Hacker News new | past | comments | ask | show | jobs | submit login
New Chip Brings Ultra-Low Power Wi-Fi Connectivity to IoT Devices (ucsd.edu)
215 points by rbanffy 43 days ago | hide | past | web | favorite | 56 comments

> The device, which is housed in a chip smaller than a grain of rice, enables Internet of Things (IoT) devices to communicate with existing Wi-Fi networks using 5,000 times less power than today’s Wi-Fi radios. It consumes just 28 microwatts of power. And it does so while transmitting data at a rate of 2 megabits per second (a connection fast enough to stream music and most YouTube videos) over a range of up to 21 meters.

Seems like ideally suited to surreptitious surveillance

It's backscatter tech, which was literally invented for surveillance (see Great Seal Bug [0]) -- modern(ish) versions of which are prominently featured in the NSA's ANT Catalog [1].

This is the same technology as UHF RFID (now Rain RFID). Using dedicated RF carriers rather than ambient WiFi signals, we've demonstrated 100Mbps at 10uW using QAM Backscatter signalling [2]. Similar tech was used to read neural data off of dragonflies in flight [3] or self-contained implantable devices [4].

[0] https://en.wikipedia.org/wiki/The_Thing_(listening_device)

[1] https://en.wikipedia.org/wiki/NSA_ANT_catalog

[2] http://people.ee.duke.edu/~sjt/publications/pdf/2013_Thomas_...

[3] https://www.wired.com/2013/06/dragonfly-backpack-neuron/

[4] http://www.travisdeyle.com/publications/pdf/2017_rfic_implan...

> The Wi-Fi radio runs on extremely low power by transmitting data via a technique called backscattering. It takes incoming Wi-Fi signals from a nearby device (like a smartphone) or Wi-Fi access point, modifies the signals and encodes its own data onto them, and then reflects the new signals onto a different Wi-Fi channel to another device or access point.

Seems like this chip is purpose built for surreptitious surveillance.

It looks like it completely depends on there being other types of radio waves and modifying them. If you were security conscious you could either check if radio signals were being modified somehow or just shield off certain rooms (faraday cage).

They already said IoT

Or IoS: Internet of Surveillance

As the saying goes "the S in IoT stands for security."

As for surveillance. Well, it is a transceiver, hence it can be localized and tracked.

And, IoPO: Internet of Planned Obsolescence.

And, TaaS: Things as a Service.

Because those three are the real reasons companies put radios in consumer widgets. Getting a data stream that can be resold to marketers, getting people to buy more garbage more often, and getting people to pay a regular subscription fee to avoid having to care about buying and disposing of garbage products.

There is no way to escape this. By advancing IoT we also advance IoS, it is as simple as that.

The way to fight IoS is not by restricting advancement of technology. Even if we wanted, it is already way past it. We can assume governments are years if not decades ahead.

Would separate (v)LANs and super strict firewall rules not help somewhat? With the addition of things like DoH it's going to be trickier to "secure", or rather, isolate IoT devices for sure, but _impossible_?

whose to say that your devices can't communicate to the outside world via it's power supply?

True that.

2Mits will be the signaling rate. You won't get 2Mbits of throughput, more like 1.

I'd love to know how they got the power consumption down so low! 28microwatts when active, that's really something.

Some maths: 0.000025W. A typical AA battery is 1.5V / 2Ah (https://www.powerstream.com/AA-tests.htm). 1.5*2 = 3Wh. 3Wh / 0.000025W = 120000 hours runtime. = 5000 days = 13 years, off a single AA battery.

That can't be right? That's incredible if true.

They talk about the technique in the article:

> The Wi-Fi radio runs on extremely low power by transmitting data via a technique called backscattering. It takes incoming Wi-Fi signals from a nearby device (like a smartphone) or Wi-Fi access point, modifies the signals and encodes its own data onto them, and then reflects the new signals onto a different Wi-Fi channel to another device or access point.

It sounds like the device doesn’t actually transmit its own signals, which I believe is what typically requires the most power.

However, I get the feeling from this press release that while the WiFi chip uses less power, this does nothing to reduce the power requirements of the processing chip that is actually using the WiFi chip to communicate. In the article they use a camera as an example in the diagram. Even though the WiFi component might require less juice, you’re still going to need a hefty power supply to power the camera.

That is true, but, there are applications where the majority of the power budget goes to the radio, such as a wireless sensor. I worked with battery powered wifi sensors in the past, and very infrequent transmissions (15 minutes) still accounted for 90+% of the power budget. Average current for a 2-3 second transmission was something like 60-80 mA, with a huge in-rush current on boot. I see this and think it's probably too good to be true in practice for something like that.

Still, the problem with IoT devices is distributing the power to them. It is cheap to have some central device in your home that can have relatively powerful transmitter and sensitive receiver if you can drop a bunch of backscattering devices powered off battery.

Ah thanks, that explains it. So i assume it won't work unless something else is also transmitting?

I'm struggling to find a single use case where this would work. You need 2 wifi APs, regular wifi usage and a device where such aggressive power saving is more important than working reliably.

Actually, just 1 AP. It can backscatter signals to the AP from other wifi devices such as phones. Regular wifi usage is not a constraint in most applications.

Doesn't it say that it has to change to a different wifi channel?

