Human echolocation
Dr_Acula
Posts: 5,484
Could the propeller help blind people to see?
An interesting article on human echolocation here http://www.southampton.ac.uk/engineering/news/2013/05/20_echolocation.page and another one here http://www.bbc.co.uk/news/magazine-19524962
Ride a bike or play basketball just with clicks. But clicking, hand claps and foot stomps have problems as described in that second article by a blind person -
So how about clicks that are ultrasonic and downconvert the echoes?
Here is a whole thesis on the topic http://libdspace.uwaterloo.ca/bitstream/10012/3878/1/TCDavies%20final%20thesis%20August22.pdf and here is another shorter article http://redwood.berkeley.edu/w/images/f/f7/SohlDickstein_et_al_echolocation_for_the_blind_icme_preprint.pdf
In that second article, the hardware is described as a windows computer with a soundcard and a wheelchair battery, and there is a photo on page 5 of the whole setup in a backpack.
Ah, the inner propellerhead says...
Of course there are smaller versions. You can buy a portable bat detector on ebay for $100. But that is just for hearing bats, not for echolocation.
The simplest downconvert is a heterodyne, or for something a bit more complex, a superhet. Some other techniques here http://en.wikipedia.org/wiki/Bat_detector and the problem with the simpler analog methods is you lose amplitude information.
The long theses referenced above used some more complex DSP techniques to do doppler shifting with some promising results. Time stretching may also be useful. One of the articles about bats talks about how some bats can close off their ears when they chirp, so they don't go deaf with the loud noise and can hear the echoes better. That can all be done digitally.
To take the technique used by the Berkeley team, produce a chirp sweeping from 25khz to 50khz modelled on a bat chirp, store the echo, then play it back 25x slower. That ought to be possible with a propeller. And with different software, you could quickly switch to doppler mode if you wanted.
Thoughts would be most appreciated!
An interesting article on human echolocation here http://www.southampton.ac.uk/engineering/news/2013/05/20_echolocation.page and another one here http://www.bbc.co.uk/news/magazine-19524962
Ride a bike or play basketball just with clicks. But clicking, hand claps and foot stomps have problems as described in that second article by a blind person -
"I believe it's discouraged because it's seen as a 'blindism' - if you're clicking then you're drawing undue or negative attention to yourself."
So how about clicks that are ultrasonic and downconvert the echoes?
Here is a whole thesis on the topic http://libdspace.uwaterloo.ca/bitstream/10012/3878/1/TCDavies%20final%20thesis%20August22.pdf and here is another shorter article http://redwood.berkeley.edu/w/images/f/f7/SohlDickstein_et_al_echolocation_for_the_blind_icme_preprint.pdf
In that second article, the hardware is described as a windows computer with a soundcard and a wheelchair battery, and there is a photo on page 5 of the whole setup in a backpack.
Ah, the inner propellerhead says...
Of course there are smaller versions. You can buy a portable bat detector on ebay for $100. But that is just for hearing bats, not for echolocation.
The simplest downconvert is a heterodyne, or for something a bit more complex, a superhet. Some other techniques here http://en.wikipedia.org/wiki/Bat_detector and the problem with the simpler analog methods is you lose amplitude information.
The long theses referenced above used some more complex DSP techniques to do doppler shifting with some promising results. Time stretching may also be useful. One of the articles about bats talks about how some bats can close off their ears when they chirp, so they don't go deaf with the loud noise and can hear the echoes better. That can all be done digitally.
To take the technique used by the Berkeley team, produce a chirp sweeping from 25khz to 50khz modelled on a bat chirp, store the echo, then play it back 25x slower. That ought to be possible with a propeller. And with different software, you could quickly switch to doppler mode if you wanted.
Thoughts would be most appreciated!
Comments
I remember when we used to ski regularly, one of the instructors used to take a blind boy (teenager IIRC) skiing. He used to clap and speak so that the boy could follow. I was amazed at his skill and his ability. I just used to think how amazing it was that the boy could follow the sounds and never thought about the actual noise bringing attention to the fact.
I would think you will have to have a number of directional antennas so that the relative direction of the echoes can be located. I wonder if a tiny microphone and speaker fitted to an earplug (one for each ear) would work with the electronics being located in a belt/pocket, much like an iPod.
I think the idea is a fairly decent one to be honest. Who knows... We have EEG (mind reading), one day we might have "mind writing", we can mount a 3D scanner on someone's head, and dump the information straight into their brain :P
Watching that, I imagined that a great combination for a blind person would be echolocation combined with a haptic belt which can give you an additional directional sense - it tells you where north is, at all times. Some guy made his own version of the belt described in the article, using an Arduino.. say no more!
-Tor
I saw that same show. I still have a pile of piezo tweeters from radio shack and lm386 op amps for this project. One of the many goal projects that are waiting for my foundation projects to be completed. Good to remember this.
http://www.youtube.com/watch?v=en2r4LeGkqA
Grant Study
http://forums.parallax.com/showthread.php/145462-How-to-digitally-record-a-10ms-pulse-then-stretch-it-out-into-audio-in-real-time?p=1158493&viewfull=1#post1158493
Another project, relatively recently (2010), that I was involved in helping one of the youth at the church, was a science fair project proving the speed of sound. A unique thing that I wanted to point out is that we were NOT using any ultrasonics to perform the test, but simply an 8 Ohm speaker and an electret microphone. This is significant, because of the misconception that you must use ultrasonics (specifically 40kHz) to measure distance by echo location, which is not true. Audible ranges are perfectly acceptable.
Proving the Speed of Sound
http://forums.parallax.com/showthread.php/122467-Proving-the-Speed-of-Sound-5th-grade-science-fair-project
Yes. Check the Penguin Superhero article in Robot magazine and the downloads. It has a PING sensor that can see in complete darkness out to about 9-feet. I've had good results with some PING sensors out to around 12-feet. The program converts the distance into a scaling sound in real time. It also works for normal eyesight, walking in total darkness, with the display of graphics representing distance. With a little practice, it's possible to determine the location of nearby objects in a rudimentary fashion. The code uses the BS2px. When using the Propeller, it could speak and return the actual distance in feet and inches.
Daniel Kish has since taken on a role of teaching blind people all over the world how to use conscious echolocation to see all sorts of different objects and obstacles and even go "sight-seeing". He's a very smart guy and has done a few TED Talks: http://www.youtube.com/watch?v=ob-P2a6Mrjs
And yes, blind people have a stigma against "using" echolocation by emitting clicks and refer to it as a "blindism" or something that provokes unwanted attention. If more sighted people were exposed to what echolocation is, however, more blind people might be tempted to want to learn.
Its not using a prop, but it might be helpful.
jt