If delivery drones become widespread, then creating aircraft control systems to stop them from crashing into one another is going to be a big challenge. (That’s why Google, NASA, and the FAA are all working on it right now.) But hey, if we’re going to enable fleets of semi-autonomous vehicles to roam the skies in incredibly complicated patterns, why stop there? Why not let them drive as well?

That’s what researchers at MIT have been looking into, creating the first “path planning” system capable of controlling transformer bots that fly and drive. Now the hardware here isn’t that significant (the drones you can seen in the video below are just toys, rather than full-sized craft), but what is interesting is the path-planning system that sets the route of each drone. This tells the drone when to fly and when to drive in order to be as energy efficient as possible.

As the researchers explain in a paper describing their work, driving and flying bots have unique advantages: “Driving robots, while stable and energy efficient, are limited to mostly flat terrain. Quadcopters, on the other hand, are agile and highly mobile but have low energy efficiency and limited battery life.” Combining the two creates a vehicle with both an “energy efficient driving mode and an agile flight mode.” And this could be useful for a number of tasks, including surveillance, disaster relief, and, yes, delivering packages.

The researchers’ software is theoretically capable of controlling up to 80 drones at a time, but they tested it using just eight vehicles navigating a simplified model town. The drones were told where the roads were and where they were allowed to fly, and then automatically plotted their own routes between predetermined start and end-points, all while avoiding collisions — both midair and on the ground. At any rate, it’s something for Amazon to think about when it’s building its delivery drone beehives.