Abstract
Kidnapping is a localization problem in which a well-localized robot is suddenly moved to somewhere else without any update to its sensors. This causes failures or faults in the autonomous robot. To solve this problem in an unknown environment, a new method for kidnapping detection in Simultaneous Localization and Mapping called Double Guarantee Kidnapping Detection (DGKD) has been proposed. It helps to determine whether kidnapping has happened, and classifies and detects which type of kidnapping it is. To verify its performance in a realistic and static environment, a mobile robot platform is constructed using a Roomba robot equipped with a laser range finder. Experimental results demonstrate the validity and accuracy of the proposed method.
