## FastSLAM 2.0: An improved particle filtering algorithm for simultaneous localization and mapping that provably converges

### Cached

### Download Links

Citations: | 165 - 8 self |

### BibTeX

@MISC{Montemerlo_fastslam2.0:,

author = {Michael Montemerlo and Sebastian Thrun and Daphne Roller and Ben Wegbreit},

title = {FastSLAM 2.0: An improved particle filtering algorithm for simultaneous localization and mapping that provably converges},

year = {}

}

### Years of Citing Articles

### OpenURL

### Abstract

In [15], Montemerlo et al. proposed an algorithm called FastSLAM as an efficient and robust solution to the simultaneous localization and mapping problem. This paper describes a modified version of FastSLAM that overcomes important deficiencies of the original algorithm. We prove convergence of this new algorithm for linear SLAM problems and provide real-world experimental results that illustrate an order of magnitude improvement in accuracy over the original FastSLAM algorithm. 1

### Citations

715 |
Tracking and Data Association
- Bar-Shalom
- 1988
(Show Context)
Citation Context ...aling SLAM algorithms to maps with more than a few hundred features. It also limits the applicability of SLAM algorithms to problems with ambiguous landmarks, which induces a data association problem =-=[2; 22]-=- . Today's most robust algorithms for SLAM with unknown data association maintain multiple hypotheses (tracks), which increase their computational complexity. Consequently, there has been a flurry on ... |

448 | FastSLAM: A Factored Solution to the Simultaneous Localization
- Montemerlo, Thrun, et al.
- 2002
(Show Context)
Citation Context ... PA 15213 mmde@cs.cmu.edu, thrun@cs.cmu.edu Daphne Koller and Ben Wegbreit Computer Science Department Stanford University Stanford, CA 94305-9010 koller@cs.stanford.edu, ben@wegbreit.com Abstract In =-=[15]-=- , Montemerlo et al. proposed an algorithm called FastSLAM as an efficient and robust solution to the simultaneous localization and mapping problem. This paper describes a modified version of FastSLAM... |

426 | New Extension of the Kalman Filter to Nonlinear Systems
- Julier, Uhlmann
- 1997
(Show Context)
Citation Context ...t is similar to the arc reversal technique proposed for particle filters applied to Bayes networks [10] , and it is similar to recent work by van der Merwe [24] , who uses an unscented filtering step =-=[9]-=- for generating proposal distributions that accommodate the measurement. While this modification is conceptually simple, it has important ramifications. A key contribution of this paper is a convergen... |

343 |
Sensor fusion in certainty grids for mobile robots,”Computer
- Moravec
- 1988
(Show Context)
Citation Context ...ge provides negative evidence. The posterior probability of landmark existence is accumulated by the following Bayes filter, whose log-odds form is familiar from the literature on occupancy grid maps =-=[-=-16] : [m] n = X t ln p(i [m] n j s [m] t ; z t ; ^ n [m] t ) 1 p(i [m] n j s [m] t ; z t ; ^ n [m] t ) (22) Here [m] n are the log-odds of the physical existence of landmarks [m] n in map m, and p(i... |

302 |
On the representation and estimation of spatial uncertainty
- Smith, Cheeseman
- 1987
(Show Context)
Citation Context ... large, accurate maps is a challenging statistical estimation problem, especially when performed in real-time. Most present-day research on SLAM originates from a seminal paper by Smith and Cheeseman =-=[21]-=- , which proposed the use of the extended Kalman filter (EKF) for solving SLAM. This paper is based on the insights that errors in the map and pose errors are naturally correlated, and that the covari... |

281 | Incremental mapping of large cyclic environments
- Gutmann, Konolige
- 1999
(Show Context)
Citation Context ...result is the first convergence result for a constant-time SLAM algorithm. It even holds if all features are arranged in a large loop, a situation often thought of as the worst case for SLAM problems =-=[8]-=- . 6 Experimental Results Systematic experiments showed that FastSLAM 2.0 provides excellent results with surprisingly few particles, including M=1. Most of our experiments were carried out using a be... |

233 |
Stochastic Models, Estimation, and Control, Volume 13
- Maybeck
(Show Context)
Citation Context ... the normalized product of two Gaussians as indicated. However, if g is non-linear, the product will not be Gaussian in general. To make the result Gaussian, FastSLAM employs the standard EKF "tr=-=ick" [13]-=- : g is approximated by a linear function (see below). Under this approximation, (7) is equivalent to the measurement update equation familiar from the EKF literature [13] . In a final step, FastSLAM ... |

