For pervasive 5G and IoT, prepare for wind turbines on cells
Boffins imagine lots of small off-grid cells to power pervasive data grid
As growing user populations and hunger for mobile capacity drive ever-smaller cell sizes (even before the advent of 5G), electricity becomes a problem.
A group of researchers from the Hong Kong University of Science and Technology reckon renewables will be the only way to power the small cell networks of the future.
In this paper, they say the lower power of small cell base stations makes them an ideal candidate for small “energy harvesting”-type power. Compared to the 43 dBm transmission power of a normal base station, they explain, a femtocell base station operates at around 17 dBm.
Additionally, femtocells have simpler wireless processing, and because they're operating at lower power, they don't need to divert energy to cooling. So instead of the typical 225 W per transceiver for a macro base station, they write, the small cell model probably runs at between 5.2 W for a femto-cell up to about 72 W for a microcell.
However, unlike the very tiny requirements of Internet of Things devices, for which even ambient radio signals can sometimes provide enough power, 5W and greater can only be provided by more conventional sources.
Hence their enthusiasm for a combination of solar and miniature-scale wind. “100 W electric power can be generated by either a 121cm x 53.6cm solar panel under rated sunlight radiation, or by a rotor with a 1m diameter under an 8m/s wind speed” they write.
Such solutions also have long lifetimes and negligible ongoing costs after deployment, they write.
Not only that, but the two approaches complement each other well, the researchers note: “On daily timescales, high pressure areas tend to bring clear skies and low surface winds, which is favourable for solar harvesters, whereas low pressure areas tend to be windier and cloudier, and thus are good for wind harvesters”.
There will, however, be inevitable variability in energy, which they say will require adjustments to both the communication protocols and transmission strategies.
For example, instead of a user device associating with the nearest base station (as happens with conventional cell deployments), the network protocol could take the energy profile of small base stations into account. That way, a base station that's running into energy constraints could be skipped and the connection made to another nearby micro-pico-femto cell.
In that model, base station selection would be based both on signal-to-noise ratio (SNR) and available energy: “each user will be associated to the SCBS that provides the highest receive SNR with the available energy”, the paper suggests.
Other design issues the paper says need research include sleep strategies, user scheduling, and channel estimation.
The paper's authors are Yuyi Mao, Yaming Luo, Jun Zhang and Khaled Letaief of the Hong Kong University of Science and Technology. Letaief is also an IEEE fellow. ®