Distributed networks linking PCs, laptops, cell phones, sensors and actuators will form the backbone of future data, communication, and control networks. Applications will range from sensor networks to precision agriculture, environment monitoring, disaster relief management, smart spaces, target localization, as well as medical applications. In all these cases, the distribution of the nodes in the field yields spatial diversity, which should be exploited alongside the temporal dimension in order to enhance the robustness of the processing tasks and improve the probability of signal and event detection. Collaborative signal processing has been advocated as a way to achieve the efficient fusion of information. Regardless of the cooperative technique adopted, it is an accepted fact that distributed processing needs to be adaptive. This is because not only the environmental conditions vary with time and space, but the network topology may vary as well.
Most available distributed techniques tend to be iterative in nature as opposed to adaptive. An adaptive network should react to spatial and temporal data in an instantaneous manner through local collaborations, and the information should flow through the network in real-time. In other words, an adaptive network should behave as an adaptive entity in its own right. The property of adaptation is fundamental in order to (1) endow the network with real-time learning abilities, (2) implement robust schemes to spatio-temporal variations, and (3) limit local processing and communications.
In designing adaptive networks, there are at least two main issues to consider. One issue relates to the topology of the interacting nodes and the other issue relates to the processing and communications constraints imposed on the nodes. The talk will illustrate designs that apply to three major ways of node collaboration (incremental, diffusion, and probabilistic diffusion).