New wave bands

The demand for communicating by wireless networks is growing fast, and device manufacturers are scrabbling to stay ahead. The evidence is in every computer shop, where 'Pre-N' devices for wireless LANs are on sale even though the Institute of Electrical and Electronics Engineers is not expected to approve the generic 802.11n standard until next July at the earliest.

Manufacturers such as Belkin and Netgear are jumping the gun with their Pre-N routers and kits because of the products' prospects for delivering data at very high rates. While the fully-ratified 802.11g standard has a headline rate of 54 megabits/sec across a range of up to 30m, 11n promises 540 megabits/sec across 50m. This is thanks to developments with Multiple-input, Multiple-output (MIMO) technology.

There are several antennas at the base station and the subscriber. They provide smart ways of transmitting more data and the technique holds so much promise that it has been included in the standards for 3G mobile phone networks.

As reported in (The Engineer, 17 January), the radio signals reflect off objects, creating multiple paths that in conventional radios cause interference and fading. MIMO technology uses these paths to carry more information, which is recombined on the receiving side by the MIMO algorithms. It has been estimated that the technology will be able to at least quadruple the data rates of existing wireless LANs.

The potential benefits are a significant reduction in the transmit power required for each packet, improved network coverage and increased data capacity of the network, resulting in an overall improvement in spectral efficiency. All this can be achieved by exploiting the diversity of the channel.

In MIMO transmission schemes, data is appropriately encoded on to each transmit antenna to exploit spatial diversity in the channel. A suitable algorithm is then applied after the receive array to decode the multiple observations and recover the original data stream. It sounds simple enough but is a complex problem.

Now researchers sponsored by Motorola and Toshiba Research Europe are to investigate the possibility of using the technique for ultra-wideband (UWB) communication systems such as personal area and home networks and for larger scale WiMAX networks. The higher data rates that could come from MIMO would enable wireless transmission for applications as data-heavy as high-definition TV.

The work is being carried out at Bristol University's Centre for Communications Research, where Dr Chris Williams and colleagues have previously assessed MIMO technology for the UK's communications regulator, Ofcom.

'Bristol has been very strong in terms of its practical equipment for characterising MIMO channels,' said Williams. 'We also have the ability to do field trials, partly because of our university's location on a hill near the centre of the city. We can build life-like devices, such as a PDA with MIMO antennas, and take them into the town centre to characterise their performance.'

The angle of signal departure, angle of arrival and its strength are the key characteristics that have to be understood so that the best designs of antennas can be implemented. Typically the department's researchers work with arrays of 4x4 antennas although this can go up to 16x16.

The first challenge for using MIMO for personal area networks and WiMAX is to get a full understanding of the channel and the scattering.

'That's very important because the channel might not support MIMO at all,' said Williams. 'With the technology you are transmitting a number of data streams in parallel at the same time and the receiver must use its knowledge of the channel to work out what that data was. MIMO communications requires sufficient scattering. If you don't have that then it won't work.'

Therefore terminals will choose between a number of modes to provide the best data rate the channel can support.

The second step will be to assess how accurately the receiving antenna can estimate the channel. 'If you find that the channel can support MIMO then you have to do the channel estimate and do it sufficiently accurately otherwise you get a lot of errors,' said Williams.

'This is complicated for mobile devices such as PDAs because one end of the link is moving and you have to be able to track it.'

The two industrial sponsors have agreed with the EPSRC to fund the MIMO research studentships at Bristol for three years.