Least-Squares Fitting of Two 3-D Point Sets
IEEE Transactions on Pattern Analysis and Machine Intelligence
High-Resolution Terrain Map from Multiple Sensor Data
IEEE Transactions on Pattern Analysis and Machine Intelligence - Special issue on interpretation of 3-D scenes—part II
An intelligent, predictive control approach to the high-speed cross-country autonomous navigation problem
Real-Time Correlation-Based Stereo Vision with Reduced Border Errors
International Journal of Computer Vision
Non-parametric Local Transforms for Computing Visual Correspondence
ECCV '94 Proceedings of the Third European Conference-Volume II on Computer Vision - Volume II
Two years of Visual Odometry on the Mars Exploration Rovers: Field Reports
Journal of Field Robotics - Special Issue on Space Robotics
Stereo Processing by Semiglobal Matching and Mutual Information
IEEE Transactions on Pattern Analysis and Machine Intelligence
Mutual Information Based Semi-Global Stereo Matching on the GPU
ISVC '08 Proceedings of the 4th International Symposium on Advances in Visual Computing
Evaluation of Stereo Matching Costs on Images with Radiometric Differences
IEEE Transactions on Pattern Analysis and Machine Intelligence
A Real-Time Low-Power Stereo Vision Engine Using Semi-Global Matching
ICVS '09 Proceedings of the 7th International Conference on Computer Vision Systems: Computer Vision Systems
Stereo camera based navigation of mobile robots on rough terrain
IROS'09 Proceedings of the 2009 IEEE/RSJ international conference on Intelligent robots and systems
Reactive navigation through rough terrain: experimental results
AAAI'92 Proceedings of the tenth national conference on Artificial intelligence
Hi-index | 0.00 |
In this paper we present a visual navigation algorithm for the six-legged walking robot DLR Crawler in rough terrain. The algorithm is based on stereo images from which depth images are computed using the semi-global matching (SGM) method. Further, a visual odometry is calculated along with an error measure. Pose estimates are obtained by fusing inertial data with relative leg odometry and visual odometry measurements using an indirect information filter. The visual odometry error measure is used in the filtering process to put lower weights on erroneous visual odometry data, hence, improving the robustness of pose estimation. From the estimated poses and the depth images, a dense digital terrain map is created by applying the locus method. The traversability of the terrain is estimated by a plane fitting approach and paths are planned using a D* Lite planner taking the traversability of the terrain and the current motion capabilities of the robot into account. Motion commands and the traversability measures of the upcoming terrain are sent to the walking layer of the robot so that it can choose an appropriate gait for the terrain. Experimental results show the accuracy of the navigation algorithm and its robustness against visual disturbances.