Small Drone Autonomously Dodges Obstacles At 30 MPH
There are phones, and there are smart phones.
In the drone world, we’re currently in the “dumb” drone phase of technology, since drones have a habit of crashing into things. In fact one can find a plethora of drone fail videos out there.
But if we want light-weight drones that can navigate the real world and fly quickly, then faster and better algorithms are needed. The problem is that current sensor technology – like laser-shooting “lidar” sensors – are too heavy for small UAV aircraft. Also, many of the current obstacle detection systems rely on an external motion-capture apparatus rather than and on-board system.
Adam Bary, a MIT student has developed a solution to this problem. He wrote an algorithm that uses stereo vision camera sensors to detect objects up to 33 feet away, and can support 30 MPH flight speeds.
It’s simple enough to be replicated by experienced DIY drone builders, and can be build inexpensively with off-the shelf components. In fact, his algorithm is open-source.
The test flight below shows how the UAV detects objects in red, and then maneuvers to evade a dangerous crash.
Watch the test flight, and let us know what you think!
Barry hopes to continue working on the algorithm to make the drone more robust in computing and versatile in various environments.
“As hardware advances allow for more complex computation, we will be able to search at multiple depths and therefore check and correct our estimates,” Barry says. “This lets us make our algorithms more aggressive, even in environments with larger numbers of obstacles.”
Other Top Posts:
- DRONE CAUGHT THIS MAN ON GIANT 200 FT. WIND TURBINE
- THIS INTERCEPTOR DRONE COULD PARALYZE OTHER NAUGHTY DRONES
- THIS DRONE EXPOSED THE MOST DISGUSTING BUSINESS OPERATION YOU’VE EVER SEEN
- DRONE FLYOVER OF TESLA’S NEW GIGAFACTORY
- THESE DRONES ACTUALLY BUILT A WORKING ROPE BRIDGE [WATCH]
To read more about Adam Barry’s research, you can read his 6-page research paper.