Hero Page

Inspiration

Flight planning can be a tedious and time-consuming task for RC pilots. What began as a simple AI bot to create flight plans for RC enthusiasts has evolved into something with greater potential. By expanding into markets such as search and rescue and building inspections, Route Runner is designed to streamline mission planning, helping people respond to, and escape emergencies more efficiently and safely.

What it does

Route Runner creates a flight plan for users, and uses our custom made quadcopter to fly and collect geographical information using pyvision. The quadcopter(using its RPI) uploads this information back to the website, where the website generates a live feed. Once the drone comes back to its original position, it then takes the video it just made and turns it into a gaussian splat(Video->3d model) using Agisoft Metashape. Then using this model, the website takes it and puts it into web ar, where an augmented reality experience is set up.

How we built it

We build a custom quad-copter with a F450 frame. For the flight controller and ESC, we used the SpeedyBee F405V3 50A stack and Flash Hobby D4215 650KV brushless motors. The construction of the quad-copter took about 2 hours and configuring it for the different missions ran throughout the night. We used a range of mission planning software starting with Betaflight, moving to AdruPilot and finally settling with iNAV.

The Fullstack webpage was built using NextJs, Spline, and GPT. Our website takes a user input, and prompts this to GPT to where GPT uses Dijkstra's algorithm to create an optimal flight path that scans an area based on user input. The drone then follows this flight path with a live feed that is being fed from the rPi on the drone. The rPi on the drone also supplies PiVision to the drone for obstacle avoidance and querying data points. Once the drone is finished with the flight path, it then takes the live feed and turns it into a gaussian splat(Video->3d model) using Agisoft Metashape. Then using this model, the website takes it and puts it into web ar, where an augmented reality experience is set up.

Challenges we ran into

The biggest challenge that we ran into was setting up the drone. We wanted a cheap way to set up autonomous flights without breaking the bank but we quickly realized that the flight controller was not supported by the popular mission planning and autonomous flight software - AdruPilot. Another challenge we ran into was the weather. We could not do proper flight testing, limiting our trail runs and opportunities to improve the quad-copter.

Accomplishments that we're proud of

We are proud of the fact that we made a drone and had it flying in under 4 hours. We were also proud of how fast we were able to get AI + prompts working, as this took us over 3-4 hours in previous years.