Haptic Glove Obstacle Avoidance

ISEF Category: Biomedical Engineering

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

The Hook

A glove can give you a new sense of space without a screen. If a tiny motor buzzes on the right finger, your brain can learn to treat that buzz like a warning sign. That idea could help people who are blind or have low vision move more safely. Your project tests whether simple vibrations can guide real obstacle avoidance.

What Is It?

This project turns distance data into touch signals. A smartphone depth camera sees objects in front of the user, then software decides how strong each vibration should feel on the glove. The glove acts like a map for your fingertips, where each motor stands for a direction or level of risk.

Think of it like a parking sensor you wear. A car beeps faster as it gets closer to a wall. Your glove does the same thing, but through vibration patterns instead of sound. A neural network, or CNN, short for convolutional neural network, can help the phone read depth images and turn cluttered scenes into a distance estimate the glove can use.

Why This Is a Good Topic

This is a strong science fair topic because you can test a clear question, collect measurable data, and compare designs. You can study how vibration pattern, motor placement, or mapping style affects obstacle avoidance. The topic connects to assistive tech, human factors, and mobile sensing, so the real-world value is easy to explain. You can also learn experiment design, signal processing, and basic machine learning evaluation.

Research Questions

• How does vibration pattern affect obstacle-avoidance accuracy in a haptic glove task?
• What is the effect of fingertip motor placement on reaction time during navigation?
• Does a distance-to-vibration mapping based on a CNN improve avoidance performance compared with a simple threshold mapping?
• To what extent does training time change user accuracy with the glove?
• Which obstacle types produce the highest missed-detection rate with haptic feedback?
• How does scene clutter affect the reliability of depth-camera guidance?

Basic Materials

• Smartphone with ARKit or ARCore depth sensing support.
• Low-cost coin vibration motors.
• Microcontroller or motor driver board.
• Lightweight glove or finger sleeve.
• Rechargeable battery pack.
• Obstacles for a simple navigation course, such as cones, boxes, or chairs.
• Measuring tape.
• Stopwatch or phone timer.
• Printed data sheets or spreadsheet for logging trials.
• Tape or Velcro for mounting motors.

Advanced Materials

• Smartphone with LiDAR or high-quality depth sensing.
• Prototype glove with multiple coin motors on fingertips and palm.
• Microcontroller with PWM motor control.
• Motor driver board with current protection.
• Arduino-compatible logging setup or Bluetooth data link.
• Motion capture or overhead video setup.
• Force or contact sensors for collision scoring.
• Computer for CNN training and evaluation.
• 3D-printed motor mounts or glove fixtures.
• Calibration targets for depth and distance testing.

Software & Tools

• Google Colab: Trains and tests a CNN without needing a powerful laptop.
• Python: Cleans sensor data, runs statistics, and plots performance.
• OpenCV: Processes camera frames and helps inspect depth-related images.
• ImageJ: Measures visual markers and helps compare obstacle positions in recorded video.
• Google Sheets: Organizes trial data and calculates simple summary statistics.

Experiment Steps

1. Define the navigation task you want to measure, such as avoiding a hallway obstacle or picking the safer path in a choice test.
2. Choose one vibration mapping rule first, then decide how you will turn distance into motor strength or pulse pattern.
3. Build a fair comparison condition, such as no glove, simple threshold feedback, or your CNN-based mapping.
4. Plan how you will score success, including collisions, near misses, reaction time, and path efficiency.
5. Set up a within-subject design so the same participant uses each condition and individual differences do not dominate the results.
6. Decide how you will analyze the data, including a statistical test and a way to compare learning across repeated trials.

Common Pitfalls

• Mounting motors too loosely, which makes vibration feel different from trial to trial.
• Using changing room lighting or reflections, which can confuse depth sensing and distort distance estimates.
• Testing only one obstacle layout, which makes it hard to tell whether the glove works beyond a single setup.
• Skipping a baseline condition, which leaves you with no way to prove the glove improves performance.
• Ignoring user fatigue from repeated buzzing, which can change reaction time and hide the real effect.

What Makes This Competitive

A stronger project goes beyond building a glove that buzzes. You can compare multiple mapping strategies, test whether the CNN really beats a simpler rule, and measure both speed and safety. A competitive version also uses a careful within-subject design, clean controls, and statistics that match the question. If you add error analysis, you can explain when the system fails and why.

Project Variations

• Test the glove with blindfolded sighted participants in a simple obstacle course to measure navigation speed and collision rate.
• Compare fingertip-only vibration with fingertip-plus-palm vibration to see whether extra spatial cues improve decisions.
• Replace the CNN mapping with a rule-based depth threshold and compare accuracy, latency, and user preference.

Learn More

• PubMed: Search for review articles on haptic feedback, assistive wearables, and navigation for blind users.
• NIH RePORTER: Look for funded projects on assistive technology and sensory substitution to find current research directions.
• NASA Open Science and data resources: Explore depth sensing, computer vision, and human factors methods through free educational material.
• MIT OpenCourseWare: Find free lectures on machine learning, computer vision, and human-computer interaction.
• IEEE Xplore abstracts: Read abstracts on haptic guidance and wearable assistive devices, then use your school library for full texts if available.