Interview - Sensor networks will change our life
Answers by Karl Aberer and Holger Karl on the challenges and opportunities of wireless sensor networks
Wireless Sensor Networks are currently one of the most promising technologies in ICT. Their potential to change everyday life appears to be high. Eurescom mess@ge asked two leading European experts, Professor Dr. Karl Aberer and Professor Dr. Holger Karl, what the challenges and opportunities of wireless sensor networks (WSN) are.
Professor Aberer from the School for Computer and Communication Science at the Ecole Polytechnique Fédérale de Lausanne, Switzerland, is director of the Swiss National Centre for Mobile Information and Communication Systems. He is supervising the project Global Sensor Networks, which aims to develop a generic platform for deploying sensor networks.
Professor Karl from the department of computer science of the University of Paderborn, Germany, has co-authored a book on “Protocols and Architectures for Wireless Sensor Systems” and chairs the Wireless Worlds Research Forum's working group 3 on "Cooperating and ad-hoc networks".
How will sensor networks change our daily life?
Aberer: In short, wireless sensor networks will allow physical reality to directly interact with the Internet, extending it to a Sensor Internet, with many unpredictable consequences for our daily lives. Similarly, as at the initial stages of the Internet we could not foresee the impact of e-mail, Web, peer-to-peer technologies, and Wikis on our daily lives, it is hard to predict today the future impact of wireless sensor networks. There are obvious applications, like fine-grained observations of the environment, allowing us to improve water usage, provide faster warnings from disasters or perform more efficient agriculture, all with important economic implications. Similarly, applications in surveillance, healthcare or habitat and structural monitoring will help to increase our daily well-being. But all these applications are still quite evolutionary in comparison to what we can already do today. Where we may expect quite disruptive and unpredictable developments is when we start to take advantage of the possibility to automatically control and coordinate sensors and actuators and also automate the subsequent information processing tasks. Also, from a more human perspective, the possibility to quite directly share physical experiences might lead to fairly exciting new community-related developments. But these are next to impossible to predict.
Karl: Sensor networks will change both private and professional life by enabling new types of IT applications. Popular examples are applications in home and office buildings, in production and automation, in logistics and ecology, where connected sensors can provide data previously not available. The degree of automation will increase, but so will efficiency and safety in many areas.
In which areas will wireless sensor networks be applied first on a larger scale?
Aberer: Car technology, logistics and healthcare appear to be among the most promising areas from the industrial perspective. In the research community, environmental monitoring is an area in which wireless sensor networks are being deployed at growing scales. For example, within the Swiss research centre for mobile information and communication systems, we are planning to deploy several thousands of sensors in a Swiss alpine region within a year as a testbed for a Swiss-wide data acquisition platform for environmental sciences.
Karl: Large-scale applications are currently appearing in environmental control applications of various kinds; for example, avalanche or landslide prediction. But this is restricted to certain niche markets. On an even larger scale, automation applications – for example, for factory floors or preventive maintenance - seem to be a likely candidate as a pilot application area. But for this to happen, some reliability challenges still must be mastered.
What are the technical challenges in the development of wireless sensor networks?
Aberer: I would roughly classify the technical challenges into three categories. At a first stage developing techniques for addressing the resource constraints of wireless sensor nodes in terms of processing speed, storage capacity and communication bandwidth was considered as a key challenge. Substantial progress has been made there and will continue to be made. As the devices are now available and larger wireless sensor network deployments are becoming a reality, at a second stage, the deployment and maintenance of such networks are increasingly recognized as a major technical challenge. Controlling large-scale and often self-organizing systems is here the inherently hard task. At a third stage I expect with a growing number of deployments that the management, interpretation and analysis of the enormous data streams generated by large numbers of sensors and sensor networks will be a big challenge in the future.
Karl: Clearly, some traditional issues like energy supply and energy efficiency, dependability, but also easy programmability by non-experts or easy and reliable deployment have to be worked out. Other issues, like, for example, miniaturization, are only relevant for some areas but can then indeed be a show-stopper.
How rapidly will the cost of wireless sensor nodes decline in the next years?
Aberer: Several species of wireless sensor nodes have recently become commercially available at 50 to 100 dollar each. Through re-engineering, Moore's Law and volume production, wireless sensor nodes are expected to drop in price to less than 5 dollar each over the next five years.
Karl: This is very hard to predict. Since these nodes are not easily commoditized but have to be customized for different applications, the price trend will strongly depend on the dynamics in the different applications. It is a bit of a chicken-and-egg problem.
When do you expect wireless sensor networks to be a reality on the mass-market?
Aberer: More research and standardization are expected to be needed to solve all existing problems, at least for the next 3-5 years. Deployment in large quantities - millions and beyond - are not to be expected before 6-8 years. Given the large number of possible applications, the diversity of wireless sensor types is expected to be important.
Karl: Within five years, such networks will be deployed for specific applications. Within ten years, I would expect a WSN to be an entirely unexceptional technical means.
The interviews were conducted by Milon Gupta.
Please send us your comments on this article.