#### DMCA

## 46 1070-9932/14/$31.00©2014IEEEt IEEE ROBOTICS & AUTOMATION MAGAZINE t JUNE 2014 By

### Citations

10881 |
A Mathematical Theory of Communication
- Shannon
- 1948
(Show Context)
Citation Context ...ar number ofscells. Then, using the distribution, we can extract statisticsssuch as the mean and variance of the estimate.sInformation-Based Control Mutual information is a concept defined by Shannon =-=[18]-=- thatsattempts to quantify the amount that can be learned aboutsone random variable by observing another and is defined ass( , ; ) ( , ; ) ( ) ( ; ) ( , ; )logI x z q p X Z q p X p Z q p X Z q Z zX x ... |

1346 |
Probabilistic Robotics
- Thrun, Burgard, et al.
- 2005
(Show Context)
Citation Context ...ion rather than the collection and return tasks. The use of Bayesian filtering to estimate unknown andsuncertain environments is well established, with many currentsmethods summarized by Thrun et al. =-=[19]-=-. In particular, thesproblem of multiobject tracking has been addressed in severalscontexts, including simultaneous localization and mappings(SLAM), computer vision, and radar-based tracking, using as... |

193 | Multitarget Bayes Filtering via First-Order Multitarget Moments, - Mahler - 2003 |

141 | The Gaussian mixture Probability Hypothesis Density filter,”
- Vo, Ma
- 2006
(Show Context)
Citation Context ...oment ofsthe distribution over RFSs of object locations. The PHD filtershas been used to effectively track an unknown number ofsmoving objects using stationary sensors by Vo et al. [21] andsVo and Ma =-=[20]-=-, among others. Recently, the use of RFSs wassadopted in mobile robotics for feature-based mapping bysMullane et al. [13], [14]. Lundquist et al. [10] used this to create an obstacle map for a vehicle... |

122 |
Statistical Multisource-Multitarget Information Fusion
- Mahler
- 2007
(Show Context)
Citation Context ...bilistic framework that naturally suits problems where the dimension of the state space,si.e., the number of objects and detections, is unknown andspossibly time varying. This was developed by Mahler =-=[11]-=- andsincludes several advantages over traditional methods, mostsnotably removing the need to explicitly consider data associations. The primary tool in this field is the probability hypothesis density... |

121 | Sequential Monte Carlo methods for multitarget filtering with random finite sets,”
- Vo, Singh, et al.
- 2005
(Show Context)
Citation Context ... tracks the first moment ofsthe distribution over RFSs of object locations. The PHD filtershas been used to effectively track an unknown number ofsmoving objects using stationary sensors by Vo et al. =-=[21]-=- andsVo and Ma [20], among others. Recently, the use of RFSs wassadopted in mobile robotics for feature-based mapping bysMullane et al. [13], [14]. Lundquist et al. [10] used this to create an obstacl... |

94 | Information based adaptive robotic exploration,”
- Bourgault, Makarenko, et al.
- 2002
(Show Context)
Citation Context ...all.sUsing this estimate of object locations, the robot thensmoves to maximize the immediate information gain, a strategy sometimes called information surfing [3]. Grocholsky [7]sand Bourgault et al. =-=[2]-=- use mutual information for objectstracking and exploration tasks, but they do not use an analytic computation of the gradient. Hoffmann and Tomlin [8]suse mutual information to localize objects, usin... |

81 | Information-theoretic control of multiple sensor platforms
- Grocholsky
- 2002
(Show Context)
Citation Context ...e number of objects is small.sUsing this estimate of object locations, the robot thensmoves to maximize the immediate information gain, a strategy sometimes called information surfing [3]. Grocholsky =-=[7]-=-sand Bourgault et al. [2] use mutual information for objectstracking and exploration tasks, but they do not use an analytic computation of the gradient. Hoffmann and Tomlin [8]suse mutual information ... |

56 | Mobile sensor network control using mutual information methods and particle filters,”
- Hoffmann, Tomlin
- 2007
(Show Context)
Citation Context ...surfing [3]. Grocholsky [7]sand Bourgault et al. [2] use mutual information for objectstracking and exploration tasks, but they do not use an analytic computation of the gradient. Hoffmann and Tomlin =-=[8]-=-suse mutual information to localize objects, using particle filters to represent object locations and an iterative method toslocally maximize mutual information around the sensorsposition. Julian et a... |

20 |
A random-finite-set approach to Bayesian SLAM,”
- Mullane, Vo, et al.
- 2011
(Show Context)
Citation Context ...ber ofsmoving objects using stationary sensors by Vo et al. [21] andsVo and Ma [20], among others. Recently, the use of RFSs wassadopted in mobile robotics for feature-based mapping bysMullane et al. =-=[13]-=-, [14]. Lundquist et al. [10] used this to create an obstacle map for a vehicle. Our approach differs fromsthese works in that we do not use the PHD filter. Instead, wesrun a Bayesian filter over the ... |

