#### DMCA

## Adapting the Sample Size in Particle Filters Through KLD-Sampling (2003)

### Cached

### Download Links

Venue: | International Journal of Robotics Research |

Citations: | 150 - 9 self |

### Citations

3855 |
A new approach to linear filtering and prediction problems
- Kalman
- 1960
(Show Context)
Citation Context ...s filters. their properties in the context of mobile robot localization. An overview of the different algorithms is given in Figure 1. Kalman filters are the most widely used variant of Bayes filters =-=[47, 32, 72, 80]-=-. They approximate beliefs by their first and second moments, i.e. mean and covariance. Kalman filters are optimal under the assumptions that the initial state uncertainty is unimodal Gaussian and tha... |

2004 | A tutorial on particle filters for online nonlinear/non-gaussian bayesian tracking
- Arulampalam, Maskell, et al.
- 2002
(Show Context)
Citation Context ...ult detection [76, 16]. A recent book provides an excellent overview of the state of the art [22]. Particle filters estimate the posterior probability density over the state space of a dynamic system =-=[23, 66, 3]-=-. The key idea of this technique is to represent probability densities by sets of samples. It is due to this representation, that particle filters combine efficiency with the ability to represent a wi... |

1503 | CONDENSATION -- conditional density propagation for visual tracking,
- Isard, Blake
- 1998
(Show Context)
Citation Context ... tracking, mobile robot navigation, and computer vision. Over the last years, particle filters have been applied with great success to a variety of state estimation problems including visual tracking =-=[41]-=-, speech recognition [77], mobile robot localization [28, 54, 43], map building [59], people tracking [69, 60], and fault detection [76, 16]. A recent book provides an excellent overview of the state ... |

1223 |
Applied Optimal Estimation,
- Gelb
- 1974
(Show Context)
Citation Context ...s filters. their properties in the context of mobile robot localization. An overview of the different algorithms is given in Figure 1. Kalman filters are the most widely used variant of Bayes filters =-=[47, 32, 72, 80]-=-. They approximate beliefs by their first and second moments, i.e. mean and covariance. Kalman filters are optimal under the assumptions that the initial state uncertainty is unimodal Gaussian and tha... |

1144 | An Introduction to the Kalman Filter
- Welch, Bishop
- 1995
(Show Context)
Citation Context ...s filters. their properties in the context of mobile robot localization. An overview of the different algorithms is given in Figure 1. Kalman filters are the most widely used variant of Bayes filters =-=[47, 32, 72, 80]-=-. They approximate beliefs by their first and second moments, i.e. mean and covariance. Kalman filters are optimal under the assumptions that the initial state uncertainty is unimodal Gaussian and tha... |

1051 | On sequential Monte Carlo sampling methods for Bayesian filtering,”
- Doucet, Godsill, et al.
- 2000
(Show Context)
Citation Context ...ult detection [76, 16]. A recent book provides an excellent overview of the state of the art [22]. Particle filters estimate the posterior probability density over the state space of a dynamic system =-=[23, 66, 3]-=-. The key idea of this technique is to represent probability densities by sets of samples. It is due to this representation, that particle filters combine efficiency with the ability to represent a wi... |

838 | Monte carlo localization for mobile robots
- Dellaert, Fox, et al.
- 1999
(Show Context)
Citation Context .... Auxiliary particle filters have been applied recently to robot localization [78]. Along a similar line of reasoning, the injection of observation samples into the posterior can be very advantageous =-=[54, 74, 31]-=-. However, this approach requires the availability of a sensor model from which it is possible to efficiently generate samples. In this paper we introduce an approach to increasing the efficiency of p... |

778 | A new extension of the kalman filter to nonlinear systems,
- Julier, Uhlmann, et al.
- 1997
(Show Context)
Citation Context ... noise. Non-linearities are typically approximated by linearization at the current state, resulting in the extended Kalman filter (EKF). Recently, the unscented Kalman filter (UF) has been introduced =-=[45, 79]-=-. This approach deterministically generates samples (sigma-points) taken from the Gaussian state and passes these samples through the non-linear dynamics, followed by a Gaussian approximation of the p... |

775 | Filtering via simulation : auxiliary particle filters.
- Pitt, Shephard
- 1999
(Show Context)
Citation Context ...ult detection [76, 16]. A recent book provides an excellent overview of the state of the art [22]. Particle filters estimate the posterior probability density over the state space of a dynamic system =-=[23, 66, 3]-=-. The key idea of this technique is to represent probability densities by sets of samples. It is due to this representation, that particle filters combine efficiency with the ability to represent a wi... |

