Optimized Routing of Intelligent,Mobile Sensors for Dynamic, Data-Driven Sampling

The report describes a Dynamic Data-Driven Application Systems (DDDAS) project in which multiple mobile sensors are routed via a data-driven sampling scheme. The long-term goal of this project is to provide a control-theoretic framework to enable intelligent, mobile systems to optimally collect sensor-based observations that yield accurate estimates of unknown processes such as aircraft formation flight and environmental monitoring. The basic research objective is to apply tools from aerospace engineering, specifically nonlinear estimation and control, to design coordinated sampling trajectories that yield the most informative measurements of estimated dynamical and stochastic systems. The technical approach to achieve this objective is to construct a framework for dynamic, data-driven sampling algorithms that (1) maximize the observability of a nonlinear dynamical system subject to time-varying perturbations; and (2) minimize the uncertainty in the estimate of a nonstationary random process that requires nonuniform sampling. The approach incorporates complementary representations of an unknown process: the first uses a deterministic, model-based parametrization, whereas the second uses a lowdimensional statistical description; both approaches apply and enable the DDDAS concept in which measurement data is used to update the model description and the updated model is used to guide subsequent measurements.