Adaptive wireless networks as an example of declarative fractionated systems

Abstract

Adaptive wireless networks can morph their topology and support information gathering and delivery activities to follow high-level goals that capture user interests. Using a case study of an adaptive network consisting of smart phones, robots, and UAVs, this paper extends a declarative approach to networked cyber-physical systems to incorporate quantitative aspects. This is done by distinguishing two levels of control. The temporal evolution of the macroscopic system state is controlled using a logical framework developed in earlier work while the microscopic state is controlled by an optimization algorithm or heuristic. This two-level declarative approach is built on top of a partially-ordered knowledge sharing model for loosely coupled distributed computing and is an example of a so-called fractionated system that can operate with any number of wireless nodes and quickly adapt to changes. Feasibility of the approach is demonstrated simulation and in a hybrid cyber-physical testbed consisting of robots, quadcopters, and Android devices.