Dynamic stochastic modeling for adaptive sampling of environmental variables using an AUV

Abstract Discharge of mine tailings significantly impacts the ecological status sea. Methods to efficiently monitor extent dispersion is essential protect sensitive areas. By combining underwater robotic sampling with ocean models, we can choose informative sites and adaptively change robot’s path based on in situ measurements optimally map distribution near a seafill. This paper creates stochastic spatio-temporal proxy model dispersal dynamics using training data from complex numerical models. The consists Gaussian process an advection–diffusion partial differential equation. Informative are chosen predictions objective function favoring areas high uncertainty expected concentrations. A simulation study real-life experiments presented.