Contact State Segmentation Using Particle Filters for Programming by Human Demonstration in Compliant Motion Tasks

Meeussen, Wim; Rutgeerts, Johan; Gadeyne, Klaas; Bruyninckx, Herman; Schutter, Joris De

doi:10.1007/978-3-540-77457-0_1

Cited by 7 publications

(6 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Gadeyne et al (2005) used a particle filter to track a hybrid discrete-continuous probability distribution over a small set of contact states (discrete) and object pose (continuous). The estimator was later scaled to the full set of possible contact states by using a pre-computed contact state graph (Xiao 1993) to generate a sparse transition model between discrete contact state (Meeussen et al 2007). Constructing the contact state graph offline mirrors the MPF's pre-computation of the contact manifold (Section 6.3).…”

Section: Contact State Estimationmentioning

confidence: 99%

“…In contrast, the algorithm presented in Meeussen et al (2007) explicitly maintains a distribution over contact states. Tracking a single distribution guarantees that MPF's pose estimate incorporates the contact state constraints imposed by contact observations.…”

Section: Contact State Estimationmentioning

confidence: 99%

“…Due to this resampling, the MPF fundamentally differs from related particle filtering techniques (Gadeyne et al 2005, Meeussen et al 2007) that track a distribution over contact formations (Xiao 1993). In the general case, estimating the distribution b(s t ∈ M j ) over manifolds is as hard as solving the filtering problem itself (Section 5.1).…”

Section: M2 Tmentioning

confidence: 99%

“…However, these models fail to accurately capture the reality of manipulation because there is a large amount of uncertainty in both the object's motion and the robot's observations. Other work has employed a Bayesian approach by using a particle filter to estimate the pose (Corcoran & Platt 2010, Zhang & Trinkle 2012, contact state (Gadeyne et al 2005, Meeussen et al 2007, and physical properties ) of an object during manipulation. However, the conventional particle filter (CPF, Section 4) suffers from a startling problem: the CPF systematically performs worse as the sensor resolution and update frequency increases (Section 4.4).…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Pose estimation for planar contact manipulation with manifold particle filters

Koval

Pollard

Srinivasa

2015

The International Journal of Robotics Research

View full text Add to dashboard Cite

We investigate the problem of using contact sensors to estimate the pose of an object during planar pushing by a fixed-shape hand. Contact sensors are unique because they inherently discriminate between "contact" and "no-contact" configurations. As a result, the set of object configurations that activates a sensor constitutes a lower-dimensional contact manifold in the configuration space of the object. This causes conventional state estimation methods, such as the particle filter, to perform poorly during periods of contact due to particle starvation.In this paper, we introduce the manifold particle filter as a principled way of solving the state estimation problem when the state moves between multiple manifolds of different dimensionality. The manifold particle filter avoids particle starvation during contact by adaptively sampling particles that reside on the contact manifold from the dual proposal distribution. We describe three techniques-one analytical, and two sample-based-of sampling from the dual proposal distribution and compare their relative strengths and weaknesses. We present simulation results that show that all three techniques outperform the conventional particle filter in both speed and accuracy. Additionally, we implement the manifold particle filter on a real robot and show that it successfully tracks the pose of a pushed object using commercially available tactile sensors.

show abstract

Section: Contact State Estimationmentioning

confidence: 99%

Section: Contact State Estimationmentioning

confidence: 99%

Section: M2 Tmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Pose estimation for planar contact manipulation with manifold particle filters

Koval

Pollard

Srinivasa

2015

The International Journal of Robotics Research

View full text Add to dashboard Cite

show abstract

“…The necessary sensing actions themselves can be planned automatically, which is known as active sensing, and Lefebvre et al (2003Lefebvre et al ( , 2005a provide the means to derive an optimal compliant motion task plan consisting of both the sequence of contact formations and the compliant path to be executed while maintaining each CF. Programming a robot to perform a compliant motion task can also be done following the programming by human demonstration paradigm, as long as the motion performed by the human demonstrator can be segmented into different CFs while registering the necessary geometric information, as done in Meeussen et al (2007).…”

Section: Low-level Contact Compliant Motion Planningmentioning

confidence: 99%

Survey on assembly sequencing: a combinatorial and geometrical perspective

Jiménez

2011

J Intell Manuf

View full text Add to dashboard Cite

A systematic overview on the subject of assembly sequencing is presented. Sequencing lies at the core of assembly planning, and variants include finding a feasible sequence -respecting the precedence constraints between the assembly operations-, or determining an optimal one according to one or several operational criteria. The different ways of representing the space of feasible assembly sequences are described, as well as the search and optimization algorithms that can be used. Geometry plays a fundamental role in devising the precedence constraints between assembly operations, and this is the subject of the second part of the survey, which treats also motion in contact in the context of the actual performance of assembly operations.

show abstract