Inspiration
It's hard to collect data through cameras to train ML models for trajectory generation because data & privacy regulations don't allow filming and observing behavior of persons and/or cars in cities. With our solution, we set the foundation for collecting data and making prediction without observing having to observe everything in a city.
What it does
It creates a digital twin which simulates cars and cameras. You can place cameras in our interactive web application and obtain both the data of every car's route and data of what the cameras are observing to then train ML models and make predictions about the traffic in a region.
How we built it
We projected cars onto a map which follow traffic rules and simple aspects like different acceleration and speeds on different types of roads. The routes and the camera information get saved in a CSV which is downloadable and can be used for all kinds of tasks.
Challenges we ran into
Projecting a large number of cars onto the map and making sure that the cars behave according to the traffic law.
Accomplishments that we're proud of
Building a complete web application which can already be used for operations like feeding data to an ML model.
What we learned
How to build a web application which handles large amounts of data and design it with interactive aspects.
What's next for brocodely
Explore further use cases by discussing the idea with people in interdisciplinary fields and possibilites made possible by the new concept of collecting data for traffic analysis.
Log in or sign up for Devpost to join the conversation.