Underwater Navigation Method Based on Side-scan Sonar Images

Ziqi Song,A. Zieliński,Hongyu Bian
2015-01-01
Abstract:In recent years, navigation techniques for underwater vehicles including autonomous long range vehicles have been actively researched. This paper proposed a novel underwater navigation method using side-scan sonar images. High-order cumulant was introduced into image analysis. Bispectrum of a real-time side-scan image was calculated and was employed as a template in the following scanning process within a prior known map of seafloor. Mean square difference was chosen to evaluate the similarity at each searching point in order to obtain the best fix estimation. Simulations were done with actual sonar data and the results suggested that the method had good robustness of rotation and noise. Accuracy of the estimate was pixel level that relied on the resolution of side-scan sonar images and could be higher than 1 meter. In flat bottom specially, the method was able to give out robust position estimate and could be used as a good supplement to traditional inertial navigation systems.
What problem does this paper attempt to address?