Cyber Systems: Adaptive Distributed Systems Based on Cooperative and Combination Strategies

University Of California-Los Angeles, Los Angeles CA

Investigators

Abstract

Adaptive Distributed Systems Based on Cooperative and Combination Strategies Scope and Intellectual Merit. The objective of this research is to develop processing algorithms that endow distributed networks with real-time adaptation and learning abilities. The distributed algorithms should be able to exploit the spatial diversity of the distributed nodes, as well as the temporal dimension of the data, in order to deliver improved levels of performance through cooperation between neighboring nodes. The approach is based on exploiting incremental and combination strategies and allows for real-time flow of information across the network. The purpose is to develop adaptive distributed strategies that require limited local communications and computation and that are robust to topology changes. Broader Benefits. The driving force for such adaptive networks is the realization that distributed systems linking PCs, laptops, cell phones, sensors and actuators will form the backbone of future data and communication networks. Applications will include sensor networks, precision agriculture, environment monitoring, target localization, disaster relief management, and medical examinations. The purpose of distributed processing is to combine the data captured locally at the distributed nodes in order to deliver good performance globally. The processing needs to be done adaptively and in real-time because not only the environmental conditions vary with time and space, but the network topology may also vary. The proposed research shows how to combine tools from across disciplines in order to develop adaptive distributed networks. Progress in this work will impact the information technology infrastructure and will contribute to the training of students in an area of fundamental relevance to future communications and networking systems.