Wireless sensor networks (wsLAN)
They will warn us about landslides, monitor traffic and watch our health. Wireless sensor networks will be an important aid in the future, and can provide a new and better understanding of the world around us.
 |
Photo: U.S. Department of Interior, U.S. Geological Survey |
By using sensors we can monitor anything from blood pressure and body temperature to volcanic activity and avalanche risks. Often we would like to collect this data at a centralised location, where someone can make sure everything is all right. While this data collection traditionally has been accomplished with wires, we now see that wireless communication has become more popular. This is mainly due to the fact that wireless sensor systems are easier to install and modify than their wired counterparts.
When removing the sensors’ wires, one simultaneously removes the most obvious way of providing the sensors with power. Thus, most wireless sensors are battery powered. While battery life is a minor issue if you have a few smoke detectors in your house, the situation is entirely different in a sensor network consisting of several thousand units. In this situation the economic gain of employing power-saving networking techniques can be substantial.
The objective of the wsLAN project was to study how wireless sensor networks can be made efficient with regard to both power and response times. One relevant issue is that many sensors regularly transmit information in order to report that they are still in operation. Here, energy may be saved by letting one sensor transmit the status of a larger group of sensors. Subsequent reports can be transmitted by another group member, thus avoiding that one sensor runs out of battery before the others. Another question is how often and how detailed the sensors shall report their observations, and whether the sensors themselves shall decide whether to report or not. The design and implementation of a sensor network should be matched to the spatial and temporal correlation characteristics of the process to be monitored, both in terms of topology and node density, and sampling instants.
In order to illustrate some of the challenges related to wireless sensor networks, the wsLAN project has developed a demonstration network consisting of various types of sensor. One aspect demonstrated here is how some sensors, based on their individual observations, can control the sampling rate of other sensors in the network.
The wsLAN (Wireless Sensors Local Area Network) project was running from 2003 to 2006 and funded by the Norwegian Research Council. The project consortium consisted of SINTEF, Telenor, Memscap and Nacre.
Contact: Knut Grythe (Beskyttet adresse)
