6D SLAM with GPGPU computation

eng Artykuł w języku angielskim DOI:

wyślij Janusz Będkowski *, Geert De Cubber **, Andrzej Masłowski * * Instytut Automatyki i Robotyki, Politechnika Warszawska ** Royal Military Academy, Brussels, Belgium

Pobierz Artykuł

Abstract

The main goal was to improve a state of the art 6D SLAM algorithm with a new GPGPU-based implementation of data registration module. Data registration is based on ICP (Iterative Closest Point) algorithm that is fully implemented in the GPU with NVIDIA FERMI architecture. In our research we focus on mobile robot inspection intervention systems applicable in hazardous environments. The goal is to deliver a complete system capable of being used in real life. In this paper we demonstrate our achievements in the field of on line robot localization and mapping. We demonstrated an experiment in real large environment. We compared two strategies of data alingment - simple ICP and ICP using so called meta scan.

Keywords

6D SLAM, parallel computation

6D SLAM wykorzystujacy obliczenia GPGPU

Streszczenie

Głównym celem jest artykułu jest usprawnienie algorytmu 6D SLAM za pomocą implementacji modułu rejestracji danych wykorzystującą obliczenia równoległe. Moduł rejestracji danych jest oparty o algorytm ICP (ang. Iterative Closest Point), który został w pełni zaimplementowany w architekturze GPU NVIDIA FERMI. W naszych badaniach koncentrujemy się na mobilnych systemach robotycznych inspekcyjno-interwencyjnych dedykowanych do pracy w niebezpiecznym środowisku. Celem jest opracowanie kompletnego systemu, który może być wykorzystany w realnej aplikacji. W tym artykule przedstawiamy nasze rezultaty w zakresie lokalizacji i budowy mapy w trybie on-line. Przedstawiamy eksperyment w rzeczywistym, rozległym środowisku. Zostały porównane dwie strategie dopasowywania danych, klasyczna oraz wykorzystująca tzw. meta scan.

