Ultrasonic Key Exchange for Nearby Devices

ISEF Category: Embedded Systems

Subcategory: Networking and Data Communications  ·  Difficulty: Advanced  ·  Setup: University Lab  ·  Time: Full Year

The Hook

Your phone speaker can do more than play music. With the right signal, it can send data that people cannot hear. That makes proximity-based key exchange possible, even in a noisy classroom. The catch is making the link survive real-world noise and device differences.

What Is It?

This project studies a hidden communication channel that uses ultrasound, sound above the range most people can hear. Two devices, like phones or laptops, can send and receive these signals to exchange a small secret, such as a key used to prove they are nearby. Think of it like passing a note in a crowded room, except the note rides on sound that most humans ignore.

OFDM, or orthogonal frequency-division multiplexing, splits data across many small frequency slices instead of one big one. That helps the signal survive echo, noise, and distortion. A learned equalizer is a model that tries to undo the damage caused by the room, the microphones, and the speakers. In plain terms, it helps the receiver guess what the original signal should have been.

A secure proximity-based key-exchange primitive is the basic handshake that lets two devices prove they are close to each other before sharing a secret. Your job is not to build a full commercial system. Your job is to test whether the channel works well enough to support that handshake under different classroom-like conditions.

Why This Is a Good Topic

This is a strong science fair topic because you can test a real engineering problem with clear metrics. You can measure packet success rate, bit error rate, range, noise tolerance, and how much a learned equalizer helps. The topic connects to secure pairing, device authentication, and privacy, which gives it real-world value. You can also make it your own by comparing signal designs, room conditions, or receiver models.

Research Questions

• How does classroom background noise affect ultrasonic key-exchange success rate?
• What is the effect of using OFDM instead of a single-tone signal on decoding accuracy?
• Does a learned equalizer improve bit error rate across different phone and laptop models?
• To what extent does distance between devices change ultrasonic handshake reliability?
• Which frequency bands give the best tradeoff between inaudibility and decoding success?
• How does room echo change performance when the same waveform is sent from different locations?

Basic Materials

• Two phones or laptops with microphone and speaker access.
• A quiet room and a noisier room for comparison.
• Headphones or small speakers for controlled playback tests.
• A way to generate and record test signals, such as a laptop audio interface or app that can play and capture audio.
• A sound level meter app or decibel meter for relative noise measurements.
• Notebook or spreadsheet for logging trial conditions.
• Tripod, box, or stand to hold devices at fixed positions.
• Basic ruler or measuring tape for distance control.

Advanced Materials

• Two or more devices with programmable audio input and output.
• USB audio interface with a known sampling rate.
• Measurement microphone with flat high-frequency response.
• Signal generator and spectrum analysis software.
• Acoustic absorber panels or foam for room condition tests.
• Access to a DSP or machine learning workflow for equalizer training.
• Calibration speaker or reference microphone.
• Network analyzer or synchronized timing setup for repeatable link testing.

Software & Tools

• Python: Processes audio traces, computes bit error rate, and plots performance across trials.
• SciPy: Filters signals, finds spectra, and helps compare OFDM parameter choices.
• NumPy: Stores and manipulates recorded waveforms for analysis.
• Matplotlib: Makes clear graphs of accuracy, noise, distance, and frequency response.
• ImageJ: Not needed for audio, but useful if you document setup photos and annotate experimental diagrams.

Experiment Steps

1. Define the security goal, the channel limits, and the performance metric you will measure.
2. Choose one baseline waveform and one OFDM-based waveform so you can compare them fairly.
3. Plan a receiver pipeline that includes synchronization, decoding, and an equalization stage.
4. Design controls that separate room noise, device hardware differences, and distance effects.
5. Build a small test matrix that changes one variable at a time, then record repeated trials.
6. Decide how you will judge success, such as key agreement rate, bit error rate, or false acceptance rate.

Common Pitfalls

• Testing only in one quiet room, which hides how badly the link fails in real classrooms.
• Ignoring device-to-device speaker and microphone differences, which makes one phone look much better than another.
• Skipping calibration, which makes your frequency response results hard to compare across trials.
• Changing distance, noise, and angle at the same time, which prevents you from telling what caused the error.
• Training an equalizer on the same recordings you later score, which inflates performance and weakens your conclusion.

What Makes This Competitive

A competitive project would compare several signal designs and show why one works better under real acoustic stress. You would get stronger results if you measured more than success or failure and included bit error rate, false matches, and confidence across different devices. A deeper entry would also test whether the learned equalizer generalizes to new rooms and new hardware. That kind of analysis shows real systems thinking, not just a demo.

Project Variations

• Test the same ultrasonic handshake across smartphones, laptops, and tablets to see how hardware changes decoding reliability.
• Compare a learned equalizer with a classic digital filter to see which one handles room echoes better.
• Measure whether ultrasound works better in quiet halls, busy classrooms, or small offices for nearby-device authentication.

Learn More

• MIT OpenCourseWare: Search for digital signal processing courses to review OFDM, filtering, and channel equalization basics.
• NIH PubMed: Search review articles on acoustic communication, ultrasound, and device authentication to find prior methods.
• NASA Technical Reports Server: Search for signal processing and communication system reports that explain modulation and noise handling.
• NOAA Acoustics resources: Look for public material on sound propagation, echoes, and measurement basics.
• IEEE Xplore and journal abstracts: Search for papers on ultrasonic communication, OFDM, and learned equalizers, then read the abstracts and methods sections.