Flying drones has become quite a popular pastime over the last few years, with companies like DJI coming up with new drone designs and innovations year after year.
However, one aspect to flying drones that has seen little improvement is the way in which we pilot these flying robots. More often than not you simply tap a screen or wiggle around a little joystick to get the drone to go where you want it to go. That is all about to change thanks to work done by researchers from MIT CSAIL.
Researchers have developed a method of flying drones that uses a combination of muscle sensors, motion detection, and advanced algorithms to give the user more ‘natural’ control over the drone during flight.
This system, which researchers have dubbed ‘Control-A-Bot’ makes use of an algorithm that detects gestures using your body’s movement as well as activity in your triceps, biceps, and forearms. This means that a simple wave of your hand or clenched fist can control the drone during flight. Best of all the system doesn’t need any environmental cues, personal training, or offline calibration, meaning you could enjoy all its features straight out the box.
Control-A-Bot isn’t quite ready for the real world just yet, however, a prototype has been made that responds to 82% of over 1,500 gestures. A promising figure, but not exactly what you’d want in a real world situation.
If the system ever makes it to store shelves, it could make the hobby of flying drones more accessible to people who would normally be intimidated by conventional control setup.
Take a look at Control-A-Bot in action below.