Słowa kluczowe

6D SLAM, obliczenia równoległe

Bibliografia

  1. Bedkowski J. (2011): Data registration module - a component of semantic simulation engine, [in:] Proceedings of 5th European Conference on Mobile Robots ECMR 2011, Orebro, Sweden, 133-138.
  2. Bedkowski J., Maslowski A. (2011): GPGPU implementation of On-Line point to plane 3D data registration, [in:] Proceedings of the 2011 International Conference on Electrical Engineering and Informatics Volume 2, 17-19 July 2011, Bandung, Indonesia, 931-936.
  3. Bedkowski J. (2011): Parallel implementation of hybridICP data registration, “Elektronika” (8), 114-118.
  4. Huber D., Hebert M. (2003): Fully automatic registration of multiple 3D data sets, “Image and Vision Computing"” 21(1), 637-650.
  5. Fitzgibbon A. W. (2001): Robust registration of 2D and 3D point sets, [in:] British Machine Vision Conference, 411-420.
  6. Magnusson M., Duckett T. (2005): A Comparison of 3D Registration Algorithms for Autonomous Underground Mining Vehicles, [in:] Proc. ECMR, 86-91.
  7. Park S.-Y., Choi S.-I., Kim J., Chae J. (2010): Realtime 3D registration using GPU, “Machine Vision and Applications”, 10.1007/s00138-010-0282-z, 1-14.
  8. Park S.-Y., Subbarao M. (2003): An accurate and fast point-to-plane registration technique, “Pattern Recogn. Lett.” 24, 2967-2976.
  9. Nuchter A., Lingemann K., Hertzberg J. (2007): Cached k-d tree search for ICP algorithms, [in:] Proceedings of the Sixth International Conference on 3-D Digital Imaging and Modeling, IEEE Computer Society, Washington, DC, USA, 419-426.
  10. Rusinkiewicz S., Levoy M. (2001): Efficient Variants of the ICP Algorithm, [in:] Proceedings of the Third International Conference on 3D Digital Imaging and Modeling (3DIM).
  11. Qiu D., May S., Nuchter A. (2009): GPU-Accelerated Nearest Neighbor Search for 3D Registration, [in:] Proceedings of the 7th International Conference on Computer Vision Systems: Computer Vision Systems, ICVS ’09, Springer-Verlag Berlin/Heidelberg, 194-203.
  12. Surmann H., Nuchter A., Lingemann K., Hertzberg J. (2004): 6D SLAM - Preliminary report on closing theloop in six dimensions, [in:] Proceedings of the 5th IFAC Symposium on Intelligent Autonomous Vehicles (IAV’04).
  13. Sprickerhof J., Nuchter A., Lingemann K., Hertzberg J. (2009): An Explicit Loop Closing Technique for 6D SLAM, [in:] 4th European Conference on Mobile Robots, September 23-25, 2009, Mlini/Dubrovnik, Croatia, 229-234.
  14. Magnusson M., Duckett T., Lilienthal A. J. (2007): 3D Scan Registration for Autonomous Mining Vehicles, „Journal of Field Robotics” 24(10), 803-827.
  15. Magnusson M., Andreasson H., Nüchter A., Lilienthal A. J. (2009): Automatic Appearance-Based Loop Detection from 3D Laser Data Using the Normal Distributions Transform, “Journal of Field Robotics” 26(11-12), 892-914.
  16. Newman P., Cole D., Ho K. L. (2006): Outdoor SLAM using Visual Appearance and Laser Ranging, [in:] Proceedings of the IEEE International Conference on Robotics and Automation (ICRA).
  17. Newman P., Ho K. L. (2005): SLAM - Loop Closing with Visually Salient Features, [in:] IEEE International Conference on Robotics and Automation, 18-22 April.
  18. Cummins M., Newman P. (2008): FAB-MAP: Probabilistic Localization and Mapping in the Space of Appearance, “The International Journal of Robotics Research” 27(6), 647-665.
  19. Williams B., Cummins M., Neira J., Newman P., Reid I., Tardos J. (2009): A comparison of loop closing techniques in monocular SLAM, [in:] Robotics and Autonomous Systems.
  20. Ho K. L., Newman P. M. (2006): Loop closure detection in SLAM by combining visual and spatial appearance, [in:] Robotics and Autonomous Systems, 740-749.
  21. Lowe D. G. (2004): Distinctive Image Features from Scale Invariant Keypoints, “Int. J. Comput. Vision” 60, 91-110.
  22. Leong K., Newman P. (2005): Combining Visual and Spatial Appearance for Loop Closure Detection in SLAM, [in:] 2nd European Conference on Mobile Robots (ECMR).
  23. Ho K. L., Newman P. (2005): Multiple Map Intersection Detection using Visual Appearance, [in:] 3rd International Conference on Computational Intelligence, “Robotics and Autonomous Systems”, Singapore.
  24. Ho K. L., Newman P. (2007): Detecting Loop Closure with Scene Sequences, “Int. J. Comput. Vision” 74, 261-286.
  25. Kaustubh Pathak N. V.Max Pfingsthorn, Birk A. (2009): Relaxing Loop-Closing Errors in 3D Maps Based on Planar Surface Patches, [in:] 14th International Conference on Advanced Robotics (ICAR), IEEE Press.
  26. Segal A., Haehnel D., Thrun S. (2009): Generalized-ICP, [in:] Proceedings of Robotics: Science and Systems, Seattle, USA.
  27. Nüchter A., Hertzberg J. (2008): Towards semantic maps for mobile robots, “Robot. Auton. Syst.” 56(11), 915-926.
  28. (2010a): NVIDIA CUDA C Programming Guide 3.2, http://www.nvidia.com/cuda.
  29. (2010b): CUDA C Best Practices Guide 3.2, http://www.nvidia.com/cuda.