15 |
A multi-robot control policy for information gathering in the presence of unknown hazards
- Schwager, Dames, et al.
- 2011
(Show Context)
Citation Context ...q v p Z X q p X p Z q p Z X q X xZ z2 2 2 2 ; ; = !! //s(7) yields a meaningful result and moves the robot in the directionsof maximum local information gain, following that for a fixedsdistance. See =-=[17]-=- for a derivation of (7). The vector, ,vscould bestaken to be the robot’s pose, ,qsbut a more natural choice for oursproblem is the location of the camera center projected onto thesground plane, i.e.,... |

13 | A scalable information theoretic approach to distributed robot coordination
- Julian, Angermann, et al.
(Show Context)
Citation Context ...e mutual information to localize objects, using particle filters to represent object locations and an iterative method toslocally maximize mutual information around the sensorsposition. Julian et al. =-=[9]-=- use the gradient of mutual information to drive multiple robots for state estimation tasks. All ofsthese previous works only consider a known number ofsobjects. Ristic and Vo [15] and Ristic et al. [... |

12 |
A Note on the Reward Function for PHD Filters with Sensor Control,”
- Ristic, Vo, et al.
- 2011
(Show Context)
Citation Context ...] use the gradient of mutual information to drive multiple robots for state estimation tasks. All ofsthese previous works only consider a known number ofsobjects. Ristic and Vo [15] and Ristic et al. =-=[16]-=- consider thesproblem of localizing an unknown number of objects using assingle robot by maximizing the expected Rényi divergence, asgeneralization of mutual information, to select between asdiscrete ... |

8 | Road intensity based mapping using radar measurements with a probability hypothesis density filter
- Lundquist, Hammarstrand, et al.
- 2011
(Show Context)
Citation Context ...tationary sensors by Vo et al. [21] andsVo and Ma [20], among others. Recently, the use of RFSs wassadopted in mobile robotics for feature-based mapping bysMullane et al. [13], [14]. Lundquist et al. =-=[10]-=- used this to create an obstacle map for a vehicle. Our approach differs fromsthese works in that we do not use the PHD filter. Instead, wesrun a Bayesian filter over the distribution of RFSs under th... |

8 | Sensor control for multi-object state-space estimation using random finite sets,”
- Ristic, Vo
- 2010
(Show Context)
Citation Context ...ition. Julian et al. [9] use the gradient of mutual information to drive multiple robots for state estimation tasks. All ofsthese previous works only consider a known number ofsobjects. Ristic and Vo =-=[15]-=- and Ristic et al. [16] consider thesproblem of localizing an unknown number of objects using assingle robot by maximizing the expected Rényi divergence, asgeneralization of mutual information, to sel... |

7 | Information surfing for radiation map building
- Cortez, Tanner, et al.
(Show Context)
Citation Context ...sumption that the number of objects is small.sUsing this estimate of object locations, the robot thensmoves to maximize the immediate information gain, a strategy sometimes called information surfing =-=[3]-=-. Grocholsky [7]sand Bourgault et al. [2] use mutual information for objectstracking and exploration tasks, but they do not use an analytic computation of the gradient. Hoffmann and Tomlin [8]suse mut... |

7 | A decentralized control policy for adaptive information gathering in hazardous environments
- Dames, Schwager, et al.
(Show Context)
Citation Context ...rse model can be thought of as the probabilitysof returning a good measurement. Therefore, maximizing thissshould lead to faster localization of the objects, a concept weshave used in other work [4], =-=[5]-=-. This differs from the approachsby Ristic and Vo [15], who use the full sensor model but samplesfrom it to achieve computational tractability.sDeriving this binary model is straightforward as the onl... |

5 |
Cooperative Multi-Target Localization With Noisy Sensors
- Dames, Kumar
- 2013
(Show Context)
Citation Context ...s coarse model can be thought of as the probabilitysof returning a good measurement. Therefore, maximizing thissshould lead to faster localization of the objects, a concept weshave used in other work =-=[4]-=-, [5]. This differs from the approachsby Ristic and Vo [15], who use the full sensor model but samplesfrom it to achieve computational tractability.sDeriving this binary model is straightforward as th... |

5 |
Random finite sets for robot mapping and slam
- Mullane, Vo, et al.
- 2011
(Show Context)
Citation Context ...smoving objects using stationary sensors by Vo et al. [21] andsVo and Ma [20], among others. Recently, the use of RFSs wassadopted in mobile robotics for feature-based mapping bysMullane et al. [13], =-=[14]-=-. Lundquist et al. [10] used this to create an obstacle map for a vehicle. Our approach differs fromsthese works in that we do not use the PHD filter. Instead, wesrun a Bayesian filter over the distri... |

1 |
R-MASTIF: Robotic mobile autonomous system for threat interrogation and object fetch
- Das, Thakur, et al.
(Show Context)
Citation Context ...e best performance for objectsdetection since the black-and-white stereo cameras are not assreliable and the LIDAR cannot detect small objects on thesground. For more information on the platform, see =-=[6]-=-.sTo test the performance of our proposed algorithm, wesconduct a series of field tests on the robot. In general, visualssensors can be very noisy, returning false positives due tosother objects in th... |