SoylentNews is people
SoylentNews
Scientists at the Technical University of Munich (TUM) have developed a holographic imaging process that depicts the radiation of a Wi-Fi transmitter to generate three-dimensional images of the surrounding environment. Industrial facility operators could use this to track objects as they move through the production hall.
Just like peering through a window, holograms project a seemingly three-dimensional image. While optical holograms require elaborate laser technology, generating holograms with the microwave radiation of a Wi-Fi transmitter requires merely one fixed and one movable antenna, as Dr. Friedenmann Reinhard and Philipp Holl report in the current issue of the scientific journal Physical Review Letters.
"Using this technology, we can generate a three-dimensional image of the space around the Wi-Fi transmitter, as if our eyes could see microwave radiation," says Friedemann Reinhard, director of the Emmy Noether Research Group for Quantum Sensors at the Walter Schottky Institute of the TU Munich. The researchers envision fields of deployment especially in the domain of industry 4.0 -- automated industrial facilities, in which localizing parts and devices is often difficult.
(Score: 2) by kaszz on Saturday May 06 2017, @07:18PM (6 children)
So they use a passive array of antennas. That equals size and cost. No data on how closely spaced they are though.
Here's the 2D math that recovers the amplitude I and phase Φ:
Homodyne scheme for phase recovery. (a) The time-domain signals of the scanning and stationary antenna are Fourier transformed and (b) normalized to each other to obtain amplitude attenuation I and the phase delay Φ imparted by the propagation for every frequency f within the wi-fi bandwidth. (c),(d) Holograms are obtained by doing this for every pixel. (e) Multipath reflections in the building lead to a strong modulation with VSWR=9.7 and correlated artifacts in the phase. fc≈2.4 GHz denotes the carrier frequency. (a)–(d) have been recorded using a 5 GHz wi-fi emitter, and (e) using a 2.4 GHz emitter.
And here's how to do the 3D thing using numerical backpropagation:
All following analysis is based on a data set, which has been recorded in the setting of Fig. 3(a). We capture a metallic cross-shaped phantom object. The illumination source is a commercial 5 GHz wi-fi emitter (TP-LINK Archer C20, 802.11ac), placed at a distance of ze=230 cm from the recorded plane, 90 cm behind the object plane zo. Image reconstruction is performed by numerical backpropagation. This scheme recovers the light field at an arbitrary depth z by propagating the recorded holographic wave front from z0 in space according to the angular-spectrum relation [29,30]]
I(x,y,z)=F−1[exp(∓i2π(z−z0)λ√1−λ2f2x−λ2f2y)×F[I(x,y,z0)]].
(2)
Here, F and F−1 denote the spatial Fourier transform of the light field and its inverse, respectively, and fx, fy the transform’s spatial frequencies. The sign in the exponential function is chosen negative by default and positive for 1−λ2f2x−λ2f2y < 0 to prevent an exponential blowup of evanescent waves. Backpropagation for a range of z yields a 3D picture of all emitters and objects traversed by the emitted beams.
One question is though how they handle any aperture that is small enough to cause diffraction? and then to calculate how that aperture looked like from the resulting waves and not seeing the wave themselves.
The computational load from a quick look for 2D seems to be two Fourier transformations per pixel and for 3D one fourier per pixel in depth. So for a pixel size of x,y,z on the order of (2*x*y)^2 * z*x*y Fourier transforms to get the whole three dimensional picture. Thus the critical performance factor computationally will be something that has some serious Fourier performance.
As for privacy. Just think about this. Most homes have a wireless thing of some kind and the walls are in many cases transparent for any practical purpose. So all it takes is a camera that has a "lens" for 2-5 GHz and a suitable antenna array as image sensor. Any idea on how to design such "lens" ?
Fused quartz glass consists of SiO2 that has a average particle size of 14 nm [researchgate.net] and a spectrum transparency at circa 210 - 2500 nm [technicalglass.com] perhaps that gives a hint on how to design a 2-5 GHz lens?
Here's a illustration of their experimental setup [www.tum.de].
Article: 2017-05-05 – DOI: 10.1103/PhysRevLett.118.183901 [aps.org]
(Score: 2, Interesting) by Ethanol-fueled on Saturday May 06 2017, @08:11PM (5 children)
Working for Boston Dynamics' Wireless Applications Division, I do know is that in practice is that each receiving element should be coupled into a set of phase-matched bandpass filters [wikipedia.org] tuned to pass between 2 and 5 GHz.
A channelized setup may increase selectivity at a cost of increased complexity.
(Score: 2) by kaszz on Saturday May 06 2017, @08:54PM (4 children)
So a filter for every second "pixel" ?
Still the how to make a GHz lens is a question. Perhaps some low-k dielectric?
DOI: 10.1023/A:1018316722377 suggests using plexiglass er=2.53 for 62.5 GHz.
(Score: 1) by Ethanol-fueled on Saturday May 06 2017, @09:06PM (3 children)
I wasn't talking about none of this glass lens sheeit, just traditional passive array design.
Too lazy to read the article and found the pic of the setup rather unhelpful, especially because the receiving array was a 1-D array. And the Christian cross was kinda weird as well.
(Score: 2) by kaszz on Saturday May 06 2017, @09:14PM (2 children)
Won't you need a (plastic?) lens to focus the 2-5 GHz waves onto said passive array design?
(Score: 1) by Ethanol-fueled on Saturday May 06 2017, @10:30PM (1 child)
Friend, it is retarded to recommend the use of an optic lens for frequencies considered to be low by modern standards, and anyway, this technology is retarded because it requires the objective to be between the radiator and the passive array. Germans should know better, this is at least 40 year-old technology.
But you're half-right about the optics: Computer vision and pattern recognition have been even more viable for this application, for decades. It seems the first German Muslim immigrant affirmative-action scientists passed physics 3 and cobbled together something cute.
(Score: 2) by kaszz on Saturday May 06 2017, @11:28PM
Ah, use software to implement the lens?
Well, that would make physical assembly even simpler. If it's possible to make the whole antenna array flat like those printed circuits antennas without too poor performance it would be bullseye.