How a Half-Inch Beetle Finds Fires 80 Miles Away (scientificamerican.com) 62
How does a half-inch beetle find fires 80 miles away? Why, it's called "stochastic resonance" of course! Slashdot reader bodog shares an excerpt from a report via Scientific American: The fire chaser beetle, as its name implies, spends its life trying to find a forest fire (because freshly burnt trees are fire chaser beetle baby food). [T]hey can sense fires from distances over which car stereos are hard pressed to pick up FM radio. In fact, because the infrared emission of a burning oil tank of known volume (in this case, 750,000 barrels) can be calculated with reasonable certainty, scientists that studied the Coalinga oil tank explosion have inferred the beetles can detect infrared radiation intensities so low that they are buried in the thermal noise around them. But ... how?
The heat eyes on the sides of fire chaser beetles are filled with about 70 infrared sensilla. Inside each sensillum is a hair-like sensor (called a dendritic tip in the diagram above) that physically deforms when the sensillum expands in response to heat, triggering a neural response. [...] A signal picked up by more than one of them can be summed up and amplified by the neurons that wire the [70-90] array. As a result, the heat eye can detect softer signals than a single sensor could.
Finally, it is also possible the beetles are better able to detect a signal buried in noise due to a spooky (to me) phenomenon called "stochastic resonance." In this scenario, added thermal noise counterintuitively helps a sensor pick up a signal. A signal below the threshold for triggering a sensor -- but still close to it -- will resonate by chance with a portion of thermal noise that is the same frequency. When there is more noise, there is more signal at that resonant frequency. Together, noise plus signal adds up to an impulse sufficient enough to trip the sensor when signal alone or signal with less noise would not. Incredibly, the measurement gets more precise in the presence of noise than without.
The heat eyes on the sides of fire chaser beetles are filled with about 70 infrared sensilla. Inside each sensillum is a hair-like sensor (called a dendritic tip in the diagram above) that physically deforms when the sensillum expands in response to heat, triggering a neural response. [...] A signal picked up by more than one of them can be summed up and amplified by the neurons that wire the [70-90] array. As a result, the heat eye can detect softer signals than a single sensor could.
Finally, it is also possible the beetles are better able to detect a signal buried in noise due to a spooky (to me) phenomenon called "stochastic resonance." In this scenario, added thermal noise counterintuitively helps a sensor pick up a signal. A signal below the threshold for triggering a sensor -- but still close to it -- will resonate by chance with a portion of thermal noise that is the same frequency. When there is more noise, there is more signal at that resonant frequency. Together, noise plus signal adds up to an impulse sufficient enough to trip the sensor when signal alone or signal with less noise would not. Incredibly, the measurement gets more precise in the presence of noise than without.
As opposed to a half-inch Beatle... (Score:2, Funny)
...or 'Ringo Starr' as he's more commonly known.
Six Organs of Attainment (Score:4, Funny)
"Stochastic Resonance" is my favorite prog-rock band.
Re: (Score:2, Informative)
And here is a handy link [bandcamp.com], for anyone interested in sampling some new music.
Re: (Score:3)
Oh my goodness, I had no idea.
Re: (Score:3)
No, we all don't know how far 80 miles is. I know how far X kilometers is.
(OK, I can convert miles to kilometers and back in my head, but I'm weird do that doesn't count)
Re:"from distances over which car stereos are hard (Score:4, Funny)
Re: (Score:2)
It's about 1400 football (not soccer) fields. HTH
American or Canadian? (There's a difference.)
Re:"from distances over which car stereos are hard (Score:4, Insightful)
I'm gonna feed you the boilerplate:
This is an American tech site aimed at people who should be off work. ;)
Miles are the common casual unit of distance that best imitates what an average American on the street is likely to say. So, get used to it. This isn't an academic institution.
Frankly, as an American, I can use SI and Imperial units just fine, and even figure out how to do dimensional analysis to convert to new units.
Man, math is amazing! FFS, someone answer me this: why are the rest of you so lazy you've got to come to an American site..to bitch about your inability to convert American units.
Re: (Score:3)
The US uses US customary units. Just ask how many miles to the gallon, you'll find that the British imperial gallon is not the same as the US gallon. And since we're talking about miles, the "survey mile" used in the US in the past (and sometimes in the present for surveying) is ever so slightly larger than the internationally agreed on definition of a mile (6336/1397 meters vs 1609.34 meters)
Re: (Score:2)
Units of measure (Score:2)
No, we all don't know how far 80 miles is. I know how far X kilometers is.
If you don't understand the units that is your problem. Most of the audience reading this is American and they are perfectly comfortable with miles as a unit of measure. Yes I'd love it if my backwards country would join the 95% of the world using metric but we haven't, we aren't, and it isn't going to change any time soon at least among the general public. If the story is written by, for, about, or regarding Americans you may as well get over the fact that they are going to use US Customary Units [wikipedia.org]. Yes
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
About .5, on average for a typical sedan.
