Corner Detection for Room Mapping of Fire Fighting Robot

In mapping of an environment, an approach to locate points or features on the map is needed to be recognized by the robot. Corner detection is one of the methods used to find the feature of a corner. In the Indonesian fire fighting robot contest (KRPAI), the method was used to recognize the room where the robot started. The starting position was randomized, so it was difficult to determine which room the robot was placed. In this paper we want to solve the issue by using corner detection to look for corner in each mapped room and look for the characteristic in every room. Characteristics of the room obtained from the distance between the two corners and the distance between the door with the nearest corner. Based on experiments used this method can recognize the room with accuracy of 87.25%. The experiment result showed that 6 out of 48 trials require two scans by turning the robot at 90 degrees because of the limitations of the laser range finder which can only scan by 270 degrees. From all experiments the average error of the reference distance is 2.29 cm.