engleski

Layout Mapping Using an Autonomous Mobile Robot Equipped with an Ultrasonic Sensor

In this paper, we deal with some preliminaries concerning the problem of creating a reasonably accurate 2-D indoor map, using an autonomous mobile robot equipped with an ultrasonic sensor. For this purpose, we utilize the robotic frame DiddyBorg, coupled with the latest Raspberry Pi model. The accuracy of the used ultrasonic sensor is tested in a simulated and a real-world environment. Ultrasonic sensor flaws cause errors in the area mapping process, in the form of precision decline and unwanted ghost points, usually around the corners of an indoor environment. We present a method for filtering these unwanted occurrences. The method has been tested in a controlled test box environment and in the real-world indoor environment. We have used a square, a triangle, and a hexagon as shapes for the testing environment. As a basis for the mapping process, ordinary trigonometry was used, whereas cluster analysis was used for the removal of errors. Experimental results suggest that the proposed method is reasonably accurate in determining correct position and distance of measured points from single-point scans.

indoor mapping ; autonomous robot ; ultrasonic sensor ; machine learning

nije evidentirano