As computers become smarter, and cars faster and much more self-reliant, it is only natural that drones should follow this path. Recently, a team of computer scientists from MIT developed a prototype of a flying drone which is actually capable of maneuvering throughout the environment without preinstalled maps. The self-flying drone ran through an obstacle course without hitting anything.
From the beginning of the computer era, man dreamt that one day he will be able to develop computers that will be capable of thinking by themselves and performing the same tasks as him. That day is much closer than we think, thanks to a team of scientists from the Massachusetts Institute of Technology.
The team of computer scientists who was working for the Computer Science and AI Lab was successful in developing the first prototype of a self-flying drone. Benoit Landry, one of the scientists working on the project told the reporters that the difficulty of constructing such a flying apparatus resides in his capacity to detect obstacles and take actions against them, all without any human interference.
There have been many attempts in the past to devise such a drone, but all of them required a pre-installed map of the environment in which he will fly. According to the team, this kind of procedure is lengthy and resource consuming. Of course, even the drone itself would have to be outfitted with a storage device big enough to retain the map. Thus, the drone, which was supposed to be light and agile, becomes clunky and slow.
But thanks to the team’s latest research, it seems that they are on the verge on solving this problem. Using advanced computer algorithms, the drone can potentially maneuver through basically any type of environment. So says the paper. But let’s take a closer look at the project itself.
In their attempt at devising a self-flying drone, the scientists addressed the issue of space. Instead of installing resource consuming maps, the scientists devised an algorithm that is capable of observing the amount of free space in front of the drone. Basically, the drone sees through its camera that there is no obstacle within a given distance and then moves on scanning the next patch.
A team of scientists from MIT has been successful in developing a self-flying drone in the AI lab. The self-flying drone ran through an obstacle course without hitting anything. The makeshift obstacles course consisted of PVC tubes, strings, and boxes, all of them being arranged in a specifical manner to resemble a forest.
The drone packed with 4 motors was successful in navigating through the 10-foot square forest without hitting anything, at a top speed of 1 meter per second.