Mr. Römer, you work at the ETH Institute for Pervasive Computing in the "Distributed Systems" team and do research and development in the area of wireless networks (1). Can you tell us what precisely such a network is about?

Kay Römer: Wireless sensor networks enable detailed and wide-ranging observations to be made of phenomena in the real world, by distributing a multitude of so-called sensor nodes in a certain area. These sensor nodes are miniature computers that can detect and record various parameters, like temperature or the presence of a living creature. They process the data and, then relay the results to neighbouring sensor nodes via wireless technology. When these sensor nodes are linked to a base station, such as a laptop, the sensor networks can be integrated into the internet.

Moore's law states thatdata density doubles approximately every 18 months. This means that we're moving here in a highly dynamic area. Where does science stand at the moment in the development of wireless sensor networks?

Kay Römer: We can already build prototypes of the hardware. But we're still faced with practical difficulties. The sensor nodes run on batteries and need to be able to run for months or years without anyone having to do anything. This poses a huge challenge to the hardware with regard to energy efficiency and durability. What's more, the networks have to be able to work in an outdoor environment, and this calls for a certain level of robustness in the sensor nodes. As far as these characteristics are concerned, efforts to improve the mini-computers are on-going.

What are the challenges concerning the software?

Kay Römer: Here, we are mainly faced with the question of programming wireless sensor networks which still calls for comprehensive expertise and lots of experience. We have to find new approaches to make the sensor networks serviceable in practice. One problem in this regard comes from the very limited processing and memory capacity of these tiny computers. In order to make any progress here the sensor networks have to be put into a context of concrete application. We are working on this within the framework of NCCR-MICS, a project from the Swiss National Science Foundation, with representatives of various areas of expertise to test their usability (2). We're now approaching the practical application of sensor networks.

What would a practical use of this technology be?

Kay Römer: Scientific observation of the environment is a classical application of wireless sensor networks. For example, small sensor nodes can be placed at the openings to bird's breeding holes in order to gather information on their breeding behaviour without disturbing them. And sensor nodes can also help in the analysis of ground motion, volcanic eruptions or avalanches. We can also envisage their use in the automation of buildings. Sensor knodes installed without impairing the building substance can determine and optimise the total energy balance. Sensor knodes can also be attached to water pipes to ascertain the quantity and temperature of the water flowing through them. Further down the line, applications in health care are conceivable. Elderly people would be able to stay longer in their familiar surroundings if sensor networks were used to recognise and relay information regarding any deviation in their normal living patterns to health care specialists. At present, though, health care applications are not very realistic.

Scientists from all over the world met at ETH Zurich to exchange information on wireless sensor networks.

Why? Because the sensor nodes are still too big and would limit the living options of the persons involved or because wireless sensor networks would infringe their privacy rights?

Kay Römer: Both. Research is working on the development of ever-smaller sensor nodes. But the scientist is also involved in protecting privacy. The danger here is very differentiated. Environmental observation, for instance, often poses no problem, but in the area of medicine we must recognise that confidential information could fall quickly into the wrong hands. Here we are working on mechanisms that control the transmission of data or debating on how to introduce fuzziness deliberately in order to protect privacy.

What was the main focus of the three-day European Workshop on Wireless Sensor Networks (EWSN 06) that took place last week at ETH Zurich?

Kay Römer: The EWSN 06 was a forum for an international exchange among scientists on the subject of sensor networks. The conference attracted a lot of attention, even beyond Europe. The contributions in the workshops came in roughly equal parts from the USA, Asia and Europe. Some of the questions on the agenda dealt with data management, energy efficiency and possibilities for application. Apart from the pure scientific exchange, participants also benefited from an overview of cutting-edge research underway in Switzerland and other European countries.

Where do you personally think this technology will be used in ten year's time?

Kay Römer: By then some of the uses of sensor networks will be part of our everyday lives. While today they cause a lot of excitement–at least among scientists–they will be an almost invisible part of our surroundings. This will mean that in ten year's time we might be a lot nearer to reaching an important goal, namely the discreet integration of computers into our everyday lives; machines that function without our constant attention and care and without distracting us from our real duties(4).