摘要：AbstractAdaptive sampling and situational awareness for autonomous underwater vehicles (AUVs) is a major improvement in ocean research. By only sampling the feature of interest in a feature-relevant domain instead of a covering a whole area expensive ship time can be saved and at the same time a more comprehensive data set can be obtained. A classical marine example where adaptive sampling is useful is sampling of boundary layers such as the thermocline because the boundary layer thickness is very small compared to the depth of the water column. These boundary layers play an important role in many ocean related disciplines such as marine biology, physical oceanography and underwater acoustics. In this paper an unscented Kalman filter (UKF) based extremum seeking control (ESC) approach is presented to detect and track such boundary layers. Simulation results for different use cases are presented to show its effectiveness.