183 | Optimization of the simultaneous localization and map building algorithm for real time implementation, in
- Guivant, Nebot
- 2001
(Show Context)
Citation Context ...maps, thereby confining most computation to small regions. Some of these approaches still maintain global correlations among those submaps, hence are quadratic but with a much reduced constant factor =-=[1; 7; 22; 25]-=- . Others restrict the update exclusively to local maps [12] , hence operate in constant time (assuming known data association). A second group of researchers has developed techniques that represent m... |

153 |
Using the SIR Algorithm to Simulate Posterior Distributions
- Rubin
- 1988
(Show Context)
Citation Context ...ure [13] . In a final step, FastSLAM corrects for the fact that the pose sample s [m] t has been generated without consideration of the most recent measurement. It does so by resampling the particles =-=[2-=-0] . The probability for the m-th particle to be sampled (with replacement) is given by the following variable w [m] t , commonly referred to as importance factor: w [m] t = Z p(z t j n t ; s [m] t ;... |

148 | Stochastic simulation algorithms for dynamic probabilistic networks
- Kanazawa, Koller, et al.
- 1995
(Show Context)
Citation Context ... for general particle filters [6] and Markov Chain Monte Carlo techniques for neural networks [4] . It is similar to the arc reversal technique proposed for particle filters applied to Bayes networks =-=[10]-=- , and it is similar to recent work by van der Merwe [24] , who uses an unscented filtering step [9] for generating proposal distributions that accommodate the measurement. While this modification is ... |

145 | The unscented particle filter
- Merwe, Doucet, et al.
- 2000
(Show Context)
Citation Context ...Carlo techniques for neural networks [4] . It is similar to the arc reversal technique proposed for particle filters applied to Bayes networks [10] , and it is similar to recent work by van der Merwe =-=[24]-=- , who uses an unscented filtering step [9] for generating proposal distributions that accommodate the measurement. While this modification is conceptually simple, it has important ramifications. A ke... |

144 | A computationally efficient method for large-scale concurrent mapping and localization
- Leonard, Feder
- 1999
(Show Context)
Citation Context ...se approaches still maintain global correlations among those submaps, hence are quadratic but with a much reduced constant factor [1; 7; 22; 25] . Others restrict the update exclusively to local maps =-=[12]-=- , hence operate in constant time (assuming known data association). A second group of researchers has developed techniques that represent maps through potential functions between adjacent landmarks, ... |

138 | Robust mapping and localization in indoor environments using sonar data, in: Int
- Tardos, Neira, et al.
(Show Context)
Citation Context ...aling SLAM algorithms to maps with more than a few hundred features. It also limits the applicability of SLAM algorithms to problems with ambiguous landmarks, which induces a data association problem =-=[2; 22]-=- . Today's most robust algorithms for SLAM with unknown data association maintain multiple hypotheses (tracks), which increase their computational complexity. Consequently, there has been a flurry on ... |

135 | Bayesian map learning in dynamic environments
- Murphy
(Show Context)
Citation Context ...he categories above. FastSLAM takes advantage of an important characteristic of the SLAM problem (with known data association): landmark estimates are conditionally independent given the robot's path =-=[17]-=- . FastSLAM uses a particle filter to sample over robot paths. Each particle possesses N low-dimensional EKFs, one for each of the N landmarks. This representation requires O(NM) memory, where M is th... |

122 | Thin Junction Tree Filters for Simultaneous Localization and
- Paskin
- 2003
(Show Context)
Citation Context ... techniques that represent maps through potential functions between adjacent landmarks, similar to Markov random fields. The resulting representations require memory linear in the number of landmarks =-=[19; 23]-=- . Under appropriate approximations, such techniques have been shown to provide constant time updating (again for known data association). Unfortunately, no convergence proof exists for any of these e... |

92 | Simultaneous localization and mapping with unknown data association using FastSLAM, in
- Montemerlo, Thrun
- 2003
(Show Context)
Citation Context ...samples and N the number of landmarks in the map. However, the number of particles M needed for convergence remains an open question. FastSLAM has been extended to SLAM with unknown data associations =-=[14]-=- . If the data association is unknown, each particlesm in FastSLAM makes its own local data association decisions^ n [m] t , by maximizing the measurement likelihood. The formula for finding the most ... |

75 | A solution to the simultaneous localisation and map building (slam) problem
- Dissanayake, Newman, et al.
- 1999
(Show Context)
Citation Context ...obot systems are plagued by control noise, but possess relatively accurate sensors. Moreover, they contradict a common belief that maintaining the entire covariance matrix is required for convergence =-=[5]-=- . 2 Simultaneous Localization and Mapping SLAM addresses the problem of simultaneously recovering a map and a vehicle pose from sensor data. The map contains N features (landmarks) and shall be denot... |

