Adaptive EMG Game Controller for Rehab Feedback

ISEF Category: Embedded Systems

Subcategory: Other  ·  Difficulty: Advanced  ·  Setup: School Lab  ·  Time: Full Year

The Hook

A video game controller can do more than move a character on screen. It can also measure how hard your muscles are working. That makes it a neat bridge between rehab, biofeedback, and embedded systems. If the controller can adjust itself when your muscles fatigue, you can test whether the game keeps people engaged longer.

What Is It?

This project combines electromyography, or EMG, with a game controller. EMG measures the tiny electrical signals your muscles make when they contract. Think of it like listening to a muscle instead of looking at it. A sensor such as MyoWare picks up that signal, and an ESP32-S3 can send game inputs over BLE-HID, which means Bluetooth Low Energy Human Interface Device. That lets your device act like a keyboard, mouse, or controller.

The cool part is the feedback loop. As a muscle gets tired, the EMG signal changes. One useful clue is spectral median frequency, which is the middle frequency in the signal’s power spectrum, or frequency breakdown. When fatigue rises, that value often drops. Your system can use that shift to lower game difficulty, change the target, or give rest prompts. That turns raw muscle data into a real-time decision.

Why This Is a Good Topic

This is a strong science fair topic because you can measure something real, change one variable, and see how the system responds. You can test signal quality, fatigue detection, response latency, and user engagement without needing a hospital lab. The topic connects to rehab therapy, gaming, and wearable tech, so it has real-world value. You also get to learn sensor design, signal processing, and embedded programming in one project.

Research Questions

• How does EMG signal quality change as muscle fatigue increases during repeated contractions?
• What is the effect of different filtering methods on the accuracy of spectral median frequency estimates?
• Does adaptive game difficulty keep users active longer than a fixed-difficulty version?
• To what extent does electrode placement change the stability of EMG-based fatigue detection?
• Which fatigue threshold gives the fastest and most reliable difficulty adjustment in the game?
• How does the latency of BLE-HID input affect player performance and biofeedback timing?

Basic Materials

• MyoWare EMG sensor kit.
• ESP32-S3 development board with BLE support.
• Disposable surface electrodes.
• Breadboard and jumper wires.
• USB cable for programming and power.
• Computer with Arduino IDE or PlatformIO.
• Simple game software or browser-based test game.
• Smartphone or camera for recording screen response.
• Digital timer or stopwatch.
• Notebook or spreadsheet for data logging.

Advanced Materials

• MyoWare EMG sensor kit or equivalent EMG front end.
• ESP32-S3 development board with BLE-HID support.
• Surface EMG electrodes and lead wires.
• Oscilloscope or data acquisition interface for signal validation.
• Reference resistor and calibration setup for sensor testing.
• Motion capture or force sensor for ground-truth comparison.
• Electrode gel, prep pads, and skin-cleaning supplies.
• Computer with Python for signal processing.
• External battery supply for wearable testing.
• 3D-printed or laser-cut controller housing.

Software & Tools

• Arduino IDE: Programs the ESP32-S3 and tests BLE-HID controller behavior.
• Python: Filters EMG data and computes median frequency features.
• NumPy: Handles numerical arrays and signal calculations.
• SciPy: Applies filtering, spectral analysis, and statistical tests.
• ImageJ: Measures response timing from recorded screen or camera footage if needed.

Experiment Steps

1. Define the user action your controller will measure, such as a grip, pinch, or forearm flex.
2. Choose the fatigue signal feature you will track, then decide how you will validate it against a baseline.
3. Design the control loop that converts EMG changes into game difficulty changes, rest prompts, or score adjustments.
4. Plan a comparison between adaptive and fixed versions of the same game so you can test adherence or engagement.
5. Build a data workflow for collecting raw EMG, processed features, and gameplay outcomes in the same session.
6. Set controls for electrode placement, posture, and trial order so your results are not driven by setup drift.

Common Pitfalls

• Using loose or inconsistent electrode placement, which changes the EMG signal more than the fatigue effect you want to study.
• Treating noise spikes as fatigue, which makes the controller react to motion artifacts instead of muscle state.
• Calibrating with one person and expecting the same threshold to work for every user, which hides big body-to-body differences.
• Changing game difficulty too fast, which makes the feedback feel random and breaks the link between muscle fatigue and system response.
• Comparing raw EMG amplitude only, which can miss fatigue changes that show up better in the signal spectrum.

What Makes This Competitive

A stronger project would not just prove that EMG works. It would compare multiple fatigue features, test more than one control strategy, and report how well the system adapts for different users. You can also add latency testing, repeatability checks, and a clear comparison between fixed and adaptive gameplay. That kind of analysis shows you understand both the sensing side and the design side.

Project Variations

• Test the same controller with different muscles, such as forearm flexors versus upper-arm muscles, to see which gives cleaner fatigue detection.
• Swap the game task for a rehab-style repetition challenge and measure whether adaptive feedback improves completion rate.
• Compare spectral median frequency with RMS amplitude or muscle activation burst timing to see which feature tracks fatigue best.

Learn More

• PubMed: Search for review articles on surface EMG, muscle fatigue, and biofeedback in rehabilitation.
• NIH Rehabilitative Sciences resources: Look for articles and summaries on therapy adherence, sensor-based rehab, and patient monitoring.
• IEEE Xplore abstracts: Search for papers on EMG interfaces, BLE-HID controllers, and adaptive game systems.
• MIT OpenCourseWare: Search for signals and systems, digital signal processing, and embedded systems lecture notes.
• Arduino documentation: Read about BLE peripherals, HID behavior, and ESP32-S3 programming support.
• NASA NTRS: Search for human factors and bioinstrumentation papers if you want more on signal processing and wearables.