663 | Sequential Monte Carlo methods for dynamical system,
- Liu, Chen
- 1998
(Show Context)
Citation Context ...o one. The basic form of the particle filter realizes the recursive Bayes filter according to a sampling procedure, often referred to as sequential importance sampling with resampling (SISR, see also =-=[57, 23, 22]-=-). A time update of the basic particle filter algorithm is outlined in Table 1. At each iteration, the algorithm receives a sample set St−1 representing the previous belief of the robot, a control mea... |

599 | FastSLAM: A Factored Solution to the Simultaneous Localization and Mapping Problem,
- Montemerlo, Thrun, et al.
- 2002
(Show Context)
Citation Context ...le filters have been applied with great success to a variety of state estimation problems including visual tracking [41], speech recognition [77], mobile robot localization [28, 54, 43], map building =-=[59]-=-, people tracking [69, 60], and fault detection [76, 16]. A recent book provides an excellent overview of the state of the art [22]. Particle filters estimate the posterior probability density over th... |

590 |
Monte Carlo filter and smoother for non-Gaussian nonlinear state space models,
- Kitagawa
- 1996
(Show Context)
Citation Context ...pplied, with Bel(xt) as target distribution 2 Resampling with minimal variance can be implemented efficiently (constant time per sample) using a procedure known under the name deterministic selection =-=[48, 3]-=- or stochastic universal sampling [5]. 6s1. Inputs: St−1 = {〈x (i) t−1 , w(i) t−1 control measurement ut−1, observation zt 〉 | i = 1, . . . , n} representing belief Bel(xt−1), 2. St := ∅, α := 0 // In... |

520 |
Estimating uncertain spatial relationships in robotics,” in Autonomous Robot Vehicles,
- Smith, Self, et al.
- 1988
(Show Context)
Citation Context |

505 | A solution to the simultaneous localization and map building (SLAM) problem
- Dissanayake, Newman, et al.
(Show Context)
Citation Context ...to the simultaneous localization and map building problem (SLAM), where they estimate full posteriors over both robot positions and landmark positions (typically consisting of hundreds of dimensions) =-=[15, 20, 56, 10, 17]-=-. However, due to the assumption of unimodal posteriors, Kalman filters applied to robot localization solely aim at tracking a robot’s location. They are not designed to globally localize a robot from... |

443 | Mathematical Statistics and Data Analysis,’’ - Rice - 1988 |

430 |
Reducing bias and inefficiency in the selection algorithm,” in
- Baker
- 1987
(Show Context)
Citation Context ...on 2 Resampling with minimal variance can be implemented efficiently (constant time per sample) using a procedure known under the name deterministic selection [48, 3] or stochastic universal sampling =-=[5]-=-. 6s1. Inputs: St−1 = {〈x (i) t−1 , w(i) t−1 control measurement ut−1, observation zt 〉 | i = 1, . . . , n} representing belief Bel(xt−1), 2. St := ∅, α := 0 // Initialize 3. for i := 1, . . . , n do ... |

361 | Markov localization for mobile robots in dynamic environments.
- Fox, Burgard, et al.
- 1999
(Show Context)
Citation Context ...position estimates in combination with high robustness to sensor noise. A grid-based method has been applied successfully for the position estimation of the museum tour-guide robots Rhino and Minerva =-=[7, 73, 30]-=-. A disadvantage of grid-based approaches lies in their computational complexity, based on the requirement to keep the typically three-dimensional position probability grid in memory and to update it ... |

355 |
Mobile robot localization by tracking geometric beacons
- Leonard, Durrant-Whyte
- 1991
(Show Context)
Citation Context ...ctive assumptions, Kalman filters have been applied with great success to mobile robot localization, where they yield very efficient and accurate position estimates even for highly non-linear systems =-=[55, 58, 2, 39]-=-. One of the key advantages of Kalman filters is their efficiency. The complexity is polynomial in the dimensionality of the state space and the observations. Due to this graceful increase in complexi... |

348 | Rao-blackwellised particle filtering for dynamic bayesian networks
- Doucet, Freitas, et al.
- 2000
(Show Context)
Citation Context ...al sub-spaces can be represented using the most adequate representation, such as continuous densities, samples, or discrete values [50]. Recently, under the name of Rao-Blackwellised particle filters =-=[21, 16, 36]-=-, the combination of particle filters with Kalman filters yielded extremely robust and efficient approaches to higher dimensional state estimation including full posteriors over robot positions and ma... |

