The first steps of my Master’s thesis research was featured today in Wired’s Danger Room in an article titled New Use for iPhones: Controlling Drones. It’s a fun read with some good quotes from my thesis advisor, but it largely ignores the research implications. What the video above shows is us controlling an autonomous quad rotor helicopter by sending incremental “move to” commands. We make the user perceive this as a first-order (or velocity-based) control system by having them tilt the iPhone in the direction they want the quad rotor to move. More extreme tilts simply send a “move to” command which is further away from the helicopter’s current position.
Most people think they want more control over flight systems than they’re actually prepared for. If I told you that moving a helicopter around was as simple as saying “go 1/3rd of a meter to the right,” you’d likely be concerned that you don’t have enough control over the system to make it do what you want. The seemingly reasonable desire “just let me tell it go right at the speed I want until I want it to stop” is actually very complicated under the hood, and involves a lot of things humans are poor at (feedback loops, integrators, etc).
However, that’s what you think you want, so I designed and built this interface to satisfy both needs- a simple control system that still makes the operator think they are fully in control of the helicopter.