Improved Motion Tomography for Measuring Ocean Flow Field and Informative Path Planning With Unmanned Surface Vehicles

Flow field models with high resolution are essential for the effective navigation of marine vehicles. This paper proposes an improved Motion Tomography (MT) method for measuring and estimating the ocean flow field with unmanned surface vehicles (USVs). The MT method relies on measuring the differences between actual and predicted USV trajectories caused by ocean flows, that is, the motion‐integration error. There are two main differences between the proposed MT and the traditional MT. First, the motion‐integration error formulation is improved by considering the impact of the ocean flow on the heading of the USV. Second, the flow field estimation performance is improved due to the readily available USV actual trajectory information. Alongside the USV traversing within the field of interest, the information of the environmental flow field is gradually revealed with the improved flow field mapping. Due to the spatial and temporal discretization of the flow field, a parameterized flow field model is employed to address the high‐dimensional solution space issue. With the estimated flow field model, the USV path is then carefully planned to maximize the sampled information along the path, known as informative path planning (IPP). We conducted sea trials in the East China Sea to validate the proposed MT method and the IPP application performance. The flow field information over a 300 m 300 m area is obtained using the proposed improved MT method and a dual‐propeller USV. The USV path is then planned to sample the surface light‐intensity information and automatically return in an energy‐efficient way. Sea‐trial results demonstrate that the measured path of the USV is highly consistent with the IPP planned path, which is not rushed away by the flow. Moreover, the energy consumption calculated based on the USV speed aligns closely with the IPP path planning predictions. Compared with the traditional MT method, the proposed approach improves energy consumption prediction accuracy by 20%. Therefore, from both the heading and speed perspectives, the results demonstrate that the proposed improved MT method can accurately estimate the ocean flow field and can improve the IPP performance.