You asked...
Re: (Score:2)
Comment removed (Score:5, Insightful)
Re: (Score:2)
Pulling very weak signals out from noise (Score:2)
In addition to boosts incurred from an actual signal added to random noise, there is another potential way to extract very low signals from noise. In image processing and radio signal detection, we call it "stacking."
The thing taken advantage of is that noise is random about its baseline level, existing as both positive and negative excursions, so over time, multiple samples of noise will average out to that baseline, whereas an actual signal is not random, so it will remain at whatever level it actually is
Re: (Score:2)
Stronger?
FM transmitter: a few kW. Several thousand tons of oil on fire: 1.21 GW...
The bandwidth of the FM transmitter is much narrower, though.
Re: (Score:2)
When was the last time you thought about distance in terms of how far a car stereo would have to be
Every time I'm driving in North Seattle and a local station gets swamped by a French language station out of Vancouver BC.
squint (Score:1)
It's the same principle that unmasks pixelated images by squinting
I know this effect first hand. (Score:5, Interesting)
I have a slight tinnitus in my right ear. It's actually better at hearing in certain situations were there is background noise. Such as having a phonecall on a noisy train. And it's more sensitive to noise in general because the baseline of noise is already moved up by the Tinnitus.
Looks like this fire beetle evolved to use that effect neatly.
Re:I know this effect first hand. (Score:5, Interesting)
We use this trick all the time to measure very low concentrations of chemicals in gas phase.
A commercial IR instrument that has an array of sensors is quite powerful and expensive. This humble beetle has 10 times more "channels" than the best instrument out there. And one hell of neural network to process the signals...
Seems like Nature almost always holds the "prior art". E.g. there are two competing theories of how the nose works; one of them claims that it is kind of a Mass Spectrometer able to detect and quantify simultaneously and continuously more than 300 molecular species (radicals, ions, molecular groups, etc.) and the combined signal, after heavy postprocessing by the most sophisticated neural network in existence is what we call sense of smell. This is spectacularly efficient and complicated solution.
Of course, our dear mother could not come up with some of our unique inventions such as the wheel. Not that it matters anyway, we are still arguing if it is commercially viable or failing that what color it should be...
Re: (Score:3)
Of course, our dear mother could not come up with some of our unique inventions such as the wheel. Not that it matters anyway, we are still arguing if it is commercially viable or failing that what color it should be...
Not quite wheels, but bacterial flagella [wikipedia.org] are rotating structures that work like screw propellers. A related article [wikipedia.org] discusses the reasons why wheeled animals would be unlikely to evolve very far.
Re: (Score:1)
Of course, our dear mother could not come up with some of our unique inventions such as the wheel..
Thanks for sharing the fascinating 411. However mom did make wheels e.g.the protein rotating motor that powers the flagellum of a bacteria.
Re:I know this effect first hand. (Score:5, Interesting)
Your ears and eyes and skin all already use stochastic resonance, although the tinnitus can certainly help under the circumstances you have described (it might be by masking higher frequencies rather than revealing lower frequencies, though).
But stochastic resonance is what allows the human retina to detect single photons under laboratory conditions (although not with 100% certainty), despite overwhelming thermal noise. It is what allows you to perceive details that are finer than the resolution of photodetectors in your eyes (although that's spatial stochastic resonance rather than temporal). It's what allows you to feel surface textures that can be only microns in height.
The term stochastic resonance is a funny one because is not accurate (there is no resonance), but if you understand how pulse width modulation can be used to encode a continuous signal, then, fundamentally, you understand stochastic resonance, because they are based on the same principle -- average over quantized values to reveal a signal smaller than the quantization increment, especially when you add a little noise of known properties to set the probability of pushing the sub-quantum input above the detector's lowest threshold.
Dithering (Score:3, Informative)
In Digital Signal Processing, this is known as dithering.
Gah! (Score:2)
"Sufficient enough"? Really, Scientific American?!
Only nearly 7 years late (Score:3, Interesting)
Discover Magazine covered this story [discovermagazine.com] almost 7 years ago. I know /. lags with posting news, but this has to be a new record.
So is it illegal to fight forest fires? (Score:5, Funny)
Re: (Score:3)
That's pretty sensitive (Score:4, Funny)
But is it as sensitive as a teenager's eye in the 80s trying to watch scrambled late night soft porn on Cinemax in the hopes of detecting a nipple or a flash of pubic hair?
Probably more to it (Score:2)
The beetles might also be responding to each other. Fireflies behave that way, and by seeing and then relaying each others' light they can "coordinate" over long distances. They see each other and blick in a given way, and the "seeing" is relayed via light. The fire beetles are probably both sensing the fire nearby (they have eyes for that) and also transmitting that information to other beetles nearby perhaps via clicks or just their wing noises, and so on down the line. Why bother? So as to ensure there a
Re: (Score:2)