Subject:

Implementation of an Algorithm for Simultaneous Localization and Mapping using RGB(-D) Data and a Wheeled Mobile Robot's Odometry; Adaptation to Indoor Navigation

Introduction:

Different open source solutions to the simultaneous localization and mapping (SLAM) problem have been released by the research community (e.g. Kintinuous, ElasticFusion, LSD-SLAM, and ORB-SLAM) over the last years. These approaches provide accurate and consistent motion tracking and mapping of the environment using standard image streams and/or depth data from e.g. a Kinect camera. Still, their performance is subject to the perceived environment, primarily by their deployment indoors or outdoors. On the one hand, indoor perception is in close-range and largely free from IR radiation. On the other hand, indoor perception is subject to repeated (ambiguous) scenes and large untextured areas.

In the current work one of the freely available SLAM frameworks above is to be adapted to indoor localization by cameras or depth sensors mounted on a wheeled robot within our institute's main building. In addition, the localization software will make use of the robot's odometry by data fusion in case of ambiguous perception and untextured areas. To conclude, a basic method for autonomous navigation will be produced. Provision is made for a Pioneer 3-DX wheeled mobile robotic platform.

Requirements:

Degree courses in engineering, computer science, maths or related with above-average grades

Very good coding skills in C++ -- desirable in Python and Matlab

Preferable: prior programming experience on robotic hardware

References:

T. Whelan, M. Kaess, M.F. Fallon, H. Johannsson, J.J. Leonard and J.B. McDonald. "Kintinuous: Spatially Extended KinectFusion." In RSS Workshop on RGB-D: Advanced Reasoning with Depth Cameras, (Sydney, Australia), July 2012.

T. Whelan, S. Leutenegger, B. Glocker, R. F. Salas-Moreno, and A. Davison. "ElasticFusion: Dense SLAM Without A Pose Graph." Robotics: Science and Systems (RSS), Rome, Italy, July 2015.

J. Engel, T. Schoeps, and D. Cremers . "LSD-SLAM: Large-Scale Direct Monocular SLAM. " In European Conference on Computer Vision (ECCV), 2014.

Raul Mur-Artal, J. M. M. Montiel and Juan D. Tardos. "ORB-SLAM: A Versatile and Accurate Monocular SLAM System." IEEE Transactions on Robotics, vol. 31, no. 5, pp. 1147-1163, October 2015.

Advisor:

Dr.-Ing. Klaus Strobl
eMail: Klaus.Strobl@dlr.de
www: http://www.robotic.dlr.de/Klaus.Strobl/