Mark Mueller’s fail-safe algorithm (via Mark Mueller)
Drones have a habit of dropping out of the sky no matter what size they are, including military UAVs such as the Predator and Reaper along with smaller civilian drones such as the hexacopter that crashed into a spectators face (Great Bull Run in Virginia). Like with all things mechanical, they fail sometimes, however when flying things fail it tends to be catastrophic, especially if they are airborne. Anyone who has ever seen a quadrocopter knows they are capable of performing incredible feats of acrobatic agility, like being able to flip on a dime, bank at incredible speeds and barrel roll two feet off the ground.
They too can suffer failure, especially when one of the four propellers malfunctions, which plummets the drone into the ground, which seldom leads to becoming airborne again afterwards. The only way a multi-propeller drone could survive the loss of a prop was through redundancy, meaning the drone needed to have more than four propellers (hexa and octocopters) to remain in flight. To help keep the quadrocopters aloft long enough to land safely, ETH’s Mark Mueller has developed an ingenious software algorithm that gives those drones a better chance of surviving propeller failure.
Known as the ‘automatic failsafe algorithm’, the software allows the drone to deal with the cost of losing a single propeller. Once the propeller malfunction is detected, the software kick’s in and drone recovers its flight and returns to its previous stable position. After which, the qudrocopter performs a controlled, soft landing at the operator’s command. Mark states that the system works on hardware that is readily available on standard quadrocopters and therefore could be used as a simple software upgrade to its existing control system. While the algorithm isn’t available yet for drone users, Mark is currently applying for a patent for his technology, so it’s likely we will see it on the market in the near future.
See more news at: