Error modeling of Doppler velocity logs using a point-based scattering model

Abstract

As Doppler Velocity Log (DVL) systems become smaller, cheaper, and increasingly capable, they are becoming standard options on even low-end undersea vehicles. DVLs are the principal sensor that provide ground-referenced velocity measurements to compensate for the bias and drift errors found in inertial navigation systems. Therefore, DVLs fill an essential role in achieving an accurate navigation solution for unmanned undersea vehicles. Although there are many modeling and simulation techniques for comparing a fused-sensor inertial navigation solution to earth-reference, there is a sparsity in the literature for models and simulation environments that are suited for estimating the performance of a standalone DVL system. A practical model must incorporate sufficient environmental fidelity so as to match with empirical results, while still remaining computationally efficient. The point-based scattering model described by Brown et al. [JASA Elect. Lett. i(2017)] may be a useful tool for modeling DVL performance. The implemented model incorporates seafloor roughness, scatterer density, textures, and is well-suited for integration into Monte Carlo simulations. This presentation will compare the DVL performance predicted by this point-based scattering model to several other error models and discuss the utility for predicting performance in natural marine environments.