Unscented Kalman Filter with a Nonlinear Propagation Model for Navigation Applications

TL;DR

The paper tackles improving UKF-based navigation by more accurately propagating sigma points through nonlinear navigation dynamics. It introduces UKF-NESPM, which runs the strapdown navigation cycle for each sigma point and extracts deviations to form propagated points, with biases held constant. Using a real AUV dataset (Snapir) for INS/DVL fusion, the method outperforms a state-of-the-art adaptive UKF baseline in VRMSE and MRMSE. The results demonstrate that integrating the navigation algorithm into sigma-point propagation enhances estimation accuracy under dynamic underwater conditions.

Abstract

The unscented Kalman filter is a nonlinear estimation algorithm commonly used in navigation applications. The prediction of the mean and covariance matrix is crucial to the stable behavior of the filter. This prediction is done by propagating the sigma points according to the dynamic model at hand. In this paper, we introduce an innovative method to propagate the sigma points according to the nonlinear dynamic model of the navigation error state vector. This improves the filter accuracy and navigation performance. We demonstrate the benefits of our proposed approach using real sensor data recorded by an autonomous underwater vehicle during several scenarios.

Figures (4)

• Figure 1: The Snapir AUV during sea experiments.
• Figure 2: Horizontal position of tracks 1-2 used in the training dataset.
• Figure 3: Horizontal position of tracks 3-4 used in the training dataset.
• Figure 4: Horizontal position of tracks 5-6 used in the testing dataset.