SoylentNews is people
SoylentNews
Tags on a fragile packages may someday be able to say whether the goods are riding safely in the back of a truck or bouncing around in a hazardous way.
If Intel follows through on an IoT research project it demonstrated at Intel Developer Forum (IDF) this week, those tags could report on shipping conditions in real time without needing a battery to stay powered. After the package is delivered, the label might even be disposable.
Intel demonstrated a prototype "smart tag" for packages at Intel Developer Forum in San Francisco on Aug. 16, 2016. The tag could detect motion and show it on a chart in real time. It's the kind of IoT that enterprises might not even be aware of as a device or system but could still benefit from every day. Meanwhile, for a shipping company, it might save on labor and specialized technology rollouts.
The key to this oversharing tag is what Intel calls a mote. It's a chip so small it could get lost in a jar of coarse-ground pepper. The IA-32 mote shown at IDF, built on Intel's Quark architecture, would be the brains of a tag that could incorporate motion, temperature and other sensors.
The tiny size of the processor is a big part of what makes the whole system viable, as it keeps both energy draw and per-chip costs low.
The mote doesn't talk directly to the cloud. Instead, it uses a short-range, low-power network – in this case, Bluetooth Low Energy – to a local gateway device that forwards the data over a longer-range network like cellular. The local gateway might also process or store the data first.
The mote is so small and efficient, using only about 1 milliwatt, that it can run on energy harvested from radio waves wafting through the air around it. For example, a Wi-Fi network on a truck might bathe a trailer full of packages with enough RF (radio frequency) energy to power motes on every one of them, Intel researcher Turbo Majumder said.
(Score: 2) by Gravis on Friday August 19 2016, @11:52PM
Hold on there, Turbo! As someone who has worked on very low power computing systems, I know you are omitting some damn important information.
If you are harvesting power from RF, unless your chip is specifically designed with it's own power system, you will need a Power Management Integrated Circuit (PMIC) which requires power too. Add on to that you need to power the MEMS sensors. You are going to end up needing a constant stream of 3mW before we even consider you chip. When you say "so small and efficient, using only about 1 milliwatt," I think you mean that's how much power it needs to simply be active and "about" probably means you are probably closer to 1.5mW and you need more power for instructions to actually run. Now, if you aren't executing from FLASH memory and have FRAM registers, whenever your chip drops below the minimum threshold, your chip will lose it's data and effectively reboot. Now, you say you want to transmit your data using BLE? That's going to be about 15mW for a momentary transaction before you need to turn off that chip. That's a lot of things that aren't integrated and it will cost more power with every addition required.
TL;DR: unless you integrate power management, a CPU with non-volatile registers and a radio that can be toggled at will, you're going to need a shitload more power than 1mW.
(Score: 0) by Anonymous Coward on Saturday August 20 2016, @02:20AM
Having worked with the truck industry for a few years. The back of a truck is not a nice environment either.
Some carriers have amazing equipment. Others? They 'pass the required laws' and the rest is marginal.
Then you get to the drivers themselves. Oh your trailer needs power? Did they hook up the pigtail? Is the pigtail still any good? Oh it was hooked up and fine? Did they hook it up correctly? Oh so the pigtail is done correctly? Oh the previous driver hit an overpass and now the rf module is broke. Oh does your RF module like +/- 14v in about 200 spikes in 2 seconds? Once out of the cab expect your fail rate to go much higher.
When we put this stuff into a truck we would try to run our own power and make it tough and extremely durable. Hopefully, with little to no interaction from the drivers. Why avoid drivers? You run the gamut from does everything by the book and no problem extra step guy to 'ive been doing this since 1940 and no fancy computer guy is going to tell me my job'. Unfortunatly the trailer is the weak point in this scheme. You have to rely on the pigtail and the driver and trailers in marginal conditions.
It is all doable. But it is a tricky problem. The company I worked at we had it down to about 50-150 per module tossed in the back of the truck. Putting them on boxes was a total non starter. None of the shipping companies wanted it. They are making 1-2 bucks per cardboard box. Pallet shipping though the margin was better and reuse more likely.
OK so now you have the data. What do you do with it? You need a service to put it into. So you now need to sell this to the carriers. The small guys follow what the big guys do. So your system better integrate into peoplenet, mcleod, omnitracs, and others. The big carriers *may* experiment with you and your extra service. But the little guys just want to buy one package.