Implementasi Machine Learning pada Robot Krsri untuk Melakukan Pendeteksian Rintangan

Abstract

The Indonesian SAR Robot Contest (KRSRI) is a development of the Fire Extinguisher Robot Contest (KRPAI), where at first the robot at KRPAI only put out fires but at KRSRI the robot was asked to store SAR mushrooms there were obstacles in this contest which later the robot had to pass in completing mission. Based on this, an obstacle detection system was designed for the robot using Machine Learning with the K-Nearest Neighbor algorithm and feature extraction of the Gray Level Co-occurrence Matrix, later the robot is expected to be able to detect accurate obstacles for the sake of saving efficiency so that no more time is wasted because the robot is wrong detect obstacles. The results of the tests that have been carried out are detection accuracy based on dataset tests, namely 80% for climbing obstacles, 100% for gravel obstacles, and 90% for stepped obstacles, and an error value of 20% is obtained for climbing obstacles, 0% for gravel obstacles, and 10% for the obstacle steps, and get the robot the ideal distance in detecting that is at a distance of 10cm and 15cm.