Integrated Navigation of SINS DVL SSS Based on FGO

Side-scan sonar (SSS) can provide a top-down view image of the seafloor. Combining with strap-down inertial navigation system (SINS) and Doppler velocity log (DVL) navigation system, a map of the seafloor can be created. However, over time, the position error from these navigation systems may grow. To address this issue, this study proposes a method that combines the SSS image information in order to reduce these errors.Current algorithms use a feature-based method to register SSS images and calculate relative position. However, this method is susceptible to disturbance from acoustic noise and viewing direction. Furthermore, while most terrain-aided navigation algorithms can suppress position error drift, they cannot correct heading errors, limiting their ability to improve accuracy.In our paper, we propose an improved method that uses Fourier-based registration technique for SSS images. Instead of determining image translation from features, this algorithm uses phase correlation to achieve more stable and accurate relative position output. Additionally, we use factor graph optimization (FGO) to integrate information from both SSS images and SINS/DVL systems. This not only reduces position estimation errors but also improves heading accuracy.Tests on physical sonar images have verified that the proposed Fourier-based method enhances stability. Moreover, field tests have shown that using the integrated navigation method has improved positioning accuracy.