343 | Monte carlo localization: Efficient position estimation for mobile robots
- Fox, Burgard, et al.
- 1999
(Show Context)
Citation Context ...Over the last years, particle filters have been applied with great success to a variety of state estimation problems including visual tracking [41], speech recognition [77], mobile robot localization =-=[28, 54, 43]-=-, map building [59], people tracking [69, 60], and fault detection [76, 16]. A recent book provides an excellent overview of the state of the art [22]. Particle filters estimate the posterior probabil... |

339 | The Spatial Semantic Hierarchy.
- Kuipers
- 2000
(Show Context)
Citation Context ...ed on symbolic, graph structured representations of the environment. The state space of the robot consists of a set of discrete, locally distinctive locations such as corners or crossings of hallways =-=[46, 70, 52, 40, 11, 53]-=-. The advantage of these approaches lies in their efficiency and in the fact that they can represent arbitrary distributions over the discrete state space. Therefore they can solve the global localiza... |

329 | Experiences with an interactive museum tour-guide robot
- Burgard, Cremers, et al.
- 1999
(Show Context)
Citation Context ...position estimates in combination with high robustness to sensor noise. A grid-based method has been applied successfully for the position estimation of the museum tour-guide robots Rhino and Minerva =-=[7, 73, 30]-=-. A disadvantage of grid-based approaches lies in their computational complexity, based on the requirement to keep the typically three-dimensional position probability grid in memory and to update it ... |

302 |
Elements of Information Theory, Wiley Series in Telecommunications
- Cover, Thomas
- 1991
(Show Context)
Citation Context ...ied threshold ε. We denote the resulting approach as the KLD-sampling algorithm since the distance between the MLE and the true distribution is measured by the Kullback-Leibler distance (KL-distance) =-=[12]-=-. The KL-distance is a measure of the difference between two probability distributions p and q: K(p, q) = � x p(x)log p(x) q(x) KL-distance is never negative and it is zero if and only if the two dist... |

293 | Probabilistic robot navigation in partially observable environments.
- Simmons, Koenig
- 1995
(Show Context)
Citation Context ...unimodal probability densities [9]. Only recently, several approaches have been introduced that can solve the global localization problem, among them grid-based approaches [9], topological approaches =-=[46, 70]-=-, particle filters [28], and multihypothesis tracking [4, 68]. The most challenging problem in mobile robot localization is the kidnapped robot problem [25], in which a well-localized robot is telepor... |

278 | Robot Pose Estimation in Unknown Environments by Matching 2D Range Scans.”
- Lu, E, et al.
- 1994
(Show Context)
Citation Context ...ctive assumptions, Kalman filters have been applied with great success to mobile robot localization, where they yield very efficient and accurate position estimates even for highly non-linear systems =-=[55, 58, 2, 39]-=-. One of the key advantages of Kalman filters is their efficiency. The complexity is polynomial in the dimensionality of the state space and the observations. Due to this graceful increase in complexi... |

239 | A probabilistic approach to collaborative multi-robot localization. Autonomous robots, 8(3):325– 344,
- Fox, Burgard, et al.
- 2000
(Show Context)
Citation Context ...les (steps 9-12; for clarity we omitted the fact that nχ can be determined only when k > 1). The bins can be implemented either as a fixed, multidimensional grid, or more compactly as tree structures =-=[49, 61, 64, 75, 29]-=-. Note that the sampling process is guaranteed to terminate, since for a given bin size ∆, the maximum number k of bins is limited, which also limits the maximum number nχ of desired samples. To summa... |

234 |
Der Merwe, The unscented Kalman filter for nonlinear estimation,
- Wan, Van
- 2000
(Show Context)
Citation Context ... noise. Non-linearities are typically approximated by linearization at the current state, resulting in the extended Kalman filter (EKF). Recently, the unscented Kalman filter (UF) has been introduced =-=[45, 79]-=-. This approach deterministically generates samples (sigma-points) taken from the Gaussian state and passes these samples through the non-linear dynamics, followed by a Gaussian approximation of the p... |

224 | Topological simultaneous localization and mapping (SLAM): Toward exact localization without explicit localization,”
- Choset, Nagatani
- 2001
(Show Context)
Citation Context ...ed on symbolic, graph structured representations of the environment. The state space of the robot consists of a set of discrete, locally distinctive locations such as corners or crossings of hallways =-=[46, 70, 52, 40, 11, 53]-=-. The advantage of these approaches lies in their efficiency and in the fact that they can represent arbitrary distributions over the discrete state space. Therefore they can solve the global localiza... |

223 |
An experiment in guidance and navigation of an autonomous robot vehicle”,
- Cox
- 1991
(Show Context)
Citation Context ...g a robot’s pose relative to a map of its environment. The localization problem is occasionally referred to as “ the most fundamental problem to providing a mobile robot with autonomous capabilities” =-=[13]-=-. The mobile robot localization problem comes in different flavors. The most simple localization problem is position tracking. Here the initial robot pose is known, and localization seeks to identify ... |

