수중 비전 및 초음파 화상정보 융합기반의 호버링 타입 무인자율 수중로봇 'Cyclops'의 개발

MasterToday, the demands for the safety inspection of the bridge column or dams are on the rapid rise. The tasks are risky and require a massive resource. AUVs are a perfect fit for the job. They are not only capable of inspecting the underwater targets in the reliable ways but they could also automate the task that consequently improves the work efficiency. For the automation, AUVs should (1) recognize objects in restricted visibility, (2) operate with a précised localization capability under the dynamic hydro current. The conditions are based on the difficulties of the performing shallow water tasks. This paper addresses a novel concept of the AUV system to accomplish the tasks under the proposed conditions. As the framework of the automation, AUV platform and automation strategy are studied.The proposed vehicle platform has the following critical advantages.1) Sensor fusionThe vehicle is designed for the optimal performance of object recognition. Using the complementary characteristics of the underwater optical camera and the acoustic sensor, the AUV recognizes and approaches to the target object with high-reliability. 2) Superior dynamic performanceThe vehicle has symmetry design as the 'Electric Train'. It minimizes the yawing motion and maximizes the mobility in sophisticate environment such as inside of underwater structures. Symmetric blades and special mounting layout 8 thrusters secure 6DOF positioning performance under the hydro current at any direction and enable the rapid and accurate localization. 3) Flexible configurationThe vehicle's design is configured for fast and easy maintenance. The cubic bare bone frame enables flexible sensor and payload attachments. Mid-ring type junction area facilitate the maintenance process In order to improve the reliability and accuracy of the system, mathematical modeling and simulation were carried out. The vehicle's pressure vessel's sealing and the system's hardware reliability were successfully tested at the AUV test facility at the University of Tokyo. Experimental results of the vehicle dynamic performance show the high accuracy, stability and rapid response for the tasks. As the demonstration of the proposed system, the AUV successfully carried out the highly accurate underwater image taking for the precision image mosaicking mission. The results confirmed the accuracy and reliability of the vehicle performance. In conclusion, the AUV platform showed excellent sensing and localization performance to accomplish the tasks. The results confirmed the proposed system's capability and potential for the automation