Predicting object interactions from contact distributions

Kroemer, Oliver; Peters, Jan

doi:10.1109/iros.2014.6943030

Cited by 13 publications

(10 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Other methods have also relied on the geometries of objects and scenes to reason about preferred object placements [11] or likely places to find an object [3]. Moreover, Kroemer and Peters used 3D object models to extract contact point distributions for predicting interactions between objects [15].…”

Section: Related Workmentioning

confidence: 99%

Metric learning for generalizing spatial relations to new objects

Mees

Abdo

Mazuran

et al. 2017

2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

View full text Add to dashboard Cite

Human-centered environments are rich with a wide variety of spatial relations between everyday objects. For autonomous robots to operate effectively in such environments, they should be able to reason about these relations and generalize them to objects with different shapes and sizes. For example, having learned to place a toy inside a basket, a robot should be able to generalize this concept using a spoon and a cup. This requires a robot to have the flexibility to learn arbitrary relations in a lifelong manner, making it challenging for an expert to pre-program it with sufficient knowledge to do so beforehand. In this paper, we address the problem of learning spatial relations by introducing a novel method from the perspective of distance metric learning. Our approach enables a robot to reason about the similarity between pairwise spatial relations, thereby enabling it to use its previous knowledge when presented with a new relation to imitate. We show how this makes it possible to learn arbitrary spatial relations from non-expert users using a small number of examples and in an interactive manner. Our extensive evaluation with realworld data demonstrates the effectiveness of our method in reasoning about a continuous spectrum of spatial relations and generalizing them to new objects.1 The Freiburg Spatial Relations Dataset and a demo video of our approach running on the PR-2 robot are available at

show abstract

Section: Related Workmentioning

confidence: 99%

Metric learning for generalizing spatial relations to new objects

Mees

Abdo

Mazuran

et al. 2017

2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

View full text Add to dashboard Cite

show abstract

“…Most such models are of qualitative effects (Montesano et al 2008;Moldovan et al 2012;Hermans et al 2011;Fitzpatrick et al 2003;Ridge et al 2010;Kroemer and Peters 2014), although metrically precise models have been learned (Meriçli et al 2014;Scholz and Stilman 2010). These learn action-effect correlations.…”

Section: The Importance Of Prediction For Manipulationmentioning

confidence: 99%

“…It has also been used to learn the dynamics of an object with a single, constant contact (such as pole balancing) (Schaal 1997;Atkeson and Schaal 1997). Finally, there has been work on affordance learning, and work on identifying which variables are relevant to predicting object motion (Montesano et al 2008;Moldovan et al 2012;Hermans et al 2011;Fitzpatrick et al 2003;Ridge et al 2010;Kroemer and Peters 2014). The restriction of these papers is that they make qualitative predictions of object motion, such as a classification of the type of motion outcome.…”

Section: Related Workmentioning

confidence: 99%

Learning modular and transferable forward models of the motions of push manipulated objects

et al. 2016

View full text Add to dashboard Cite

The ability to predict how objects behave during manipulation is an important problem. Models informed by mechanics are powerful, but are hard to tune. An alternative is to learn a model of the object's motion from data, to learn to predict. We study this for push manipulation. The paper starts by formulating a quasi-static prediction problem. We then pose the problem of learning to predict in two different frameworks: (i) regression and (ii) density estimation. Our architecture is modular: many simple, object specific, and context specific predictors are learned. We show empirically that such predictors outperform a rigid body dynamics engine tuned on the same data. We then extend the density estimation approach using a product of experts. This allows transfer of learned motion models to objects of novel shape, and to novel actions. With the right representation and learning method, these transferred models can match the prediction performance of a rigid body dynamics engine for novel objects or actions.

show abstract

“…We initialized the model using spectral clustering [8]. Given that phase transitions often correspond to the making or breaking of contacts [12], [19], the similarity between samples was computed using contact distribution kernels [24]. Thus, samples were clustered together if they had similar contacts between the hands and object, as well as the object and the table.…”

Section: Model Learning Using Expectation-maximizationmentioning

confidence: 99%

Towards learning hierarchical skills for multi-phase manipulation tasks

Kroemer

Daniel

Neumann

et al. 2015

2015 IEEE International Conference on Robotics and Automation (ICRA)

Self Cite

103

View full text Add to dashboard Cite

Abstract-Most manipulation tasks can be decomposed into a sequence of phases, where the robot's actions have different effects in each phase. The robot can perform actions to transition between phases and, thus, alter the effects of its actions, e.g. grasp an object in order to then lift it. The robot can thus reach a phase that affords the desired manipulation.In this paper, we present an approach for exploiting the phase structure of tasks in order to learn manipulation skills more efficiently. Starting with human demonstrations, the robot learns a probabilistic model of the phases and the phase transitions. The robot then employs model-based reinforcement learning to create a library of motor primitives for transitioning between phases. The learned motor primitives generalize to new situations and tasks. Given this library, the robot uses a value function approach to learn a high-level policy for sequencing the motor primitives. The proposed method was successfully evaluated on a real robot performing a bimanual grasping task.

show abstract

Predicting object interactions from contact distributions

Cited by 13 publications

References 25 publications

Metric learning for generalizing spatial relations to new objects

Metric learning for generalizing spatial relations to new objects

Learning modular and transferable forward models of the motions of push manipulated objects

Towards learning hierarchical skills for multi-phase manipulation tasks

Contact Info

Product

Resources

About