## Real-Time Self-Localization in Unknown Indoor Environments using a Panorama Laser Range Finder (1997)

Venue: | In IEEE/RSJ International Workshop on Robots ans Systems, IROS 97 |

Citations: | 17 - 0 self |

### BibTeX

@INPROCEEDINGS{Einsele97real-timeself-localization,

author = {Tobias Einsele and Prof Dr. --ing and Georg Farber},

title = {Real-Time Self-Localization in Unknown Indoor Environments using a Panorama Laser Range Finder},

booktitle = {In IEEE/RSJ International Workshop on Robots ans Systems, IROS 97},

year = {1997},

pages = {697--703},

publisher = {IEEE Press}

}

### OpenURL

### Abstract

This paper deals with self-localization of a mobile robot on the condition that no a-priori knowledge about the environment is available. The applied method features to be accurate, robust, independent of any artificial landmarks and feasible with such a moderate computational effort that all necessary tasks can be executed in real-time on a standard PC. The perception system used is a panorama laser range finder (PLRF) which takes scans of its present environment. A modified Dynamic Programming (DP) algorithm provides pattern matching and pattern recognition on the preprocessed panorama scans and thereby renders a qualitative fusion of the sensory data. For an exact quantitative estimate of the robot's current position, a robust localization module is employed. The knowledge gained about the environment along that way is stored in a self-growing, graph based map which combines geometrical information and topological restrictions. Preliminary experiments in a common office environment ...

### Citations

2611 |
Dynamic Programming
- Bellman
- 1957
(Show Context)
Citation Context ... the scans were taken. In contrast to related work in which neural network models [10], correlation of the laser scans [11] or least squares methods [2] fulfil these tasks, a Dynamic Programming (DP) =-=[1]-=- technique is proposed here. When comparing matching algorithms, the computational effort is always a very decisive criterion. The DP algorithm is of order O(n 2 ) and can therefore be put on the same... |

79 | A Robust, Qualitative Method for Robot Spatial Reasoning
- Kuipers, Byun
- 1988
(Show Context)
Citation Context ...evance. The only thing that matters in this situation is that the robot knows in which direction it has to go and recognizes the place at the end of the corridor in order to re-localize exactly again =-=[6, 7]-=-. So one can conclude that a geometrically exact localization is only required when arriving at or moving within such a `distinctive' place. Therefore, instead of an environmental map referring to a g... |

21 |
A Data Driven Organization of the Dynamic Programming Beam Search for Continuous Speech Recognition
- Ney, Mergel, et al.
- 1987
(Show Context)
Citation Context ...r, what makes DP superior to those is that it can be massively accelerated, which of course does not mean decreasing its order, but can reduce the de facto processing effort in a very significant way =-=[8]-=-. The DP is a classical pattern matching algorithm which establishes and evaluates correspondences between a reference and a test pattern. In this context the term pattern denotes the set of line segm... |

16 |
Navigating with a rat brain: A neurobiologically inspired model for robot spatial representation
- Mataric
- 1991
(Show Context)
Citation Context ...evance. The only thing that matters in this situation is that the robot knows in which direction it has to go and recognizes the place at the end of the corridor in order to re-localize exactly again =-=[6, 7]-=-. So one can conclude that a geometrically exact localization is only required when arriving at or moving within such a `distinctive' place. Therefore, instead of an environmental map referring to a g... |

7 |
Keeping track of position and orientation of moving indoor systems by correlation of range-finder scans
- Wei��, Wetzler, et al.
- 1994
(Show Context)
Citation Context ...= 2:5 mm. 3. Range data preprocessing The task of the preprocessing unit is to extract line segments from the acquired range data. In other approaches classical clustering techniques [10], histograms =-=[11]-=- or the Hough transform [4] are used to solve this problem. All of these methods have the property that the range data has to be transformed into a feature space. This requires that from each range sa... |

6 |
Segmentation of planar curves
- Pavlidis, Horowitz
- 1974
(Show Context)
Citation Context ...whole laser scan is segmented into piecewise linear sections. This is achieved by an iterative algorithm which originates from image processing and for which an effort minimized implementation exists =-=[9]-=-. It is therefore capable to process the rather few sampling points of a laser scan (compared to the large amount of pixels in an image) faster than e. g. a cluster algorithm -- even if the simplifica... |

5 | Exploration, Navigation and Self-Localization in the autonomous mobile robot
- Edlinger, Weiss
- 1996
(Show Context)
Citation Context ... to preserve a useful database, the map needs to be updated and extended continuously. This is done by the mapping unit that shows the following behaviour which is different from what can be found in =-=[3]-=-: If it is possible to match a presently taken and preprocessed scan with an already stored scan, producing costs that fall below a given threshold, which means that the similarity measure exceeds a c... |

3 |
Navigation in cluttered rooms using a range measuring laser and the hough transform
- Forsberg, Larsson, et al.
- 1993
(Show Context)
Citation Context ...rocessing The task of the preprocessing unit is to extract line segments from the acquired range data. In other approaches classical clustering techniques [10], histograms [11] or the Hough transform =-=[4]-=- are used to solve this problem. All of these methods have the property that the range data has to be transformed into a feature space. This requires that from each range sample to e. g. its direct su... |

3 |
Learning control and localisation of mobile robots
- Vestli, Tschichold-Gürman, et al.
- 1994
(Show Context)
Citation Context ...ch amounts to oe = 2:5 mm. 3. Range data preprocessing The task of the preprocessing unit is to extract line segments from the acquired range data. In other approaches classical clustering techniques =-=[10]-=-, histograms [11] or the Hough transform [4] are used to solve this problem. All of these methods have the property that the range data has to be transformed into a feature space. This requires that f... |

1 |
Determination of Position and Orientation of Autonomous Vehicles in Production Type Environments Using a 94 GHz Radar Sensor
- Detlefsen, Rozmann
- 1994
(Show Context)
Citation Context ...on 6) as well as recognition of the places in which the scans were taken. In contrast to related work in which neural network models [10], correlation of the laser scans [11] or least squares methods =-=[2]-=- fulfil these tasks, a Dynamic Programming (DP) [1] technique is proposed here. When comparing matching algorithms, the computational effort is always a very decisive criterion. The DP algorithm is of... |