Adaptive navigation of a high speed autonomous underwater vehicle using low cost sensors for low-altitude survey

In order to get highly detailed information of seafloor with low-cost, the authors have been developing high speed autonomous underwater vehicle (AUV) using low cost scanning sonar. It is difficult to design the AUV navigation system functioning properly in various seafloor environment because the sonar echo levels from the seafloor are relatively weak and differ with the survey area. In this study, the authors propose the AUV navigation method adaptive to the various seafloor environment using reinforcement learning in which AUV learns how to move by itself online and test the algorism in a simulation.