Minimizing sum distortion for static and mobile fusion center placement in underwater sensor networks

The problem of optimizing the position of a mobile fusion center in a sensor network is considered. The optimization criterion of interest is the sum distortion for the communication of all the information from each of the nodes to the fusion center. Transmission losses along underwater links are modeled and a time-multiplexing architecture imposed on the sensor nodes for communication with the fusion center. Classical results from the information theory literature are leveraged and the optimization problem is formulated and solved analytically for the case when the fusion center is at one fixed location and numerically for the case when the fusion center is mobile. It is observed that in the low node power regime, with the fusion center at a fixed location, the fusion center selectively communicates with a few nodes while turning the others off. In the high power regime, the time allocated to the nodes is a function of the information they need to transmit to the destination and the distance from the node to the fusion center. Further, when the fusion center is mobile, in the low power regime, the fusion center is placed close to the node with the largest information content while for higher powers the difference from a fixed fusion center declines.