219 | Particle filters for positioning, navigation, and tracking.
- Gustafsson, Gunnarsson, et al.
- 2002
(Show Context)
Citation Context ...s exponentially with the number of dimensions, it is doubtful whether they can be applied to higherdimensional state spaces. Sample-based approaches represent beliefs by sets of samples, or particles =-=[28, 19, 31, 54, 43, 36]-=-. A key advantage of particle filters is their ability to represent arbitrary probability densities, which is why they can solve the global localization problem. Furthermore, 1 Topological and in part... |

218 |
Following a moving target - Monte Carlo inference for dynamic Bayesian models.
- Gilks, Berzuini
- 2001
(Show Context)
Citation Context ...re most important. 1sSince the complexity of particle filters depends on the number of samples used for estimation, several attempts have been made to make more efficient use of the available samples =-=[34, 66, 35]-=-. So far, however, an important source for increasing the efficiency of particle filters has only rarely been studied: Adapting the number of samples over time. While sample sizes have been discussed ... |

218 | Bayesian statistics without tears: A sampling-resampling perspective.
- Smith, Gelfand
- 1992
(Show Context)
Citation Context ...for i := 1, . . . , n do // Normalize importance weights 10. w (i) t := w (i) t /α 11. return St Table 1: The basic particle filter algorithm. and p(xt | xt−1, ut−1)Bel(xt−1) as proposal distribution =-=[71, 31]-=-. By dividing these two distributions, we get p(zt | x (i) t ) as the importance weight for each sample (see Eq.(2)). Step 7 keeps track of the normalization factor, and Step 8 inserts the new sample ... |

202 | Acting under Uncertainty: Discrete Bayesian Models for Mobile-Robot Navigation.
- Cassandra, Kaelbling
- 1996
(Show Context)
Citation Context ...unimodal probability densities [9]. Only recently, several approaches have been introduced that can solve the global localization problem, among them grid-based approaches [9], topological approaches =-=[46, 70]-=-, particle filters [28], and multihypothesis tracking [4, 68]. The most challenging problem in mobile robot localization is the kidnapped robot problem [25], in which a well-localized robot is telepor... |

200 | Estimating the absolute position of a mobile robot using position probability grids,
- Burgard, Fox, et al.
- 1996
(Show Context)
Citation Context ...but instead has to determine it from scratch. The global localization problem is more difficult, since the robot’s position estimate cannot be represented adequately by unimodal probability densities =-=[9]-=-. Only recently, several approaches have been introduced that can solve the global localization problem, among them grid-based approaches [9], topological approaches [46, 70], particle filters [28], a... |

196 | Probabilistic algorithms and the interactive museum tour-guide robot minerva,”
- Thrun, Beetz, et al.
- 2000
(Show Context)
Citation Context ...position estimates in combination with high robustness to sensor noise. A grid-based method has been applied successfully for the position estimation of the museum tour-guide robots Rhino and Minerva =-=[7, 73, 30]-=-. A disadvantage of grid-based approaches lies in their computational complexity, based on the requirement to keep the typically three-dimensional position probability grid in memory and to update it ... |

185 | An experimental comparison of localization methods continued.
- Gutmann, Fox
- 2002
(Show Context)
Citation Context ... designed to globally localize a robot from scratch in arbitrarily large environments. Some of the limitations of Kalman filters in the context of robot localization have been shown experimentally in =-=[37]-=-. 4sMulti-hypothesis approaches represent the belief state by mixtures of Gaussians [4, 68, 42, 1]. Each hypothesis, or Gaussian, is typically tracked by an extended Kalman filter. Due to their abilit... |

161 | A computationally efficient method for large-scale concurrent mapping and localization,”
- Leonard, Feder
- 2000
(Show Context)
Citation Context ...to the simultaneous localization and map building problem (SLAM), where they estimate full posteriors over both robot positions and landmark positions (typically consisting of hundreds of dimensions) =-=[15, 20, 56, 10, 17]-=-. However, due to the assumption of unimodal posteriors, Kalman filters applied to robot localization solely aim at tracking a robot’s location. They are not designed to globally localize a robot from... |

158 | Sensor resetting localization for poorly modelled mobile robots.
- Lenser, Veloso
- 2000
(Show Context)
Citation Context ...Over the last years, particle filters have been applied with great success to a variety of state estimation problems including visual tracking [41], speech recognition [77], mobile robot localization =-=[28, 54, 43]-=-, map building [59], people tracking [69, 60], and fault detection [76, 16]. A recent book provides an excellent overview of the state of the art [22]. Particle filters estimate the posterior probabil... |

