Drones are known to be noisy. The buzzing sound of a drone sounds like an angry oversized bee, which can be horribly disturbing for human ears. As per NASA’s report on “Initial Investigation into the Psychoacoustic Properties of Small Unmanned Aerial System Noise”, it was observed that most people find noises from a drone to be more annoying than noises from other ground vehicles like a car or a truck.

Now, Amazon has come up with a drone technology that will use machine learning to reduce the noise produced by the drones. They have filed a patent application to protect this technology.

Amazon is developing drone technology for delivering products to buyers quickly and efficiently. In our past posts, we covered drone-ports and universal flying machine from Amazon. The new less noisy drones will allow Amazon to serve their customers better.

 PATENTED TECHNOLOGY

According to the patent application, Amazon’s drones are equipped with sensors and processing devices that are designed to sense noise signals from a region surrounding the drone and produces an anti-noise signal to reduce the net effect of such noises. The anti-noise signals are actually signals that have the same magnitude and frequency as that of the noise signals sensed from the region surrounding the drone, but are completely out-of-phase. Two signals that are of the same magnitude with the same frequency but out-of-phase result in a destructive interference. Accordingly, the sum of the two signals will be minimum or may tend to zero. Therefore, the net noise from the drone is significantly reduced.

Further, the drone, while flying from one location to another, captures intrinsic data and extrinsic data with the help of multiple sensors that are embedded within the drone. The intrinsic data provides information about acoustic energies (e.g., sound pressure levels or intensities, or frequencies) captured by the drone at different operating conditions (e.g., altitudes, courses, speeds, rates of climb or descent, turn rates, or accelerations, or dimensions of structures or frames, numbers of propellers or motors, operating speeds of such motors) or tracked positions (e.g., latitudes and/or longitudes) etc. Similarly, the extrinsic data corresponds to the acoustic energies captured by the drone at different environmental conditions (e.g., temperatures, pressures, humidity, wind speeds, and directions), times of day or days of a week, month or year when the drone is operating etc.

Further, the intrinsic data and the extrinsic data captured from the drone during the flight is then used to train a machine learning system to find a pattern between a drone’s operation and location with the acoustic energies. Further, the trained machine learning system is then used to predict noises that can be expected when the drone operates in a predetermined location or operates at predetermined conditions such as velocity or position. Further, once the noise is predicted by the drone, the anti-noise signal can be generated by the drone to cancel out the net effect of such noises.

WHAT IS YOUR TAKE?

So what do you think about this technology from Amazon? Let us know in the comment section.

US20180286372