Monte Carlo localization: efficient position estimation for mobile robots
AAAI '99/IAAI '99 Proceedings of the sixteenth national conference on Artificial intelligence and the eleventh Innovative applications of artificial intelligence conference innovative applications of artificial intelligence
| Hi-index | 0.00 |
We propose a method for performing localization of a mobile robot using a Microsoft Kinect RGB-D sensor. We make use of artificial landmarks placed at known locations in the environment. The robot tries to locate these landmarks based upon their color and shape and in turn, measure the range and bearing to the same. The range and bearing information so received is noisy. The proposed method estimates the robot's pose in the world coordinates using Monte Carlo Localization. The total count of the landmarks and their individual locations (in world coordinate system) is to be provided apriori to the robot, which is akin to using a map of the environment, but having very low resource requirements. The proposed method has been implemented and experimentally tested. The timing performance of the algorithm is proportional to the number of landmarks in the view of Kinect sensor. Thus landmarks need to be placed in such a way as to ensure that one or at the max two landmarks remain in the view of Kinect at all times. We present results of a typical practical application for use in indoor office-like environment.