Discussion in 'Article Discussion' started by Gareth Halfacree, 24 Feb 2016.
10,000 times lower power draw.
I never considered WiFi to be very power hungry but having thought about it i guess it must be seeing as (i think) they're mini radio transmitters and how big some commercial transmitters were.
When you've got a remote sensor powered by an ickle lithium polymer battery that you don't want to be disturbing for the next five years, Wi-Fi is plenty power hungry. Passive Wi-Fi requires 59.2µW at 11Mb/s dropping to 14.5µW at 1Mb/s (which would be more than fast enough for real-time readings from a temperature sensor or similar). A 802.11b radio requires 670mW - or 670,000µW - to do the same thing at the same speed. Let's say your radio is always running - perhaps you're streaming video or audio or something - and that the power draw of your sensor is negligible: that means a 280mAh 3.7V lithium polymer battery (standard for small embedded devices) would run dry in an hour and a half, whereas under Passive Wi-Fi it would run for 17,559 hours.
Now, if you're talking a tablet, laptop, or smartphone, those figures look different: the biggest power draw will always be the screen, closely followed by that oh-so-hungry processor, meaning that the Wi-Fi module makes up a tiny fraction of the overall power draw. For embedded, though? The radio is the hungriest part, and dropping that by a few orders of magnitude is good news indeed.
There's no such thing as a free lunch in physics.
Less power used means less power transmitted, which consequently translates into worse SNR (assuming one can establish a bidirectional connection in the first place ...).
Plus there's the bit about "passive WiFi" re-modulating and re-radiating received signals, which does nothing for power savings at the remote end (so only one end of a connection can employ this "passive WiFi").
Interesting achievement from technical point of view though although I don't see it ever making a significant impact (I'm not into "smart" houses due to security concerns and this is pretty much useless for any other use).
Powering cat cams for longer would be a boon for all them YouTube videos, how can that not be considered a significant impact.
...You are aware that the market for wireless sensor networks for industrial, rather than consumer, use is absolutely massive, right?
Alecto - have a read about M2M and IoT. It's MASSIVE
Perhaps if they develop some custom bulletproof protocol that gives 100% priority to reliability - but to me, reliability and low power consumption do not often go hand in hand. Not in industrial areas. Pretty much every single sensor/actuator/etc is connected by wire and higher level control systems do not really like missing or erratic signals.
Also in industry where you have motors, big inverters, long power cables carrying massive currents and generally a lot of EM noise, low power wireless might not be the best solution.
Personally I would expect this in IoT, which has been picking up speed nicely in last few months.
Not all 'industry' is chunky wires and big motors. Passive NFC sensors buried in the soil have shown considerable promise in allowing farmers to monitor moisture and temperature level as they drive their tractors over their fields, as but one example. Having done plenty of work in this area, I can confidently say that the industry is going to flip if they can use proven, effective 802.11b at one-10,000th the power draw.
You know the Internet of Things is just a new name for Machine To Machine (M2M), right? As for the "last few months," try years - the IoT, in both industry (where it's the new name for M2M) and consumer (where it's smart thermostats and really crappy security systems), has been growing for a long time, now.
Ah, did not know that, seems I spend too much time with my type of machines and it starts to show
Thanks for the heads up!
Surely anything utilising the wifi would be powered anyway?
Wifi is incredibly power hungry - many people aren't aware, because they never turn it off. For my phone, my battery might last 3 days with wifi on. With it off, it lasts a full week. My laptop might last 3 hours with wifi, and around 4 or 5 hours with wifi off.
I would like to clarify that it doesn't matter if I'm actively using the internet connection. If I do, the lifespan shrinks even more.
As for practicality, I definitely see a lot of great purposes for this (particularly in robotics) but generally speaking, passive wifi doesn't have much leverage in such purposes:
* If you need the bandwidth, 802.11b speeds might not be good enough. Other forms of wifi will outperform it; maybe even non-passive b will outperform, since it's a more "solid" connection.
* If you need good range at low energy consumption, there are plenty of low-power radios that have ranges of hundreds of meters. Slow data rate (usually 9600bps or less) though.
* Depending on the version of Bluetooth you're looking at, you can effectively get the same speed, range, reliability, and security with BT. BT is usually a much cheaper technology, but, I'm not sure if it's more power efficient than passive wifi.
I'm not trashing this feature though. There are definitely uses for passive wifi, but it's definitely a niche.
It's not; the article clearly states that Passive Wi-Fi is (claimed to be) 10,000 times more power efficient than standard Wi-Fi and 1,000 times more power efficient than Bluetooth Low Energy. Plus has higher data rates, so that's good news for future headsets: better quality audio and massively improved battery life a-go!
I understand that, and I did read that part. But under what conditions? Even when bluetooth has an established connection, it uses almost no energy at all if there is no data being transmitted. Wifi, however, will use more energy than Bluetooth even if there isn't an established connection (in my experience, anyway).
So, it's not that I don't think passive wifi can't use 1000x less energy, but it could be at the cost of reliability, speed, and/or security (in comparison to bluetooth).
Read the linked paper; it's all explained there, complete with pretty graphs. (I'm a sucker for a pretty graph.)
Separate names with a comma.