It sounds like this technique won't work for realtime communication, but that doesn't rule out much of the IoT use cases. As for the 2 APs requirement: when's the last time you've been in range of just one AP, with the 2.4GHz band otherwise quiet?

I don’t see anything that says you have to control both APs. Almost everyone in populated areas lives within range of a few neighbour APs.

Marketing is the use case.

Uses New ultra green eco friendly low power wifi technology that is micro sized. Your internet will be FASTER! BUY NOW

Automated factories

If you have a factory you run wires to the things that are important for your factory to run.

Collecting data from those processes is arguably easier done with Wi-Fi

Easier but factories are unfriedly to WiFi with all the giant metal structures, high,medium,low voltage power cables, emi noise from motors and VFDs, etc.

WiFi is more difficult to troubleshoot whereas Pretty much anybody can use a cat5 cable tester.

And there are already cable trays and raceways going to all of the different equipment because it uses electricity.

So in the industrial processes I’ve been involved in radio was only used outdoors and when communications cable would have had to have been strung up on poles.

Reliability and ease of troubleshooting are characteristics industry values more than labor saved running cat5, in my experience.

Cameras are not entirely out of the question:


This is the trick, it's using backscatter to piggy-back off nearby wi-fi signals:

> The Wi-Fi radio runs on extremely low power by transmitting data via a technique called backscattering. It takes incoming Wi-Fi signals from a nearby device (like a smartphone) or Wi-Fi access point, modifies the signals and encodes its own data onto them, and then reflects the new signals onto a different Wi-Fi channel to another device or access point.

"Knows nothing beyond what a patch antenna is" here, and I have a couple of question if you could please answer them:

- Doesn't this degrade the quality of the original signal ? (i.e. some sort of destructive interference)

- Isn't it illegal to interfere in signals this way ?

If this is reflecting onto a different channel (frequency) then it shouldnt interfere with the original signal (except perhaps by reducing its effective power output by a small amount).

It is a general FCC rule that radios and certain other types of devices are not supposed to intentionally cause interference, but in this case there shouldn't be any. Even if there was same channel interference, its not necessarily considered intentionally harmful - for example, its not illegal to use the same WiFi channel as your neighbor as long as you are within other specified technical limits such as transmit power or antenna gain. But doing so for the express purpose of interfering with them would be.

Thanks for the explanation !

Most batteries have shelf lives under 10 years so the typical AA probably wouldn’t last for 13 years. That said, it would be pretty incredible.

From the article it says it uses backscattering and the range is not very high. I guess it would work for some devices in areas that are already blanketed with WiFi signals, though?

Backscatter. It sounds like it needs two adjacent networks. One network transmits something, and the device backscatter-modulates its signal to be received by the other network.

It is incredible, but you also need to remember to be able to transmit the device needs more, sensors, processing power, etc.

It looks like some people at UW did something similar a few years ago.


Just a reflection about this: Could a modification of this backscattering technique be used so that hundred of this devices use the same channel to transmit a set?, I am thinking about collecting the information of hundred of devices each sending a single bit in the same channel. This must be something like dividing a channel in subchannels but I am thinking about using the same frequency in the final output. Yes, it should be better if I read something about transmission of information.

I am thinking about an API in which you assing a number (index of device) to each device and you receive a set of bits ( for example danger/no danger) without having to make any modification to wifi signals or having to select, create or buy new channels and without danger of interfering with other signals because use a lot of bandwidth.

You (me) must read something about backscattering, a quick suggestion is: https://www.sciencedirect.com/topics/computer-science/backsc...

It's about time that someone harnesses the free power that everyone is broadcasting....

This is really nice. I could imagine smaller MCU integrating this and perhaps we could get rid of bluetooth. Don't use the latter very often, but the one advantage I know of is low power consumption.

So this needs a powerful wifi transmitter nearby? Does a normal router continuously blast out enough power for these backscatter transmitters to harness, so no other transmitters are present?

Battery is the biggest problem of IoT, this could be a tipping point.

They make a big deal about the small size of the chip (less than a grain of rice). But what's the smallest size antenna this thing could be plugged into and remain effective?

Couple it with a low power camera and suddenly wireless security camera can last for years.

Does anyone know if you need special router firmware for this to work ?

You need special router hardware for this to work.

There's no suggestion that you do and the abstract says "commodity" routers.

What's the lowest power WiFi device we have now?

It depends what your are doing. Some chips can get down to about 100uw to maintain associated and send small amounts of data. I'm not sure how they would compare at 2mbit or how the range would compare.

I'd say about 10mA (about 30-40mW at 3.3V) is the lower limit if you go a bit out of Wifi spec to save some power with current wifi modules in the common market for Microcontrollers and IOT applications.

What devices have that?

I've measured that on an ESP32, specifically, using various power saving measures and substracting the chip's own consumption. Though Wifi is atleast 99% of power consumption on a sensor device with not much else going on.

Interesting. Thanks for replies. I was hoping to find something I could add a small battery to and it could run a small vpn client as long as it had wifi connection. Ideally for months or years without maintenance if possible.

To save power, it could even wake up at certain times, say 5 min a day, then go back to sleep until that time window arrives again.

I can do all of the software I just don't know what hardware I should be looking at.

Computing antennas are fun

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact