Underwater Tightly-Coupled Method for Heading Estimation Against Uncertain Disturbances: Using Differential Light Intensity

Polarization navigation has been proven to be a remarkably effective strategy for unmanned underwater vehicles (UUVs) in heading determination. However, it is difficult for existing schemes to deal with dynamic refraction and uncertain multiple scattering, deteriorating navigation accuracy. This article proposes an underwater tightly-coupled method using original differential light intensity to determine heading in the presence of hybrid uncertain disturbances. In order to reduce the modeling error caused by underwater disturbances, we develop a tightly-coupled model employing original differential light intensity, in which the direct relationship between differential polarization and attitudes is investigated. Meanwhile, the compensation factor is introduced into the tightly-coupled model to deal with nonvertical errors induced by underwater hybrid optical effects. In view of measurement uncertainty in underwater polarization, adaptive parameter adjustment is implemented by variational Bayesian (VB) to optimize the state estimation, which benefits the reliability and robustness of the system. Both simulation and ocean tests are carried out to demonstrate the feasibility and superiority of the proposed method.