#### DMCA

## Conditions for suboptimal filter stability in SLAM (2004)

Venue: | In Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems |

Citations: | 5 - 3 self |

### Citations

735 |
Estimation with Applications to Tracking and Navigation
- Bar-Shalom, Li, et al.
- 2001
(Show Context)
Citation Context ...The Kalman gains for the revision of the landmark estimates rapidly tend to zero, the reason being that these states are uncontrollable. definite pseudo-noise covariance to those uncorruptible states =-=[7]-=-. A. O(N) but unstable partially observable SLAM A clever way to add pseudo-noise to the model is by diagonalizing the state error covariance matrix [4], [5], [8]. The result is a suboptimal filter th... |

505 | A solution to the simultaneous localization and map building (SLAM) problem
- Dissanayake, Newman, et al.
(Show Context)
Citation Context ...Localization and Map Building (SLAM) in mobile robotics has been an active research topic for over fifteen years. A Kalman filter (KF) based approach to the SLAM problem is adopted in this paper [1], =-=[2]-=-. One advantage in using such optimal state estimator, is that it is possible to show the convergence properties of SLAM; at least, for the linear case. Under the typical fully correlated SLAM measure... |

15 |
Solving computational and memory requirements of feature-based simultaneous localization and mapping algorithm,”
- Guivant, Nebot
- 2003
(Show Context)
Citation Context ...ive map with zero uncertainty [2]. To speed up the performance of the algorithm, some authors have proposed the use of covariance inflation methods for the decorrelation of the state error covariance =-=[4]-=-, subject to suboptimality of the filter. Adding pseudonoise covariance for the landmark states is equivalent to making the system controllable. However, full decorrelation of a partially observable s... |

14 | Mobile robot self-localization and measurement of performance in middle scale environments
- Duckett, Nehmzow
- 1998
(Show Context)
Citation Context ...eous Localization and Map Building (SLAM) in mobile robotics has been an active research topic for over fifteen years. A Kalman filter (KF) based approach to the SLAM problem is adopted in this paper =-=[1]-=-, [2]. One advantage in using such optimal state estimator, is that it is possible to show the convergence properties of SLAM; at least, for the linear case. Under the typical fully correlated SLAM me... |

10 |
The effects of partial observability
- Andrade-Cetto, Sanfeliu
- 2004
(Show Context)
Citation Context ...LAM; at least, for the linear case. Under the typical fully correlated SLAM measurement model, it is not possible to obtain a zero mean state error estimate, unless partial observability is corrected =-=[3]-=-. Instead, one may obtain a constant bounded state estimate, dependant on the initial filter conditions. The reason being, that the filter used, one in which the vehicle and landmark estimates are sta... |

8 |
2003: The stability of covariance inflation methods for SLAM
- Julier
(Show Context)
Citation Context ...r. Adding pseudonoise covariance for the landmark states is equivalent to making the system controllable. However, full decorrelation of a partially observable system might lead to filter unstability =-=[5]-=-. In this communication we show how to diagonalize only part of the state error covariance to obtain a suboptimal filter that is both linear in time, and stable, at the same time. The paper is structu... |

2 |
Optimization of simultaneous localization and map-builidng algorithm for real-time implementation
- Guivant, Nieto
- 2001
(Show Context)
Citation Context ...ariance to those uncorruptible states [7]. A. O(N) but unstable partially observable SLAM A clever way to add pseudo-noise to the model is by diagonalizing the state error covariance matrix [4], [5], =-=[8]-=-. The result is a suboptimal filter that will compute inflated estimates for the vehicle and landmark covariances, that has the computational advantage of being uncorrelated. The addition of a covaria... |

2 |
experimental outdoor dataset,” 2002. [Online]. Available: http://www.acfr.usyd. edu.au/homepages/academic/enebot/dataset.htm [10
- Nebot, Guivant, et al.
- 2001
(Show Context)
Citation Context ... observable case. D. Experimental Results We show now results on a series of experiments for nonlinear vehicle with an also nonlinear measurement model, using the ACFR - University of Sydney database =-=[9]-=-. The test run used corresponds to a car-like vehicle at the University Car Park. The landmarks used are tree trunks, as measured with a laser range finder. The reconstructed maps are compared to GPS ... |

1 | A solution lo the simullaneous loeali,.alion and map building (SLAM) problem - Uisranay&c, I |

1 | 131 1. Andmde-Qtlo and A. Sanleliu, "Tlx effects of panid ohsmahility - no - 2001 |

1 | Solving compulalional and mcmory requim"s of feaux~hased simuhancow lwaliruion and mapping algoriihmr - Guivani, Nielo - 2003 |

1 | lulim, 'The s~ahility of covariance inllaion mahods far SLAM - I - 2003 |

1 | Esrimiioii wirh Applicotiom IO hocking ond Noi,igotio!l - Rar-Shalom, Li, et al. - 2001 |

1 | Oplimiution of simultaneous localication and map-huilidnp algorithm for real-cimc implemenmlion - Guivanl, Nieto - 2001 |