155 | Tracking multiple moving targets with a mobile robot using particle filters and statistical data association.
- Schulz, Burgard, et al.
- 2001
(Show Context)
Citation Context ...pplied with great success to a variety of state estimation problems including visual tracking [41], speech recognition [77], mobile robot localization [28, 54, 43], map building [59], people tracking =-=[69, 60]-=-, and fault detection [76, 16]. A recent book provides an excellent overview of the state of the art [22]. Particle filters estimate the posterior probability density over the state space of a dynamic... |

154 | The spmap: A probabilistic framework for simultaneous localization and map building.
- Castellanos, Montiel, et al.
- 1999
(Show Context)
Citation Context ...to the simultaneous localization and map building problem (SLAM), where they estimate full posteriors over both robot positions and landmark positions (typically consisting of hundreds of dimensions) =-=[15, 20, 56, 10, 17]-=-. However, due to the assumption of unimodal posteriors, Kalman filters applied to robot localization solely aim at tracking a robot’s location. They are not designed to globally localize a robot from... |

152 |
Branching and interacting particle systems approximations of Feynman–Kac formulae with applications to non-linear #ltering.
- Moral, Miclo, et al.
- 2000
(Show Context)
Citation Context ...ticle filters has only rarely been studied: Adapting the number of samples over time. While sample sizes have been discussed in the context of genetic algorithms [65] and interacting particle filters =-=[18]-=-, most existing approaches to particle filters use a fixed number of samples during the entire state estimation process. This can be highly inefficient, since the complexity of the probability densiti... |

139 | A review of statistical data association techniques for motion correspondence,
- Cox
- 1993
(Show Context)
Citation Context ...sed in [19]. In addition to pure Kalman filtering, multi-hypothesis approaches require sophisticated heuristics to solve the data association problem and to determine when to add or delete hypotheses =-=[6, 14]-=-. Topological approaches are based on symbolic, graph structured representations of the environment. The state space of the robot consists of a set of discrete, locally distinctive locations such as c... |

137 | Using the condensation algorithm for robust, vision-based mobile robot localization
- Dellaert, Burgard, et al.
- 1999
(Show Context)
Citation Context ...is approaches have been shown to be very robust to violations of these assumptions. So far it is not clear how these methods can be applied to extremely non-linear observations, such as those used in =-=[19]-=-. In addition to pure Kalman filtering, multi-hypothesis approaches require sophisticated heuristics to solve the data association problem and to determine when to add or delete hypotheses [6, 14]. To... |

132 | Conditional particle filters for simultaneous mobile robot localization and people-tracking”,
- Montemerlo, Thrun, et al.
- 2002
(Show Context)
Citation Context ...pplied with great success to a variety of state estimation problems including visual tracking [41], speech recognition [77], mobile robot localization [28, 54, 43], map building [59], people tracking =-=[69, 60]-=-, and fault detection [76, 16]. A recent book provides an excellent overview of the state of the art [22]. Particle filters estimate the posterior probability density over the state space of a dynamic... |

111 | Particle Filters for Mobile Robot Localization. In
- Fox, Thrun, et al.
- 2001
(Show Context)
Citation Context ...s exponentially with the number of dimensions, it is doubtful whether they can be applied to higherdimensional state spaces. Sample-based approaches represent beliefs by sets of samples, or particles =-=[28, 19, 31, 54, 43, 36]-=-. A key advantage of particle filters is their ability to represent arbitrary probability densities, which is why they can solve the global localization problem. Furthermore, 1 Topological and in part... |

96 |
Modeling a dynamic environment using a Bayesian multiple hypothesis approach
- Cox, Leonard
- 1994
(Show Context)
Citation Context |

94 | Bootstrap learning for place recognition,” in
- Kuipers, Beeson
- 2002
(Show Context)
Citation Context ...ed on symbolic, graph structured representations of the environment. The state space of the robot consists of a set of discrete, locally distinctive locations such as corners or crossings of hallways =-=[46, 70, 52, 40, 11, 53]-=-. The advantage of these approaches lies in their efficiency and in the fact that they can represent arbitrary distributions over the discrete state space. Therefore they can solve the global localiza... |

94 |
Acting under uncertainty: Discrete bayesian models for mobile robot navigation,”
- Kaelbling, Cassandra, et al.
- 1996
(Show Context)
Citation Context ...unimodal probability densities [9]. Only recently, several approaches have been introduced that can solve the global localization problem, among them grid-based approaches [9], topological approaches =-=[46, 70]-=-, particle filters [28], and multihypothesis tracking [4, 68]. The most challenging problem in mobile robot localization is the kidnapped robot problem [25], in which a well-localized robot is telepor... |

