by Ben Fisher
Posted on November 06, 2015
The commercial drone industry has seen huge growth this year and it seems one of the next chapters is also starting to unravel. MIT students have been working on a autonomous drone avoidance systems for the past few years with impressive results.
In the video below you can see how the autonomous drone flying at 30 MPH is capable of taking avoidance manoeuvres through the landscape.
Andrew Barry, a PhD student at MIT’s Computer Science and Artificial Intelligence Lab (CSAIL), has developed the technology as part of his thesis. Together with Professor Russ Tedrake, they have developed software which runs 20 seconds faster than existing detection obstacle detection software. Operating at 120 frames per second, the open-source software allows the drone to detect objects and map its environment in real time, extracting depth information at 8.3 milliseconds per frame.
One of the secrets to their success is how they have approached the challenge. You can gather some insight from their comments below:
“Sensors like lidar are too heavy to put on small aircraft, and creating maps of the environment in advance isn’t practical,” Barry says. “If we want drones that can fly quickly and navigate in the real world, we need better, faster algorithms.”
“Traditional algorithms focused on this problem would use the images captured by each camera, and search through the depth-field at multiple distances – 1 meter, 2 meters, 3 meters, and so on – to determine if an object is in the drone’s path.
“Such approaches, however, are computationally intensive, meaning that the drone cannot fly any faster than 5 or 6 miles per hour without specialized processing hardware.
“Barry’s realization was that, at the fast speeds that his drone could travel, the world simply does not change much between frames. Because of that, he could get away with computing just a small subset of measurements – specifically, distances of 10 meters away.
“You don’t have to know about anything that’s closer or further than that,” Barry says. “As you fly, you push that 10-meter horizon forward, and, as long as your first 10 meters are clear, you can build a full map of the world around you.”
This approach has also enabled the development of the hardware parts to total a mere $1,700 with a full list available here.
Part of the reason drones have become so successful in the past few years has been due to their increased capability in control. Flying remotely pilot aircraft is a highly skilled operation and the more capable drones are becoming in handling flight, the more we will find uses for them.
One of the next barriers for the industry is integrating them further into airspace. The point at which technology meets regulation is always quite contentious as motivations and arguments collide. However, the UK CAA has clear guidelines for aircraft capabilities to have sense and avoid capabilities. So there is a rough roadmap in place to help the technology reach those hurdles until they are accepted as a safe vehicle which can operate more freely in airspace.