MIT researchers develop algorithms to improve delivery drones
In the near future, you won’t have to wait days for your online product to get delivered. Online retailer Amazon announced plans to deliver its products by drones back in December. Named the Octocopter, the flying robots will drop purchase packages to customers within 30 minutes of an order. However, there are many factors such as high winds, fuel, mechanical issues and sensor measurement errors that the drones would need to deal in order to make safe, timely and accurate delivery of the products possible. Stepping up for the technical challenge, MIT researchers have come up with a two-pronged method that considerably reduces the computation associated with lengthy delivery missions.
The team first developed an algorithm that allows the drone to monitor its own health in real-time. The algorithm can measure and predict the drone’s fuel levels and condition of propellers, cameras and other sensors which enable it to fly and navigate accurately. It can also instruct the drone to take any essential proactive measures such as changing its programmed route to a charging station, if needed.
Tests conducted by researchers revealed that with the new algorithm, the drones are able to deliver multiple orders under various environmental conditions with far fewer failures or breakdowns. Also, the navigational capabilities of the drones are improved too. Therefore, a drone can efficiently compute its possible future locations offline, before it takes off. Thus, the drone can then simplify all its potential routes to reach a destination without colliding with any obstacles.