90 | An introduction to hidden markov models and bayesian networks.
- Ghahramani
- 2001
(Show Context)
Citation Context ...nces in the structure of the state space to break the state into lower-dimensional sub-spaces (random variables). Such structured representations are known under the name of dynamic Bayesian networks =-=[33]-=-. The individual sub-spaces can be represented using the most adequate representation, such as continuous densities, samples, or discrete values [50]. Recently, under the name of Rao-Blackwellised par... |

89 |
Rao-Blackwellised particle filtering for dynamic Bayesian networks.
- Murphy, Russell
- 2001
(Show Context)
Citation Context ...he combination of particle filters with Kalman filters yielded extremely robust and efficient approaches to higher dimensional state estimation including full posteriors over robot positions and maps =-=[62, 59]-=-. 2.3 Particle filters Particle filters are a variant of Bayes filters which represent the belief Bel(xt) by a set St of n weighted samples distributed according to Bel(xt): St = {〈x (i) t , w (i) t 〉... |

88 |
Markov localization: A probabilistic framework for mobile robot localization and navigation
- Fox
- 1998
(Show Context)
Citation Context ...can solve the global localization problem. Furthermore, 1 Topological and in particular grid-based implementations of Bayes filters for robot localization are often referred to as Markov localization =-=[26]-=-. 5sparticle filters can be shown to converge to the true posterior even in non-Gaussian, nonlinear dynamic systems [18]. Compared to grid-based approaches, particle filters are very efficient since t... |

88 | KLD-sampling: Adaptive particle filters
- Fox
- 2001
(Show Context)
Citation Context ...possible to efficiently generate samples. In this paper we introduce an approach to increasing the efficiency of particle filters by adapting the number of samples to the underlying state uncertainty =-=[27]-=-. The localization example in Figure 3 illustrates when such an approach can be very beneficial. In the beginning of global localization, the robot is highly uncertain and a large number of samples is... |

88 | Efficient locally weighted polynomial regression predictions
- Moore, Schneider, et al.
- 1997
(Show Context)
Citation Context ...o derive a bound on the approximation error, we assume that the true posterior is given by a discrete, piecewise constant distribution such as a discrete density tree or a multi-dimensional histogram =-=[49, 61, 64, 75]-=-. For such a representation we show how to determine the number of samples so that the distance between the sample-based maximum likelihood estimate (MLE) and the true posterior does not exceed a pre-... |

79 | Integrating global position estimation and position tracking for mobile robots: The dynamic markov localization approach,
- Burgard, Derr, et al.
- 1998
(Show Context)
Citation Context ...f the environment can be used, and adequate features might not be available in arbitrary environments. Grid-based, metric approaches rely on discrete, piecewise constant representations of the belief =-=[9, 8, 51, 63]-=-. For indoor localization, the spatial resolution of these grids is usually between 10 and 40 cm and the angular resolution is usually 5 degrees. As the topological approaches, these methods can repre... |

71 |
Continuous Univariate Distributions volume 1
- Johnson, Kotz, et al.
- 1994
(Show Context)
Citation Context ... densities). In what follows, we will first determine the number of samples needed to achieve, with high probability, a good approximation of an arbitrary, discrete probability distribution (see also =-=[67, 44]-=-). Then we will show how to modify the basic particle filter algorithm so that it realizes our adaptation approach. To see, suppose that n samples are drawn from a discrete distribution with k differe... |

68 |
Error correction in mobile robot map learning,”
- Engelson, McDermott
- 1992
(Show Context)
Citation Context ... approaches [9], topological approaches [46, 70], particle filters [28], and multihypothesis tracking [4, 68]. The most challenging problem in mobile robot localization is the kidnapped robot problem =-=[25]-=-, in which a well-localized robot is teleported to some other position without being told. This problem differs from the global localization problem in that the robot might firmly believe to be somewh... |

66 | Markov Localization Using Correlation.
- Konolige, Chou
- 1999
(Show Context)
Citation Context ...f the environment can be used, and adequate features might not be available in arbitrary environments. Grid-based, metric approaches rely on discrete, piecewise constant representations of the belief =-=[9, 8, 51, 63]-=-. For indoor localization, the spatial resolution of these grids is usually between 10 and 40 cm and the angular resolution is usually 5 degrees. As the topological approaches, these methods can repre... |