58 |
On The Solution to the Simultaneous Localisation and Map Building Problem
- Newman
- 1999
(Show Context)
Citation Context ...ng SLAM. This paper is based on the insights that errors in the map and pose errors are naturally correlated, and that the covariance matrix maintained by the EKF expresses such correlations. Newmann =-=[18]-=- recently proved that the EKF converges for linear SLAM problems, where the motion model and observation model are linear functions with Gaussian noise (see below). Unfortunately, EKF covariance matri... |

47 | Mobile robot localisation and mapping in extensive outdoor environments
- Bailey
- 2002
(Show Context)
Citation Context ...maps, thereby confining most computation to small regions. Some of these approaches still maintain global correlations among those submaps, hence are quadratic but with a much reduced constant factor =-=[1; 7; 22; 25]-=- . Others restrict the update exclusively to local maps [12] , hence operate in constant time (assuming known data association). A second group of researchers has developed techniques that represent m... |

46 | Simultaneous mapping and localization with sparse extended information filters: Theory and initial results
- Thrun, Koller, et al.
(Show Context)
Citation Context ... techniques that represent maps through potential functions between adjacent landmarks, similar to Markov random fields. The resulting representations require memory linear in the number of landmarks =-=[19; 23]-=- . Under appropriate approximations, such techniques have been shown to provide constant time updating (again for known data association). Unfortunately, no convergence proof exists for any of these e... |

44 | An efficient approach to the simultaneous localisation and mapping problem
- Williams, Dissanayake, et al.
- 2002
(Show Context)
Citation Context ...maps, thereby confining most computation to small regions. Some of these approaches still maintain global correlations among those submaps, hence are quadratic but with a much reduced constant factor =-=[1; 7; 22; 25]-=- . Others restrict the update exclusively to local maps [12] , hence operate in constant time (assuming known data association). A second group of researchers has developed techniques that represent m... |

34 | Sequential Monte Carlo Methods to Train Neural Network Models
- Freitas, Noranjan, et al.
(Show Context)
Citation Context ...ed form. This extension parallels prior work by Doucet and colleagues, who proposed a similar modification for general particle filters [6] and Markov Chain Monte Carlo techniques for neural networks =-=[4]-=- . It is similar to the arc reversal technique proposed for particle filters applied to Bayes networks [10] , and it is similar to recent work by van der Merwe [24] , who uses an unscented filtering s... |

2 |
An E cient Approach to the Simultaneous Localisation and Mapping Problem
- Williams, Dissanayake, et al.
- 2002
(Show Context)
Citation Context ...maps, thereby confining most computation to small regions. Some of these approaches still maintain global correlations among those submaps, hence are quadratic but with a much reduced constant factor =-=[1; 7; 22; 25]-=-. Others restrict the update exclusively to local maps [12], hence operate in constant time (assuming known data association). A second group of researchers has developed techniques that represent map... |

1 |
Sequential montc carlo methods to train neural network models
- Freitas, Niranjan, et al.
(Show Context)
Citation Context ...ed form. This extension parallels prior work by Doucet and colleagues, who proposed a similar modification for general particle filters [6] and Markov Chain Monte Carlo techniques for neural networks =-=[4]-=-. It is similar to the arc reversal technique proposed for particle filters applied to ROBOTICS 1151Bayes networks [10], and it is similar to recent work by van der Merwe [24], who uses an unscented ... |

1 |
A new extension of the Kalman fi Iter to nonlinear systems
- Julier, Uhlmann
- 1997
(Show Context)
Citation Context ... to the arc reversal technique proposed for particle filters applied to ROBOTICS 1151Bayes networks [10], and it is similar to recent work by van der Merwe [24], who uses an unscented filtering step =-=[9]-=- for generating proposal distributions that accommodate the measurement. While this modification is conceptually simple, it has important ramifications. A key contribution of this paper is a convergen... |

1 |
Stochastic Models, Estimation, and Control, Volume I
- beck
- 1979
(Show Context)
Citation Context ...is the normalized product of two Gaussians as indicated. However, if is non-linear, the product will not be Gaussian in general. To make the result Gaussian, FastSLAM employs the standard EKF "trick" =-=[13]-=-: g is approximated by a linear function (see below). Under this approximation, (7) is equivalent to the measurement update equation familiar from the EKF literature [13]. In a final step, FastSLAM co... |

1 | Thin junction tree fi Iters for simultaneous localization and mapping - Paskin - 2002 |

1 |
The unscented particle fl Iter
- Merwe, Freitas, et al.
- 2001
(Show Context)
Citation Context ...ques for neural networks [4]. It is similar to the arc reversal technique proposed for particle filters applied to ROBOTICS 1151Bayes networks [10], and it is similar to recent work by van der Merwe =-=[24]-=-, who uses an unscented filtering step [9] for generating proposal distributions that accommodate the measurement. While this modification is conceptually simple, it has important ramifications. A key... |