New technologies bring with them new challenges in testing. The use of multiple antennas at both ends of a radio link (MIMO or Multiple-Input Multiple-Output) is one of the fundamental building blocks in LTE to increase data throughput and cell capacity. Even in today’s HSDPA systems, simulations have shown that use of two antenna diversity in terminals could increase the network capacity by almost 50%. A smaller additional gain then comes from adding a second transmit antenna at the base station. It is clearly in the interest of a network operator to encourage the use of terminals that will maximise the capacity of the network and improve the experience of the end users. But how can we tell a good product from a bad product? Because of the difficulty of testing over-the-air performance, radio systems are traditionally tested using “conducted tests”. In such a test, the antenna in the terminal is bypassed and the internal electronics is connected directly to test equipment. While such a test is important, it only tells part of the story. It is also increasingly difficult to do, as modern phones seldom have external antenna connectors (mainly because hands free car kits have moved over to using Bluetooth and no longer need a physical connection to the phone). Radiated testing is generally harder to do than a conducted test, because of the need to recreate a controlled radio environment.A story from the early days of GSM illustrates both the importance and difficulty of radiated testing. In order to compare the performance of a range of new GSM phones, we set up a special base station designed to be used only by our own phones. We set up attenuators in both transmit and receive paths at the base stations. We then walked up and down the same street over and over again adjusting the attenuators and recording a subjective measure of the speech quality of each phone. This is not something you could do as a regular measure - we got funny looks from the local residents and a visit from the boys in blue to ask what we were doing. However, the result was very revealing. The worst phone at the time had receiver sensitivity 10dB (=10 times) worse than the best. To achieve the same degree of coverage for such a phone, you would have to build four times as many base stations as if everyone was using the best. That’s a lot of network investment you can avoid by controlling the quality of the terminal.

There is an even bigger testing challenge in the case of MIMO systems. The gain achieved by using multiple antennas is fundamentally reliant on multipath propagation, where the signal arriving at the receiver not only in a direct line from the transmitter but also via delayed reflections from surrounding terrain and buildings, at distances of up to a few kilometres. In the past, most radio systems have treated multipath as an enemy, as it causes signals to fade and drop out. With MIMO we can finally exploit the multiple paths to carry more data. The challenge in testing of MIMO systems is thus to reproduce a realistic and controlled multipath radio channel in a small space and via the built-in antenna of the device being tested.

A year ago, there had been little focus on how MIMO systems would be tested. We started talking to a few interested parties to encourage people to think about how this might be achieved. Many of the technical experts in this field meet up through the European group COST2100. Within the LTE standards process, the responsibility for testing lies with 3GPP RAN4. These two groups are now cooperating to develop an Over-The-Air (OTA) test for HSPA and LTE devices. In parallel, US-based industry group CTIA is also gathering its technical experts in this field and has set up two sub-working groups to study this topic. From a starting point where nobody knew how to test MIMO terminals, we have now reached a situation where there are several different techniques on the table. There is now a good likelihood of agreement on a standardised test in the near future. This will make it easier for us to ensure that our customers have access to good quality products which will also maximise the capacity of our networks.

The yearly flagship academic conference on wireless communications IEEE ICC took place in June 2009 in Dresden, Germany. With 2000 people it was well attended, including CEOs and CTOs from key operators and equipment vendors. The programme of the conference was organized by Professor Gerhard Fettweis, who holds a university chair in Dresden sponsored by Vodafone (for more than 15 years now).
What was special about this year’s ICC was that we not only had academic papers on specialist technical topics, but also a large number of exhibitions and demonstrations. I think it is very important to test or trial innovations in realistic environments before sinking money in new technologies or even worse before unrealistic expectations are raised with end-customers. At the CTO panel discussion, Vodafone’s Steve Pusey made the point that the whole industry made the experience that “early 3G over-promised and under-delivered” - something we have to avoid with 4G.

At ICC we could see pre-commercial LTE with some applications like gaming running over it from one vendor. There were also demonstrations of technologies which are not even standardized yet - mainly within the EASY-C research project cluster. Some people have the opinion that cooperative MIMO (cooperative transmission between several base stations or cooperative multipoint / COMP) is an academic phantasm - but at ICC in Dresden it was shown first time that it works in a real network layout for both the uplink and the downlink. COMP was shown between two sites (provided by Vodafone and T-Mobile) and two test terminals with a LTE-like hardware - but the plan is to extend the field trial to many more sites and terminals. The anticipated benefits of COMP would be significantly increased user experience at the cell-edge - i.e. providing a more homogeneous user experience. The drawback is more backhaul investments - but they have to be done anyway if operators go to HSPA+/LTE.
Apart from the live testbed in Dresden, which is run by the Vodafone Chair of TU Dresden; Fraunhofer HHI demoed also their COMP testbed which is based in Berlin. Another COMP demo was shown by Alcatel-Lucent, together with a very neat LTE antenna panel from Kathrein (multi-column x-pol) which communicated to the test terminal on a Rickshaw.

Another noteworthy demo in the LTE / LTE-Advanced space was LTE relay shown by NSN - it is quite impressive to see the relaying working for an FDD technology. This could eventually enable very cost-attractive coverage hole fill-in solutions with self-backhauling micro-base stations. Coverage is as ever before one of the biggest challenges for mobile operators. Last week’s OFCOM comments and the media noise around pointed to coverage gaps in all UK networks. The real issue is to manage the costs for providing coverage - and Relays might help to achieve this.

As with COMP, standardization of Relaying is in early days within the LTE-Advanced study item. The good thing is that we have already lab and field trials going on in order to gain experience, drive standardization and to discover technical issues before substantial commercial activities.

Early in the evening of the 21st May some 160 members of the Berkshire Branch of the IET gathered in the pavilion at Vodafone’s headquarters in Newbury for the annual Vodafone Lecture. As usual the audience spanned all ages, from youths to pensioners, as well as many different engineering backgrounds. And as usual the theme of the lecture was the outlook for the technology of mobile communications. The lectures have been running for nearly ten years, and have consistently attracted a pretty large audience - well large for evening lectures on technology - with many coming year on year. In part the attendance may be attributable to the excellent tea that we always provide, but occasionally the lecture also includes something special. The 2001 lecture was notable in that respect since it was given just after we had paid £6 billion for our 3G license - an eye watering sum of money at the time (but trivial in comparison to the sums our Government is prepared to spend today on floundering businesses). The lecture was particularly memorable to me since I had designed our bid strategy and led our bidding team, so the auction figured large in it. This year’s lecture coincided with what for me is another seminal moment in the history of mobile communications. I was able to show the audience a live demonstration of LTE - and I believe this was probably the first time LTE has been shown live in public.

Coverage for the demonstration was provided across the Vodafone campus by two splendid Alcatel-Lucent base stations operating in the 2.6 GHZ band, connected through the Vodafone UK network, and delivering breathtaking bit rates to LG modems. I hope the audience was impressed by the performance, I know I was - and I had experienced it several times beforehand. Of course the demonstration wasn’t set up especially for my lecture, the lecture just happened to coincide with a friendly user trials of the technology on our campus, the friendly users being Vodafone staff - from executives to staff. The trials represent the culmination of a series of intensive LTE technology trials across Europe and the US, trials that pushed LTE to its limits, and proved that it lives up to expectation. But we will return to that later. In the meantime, you can see the the slides from my IET presentation. - If you want to see the “movie”, then go to IET.tv, where it will be posted shortly.