EFFECTIVE LOCALIZATION OF MOBILE ROBOTS USING A SONAR SENSOR RING

摘要

This study introduces the effective recognition method using a sonar sensor ring and activating the localization of mobile robots. To localize robot poses effectively with the sonar sensor ring, two kinds of observations are defined; the scanned feature and the tentative feature. The scanned features such as lines, points and even circles are extracted from sets of the accumulated sonar data. However, the performance of the localization with only scanned features should not be satisfactory because several steps of a robot are taken to extract the scanned features For these reasons, we adopt the current raw data as a tentative feature as well as the scanned feature. The tentative features are obtained by matching the current data to the given feature landmarks Both the feature observations are concurrently applied to extended Kalman filter (EKF)-based localization of robot poses. We demonstrate the validity of the proposed method with the results of real experiments