Breaking News
News & Analysis

Technion: Compressed sensing ups data acquisition resolution, cuts cost

8/18/2010 01:44 PM EDT
3 comments
NO RATINGS
Page 1 / 2 Next >
More Related Links
View Comments: Newest First | Oldest First | Threaded View
kendallcp
User Rank
Rookie
re: Technion: Compressed sensing ups data acquisition resolution, cuts cost
kendallcp   8/22/2010 4:24:04 PM
NO RATINGS
An earlier paper on the process can be found at http://arxiv.org/PS_cache/arxiv/pdf/0902/0902.4291v3.pdf, a colleague pointed it out to me last year It's interesting, but there's no "free lunch" in there. The signal path still needs to deliver the required performance at the required bandwidth. It's a little different from the commonplace (in radio) application of undersampling, where the sampling process targets a higher order band of alias frequencies. Here, sparseness in the spectrum is exploited to allow the wanted aliases to be watermarked as belonging to a particular alias zone. But it's a bit disingenuous of one of the originators to say that they are doing something that EE's are normally taught not to do.

eryksun
User Rank
Rookie
re: Technion: Compressed sensing ups data acquisition resolution, cuts cost
eryksun   8/22/2010 11:05:29 AM
NO RATINGS
Each channel gets modulated separately by a PWM signal that has a spectral density specific to that channel. The modulated spectrum for each channel thus consists of scaled and shifted copies of the input signal's spectrum at the harmonics of the PWM signal. The important part here is that the resulting spectra of the 4 channels are all unique in a mathematically precise way. Thus the aliased output of undersampling is unique for each channel and mathematically related in such a way that it's possible to compute the original signal from the combined information. Also, since they claim compression, I assume each of the 4 ADCs is running at less than 1/4 the nyquist rate. If you record to an intermediate media, you could use this technique to sample at the bandwidth of the analog media using a much lower bandwidth ADC (but precise and accurate) -- trading off speed and real-time processing vs. the cost of having a separate ADC/PWM for each channel.

yosko
User Rank
Rookie
re: Technion: Compressed sensing ups data acquisition resolution, cuts cost
yosko   8/19/2010 12:07:35 PM
NO RATINGS
They suggest swapping one high-frequency sampler, with 4 lower-frequency samplers, each working on a different (ie. shifted) 'clock' (PWM signal). By modulating the high-frequency input signal, you "copy" it's frequency content to the lower frequencies (ie aliasing), which usually "harms" the input signal and causes data-loss. I'm assuming the input signal has a very narrow band frequency content, so the aliasing to lower frequencies doesn't harm it.

Top Comments of the Week
August Cartoon Caption Winner!
August Cartoon Caption Winner!
"All the King's horses and all the KIng's men gave up on Humpty, so they handed the problem off to Engineering."
5 comments
Like Us on Facebook

Datasheets.com Parts Search

185 million searchable parts
(please enter a part number or hit search to begin)
EE Times on Twitter
EE Times Twitter Feed
Radio
LATEST ARCHIVED BROADCAST
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.
Flash Poll