Bayesian Optimization-Based Tasking Method for an Ozone Observation Satellite with Maneuverable Sensors

Abstract

This paper presents a novel method to guide an Earth observation satellite equipped with maneuverable sensors for remote sensing using the measurement of the total column ozone (TCO) as the application. The resulting task planning can interpret the measurements in real time and balance the exploration and exploitation of the area of interest intelligently. A Bayesian optimization (BO)-based method is developed to achieve this goal. For comparison, by the nadir-pointing strategy, the satellite always points to and measures the nadir points; whereas by the offset-pointing strategy, the satellite is guided by the BO-based tasking method to measure the location with either larger TCO values or higher model uncertainties. As demonstrated by the simulation results, the proposed tasking method leads to both better prediction accuracy and better precision of the TCO distribution. This reveals that the proposed observation strategy can enhance the efficiency and accuracy of the remote sensing satellite. A comparison with a deterministic offset-pointing method is also presented to demonstrate that the BO-based tasking method is indeed superior.