Flow parameter estimation based on on-board measurements of air vehicle traversing turbulent flows
Inspired by principles from particle transport theory in fluid dynamics, we recently developed a new energy-efficient control approach via implicit model following (IMF) for air vehicles traversing turbulent flows. However, the control design requires prior knowledge of the vortex timescale of the turbulent flow. In this paper, we propose an approach to estimate the turbulent flow parameters based on noisy on-board measurements without prior knowledge of the exact flow conditions, and validate it on a two-dimensional cellular flow model. By sparse identification of nonlinear dynamics (SINDy), the nonlinear vehicle dynamics with wind effects introduced are approximated as a sparsely weighted combination of user-defined candidate functions. Accordingly, separate optimization problems are set up to determine the weights representing the active level of candidate functions in the unknown nonlinear dynamics. We then identify the flow parameters by analyzing the weights statistically. Finally, the ability of the proposed method to estimate the flow parameters, including the mean velocity, the vortex length scale and the vortex timescale, is demonstrated by testing the algorithms on different measurement data sets with various initial conditions and flow parameters. We show that this method can accurately estimate the flow parameters within a permissible range of error.
