Inspiration
Imagine walking down a hallway you know by heart, only to find a box suddenly in your path. For a visually impaired person, once familiar spaces can change unexpectedly.
We wanted to build something that does more than warn you about obstacles. What if your glasses could explain what's ahead, remember what you've encountered, and notice when something changes?
That's how Ankh started: smart glasses designed to make navigating the world feel more intuitive and independent.
What it does
Ankh combines real-time obstacle detection, AI-powered vision, voice guidance, and environmental memory in a pair of smart glasses.
- Obstacle detection: A forward-facing camera and proximity sensor mounted on the glasses help detect nearby obstacles.
- Vibration and voice alerts: Receive feedback about hazards and information about your surroundings.
- Route Memory: Compare current observations with previous walks to help identify changes, like a box appearing in a previously clear hallway.
- Object lookup: Ask questions like "Where did I last see my keys?" using previously recorded observations.
- Companion app: Our Base44 web app brings together route history, hazards, saved routes, and previously observed objects.
The big idea is simple: don't just detect the world. Remember it.
How we built it
We combined hardware, AI, voice technology, data storage, and web development to bring Ankh to life.
- Arduino and proximity sensor: Handle immediate distance sensing on our glasses prototype.
- Google Gemini API: Analyzes camera images to help identify objects, understand surroundings, and support Route Memory.
- ElevenLabs: Powers voice interaction and spoken guidance.
- Tiger Data: Supports storing timestamped observations, route checkpoints, and hazard history.
- Base44: Powers our companion web app, where users can review routes, check hazards, and look up previously observed objects.
Our goal was to let the hardware focus on immediate awareness while AI adds context. We also wanted to avoid overwhelming users with constant alerts.
Challenges we ran into
Getting hardware, AI, voice, data storage, and a web app to work together was a challenge in itself. We troubleshot Arduino connectivity and sensor integration while thinking about the trade-off between fast feedback and more detailed AI analysis.
Route Memory introduced another challenge. Detecting an object is one thing, but figuring out whether it's new is another. We had to think beyond individual images and consider how observations would be stored and compared over time.
Accomplishments that we're proud of
We're proud to bring together a physical wearable prototype and a companion application around a real accessibility challenge.
Route Memory is what makes Ankh especially exciting to us. Instead of treating every walk as a fresh start, we're exploring how technology can remember previous surroundings and help users recognize what's changed.
We're also proud that the project started with a problem that matters to people we know personally. We wanted to build something that supports independence rather than adding another barrier.
What we learned
Building Ankh taught us that creating a useful AI product isn't about adding as much AI as possible. It's about choosing the right tool for the right moment.
Sensors provide immediate distance information, Gemini adds visual context, voice makes feedback accessible, and persistent data supports environmental memory. Bringing these pieces together also taught us how important integration and thoughtful design are when building across hardware and software.
What's next for Ankh
We want to make Ankh more responsive, strengthen Route Memory's ability to recognize environmental changes, and make the glasses smaller, more comfortable, and more affordable.
We also hope to gather feedback directly from visually impaired users and accessibility experts to gain a better understanding of what could make Ankh more useful and seamless for everyday use.
Ultimately, Ankh is about more than knowing what's in front of you. It's about remembering what changed, so you can move through the world with greater confidence and independence.
Built With
- arduino
- base44
- c++
- css
- elevenlabs
- gemini
- javascript
- postgresql
- python
- sql
- tiger-data
- typescript
Log in or sign up for Devpost to join the conversation.