<![CDATA[The condition of an earthquake in a high-rise building often makes visitors panic when evacuating. This requires assistance to direct visitors to the evacuation route with voice commands or text written in the building. Therefore in this research a navigation system on wheeled robots is designed which aims to help the evacuation process at the time of an earthquake in a high-rise building by directing it to the evacuation route based on text orders. In the navigation system is designed to use a camera to recognize text. The camera detection results are used as input to be processed in the Raspberry Pi using the optical character recognition method. The output results are used as parameters to make the movement of the wheeled robot that is moving forward, turn right, turn left, walk backward and stop. Based on testing the detection accuracy obtained page segmentation modes value 6 with an average accuracy of 98.11%. The average result of time needed to detect an object is 48.5 seconds at a test distance of 3 meters and 185.75 seconds at a test meter of 6 meters. Wheeled robot speed obtained high percentage duty value of 13% and 14% with an average detection distance of 193.3cm. The accuracy of the motion results obtained by a percentage of 100% on all four objects except at the RUN object can make movements at a test distance of 50-250cm.]]>
Copyrights © 2020
