We proposed the HS (Hand-in-hand System) for the autonomous inspection of structures. The HS addresses cable-based modular robotics as an application. In this paper, 3 representative experiments are presented to demonstrate the efficiency of the HS. The first experiment is navigation by HS. AUV (Autonomous Underwater Vehicles) take images of all inspection targets and navigational accuracy is estimated. The second experiment deals with obstacle avoidance. In case the smart cable is entangled within obstacles, AUV disentangle the cable and modify the plan for avoidance. The last experiment is the high-speed scanning by HS. AUV make a formation to generate the panoramic image and scan the tank wall surface. Finally, we estimate the accuracy of the mosaicked image they have developed.


Recently, underwater structures have rapidly increased in number, and their periodical safety inspections have become important. For the improvement of these inspection's efficiency and reliability, their automation is required. For autonomous visual safety inspection, we proposed multi-AUV(Autonomous Underwater Vehicles) based on the HS (Hand-in-hand System; Yu, 2004). It is a multi-AUV system interconnected with a smart cable (Shape tape website) which provides its shape. In the smart cable, many fiber-optics sensors provide not only the position of the endpoints but also information concerning its own shape in real-time. The HS can accomplish the inspection task by these 3 processes:

  • Navigation by HS: AUV move to target area

  • Obstacle avoidance: When the smart cable is entangled with obstacles, they have to disentangle it.

  • Scanning by HS: When AUV reach the target area, they have to take images of a target object or scan a target area.

This paper presents experiments and results of these 3 essential processes to demonstrate the efficiency of HS. HS EXPERIMENTS Navigation by HS The navigation experiment was carried out to estimate the accuracy and performance of the HS. Fig. 1 illustrates the AUV coordinates.

This content is only available via PDF.
You can access this article if you purchase or spend a download.