Adaptive Sampling

Being able to sense and localize abnormalities in an aquatic environment efficiently and quickly is crucial for both the environmental scientists and government authorities to study the problem and take remediation measures. Conventional, statically-deployed buoy stations can only provide a sparse, low-resolution snap shoots of the environmental parameters of interest within a predefined area. Data interpolation between the stations, and extrapolation outside the area introduce estimation errors at best, and for the worst case, completely miss the environmental abnormality. In collaboration with Texas AM (D. Shell) and ASU (S. Saripalli), we developed an adaptive sampling strategy using receding-horizon cross-entropy trajectory optimization on a Gaussian Process (GP). Specifically, we employ Upper-Confidence-Bound (UCB) search method to adaptively plan the vehicle’s path that exhibit an exploitation-exploration trade-off. Path planning at the initial stage is focus on exploring and learning a model of the environment, and later, to exploit the learned model to plan paths that increase sampling frequency around regions of “high interest”.

 Upper-confidence bound (UCB) stochastic trajectory optimization for extrema search on a Gaussian Process environmental model with single hotspot.

Upper-confidence bound (UCB) stochastic trajectory optimization for extrema search on a Gaussian Process environmental model with multiple hotspots.


Standard lawnmower sampling strategy (a) and measured values (b), in comparison to adaptive sampling (c) which focuses resources into informative regions of interest (ROI) with high concentration values (d).

Field Experiment with Autonomous Surface Vehicle (USV)

We have developed a static 2-d version of the modeling algorithm, integrated and deployed it on an unmanned surface vehicle (USV). The system was validated by mapping adaptively a coastal region and demonstrating that the vehicle can quickly explore the environment and subsequently focus on the more “interesting” parts of the field being estimated. We performed field validation using a USV over a small area with interesting topography. Bathymetry measurements were used for representing the field and the goal was to quickly discover the deepest part of the area of interest, as a surrogate for e.g. the extremal concentration of a pollutant in the environment.


Adaptive sampling path (a) computed and executed on-board the JHU USV (b) using bathymetry as a surrogate for concentration measurements.

JHU USV in performing adaptive sampling at the Loch Raven reservoir, MD.