As lane-keep assist and adaptive cruise control features are becoming more available in commercial vehicles, we wanted to explore the potential of a dedicated collision avoidance system

What it does

We've created an adaptive, small-scale collision avoidance system that leverages Apple's AR technology to detect an oncoming vehicle in the system's field of view and respond appropriately, by braking, slowing down, and/or turning

How we built it

Using Swift and ARKit, we built an image-detecting app which was uploaded to an iOS device. The app was used to recognize a principal other vehicle (POV), get its position and velocity, and send data (corresponding to a certain driving mode) to an HTTP endpoint on Autocode. This data was then parsed and sent to an Arduino control board for actuating the motors of the automated vehicle

Challenges we ran into

One of the main challenges was transferring data from an iOS app/device to Arduino. We were able to solve this by hosting a web server on Autocode and transferring data via HTTP requests. Although this allowed us to fetch the data and transmit it via Bluetooth to the Arduino, latency was still an issue and led us to adjust the danger zones in the automated vehicle's field of view accordingly

Accomplishments that we're proud of

Our team was all-around unfamiliar with Swift and iOS development. Learning the Swift syntax and how to use ARKit's image detection feature in a day was definitely a proud moment. We used a variety of technologies in the project and finding a way to interface with all of them and have real-time data transfer between the mobile app and the car was another highlight!

What we learned

We learned about Swift and more generally about what goes into developing an iOS app. Working with ARKit has inspired us to build more AR apps in the future

What's next for Anti-Bumper Car - A Collision Avoidance System

Specifically for this project, solving an issue related to file IO and reducing latency would be the next step in providing a more reliable collision avoiding system. Hopefully one day this project can be expanded to a real-life system and help drivers stay safe on the road

Built With

Share this project: