduce, modify, adapt.
Reference material
- This tutorial expands the live 30-minute demo delivered at École Centrale Nantes on 2025-12-04 (Model-Based Design: From Laboratory to Embedded Deployment) into a self-paced, reproducible walkthrough — every figure and dataset from that demo is re-used in the six pages below.
- Rapid Control Prototyping course, INSA Lyon 2022 — 3-hour teaching module including fixed-point deep-dive and multi-rate scheduling (reveal.js slides, French).
- ctrl-elec project — advanced motor-control algorithms and educational material.
Ready? → Start with 1. Hardware .
1. Hardware
What is the Picooz? The Picooz is an inexpensive IR-controlled toy helicopter (~€15). Inside it you find:
Read more →2. Acquisition with External Mode
Goal. Record how the motor reacts to a known PWM input signal, so we have the data needed to fit a mathematical model in the next step. 💻 Try the …
Read more →3. Identification — Five Methods
Goal. Turn the recorded pair \((u(t), \omega(t))\) into a mathematical model we can simulate — and do so without relying on an expensive add-on …
Read more →4. Controller Design
Goal. Design a controller that drives the rotor to a commanded speed setpoint, rejecting disturbances (wind, battery droop, rotor wear), and is cheap …
Read more →5. Simulate & Deploy
Goal. Verify on the PC that the identified model + designed controller behave correctly, then generate embedded C code and flash the target — in one …
Read more →6. Validate & Beyond
Goal. Verify that the real hardware response matches the simulated prediction — and draw the conclusion that gives MBD its name. The final check — …
Read more →