Professor Richard Langley, head of the Communications Research Group at the University of Sheffield, began the session by presenting the work his team has done as part of the Wireless Friendly Energy Efficient Buildings (WiFEEB) project. He highlights the need to factor in human occupants when modelling or designing buildings for wireless coverage, given the way we ab- sorb signals. As part of the project, he and his team looked at how well signals propagate in a typical Victorian house from a smart meter perspective. One of the key findings was that coverage fell as the frequency used increased. In his view systems such as Zigbee operating at 2.4 GHz “made life difficult for themselves, as these aren’t high data rate systems”.
Langley also explored the use of frequency selective surfaces (FSS) and intelligent wall units (IWUs) to create a smart environment “that we can change dynamically”, like a conference centre. FSS are printed panels with a window made up of a printed circuit board that switches between reflecting and transmitting the signal. However, he explains that the attenuation they can produce is only 20 dB and for many signals that isn’t enough. IWUs are easy to design, much cheaper than FSS and can do multi-band very easily. They consist of a receiving antenna, a filter, an amplifier and a transmitting antenna.
He notes that in some cases, e.g. prisons, hospitals and military buildings, there’s a need for communication signals to not enter or leave the premises. Langley explains that he and his colleagues have looked into stopping GSM mobile phone signals using a time varying FSS approach and found that this concept “is very robust for secure locations”. However, completely securing a building to prevent RF signals escaping or entering is very difficult.
Langley also discusses a scenario in which a UK residential area was made up of detached, semi-detached and terraced houses, each with a Zigbee Coordinator at 2.4 GHz in the meter box. He stressed the need for software that ensures the meters and the Wi-Fi access points (also on 2.4 GHz) are not all on the same channel to prevent interference.
Rob Piechocki, senior lecturer at the University of Bristol, introduced SPHERE – Sensor Platform for Healthcare in a Residential Environment – a £15 million project investigating the use of wearable devices to assess the movements of people with long-term conditions. SPHERE has produced its first ultra-low power wrist-worn sensor, is currently receiving data from the occupants of its first house, and aims to install its technology across 100 homes in Bristol from 2017.
Piechocki explains some of the challenges around the use of battery-less IoT devices. “There’s been quite a lot of stories on energy harvesting and lots of organisations claiming that they’ve cracked the problem... It’s not that easy. It depends on what you want to power and the sensor’s information rate. If you’re running a CO2 sensor that samples once every five minutes it’s not that big a deal, but if you want to power a wearable device that will do activity recognition it has a couple of tri-axial accelerometers that will easily use in the region of two gigabytes of data in about a month.
“Another big issue is that the only sensible way of powering these things is with RF, anything else in the indoor environment doesn’t really work,” he continues. “That’s challenging as RF energy drops off very quickly as you move away from the transmitter.”
Martin Ganley, director of smart homes and buildings at built environment consultancy BRE, highlights the wide spread of results within propagation models. He attributes this to both the variation in the type and ages of buildings, together with differences in how the measurements were carried out.
“Even when you narrow it down to suburban residential you’ve still got a large spread; 20 dBs or so across the frequency range. That’s very difficult for people planning networks. It’s made more difficult by the extensive use of energy efficient building materials that have metallic coatings or layers.”
He adds that recent years have seen the introduction of materials such as insulation board, which has foil on both sides, and multi-foil, which has around 14 layers of foil between other insulation materials. According to Ganley, manufacturers are now trying to use these materials to meet fire resistance requirements in building regulations.
He notes that as annual new build construction is typically less than one per cent a year of the total housing stock and 86 per cent of housing was built before 1967, retrofitting solutions into existing homes is going to be a big part of improving buildings’ performance.
Ganley explains the work BRE has done on in-building propagation for Ofcom, with the help of Aegis. This focused on outdoor/indoor propagation for a range of frequencies – 88 MHz, 217 MHz, 698 MHz, 2.4 GHz and 5.7 GHz. It also involved testing on two houses (one of which was retrofitted with energy efficient materials as part of the study) and some anechoic chamber testing. The latter gave propagation loss measurements of ~15-24dB for Low-E glazing, ~30-45dB for foil-backed plasterboard, and ~43-62dB for insulation board (foil on both sides). However, in situ measurements [See the graph below – Ed] differed considerably to those obtained in the anechoic chamber.
Graph of building entry loss, whole building averages. Taken from "Building Materials and Propagation", Ofcom 2014. Win = metalised windows, type 1 glass; Win2 = metalised windows, type 2 glass; FBP = foil backed plasterboard; foil = foil over windows. Credit: Ofcom, report produced by Dr Richard Rudd (Aegis), Dr Ken Craig (Signal Science), Dr Martin Ganley (BRE), Richard Hartless (BRE)
In response to Langley’s observation that metallicised glass “doesn’t seem to make much difference to signal propagation up to about 10 GHz”, Ganley counters that there are different types of this glass; “there are hard processing and soft processing versions. One uses tin oxide, one uses silver, and they use different thicknesses. I think one of them is probably winning out over the other.”
Ganley says that Richard Rudd at Aegis’ work found that Ofcom’s building has a propagation loss around 30-40 dBs and that the glass’ contribution was around 20 dB “or maybe more”. He adds that the next step is for BRE, Aegis and the University of Surrey to conduct follow-up work by modelling buildings to better understand these issues, and to look at wider housing stock.
Stephen Lowe, knowledge transfer manager for the modern built environment at The Knowledge Transfer Network, points out that consumers have long accepted the need for external aerials for both satellite and terrestrial TV, “so why do we need building penetration? Why can’t we bring [mobile signals] into the building by putting in an external aerial? Perhaps we need to think about what we’re trying to deliver.”