62 | Rao-Blackwellised Particle Filtering for Fault Diagnosis,”
- Freitas
- 2002
(Show Context)
Citation Context ... variety of state estimation problems including visual tracking [41], speech recognition [77], mobile robot localization [28, 54, 43], map building [59], people tracking [69, 60], and fault detection =-=[76, 16]-=-. A recent book provides an excellent overview of the state of the art [22]. Particle filters estimate the posterior probability density over the state space of a dynamic system [23, 66, 3]. The key i... |

61 | Probabilistic self-localization for mobile robots.
- Olson
- 2000
(Show Context)
Citation Context ...f the environment can be used, and adequate features might not be available in arbitrary environments. Grid-based, metric approaches rely on discrete, piecewise constant representations of the belief =-=[9, 8, 51, 63]-=-. For indoor localization, the spatial resolution of these grids is usually between 10 and 40 cm and the angular resolution is usually 5 degrees. As the topological approaches, these methods can repre... |

60 | Using learning for approximation in stochastic processes
- Koller, Fratkina
- 1998
(Show Context)
Citation Context ...on process. This can be highly inefficient, since the complexity of the probability densities can vary drastically over time. Previously, an adaptive approach for particle filters has been applied by =-=[49]-=- and [28]. This approach adjusts the number of samples based on the likelihood of observations, which has some important shortcomings, as we will show. In this paper we introduce a novel approach to a... |

51 | Bumptrees for efficient function, constraint, and classification learning.
- Omohundro
- 1991
(Show Context)
Citation Context ...o derive a bound on the approximation error, we assume that the true posterior is given by a discrete, piecewise constant distribution such as a discrete density tree or a multi-dimensional histogram =-=[49, 61, 64, 75]-=-. For such a representation we show how to determine the number of samples so that the distance between the sample-based maximum likelihood estimate (MLE) and the true posterior does not exceed a pre-... |

51 | Robust vision-based localization for mobile robots using an image retrieval system based on invariant features”, in - Wolf, Burgard, et al. - 2002 |

49 | Bayesian estimation and Kalman filtering: A unified framework for mobile robot localization,
- Roumeliotis, Bekey
- 2000
(Show Context)
Citation Context ...roaches have been introduced that can solve the global localization problem, among them grid-based approaches [9], topological approaches [46, 70], particle filters [28], and multihypothesis tracking =-=[4, 68]-=-. The most challenging problem in mobile robot localization is the kidnapped robot problem [25], in which a well-localized robot is teleported to some other position without being told. This problem d... |

49 | Particle methods for Bayesian modelling and enhancement of speech signals,”
- Vermaak, Andrieu, et al.
- 2002
(Show Context)
Citation Context ...avigation, and computer vision. Over the last years, particle filters have been applied with great success to a variety of state estimation problems including visual tracking [41], speech recognition =-=[77]-=-, mobile robot localization [28, 54, 43], map building [59], people tracking [69, 60], and fault detection [76, 16]. A recent book provides an excellent overview of the state of the art [22]. Particle... |

47 |
Feature-based multi-hypothesis localization and tracking using geometric constraints. Robotics and Autonomous Systems,
- Arras, Castellanos, et al.
- 2003
(Show Context)
Citation Context ...ome of the limitations of Kalman filters in the context of robot localization have been shown experimentally in [37]. 4sMulti-hypothesis approaches represent the belief state by mixtures of Gaussians =-=[4, 68, 42, 1]-=-. Each hypothesis, or Gaussian, is typically tracked by an extended Kalman filter. Due to their ability to represent multi-modal beliefs, these approaches are able to solve the global localization pro... |

46 | Active Global Localisation for a Mobile Robot Using Multiple Hypothesis Tracking.
- Jensfelt, Kristensen
- 1999
(Show Context)
Citation Context ...ome of the limitations of Kalman filters in the context of robot localization have been shown experimentally in [37]. 4sMulti-hypothesis approaches represent the belief state by mixtures of Gaussians =-=[4, 68, 42, 1]-=-. Each hypothesis, or Gaussian, is typically tracked by an extended Kalman filter. Due to their ability to represent multi-modal beliefs, these approaches are able to solve the global localization pro... |

46 | Bayesian optimization algorithm, population size, and time to convergence
- Pelikan, Goldberg, et al.
- 2000
(Show Context)
Citation Context ...e for increasing the efficiency of particle filters has only rarely been studied: Adapting the number of samples over time. While sample sizes have been discussed in the context of genetic algorithms =-=[65]-=- and interacting particle filters [18], most existing approaches to particle filters use a fixed number of samples during the entire state estimation process. This can be highly inefficient, since the... |

