Field experimental results of path re-planning method for an AUV to visualize complicated surface in 3D

Although AUVs can observe seafloor without tether cables nor human control, it is impossible for operators to confirm in real time that there are no unscanned areas caused by occlusions, positioning errors, and so on. If unobserved areas are found by post processing, another deployment of the AUV is necessary to cover them. Therefore, we have proposed a recursive observation method of rough terrain by an AUV to realize a high coverage observation at one deployment. In the method, the AUV evaluates unobserved regions after completing the preplanned path. After that, the vehicle generates a new path to cover them. This paper reports the results of sea experiments held at Suruga Bay in Japan, in November 2012. An artificial target was set up on the seafloor. Then the AUV Tri-TON observed the target based on the proposed method. As the generated path was concentrated on the target, the performance of the proposed method at the real sea environment was verified. In addition, mapping accuracy was evaluated by comparing the observational results and ground truth.