İleri Mühendislik Çalışmaları ve Teknolojileri Dergisi
Yazarlar: Serhat YILMAZ, Sadettin Burak KILCI
Konular:Mühendislik
Anahtar Kelimeler:Underwater Vehicles,Biaxial Manipulator,RGB-HVS Transformation
Özet: Robotics applications have been gaining considerable prominence in military and industrial fields. Researches on marine robotics and underwater vehicles focus on the benefits of use of the robot arm and its underwater applications. For instance; underwater vehicles can be used for shipwreck research, environmental analysis, archaeological research, underwater specimen collection and sub-ship inspection. In the study, the 2-axis robot arm (manipulator) prototype has been implemented for underwater vehicles that are intended to be used in underwater operations that people cannot reach or have difficulty in reaching. The prototype performs Rotational-Rotational (RR) movement. It is integrated into a 4 degree of freedom underwater vehicle and can grasp by the gripper end effector located in front of the vehicle. The manipulator movement is fulfilled by visual feedback of the error from the target to be grasped to the end-effector with a camera. The image processing software required for this process has been prepared by using OpenCV libraries in the C programming language on the Raspberry Pi 3 development board. The image taken from the camera has been converted by RGB-HSV conversion into the format on which the image will be processed. The vehicle is aimed to perform an assigned task autonomously refer to real-time data from the camera. Appropriate algorithms have been prepared for searching, discovering and gripping an object. In surface tests, an object and it’s location were determined based on color information with real time image taken from the camera. The manipulator moved automatically with the proportional control method and capture the object 20 cm away from itself within 20 seconds. There have been problems with object capturing during underwater tests. However, it is possible to increase the performance by using more suitable cameras on the prepared system infrastructure.