46 | Auxiliary particle filter robot localization from high-dimensional sensor observations
- Vlassis, Terwijn, et al.
- 2002
(Show Context)
Citation Context ... p(zt | xt) describes the likelihood of making the observation zt given that the robot is at location xt. Particle filters have been applied to a variety of robot platforms and sensors such as vision =-=[19, 54, 24, 81, 78, 38]-=- and proximity sensors [28, 31, 43]. Figure 3 illustrates the application of particle filters to mobile robot localization using sonar sensors. Shown there is a map of a hallway environment along with... |

42 | Improvement strategies for Monte Carlo particle filters,” in Sequential Monte Carlo Methods in - Godsill, Clapp - 2001 |

39 | Feature based condensation for mobile robot localization
- Jensfelt, Wijk, et al.
- 2000
(Show Context)
Citation Context ...Over the last years, particle filters have been applied with great success to a variety of state estimation problems including visual tracking [41], speech recognition [77], mobile robot localization =-=[28, 54, 43]-=-, map building [59], people tracking [69, 60], and fault detection [76, 16]. A recent book provides an excellent overview of the state of the art [22]. Particle filters estimate the posterior probabil... |

29 | Landmark-based autonomous navigation in sewerage pipes
- Hertzberg, Kirchner
- 1996
(Show Context)
Citation Context |

26 | Nonparametric fault identification for space rovers
- Verma, Langford, et al.
- 2001
(Show Context)
Citation Context ... variety of state estimation problems including visual tracking [41], speech recognition [77], mobile robot localization [28, 54, 43], map building [59], people tracking [69, 60], and fault detection =-=[76, 16]-=-. A recent book provides an excellent overview of the state of the art [22]. Particle filters estimate the posterior probability density over the state space of a dynamic system [23, 66, 3]. The key i... |

25 | Using Multiple Gaussian Hypotheses to Represent Probability Distributions for Mobile Robot Localization.
- Austin, Jensfelt
- 2000
(Show Context)
Citation Context ...roaches have been introduced that can solve the global localization problem, among them grid-based approaches [9], topological approaches [46, 70], particle filters [28], and multihypothesis tracking =-=[4, 68]-=-. The most challenging problem in mobile robot localization is the kidnapped robot problem [25], in which a well-localized robot is teleported to some other position without being told. This problem d... |

25 | Sampling in factored dynamic systems. - Koller, Lemer - 2001 |

24 | Monte carlo hidden markov models: Learning non-parametric models of partially observable stochastic processes
- Thrun, Langford, et al.
- 1999
(Show Context)
Citation Context ...o derive a bound on the approximation error, we assume that the true posterior is given by a discrete, piecewise constant distribution such as a discrete density tree or a multi-dimensional histogram =-=[49, 61, 64, 75]-=-. For such a representation we show how to determine the number of samples so that the distance between the sample-based maximum likelihood estimate (MLE) and the true posterior does not exceed a pre-... |

23 |
robust self-localization in polygonal environments
- Fast
- 1999
(Show Context)
Citation Context ...ctive assumptions, Kalman filters have been applied with great success to mobile robot localization, where they yield very efficient and accurate position estimates even for highly non-linear systems =-=[55, 58, 2, 39]-=-. One of the key advantages of Kalman filters is their efficiency. The complexity is polynomial in the dimensionality of the state space and the observations. Due to this graceful increase in complexi... |

10 | Maximally informative statistics for localization and mapping,”
- Deans
- 2002
(Show Context)
Citation Context |

8 | Soccer-robot locatization using sporadic visual features - Enderle, Ritter, et al. - 2000 |

7 |
Error Correction in Mobile Robot Map Learning
- Engleson, McDermott
- 1992
(Show Context)
Citation Context ... approaches [9], topological approaches [46, 70], particle filters [28], and multihypothesis tracking [4, 68]. The most challenging problem in mobile robot localization is the kidnapped robot problem =-=[25]-=-, in which a well-localized robot is teleported to some other position without being told. This problem differs from the global localization problem in that the robot might firmly believe to be somewh... |

6 |
high-precision localization for the mail distributing mobile robot system MOPS
- Hybrid
- 1998
(Show Context)
Citation Context |

6 |
Auxiliary particle filter robot localization from high-dimensional sensor observations
- Vlassis, Terwijn, et al.
- 2002
(Show Context)
Citation Context ...nimizing the variability of the importance weights, which in turn determines the efficiency of the importance sampler [66]. Auxiliary particle filters have been applied recently to robot localization =-=[78]-=-. Along a similar line of reasoning, the injection of observation samples into the posterior can be very advantageous [54, 74, 31]. However, this approach requires the availability of a sensor model f... |