Distributed motion tomography for time-varying flow fields

Knowledge of a flow field is crucial to guide oceanic mobile sensing platforms; yet, existing regional ocean models provide insufficient spatial and temporal resolutions for precise guidance of mobile platforms. By combining computational ocean models with real-time data streams collected from mobile platforms, generic environmental models (GEMs) provide high-resolution predictions of ocean currents near the mobile platforms. Motion tomography (MT), which is a novel method for constructing GEMs, employs trajectory information of multiple mobile platforms to create a high-resolution spatial map of ocean flow in the region traversed by the mobile platforms. This paper extends the MT method to resolve the coupling between temporal variations and spatial variations in flow modeling. To incorporate temporal variability of flow into MT mapping, MT employs a parametric flow model and constructs a time-varying flow field by estimating the parameters of the flow model. The original MT problem deals with collective information of all the mobile platform trajectories. To address the communication limitation for data collection, distributed MT is developed in which mobile platforms share their estimated model parameters with other mobile platforms nearby. The paper demonstrates that the proposed methods in both non-distributed and distributed fashion successfully construct an underlying time-varying flow field that affects the trajectories of mobile platforms.