When people work in the water, it must be restricted to dive into a certain depth and work over time. Therefore, underwater robots are applied in various fields. Most of them are Remote Operated Vehicle (ROV) type which is controlled by human resources on the water. The underwater robots, which can be operated autonomously, are equipped with expensive sensors, which restrict them to be used only in high-priced industries such as military, ocean exploration, and oil/gas mission.
In this study, we suggest an underwater robot with a camera system instead of expensive sensors, which can be worked autonomously in artificial submerged structures, such as aqua farms, in which the vision information can be applied.
A sensor part, which is important among the robot hardware, is not composed of expensive sensors, but a camera, an IMU, and a depth sensor, which lowers the system price. The proposed control system consists of two parts; Navigation & Vision processing Controller (NVC), and Thruster & Sensor Controller (TSC). NVC performs decision making operations, such as mission management, vision processing, and system error detection. TSC controls thrusters and acquires data from sensors. Additionally, a mission planner is proposed to drive the composed robot system autonomously.
The mission planner plans missions by defining tasks, which composes the missions, and setting up behaviors, which composes the tasks, by a robot operator. The mission planner consists of three editors which have different purposes individually. Control Parameter Editor (CPE) revises the created mission files briefly, and creates new small missions. Behavior / Task Editor (BTE) manages behavior/task information in the previous missions. Artificial Landmark Editor (ALE) creates missions efficiently using artificial landmark information.
Thus, the proposed robot hardware is implemented to satisfy appropriate requirements to be applied to artificial submerged structures. The mission planner, whi...