The roots that feed electronics

The roots that feed electronicsThe Department of Electrical and Electronic Engineering at Bristol University is recognised as one of the academic world’s leading communications and medical-electronics laboratories. Ideas that began in the lab 15 years ago and are only now appearing in real-world applications. Peter Mitchell reports. Electronics moves so fast that one sometimes feels today’s technology has just popped out of nowhere – or at least that the manufacturer thought it all up only a couple of years ago. That is rarely true. The roots usually go back much further, into academic research. For one of today’s hottest sectors – mobile communications – quite a few of the roots can be traced to Professor Joe McGeehan’s Department of Electrical and Electronic Engineering at Bristol University. Many ideas that emerged in his lab more than 15 years ago are now appearing in the second and third generations of cellular radio equipment, especially in North America and Japan. And more are on the way. Muscle control signals in multiple sclerosis or stroke patients are often too weak to hold the foot up whilst walking. This results in the toe dragging inconveniently on the floor. The University of Bristol is working on electronic ways to minimise ‘foot-drop’. The department has specialised in communications for many years – especially cellular and mobile radio. “We pioneered a method called linear modulation which is now sweeping the world in terms of spectrum-efficient communications for cellular and portable telephone systems,” says McGeehan. “We helped develop a form of linear modulation called TTIB (transparent tone in band), and helped establish a company Linear Modulation Technology, now part of Intek in the US. Now LMT is making equipment for a 2MHz radio slot operating nationwide across the States.” Bristol is recognised as one of the academic world’s leading communications laboratories. “A lot of what we do is fundamental, but what’s important to us is that the work we do goes, five ten or fifteen years later, into real products. Parts of our work are now informing many of the world’s cellular standards – for example sub-system components like basestations.” One idea that McGeehan claims he was promoting 12 years ago, but only now has become a hot topic, is the ‘software radio’ – a mobile radio that can be configured through a built-in DSP chip, with all the modulation, coding and filtering, in downloadable software. The goal is a universal piece of hardware, adaptable to either GSM, FM/AM/SSB, linear modulation, or any other standards, which could replace today’s “crazy idea of needing multiple kinds of handset for use with different wireless networks”, he says. “With a software radio, when you go to Japan, you connect to the radio network, download some software, and you have a mobile phone appropriate for communicating in Japan. We are now building and demonstrating these multi-mode radios.” But there are still problems, because the concept requires very broadband RF architectures, able to exactly replicate the incoming spectrum at baseband. “We are looking at developing these broadband front ends – bandwidths in excess of a gigahertz – so the front end can handle, say, GSM at 900MHz and PCS at 1.8GHz,” adds McGeehan. Wireless telephony bit rates have to date been relatively low. Bristol’s work in an EC-funded project, Laura, led to a 23.5Mbit/s core network architecture standardised as HiPerLAN in Europe. “Now we are working with Alcatel and NTT, on a prototype called AWACS, to get transmission at 17GHz,” says McGeehan. “That’s pushing the data rates to 70Mbit/s and on to 155Mbit/s – very high bit rates.” Another of McGeehan’s babies – intelligent ‘adaptive’ antennas that can modify their behaviour according to information about local obstructions like trees and buildings – is now a darling of the mobile telecommunications industry, he says. His group has designed a suite of linearised power amplifiers and high dynamic range receivers for the European Commission-funded TSUNAMI project, proving that adaptive antennas were feasible for third-generation mobile telephones. It has since developed simulation models for Orange PCS, to evaluate adaptive antennas in DCS1800 based networks. The department’s practical expertise is not confined to mobile radio. Its applied electromagnetics team played a significant part in the Ministry of Agriculture, Fisheries and Food’s investigation into the safety of microwave food heating a few years ago. More recently it became involved in a contract from the defence industry to develop EM detection methods for plastic mines. And the GEC Marconi-made speech scrambler now used by UK police forces was developed at Bristol too. DSP has also moved out into the medical realm. “Many of the signal processing techniques developed for telecommunications can be applied in some form or another in the medical field. We have established a research centre between the engineering faculty and the school of medicine. That, combined with the University teaching hospital, is a very powerful combination,” says McGeehan. One medical electronics team is developing signal processing algorithms for hearing aids that will counter the ‘cocktail party’ effect – in which partially deaf people find it hard to pick out speech against a constant low-level hum of background conversation. Another is working on ways of compressing medical images so they can be transmitted faster, more accurately, and at less cost. A particularly exciting project is investigating functional electrical stimulation to restore limb control to people with neuromuscular diseases – by amplifying their own nerve impulses. The idea is to build arrays of sensors that measure the motor ‘thoughts’ (electromyograms) that the patient is trying to send to their muscles (in multiple sclerosis or stroke patients these EMGs are often too weak to cause motion, resulting in symptoms like ‘foot drop’). The output of these sensors can then be combined with information from positional feedback switches in the foot, and delivered to neuromuscular stimulators to amplify the impulses and so make the muscles produce the desired movement (see diagram). The problem is complex because the act of measuring the nerve impulses introduces artifacts and muscle movements that have to be filtered out before the the true ‘desire to move’ can be identified. But 16-element integrated arrays have already been tested and the department is working with larger 60-element sensors and new algorithms to make the movement actually produced match more closely the one intended.
“Academic life is very pressured now,” says McGeehan. “But if we can solve some of these medical projects, it will bring a lot of quality to people’s lives. I believe passionately that we have to marry fundamental work to applications. That is what we regard as success.”

Leave a Reply

Your email address will not be published. Required fields are marked *