Backscatter is (indeed) phenomenally interesting. We used it to build "cyborg" dragonflies [1] and to build battery-free tags capable of transmitting video and audio [2]. We've demonstrated fully-passive tags (at 915MHz UHF) capable of 100Mbps communications. Using ambient signals as the carrier is a pretty cool extension.
I believe you re:ambient signals (or rather, no reason to disbelieve you).
For those who are unfamiliar with backscatter modulation / passive UHF RFID... most systems are forward-link limited: the limitation on range is dictated by the ability to harvest enough energy to powerup the tag.
At a fundamental level, backscatter modulation is little more than a reflector that modulates its radar cross section. You could imagine building massive spinning reflectors to serve the same purpose at very low datarates. As for conventional tags, I've seen read ranges on the order of 30-100 meters using directional antennas.
You can build systems that are limited by receiver sensitivity, eg. by using different RX-TX antennas or by using a battery (not for transmission, but for very low power load-modulating the antenna). This is a classic radar problem -- but you can get pretty substantial ranges.
Backscatter? Don't you mean just the ambient radio waves?
Not quite sure I understand the backscatter concept... but here's an analogy I hope you can help me with.
Imagine you and I are sat next to a round pool of water. In the middle is bird that's swimming around, flapping it's wings, disturbing the water. We both observe very similar ripples in the water. However, if I place a very thin sheet of plastic in the water, a baffle, the pattern that you observe is now different! (It's thin so it doesn't cause it's own ripples).
If I take this baffle in and out of the water in a discernible pattern, I could send you a message. The hard part is for you to decode when my baffle is in or out of the water. You don't just notice one or two ripples are affected, but over a few you can tell when the baffle has been placed and removed.
What I don't quite understand is how you can tell. In the analogy, you can make assumptions about the normal type of pattern, but realistically, how does it work?
The water analogy is hard to understand. As I said below:
At a fundamental level, backscatter modulation is little more than a reflector that modulates its radar cross section. You could imagine building massive spinning reflectors to serve the same purpose at very low datarates.
Basically, you're looking for the presence (or absence) of reflections over time. Your "baffle" needs to be able to change it's state to change how much of the wave it's reflecting, and the receiver needs to be able to detect the minute changes in the reflections. On the receiving side, you'd have another "duck" that generates waves to exactly cancel out the incoming "big" (direct) waves through destructive interference (but this new duck's waves don't propagate to the rest of the system). Then, all you're left with at the receiver are the small reflection waves, which you measure changes in over time. [See, it's hard to visualize for water.]
For RF signals, this is done by mixing with a local copy of the carrier, low-pass (or bandpass) filtering the baseband signals, and then running symbol correlators. For ambient backscatter... I presume they have to do carrier recovery for their own LO.
Thanks... I appreciate the explanation. It's kind of what I thought.
I still don't quite understand how I can cancel out the carrier. I have an incoming signal, that includes the backscatter.
Do I average out the signal to get the carrier, use that averaged signal to cancel out the non-averaged signal - ending up with a version of the signal, and then apply a filter to this signal (to remove anything else that is causing weaker back scatter/multi-path interference), and hope that the intentional back scatter survives?
This is very cool. There was an interesting project at a former employer which used ambient light to power conference room schedule displays, attached to the outside of conference rooms.
Something like that with e-ink would be even nicer (you can save up energy for changing things rather than keeping an LCD on)
There was also some interesting robotics work out of Mark Tilden's lab called BEAM which was primarily solar powered things but it would be really cool to do those things with backscatter powered energy harvesting devices.
Short answer: Probably okay if you have only a few of them, but you will probably see degradation if you have lots of them in a cluster.
Longer answer: The Shannon–Hartley theorem tells us how much data can be pumped though a communications channel before we get unrecoverable errors. The data capacity for modulating the backscatter will come out of the same "bucket" as the capacity for the active transmitter, meaning it will slightly reduce the channel capacity available to the active transmitter. In practice, the active transmitter is probably not near to the theoretical limit and has a rate much greater than the backscatter transmitter, so it can easily correct the errors caused by the backscatter modulation. Increase the amount of data on the backscatter channel, by upping the rate or number of transmitters, and the active transmitter probably won't be able to recover from the loss of available capacity, and the user will see an increase in bit error rate.
I'd like to see a MIMO version of the backscatter transmitter!
This communication method creates a type of interference called inter-symbol interference, or ISI, for any intended receivers of the carrier transmission. ISI is basically what happens when echoes cause a signal to interfere with itself. Since this communication method modulates the amount of echo in the carrier transmission, I can only assume that any typical anti-ISI techniques used by most receiver designs (e.g. rake receiver) would be confused by the rapidly-changing ISI caused by these devices, increasing the total noise present in the received signal.
This was my question, too! According to the paper:
>"Finally, we test the interference of ambient backscattering and find that, even in less favorable conditions, it does not create any noticeable glitches on an off-the-shelf TV, as long as the device is more than 7.2 inches away from the TV antenna."
It's discussed in more detail in 6.4. So far they've looked mostly at TV interference. I'd be surprised if it interfered with anything else at all, but that's all I could find for now.
RF is nonionizing radiation, meaning that it isn't energetic enough to knock electrons off of individual molecules. Obviously it's impossible to say "it can't cause cancer" with 100% certainty, but in all the studies that have been conducted on RF over the last 70 years, there has never been any indication that nonionizing radiation causes an increased incidence of cancer.
