Wireless sensor networks typically consist of a large number of sensors transmitting their measurements to some central fusion point where these data are processed. In the case of simple functions, like the arithmetic mean, it seems rather inefficient for all sensors to transmit their data separately. Instead, some form of “online function computation” should be realized which performs the desired computation in the process of transmission. In the context of the LOEWE project COCOON, the “Communication Systems Group” is investigating efficient algorithms for “online function computation” in wireless sensor networks.
In the context of online function computation in wireless sensor networks the wireless channel offers attractive characteristics in the form of its broadcast and superposition nature so that simple functions like addition, arithmetic mean or polynomials could be realized. On the other hand, especially the multipath and pathloss characteristics lead to a superposition of the transmit signals that is distorted in phase and amplitude. These problems have to be eliminated by proper algorithms. In the design of such algorithm emphasis has to be put to complexity as well as low and constant energy consumption, to minimize the sensor costs and to maximize the sensor lifetime.
Algorithms for the efficient realization of “online function computation” techniques
Energy efficiency, constant energy consumption among the sensors, complexity