“Multi-vehicle estimation and control for the optimization of spatiotemporal sampling”
By: Dr. Derek Paley
Associate Professor, Department of Aerospace Engineering and Institute for Systems Research
University of Maryland
This talk will present recent research in the area of multi-vehicle control and optimization for sampling spatiotemporal processes in the atmosphere and ocean. Sampling performance can be optimized by path-planning algorithms that drive vehicles to specific regions of the operational domain containing the most informative data. I will describe two planning approaches: one using nonlinear observability to optimize sampling trajectories and the other using optimal interpolation of nonstationary processes. In the first approach, coordinated trajectories for sampling a parametrized flowfield are optimized using the empirical observability gramian. Parameters of the flowfield are reconstructed from noisy flow measurements collected along the sampling trajectories using a recursive Bayesian filter. The second approach is designed for sampling nonstationary fields, in which the spatial and temporal statistics may vary in space and time. A coordinate transformation is designed under which uniform sampling is optimal because the unknown field is stationary in the new coordinates. Both approaches use tools from nonlinear control, specifically Lyapunov-based control, to design decentralized algorithms for stabilization of multi-vehicle sampling formations. Time permitting, applications including hurricane forecasting and underwater vehicle navigation will be discussed.