Ionizing radiation (X-rays, gamma radation, etc.) is on the other side of the electromagnetic spectrum relative to visible light. These types of radiation are energetic enough to interfere with and ionize molecules, such as your DNA. This can directly lead to cancer.
The only thing dangerous about RF is that if you disperse too much RF energy into something (ie. your body) it usually comes out as an increase in temperature or as an influx of electrical current, which can cause internal/external burns, etc.
Source: mandatory yearly IEEE mandated RF safety training
If I understand this correctly, the backscatter is always there. It's residual radiation from the Big Bang. It's also why analogue tv sets show static when not tuned to a station. TV and Radio stations output a signal that's much stronger than the backscatter, so that's what gets received by the TV or radio set.
I could be wrong about any of this though...
Am I the only concerned that once a few billion of these energy parasites are deployed everyone will have to have to use more energy to broadcast the same signal the same distance?
That's not how radio travels though. Instead of thinking of it as a beam that may be diverted by these, think of it as a bright light (omnidirectional) at a distance.
If there are "things" in the way, some of that light will bounce off and cast a shadow behind those things. Or they may reflect some of the light against the neighbors and you'll get some illumination behind obstacles. Some of these obstacles can be transparent to the light, translucent, semi-opaque or completely opaque (I.E. Paper, Wood, Concrete, Concrete with rebar, Steel facade).
But since the signal travels in all directions, there won't be enough attenuation caused by these that are any more significant than actual obstacles to reception.
A bigger problem may be interference from neighboring devices that function the same.
Actually if you are close enough to the transmitter you can cause issues. I remember at school we went on a tour of the local transmitting station. It was used last century for worldwide transmission of BBC World Service. At the peak of operation in the 80s, it had ten 500 kW transmitters. Anyway, so apparently one of the local farmers decided he wanted to cut his electricity costs. I'm not sure exactly what he did, but it was something like this (on a bigger scale), which did actually cause detectable issues. I think he ended up in prison after they caught him :s
Observing the situation when I'm in the densely populated area where I have problem in communicating with my own WiFi base because of the noise from other people's stations, I can imagine exactly two potential problems:
1) In densely packed areas, the devices first installed stop working as soon as there are more devices, due to noise.
2) In far-away places, not enough energy for devices to work.
It's not noise from other WiFi devices that cause problems in denay packed environment. The problem is the available bandwidth is consumed by other devices. Bluetooth uses the same frequencies as WiFi, but overcomes the problem by using frequency hopping and a method of adapting to the environment.
I did an experiment a while ago on whether I could use a crystal radio to light up an LED. Short answer, yes. Long answer, only in very short pulses and that still required a fairly strong AM signal. I guess, this absorbs frequencies of a much wider range.
There was a time when battery-less electronics might have excited me, but these days dis-connectable batteries/power supplies and real, hardware off switches seem like important, useful features for just about anything.
I know, its a joke, but just the other day, I was thinking about faraday cages and how maybe it wouldn't be terrible to have one around the house and if I could pull it off with tin foil...
I was seriously considering tin foil to stop snooping from government rays. What. A. World.
I think it has more to do with Ubiquitous computing. Turning everything into a potential computer or sensor and then connecting them all together into a giant invisible network.
Shaping the physical world to be more intelligent or interactive seems a lot more interesting that just tracking things that move around for logistics purposes.
From my first read it seems that I will not be able to pay with my "backscsatter creditcard" in a Faraday cage or anywhere, where background radiation is strongly shielded. But from a general stand point, I really like the idea.
I remember, as a small child, building a powerless radio that vibrated the speaker using the strength of the radio transmission itself. Is ambient backscatter is the same thing, except across much wider frequencies ranges?
:) Yes, I am. It was part of an electronics kit my father bought for me. I remember carefully placing the pieces between spring-end contacts to construct the basic circuits.
Building electronics from scratch is very satisfying. I think I may get into DIY drones soon to scratch that itch.
Backscatter refers to reflecting RF power. Ambient backscatter is modulating the reflections of the ambient RF signals as opposed to beaming a signal in from a source as RFID does.
The device you built uses the incoming signal to power the device as well as drive the speaker. The devices in this article are using the incoming RF signal to both power the device as well as reflect for wireless communication.
This needs to be in any 2013 year in review article. The implications for this with existing and new technologies is unique and doesn't happen often enough, presuming my understanding is correct :)
There are those times when the device you're powering isn't supposed to be visible and would benefit greatly by the reduction in size from shucking the battery.
wouldn't deploying this in big numbers work against the solution itself? you would have so much noise in the existing signals that the main signal, as well as the bounced signals, would be useless?
Also, can we just say this is a cat's whisker receiver[1] that instead of powering a speaker/headphone powers another transmitter?
This is the same principle as long-range UHF RFID tags. There is a communications protocol (Gen2) that deals with collision detection and multiple access to allow for hundreds (thousands?) of tags in the same space. Part of the protocol involves a tunable value for "the probability that an individual tag should respond."
i imagine the signal would have to be reflected in a highly directional manner or at very close distances. it's cool nevertheless, but i think storing some solar power in a capacitor/coil would produce more powerful rf communication bursts at least for outdoors applications.
[1] http://www.wired.com/wiredscience/2013/06/dragonfly-backpack...
[2] http://www.travisdeyle.com/publications/pdf/2013_rfid_rich_m...