Water Rocket Attitude Control With Reaction Wheels

ISEF Category: Robotics and Intelligent Machines

Subcategory: Control Theory  ·  Difficulty: Advanced  ·  Setup: School Lab  ·  Time: 1 to 2 Months

The Hook

A small push can throw a spinning system off balance fast. That is the whole challenge of attitude control, the art of keeping a body pointed where you want it. With one reaction wheel and an IMU, you can test three classic control methods and see which one recovers best after a sudden nudge.

What Is It?

This project studies attitude control, which means keeping an object pointed in a chosen direction. Think of it like trying to balance a broom on your hand, except your hand is a motorized wheel and your eyes are an IMU, a sensor that measures motion and orientation. When the system gets disturbed, the controller decides how hard to spin the reaction wheel so the body comes back to level.

You will compare PID, LQR, and sliding-mode control. PID uses present, past, and future error trends to correct motion. LQR picks motor commands by balancing stability against effort. Sliding-mode control pushes the system toward a target path even when conditions change, which can make it tough against disturbances. The fun part is that each controller can look good on paper, but the IMU data will tell you which one really handles a sudden side push.

Why This Is a Good Topic

This makes a strong science fair topic because you can change one control method at a time and measure clear outputs like overshoot, recovery time, and steady-state error. It connects to drones, satellites, robotics, and self-balancing systems, so the real-world use is easy to explain. You can learn control design, sensor fusion basics, data plotting, and fair testing without needing a full university lab.

Research Questions

• How does PID, compared with LQR, change recovery time after a sudden lateral nudge?
• What is the effect of sliding-mode control on overshoot during attitude recovery?
• Does increasing the reaction wheel response limit improve disturbance rejection across all three controllers?
• To what extent does IMU noise change the measured difference between controller types?
• Which controller uses the least motor effort while still returning the system to level?
• How does the size of the disturbance affect which controller performs best?

Basic Materials

• Single-board microcontroller such as Arduino or ESP32.
• IMU sensor module with accelerometer and gyroscope.
• Small DC motor or brushless motor for a reaction wheel.
• Motor driver matched to your motor.
• 3D-printed or laser-cut gimbal stand.
• Assorted fasteners, nuts, bolts, and standoffs.
• Bench power supply or protected battery pack.
• Digital kitchen scale or small balance scale.
• Tape measure or ruler.
• Laptop for coding and data logging.
• Safety glasses.

Advanced Materials

• Higher-quality IMU with lower drift and known calibration data.
• Precision motor controller or ESC with closed-loop support.
• Load cell or force sensor for disturbance input measurement.
• Motion capture system or high-speed camera for independent angle checks.
• Vibration-isolated mounting plate.
• Torsion springs or calibrated disturbance mechanism.
• 3D printer with stronger filament options.
• Power analyzer or current sensor for motor effort measurement.
• Soldering tools and test equipment.
• University machine shop access for rigid mounts.

Software & Tools

• Arduino IDE: Programs the microcontroller and logs sensor data from the IMU.
• Python: Plots angle, error, and recovery metrics for each controller.
• Jupyter Notebook: Keeps your analysis, notes, and graphs in one place.
• ImageJ: Measures angle from video frames if you cross-check the IMU with camera data.
• Excel: Organizes trial results and helps you compute summary statistics.

Experiment Steps

1. Define the one motion you want to control, then turn it into a measurable angle or orientation error.
2. Build a simple dynamics model so you know what each controller should try to correct.
3. Choose fair performance metrics, such as recovery time, overshoot, steady-state error, and motor effort.
4. Design the test stand so each controller faces the same disturbance and the same starting position.
5. Plan a tuning process for each controller that uses the same goal, not guesswork.
6. Set up your data pipeline before testing so IMU readings, timestamps, and trial labels stay clean.

Common Pitfalls

• Tuning PID by eye while letting LQR and sliding-mode use different performance goals, which makes the comparison unfair.
• Mounting the reaction wheel with too much flex, which adds vibration that looks like bad controller behavior.
• Trusting raw IMU angle output without checking drift, which can hide real recovery differences.
• Using a disturbance push that changes strength from trial to trial, which makes the results noisy and hard to defend.
• Measuring success only by final angle and ignoring overshoot, recovery time, and motor effort.

What Makes This Competitive

A stronger version of this project goes beyond picking a winner. You can build a fair comparison framework, then test each controller across several disturbance sizes and starting angles. You can also report more than one metric, such as error, energy use, and robustness to sensor noise. That kind of analysis shows you understand control theory, not just coding.

Project Variations

• Test the same controllers on a two-axis gimbal instead of a single-axis stand to see how coupling changes the results.
• Compare a cheap IMU with a higher-grade IMU to measure how sensor quality changes controller performance.
• Add a disturbance-rejection metric based on recovered angle area under the curve, then rank the controllers by stability and efficiency.

Learn More

• MIT OpenCourseWare: Search for feedback control and dynamic systems lectures to learn the math behind PID, LQR, and state-space control.
• NISE Network: Search for free classroom materials on control systems and robotics to build intuition before you tune your own system.
• NASA Technical Reports Server: Search for reaction wheel control and attitude stabilization papers to see how real spacecraft handle similar problems.
• PubMed: Search for review articles on sensor fusion, IMU drift, and motion estimation if you need background on measurement error.
• IEEE Xplore: Search for open-access or freely available conference papers on sliding-mode control and reaction-wheel platforms.