“Propagation in the house is obviously an issue but an even bigger issue will be interference between different systems,” says Piechocki. “Within buildings you will have multiple systems, you’ll have Wi-Fi that very likely won’t like your IoT system and those won’t be co-ordinated. With the smart home coming along the amount of traffic is going to be big and lots of vendors, lots of systems don’t co-operate well with each other. I think that’s going to be a major issue.”
However, Ganley is confident that these issues will be addressed. “I think smart homes will drive the need for improvements in reliability because the markets for smart home services are potentially so big.
“[Regarding] 5G and IoT, I think there’s a chance to shape where we’re heading by understanding the needs of homes and buildings. Having better data on how they perform should help this thinking,” he adds. He also notes that the construction industry is “very detached at the moment” from the implications of its use of materials on the wireless sector.
Ganley says that given the sheer variety of buildings in the UK, simply understanding their wireless properties would be a big aid to the radio comms industry. “You don’t want to totally over-engineer everything, but equally when you start to deliver critical services in homes you need to make sure they’re reliable.” He adds that there is an opportunity to use propagation modelling in conjunction with measurements to produce a more sophisticated building entry loss model.
The panel was asked for their thoughts on the “wisdom or otherwise” of the government’s decision to migrate the emergency services from the Airwave TETRA network to the 4G Emergency Services Network (ESN), and how this will affect in-building coverage.
“It seems to me like a very unwise move,” says Langley, citing the currently poor 4G coverage in some areas of the country.
However, John Haine, visiting fellow in the University of Bristol’s faculty of engineering, responds: “A transition to 4G might be better simply because the infrastructure is denser and designed for indoor coverage in urban areas, but as you say – out in the sticks it might be a retrograde move.”
“I’m involved in putting 4G into the station environments underground,” says Matthew Griffin, head of digital (commercial development) at TfL. “From my perspective [the ESN] is great because it’s giving me another use case, building on the business case, for putting cellular into the station and tunnel environments.
“But it is not an easy thing to do at all. I think it will be great for the consumer in the long term because it will force greater coverage requirements and mean that we get better 4G coverage indoors,” he adds.
A potential solution to in-building coverage issues is the use of small cell technology. Nick Johnson, CTO and head of PLM at ip.access, a company specialising in small cell deployment, notes that the 3GPP multi- operator core network (MOCN) feature allows small cells to be operator agnostic, enabling excellent spectrum use, low costs and little disruption. However, this creates a headache for MNOs as the infrastructure can be used by their competitors.
“This is very doable, technically feasible, and there are already a few deployments around the world. However, commercially it doesn’t hang together because you suffer a disadvantage from being the first to deploy this sort of thing,” he says. “You end up enabling all of your competitors to come and join in.”
In Johnson’s view this problem can be solved through putting a neutral host service wrapper around it. This approach, which he dubs Super MOCN or SUMO, allows sharing of radio resources in “a very controlled and monetisable way. If you are the spectrum holder you can measure how much throughput you are providing to your competitors’ subscribers and you can charge for it in a very confident way, as you’re measuring the data volumes and the calls that you’re carrying.”
Johnson explained that in the US, it is now possible to deploy LTE “on your own”, in the 3.5 GHz band via the Citizens Broadband Radio Service as long as it has an interface to a spectrum access system (SAS), such as that provided by Google, which allows you to “steer around some of the military incumbents, the coastguard, the US Navy, which already uses it in certain locations...”
The world’s first licensed shared access (LSA) pilot was launched in France on 7 January this year, in which spectrum allocated to the French Ministry of Defence in the 2.3-2.4 GHz band will be shared using Ericsson’s radio access network. This will be supported by additional technology from Qualcomm and RED Technologies.
“Google’s initiative with SAS will make a big difference [in the adoption of small cells]. I think [this approach] is going to come over to Europe with LSA. Once that comes to pass and people are able to freely deploy small cells just like they’re able to deploy Wi-Fi that’s when [exponential growth] will [begin].”
Johnson also believes that the spectrum sharing approach could overtake or avoid the controversy over LTE-U. LTE-U will allow the use of LTE in the unlicensed 5 GHz band currently used by Wi-Fi and is a concern for some parties, particularly given the lack of coexistence requirements in the US.
Another factor Johnson sees as a driver towards greater use of small cells is the growth in demand for data coupled with spectrum being a finite resource. “People seem much more interested in buying more spectrum than in deploying small cells. The catalyst will be when we’ve run out of spectrum.”
Speaking with Land Mobile after the Cambridge Wireless event, Johnson says that ip.access is increasing its focus on a service- led deployment model, in which MNOs pay a monthly figure for an installation that it deploys, manages and monitors, backed with guarantees around the quality of service provision and availability.
“It simplifies the whole model, allows us to deploy faster, more efficiently and more responsively to the operator’s own customers. Quite a lot of them do a similar thing with their macro deployments – they already get a managed service provider to help them run their network – so we’re extending that model down to small cells. That works very well with the neutral host model; when you’re providing the service for one operator you can deploy the same kit in a neutral host model and provide the same service to multiple operators at the same time.”
We’ve seen that considerable work has been done to understand the complex in- building wireless conditions radio engineers have to grapple with. With the expectation of greater demand for reliable connectivity, and the new business models enabled by new technology, the next few years could be full of surprises.
Further reading
For more information on wireless in-building connectivity, and the projects discussed in this article, visit these websites:
• stakeholders.ofcom.org.uk/market-data-research/other/technology-research/2014/buildingmaterials/
