A clever way forward for powering the internet of things.
Many prophets of information technology (IT) believe that the next big movement in their field will be the “internet of things”. This, they hope, will connect objects hitherto beyond the reach of IT’s tendrils so that, for example, your sofa can buzz your phone to tell you that you have left your wallet behind, or your refrigerator can order your groceries without you having to make a shopping list. That, though, will mean putting chips in your sofa, your wallet and your fridge to enable them to talk to the rest of the world. And those chips will need power, not least to run their communications.
Sometimes, this power will come from the electricity grid or a battery. But that is not always convenient. However Shyam Gollakota and his colleagues at the University of Washington, in Seattle, think they have at least part of an answer to the problem. They propose to reconfigure a chip’s communications so that they need almost no power to work.
Most conceptions of the internet of things assume the chips in sofas, wallets, fridges and so on will use technologies such as Wi-Fi and Bluetooth to communicate with each other—either directly, over short ranges, or via a base-station connected to the outside world, over longer ones. For a conventional chip to broadcast a Wi-Fi signal requires two things. First, it must generate a narrow-band carrier wave. Then, it must impress upon this wave a digital signal that a receiver can interpret.
Following Moore’s law, the components responsible for doing the impressing have become ever more efficient over the past couple of decades. Those generating the carrier wave, however, have not.
Dr Gollakota and his team reasoned that it should be possible to separate the jobs of generation and impression. The system they have designed has a central transmitter (which might be built into a Wi-Fi router) that broadcasts a pure carrier wave. Dr Gollakota’s new chips then impress binary data on this carrier wave by either reflecting it (for a one) or absorbing it (for a zero). Whether a chip reflects or absorbs the signal depends on whether or not its aerial is earthed, which is in turn controlled by a simple switch.
Not having to generate its own carrier wave reduces a chip’s power consumption ten-thousandfold, for throwing the switch requires only a minuscule amount of current. Moreover, though Dr Gollakota’s prototypes do still use batteries, this current could instead be extracted from the part of the carrier wave that is absorbed.
The chips in this system, which Dr Gollakota plans to unveil on March 17th at the USENIX Symposium on Networked Systems Design and Implementation in Santa Clara, California can, he claims, transmit data at a rate of up to 11 megabits a second to smartphones or laptops over 30 metres (100 feet) away, and through walls. Though that rate is worse than standard Wi-Fi it is ten times better than the low-energy form of Bluetooth which is the current favourite for the internet of things.
Via Levin Chin, Dr. Stefan Gruenwald, THE *OFFICIAL ANDREASCY*