Meet Drone Shield, an Ambitious Idea For a $70 Drone Detection System

An anonymous reader writes "Here's an Interesting idea of how to use a Raspberry Pi and a few other inexpensive items to make a low cost detection system. From the article: 'The Drone Shield would combine a Raspberry Pi, a signal processor, a microphone, and analysis software to scan for specific audio signatures and compare them against what known drones sound like. (Because obviously a Predator drone is going to sound very different than a small quadcopter.) Once a match is found, the Drone Shield then sends an e-mail or SMS to its owner...'"

Comments

Re:Interesting...

[Anonymous Coward]

Ofcourse it's been done. There's a lot of military equipment that works like this.

Re:Range

[DerekLyons]

Disclaimer, I was a submarine crewman and though I was a qualified sonar watchstander I was not a sonar tech. I have also studied fairly widely in the unclassified literature.

So now the highly directional microphone has to be pointed toward the undetected drone in order to detect it? That makes perfect sense.

If you're sophisticated and have multiple microphones in a well planned array, then you can aim your microphone in software and sweep the sky looking for the signature. Look up acoustic beam-forming. If the array is large enough you can estimate distance as well as angle. The bonus is that you get actual tracking instead of just detection.

It's certainly simple - in theory. In reality, picking out such a faint noise from the background is Very Difficult Indeed.
 

The problem would be processing power though. Simple implementations could range from 4 microphones that you sum/subtract to look at quandrants, up the way to something approaching what the US Navy does with its towed arrays. I doubt the PI could handle the processing of the signals in both the time domain to get tracking, and the frequency domain to do target qualification.

Processing power, both for signal analysis (finding the faint signal) and for beamforming is on the beginning of your problems. Let's just hit the high spots:

• The accuracy of your track is only as good as the accuracy of your microphone positioning. (You won't need surveyor grade accuracy, but you will probably need better than the three meter accuracy that WAAS/GPS provides.) You can't beamform if you don't know the relative locations of your microphones. Oh, and did I mention that sound is refracted as the temperature of the air changes? You'll have to account for that too - assuming you can get accurate enough data on current conditions.
• You'll need some fairly clever filtering and processing to avoid the microphones being swamped by unrelated and louder background noise.
• You also need high quality low noise amplifiers to bring the sound of Predator up to useable levels. (The highest quality commercial audiophile amplifier isn't even close. You need a supercomputer and audiophile gear by comparison isn't even as good as the throwaway calculators you get with your breakfast cereal.)
• Speaking of the sound... different frequencies get attenuated and refracted differently. You'll have to account for that too.

Etc... etc... The very definition of a non trivial project. You're essentially trying to replicate what the USN does with it's passive sonar systems, with dull and chipped stone knives. (You don't even have a bearskin. You don't even get a bearskin, just the aforementioned knives.)