The development of modern times in robotics and mechanization technology has increased significantly in the past few decades due to their high efficiency in time and energy. In the goods mobilization system for companies’ use, particularly the industrial and warehousing divisions, one of the robots that are used for transporting goods is an automatic guided vehicle (AGV). One of the old navigation methods in AGV is the use of a sensor to follow the line pattern on the detected object, namely the line on the floor. This method is rather ineffective because, gradually, these line pattern objects on the floor will fade caused by the effect of AGV wheels’ frictional forces, causing the camera sensor can no longer detect them. Therefore, it is necessary to improve the AGV navigation method so that it can be a sustainable innovation. This navigation method used four image objects positioned in the area traversed by the AGV robot and the camera served as a forward-facing sensor so that the AGV could detect the pattern of image objects with the help of computer vision using the OpenCV software library. The pattern of the detected image object was processed by a program designed on the Raspberry Pi 4 Model B minicomputer. The test results prove that this method can detect image objects within the camera’s field of view and successfully display the output of the image object. The system managed to recognize objects quite accurately, with parameters of 10–95 cm, and through several experiments. The analysis of the rotational speed of the front and rear wheels of the AGV was carried out using an oscilloscope and tachometer as a means of measuring wheel speed or rotation.
Copyrights © 2023