-
-
Example of Pikipark usage
Inspiration
Finding parking in busy areas is often a challenge, and full parking lots commonly cause a rerouting ordeal. If there was a way for drivers to avoid these occurrences, it would make for a much easier trip.
What it does
Our tool analyzes a live feed from parking lots to show users the location of available parking spaces near their destination. Users are able to see the exact number of available spots in each parking lot, relieving stress and giving parking lots more optimal business.
How we built it
We used the TRAE IDE to streamline the development process. Each team member had an area of focus as well:
- Ethan Enriquez: planning and API integration
- Sam Miller: comprehensive frontend and UI
- Natnael Tegegne: backend and systems Integration
- James Widmer: machine learning model to identify individual parking spots
Challenges we ran into
- Difficulty with setting up Overshoot to work at our desired accuracy
- Compute bottlenecks (AI Model training)
- Integration
- Environment & Dependency Conflicts
- Limited access to suitable CCTV footage of parking lots
Accomplishments that we're proud of
- Implementing a flexible machine learning algorithm
- Designing a comprehensive flow of information between APIs and other code
- Pleasing user interface that is easy to read and use
- Systems Integration and Infrastructure Management -AI Logic Optimization
What we learned
- API integration
- Machine learning training and sample data
What's next for Pikipark
- Continue training the AI model for more accurate results
- Create a navigable and easy to use landing website
- Display a graphic visualization of the parking lot and available spots to the user
- Function to open the selected parking location in another app with more advanced routing (Google Maps, Apple Maps, etc.)
- Option to reroute to the nearest available parking space if the current destination is full or close to full
Log in or sign up for Devpost to join the conversation.