Visually built task models for robot teams in unstructured environments

Sujan, Vivek A.; Dubowsky, Steven

doi:10.1109/robot.2002.1014800

Cited by 12 publications

(7 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We show that, using this criterion, the robot can efficiently recover its location when the localization system is in the initial stages or when a failure forces the robot to re-localize. The active localization problem has been addressed by some authors before (Kaelbling et al, 1996;Kleinberg, 1994;Kröse and Bunschoten, 1999) and different entropy-based criteria for action selection similar to the one we describe can be found in the literature for object recognition (Arbel and Ferrie, 1999), environment modeling (Sujan and Dubowsky, 2002) or even for robot localization (Burgard et al, 1997;Fox et al, 1998a;Davison, 1999;Maeda et al, 1997). The difference is that, in our case, the entropy-based evaluation can be computed more efficiently than in existing approaches.…”

Section: Introductionmentioning

confidence: 94%

See 1 more Smart Citation

Active Appearance-Based Robot Localization Using Stereo Vision

2005

View full text Add to dashboard Cite

Abstract.A vision-based robot localization system must be robust: able to keep track of the position of the robot at any time even if illumination conditions change and, in the extreme case of a failure, able to efficiently recover the correct position of the robot. With this objective in mind, we enhance the existing appearance-based robot localization framework in two directions by exploiting the use of a stereo camera mounted on a pan-and-tilt device. First, we move from the classical passive appearance-based localization framework to an active one where the robot sometimes executes actions with the only purpose of gaining information about its location in the environment. Along this line, we introduce an entropy-based criterion for action selection that can be efficiently evaluated in our probabilistic localization system. The execution of the actions selected using this criterion allows the robot to quickly find out its position in case it gets lost. Secondly, we introduce the use of depth maps obtained with the stereo cameras. The information provided by depth maps is less sensitive to changes of illumination than that provided by plain images. The main drawback of depth maps is that they include missing values: points for which it is not possible to reliably determine depth information. The presence of missing values makes Principal Component Analysis (the standard method used to compress images in the appearance-based framework) unfeasible. We describe a novel Expectation-Maximization algorithm to determine the principal components of a data set including missing values and we apply it to depth maps. The experiments we present show that the combination of the active localization with the use of depth maps gives an efficient and robust appearance-based robot localization system.

show abstract

Section: Introductionmentioning

confidence: 94%

“…(8) can be fully precomputed and the cost reduces to O(U I 2 ). In any case, the use of the particles to approximate the entropy provides a lower cost than other approaches that discretize the whole configuration space of the robot (Fox et al, 1998a) or the whole field of view of the cameras (Sujan and Dubowsky, 2002).…”

Section: Inputmentioning

confidence: 99%

Active Appearance-Based Robot Localization Using Stereo Vision

2005

View full text Add to dashboard Cite

show abstract

“…Other researchers assume a circle shaped local observation around each agent's current position. We find many examples in map building [66,69,73,75] and search problems [65]. In [52], the authors use a scheme whereby each agent observes the state of the agents located within a certain radius R, and prove stability in the sense that all agents reach a consensus on the heading to adopt.…”

Section: Local Observationmentioning

confidence: 99%

Optimal Agent Cooperation with Local Information

Mot

Féron

2007

Cooperative Control of Distributed Multi‐Agent Systems

View full text Add to dashboard Cite

“…Since the environment (target) motions are known with respect to the sensors, pixel-level data can be fused into a probabilistic map of the target's shape using appropriate sensor noise models. Numerous methods can be used to do this [16,17]. One simplistic approach will be illustrated here.…”

Section: Shape Estimationmentioning

confidence: 99%

State, shape, and parameter estimation of space objects from range images

Lichter

Dubowsky

2004

IEEE International Conference on Robotics and Automation, 2004. Proceedings. ICRA '04. 2004

Self Cite

View full text Add to dashboard Cite

-An architecture for the estimation of dynamic state, geometric shape, and model parameters of objects in orbit using on-orbit cooperative 3-D vision sensors is presented. This has application in many current and projected space missions, such as automated satellite capture and servicing, debris capture and mitigation, and large space structure assembly and maintenance. The method presented here consists of three parts: (1) kinematic data fusion, which condenses sensory data into coarse kinematic surrogate measurements; (2) Kalman filtering, which filters these surrogate measurements and extracts the full dynamic state and model parameters of the target; and (3) shape estimation, which uses filtered pose information and the raw sensory data to build a body-fixed probabilistic map of the target's shape. This method does not rely on feature detection, optical flow, or model matching, but rather exploits the well-modeled dynamics of objects in space using the Kalman filter. The architecture is computationally fast since only coarse measurements need to be provided to the Kalman filter. This paper will illustrate the three steps of the architecture in the context of rigid body (satellite and debris) estimation and flexible structure estimation.

show abstract

Visually built task models for robot teams in unstructured environments

Cited by 12 publications

References 19 publications

Active Appearance-Based Robot Localization Using Stereo Vision

Active Appearance-Based Robot Localization Using Stereo Vision

Optimal Agent Cooperation with Local Information

State, shape, and parameter estimation of space objects from range images

Contact Info

Product

Resources

About