Signal Processing Algorithms to Decipher Brain Function

these capabilities is limited severely by the scarcity of the two principal resources in wireless networks, namely energy and bandwidth. Much of the capacity growth of the past two decades has been enabled by major advances in the wireless physical layer. However, in recent years, attention has turned increasingly to the higher network layers to examine interactions among nodes that can lead to even greater efficiencies in the use of wireless resources. This talk examines two types of such interactions: competition among nodes in multiple-access communication networks, and collaboration among nodes in wireless sensor networks. In the first context, the network is viewed as an economic system, in which terminals behave as agents competing for radio resources to optimize the energy efficiency with which they transmit messages. A game theoretic formalism is used to analyze the effects of various design choices and quality-of-service constraints on energy efficiency. In the second context, collaborative techniques for optimizing the use of radio resources in sensor networks are considered. Here, the focus is primarily on distributed inference, in which distinctive features of wireless sensor networks can be exploited through collaboration among nodes to effect a tradeoff between inferential accuracy and energy consumption.