Reinforcement learning of clothing assistance with a dual-arm robot

Tamei, Tomoya; Matsubara, Takamitsu; Rai, Akshara; Shibata, Tomohiro

doi:10.1109/humanoids.2011.6100915

Cited by 75 publications

(74 citation statements)

References 7 publications

(9 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Previous work with such representations has only tested basic approaches for inverse mapping of fixed topological trajectories to the joint configuration [4,5,16]. In contrast, in this paper we presented methods that combine the different representations at the abstract and lower level for motion synthesis.…”

Section: Discussionmentioning

confidence: 99%

“…Bodies are mapped from the new representation to the joint angles by the least squares principle. In [16], the same representation is applied for controlling the movement of a robot that puts a shirt on a human. The mapping from the new representation to the low level representation is learned through demonstrations by humans.…”

Section: A Previous Workmentioning

confidence: 99%

“…Alternate topological representations that describes the relationships between 1D curves using their original configurations are proposed for motion synthesis [5,16] and classifying paths into homotopy groups [1]. In [5], a representation called Topology Coordinates that is based on the Gauss Linking Integral is proposed to synthesize tangling movements.…”

Section: A Previous Workmentioning

confidence: 99%

See 2 more Smart Citations

Hierarchical Motion Planning in Topological Representations

Zarubin

Ivan²,

Toussaint

et al. 2012

Robotics: Science and Systems VIII

View full text Add to dashboard Cite

Abstract-Motion can be described in alternative representations, including joint configuration or end-effector spaces, but also more complex topological representations that imply a change of Voronoi bias, metric or topology of the motion space. Certain types of robot interaction problems, e.g. wrapping around an object, can suitably be described by so-called writhe and interaction mesh representations. However, considering motion synthesis solely in topological spaces is insufficient since it does not cater for additional tasks and constraints in other representations. In this paper we propose methods to combine and exploit different representations for motion synthesis, with specific emphasis on generalization of motion to novel situations. Our approach is formulated in the framework of optimal control as an approximate inference problem, which allows for a direct extension of the graphical model to incorporate multiple representations. Motion generalization is similarly performed by projecting motion from topological to joint configuration space. We demonstrate the benefits of our methods on problems where direct path finding in joint configuration space is extremely hard whereas local optimal control exploiting a representation with different topology can efficiently find optimal trajectories. Further, we illustrate the successful online motion generalization to dynamic environments on challenging, real world problems.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: A Previous Workmentioning

confidence: 99%

Section: A Previous Workmentioning

confidence: 99%

See 1 more Smart Citation

Hierarchical Motion Planning in Topological Representations

Zarubin

Ivan²,

Toussaint

et al. 2012

Robotics: Science and Systems VIII

View full text Add to dashboard Cite

show abstract

“…This analysis assumed that the masses of the i-th and j-th fingers were identical. Equation (12) shows that a pressing force F Nj is required to be large enough to hold the fabric onto the floor; however, if Equation (9) is satisfied, a wrinkle must be generated during wiping deformation. According to Equation (9), the frictional coefficients of the manipulated object are significant for this strategy.…”

Section: Mechanical Analyses Of the Wiping Deformationmentioning

confidence: 99%

“…Indirect simultaneous positioning of an extensible fabric has been demonstrated and simulated using Proportional-Integral-Derivative (PID) control for desired locations [9]. Control of a few manipulated points was shown to regulate several positioned points to their desired locations [10,11], and a dual-armed robotic system has been proposed for dressing activity [12]. Furthermore, the folding motion of clothes using a robotic laundry system [13] and a visual perception for grasping/folding messy fabrics [14] has been assessed.…”

Section: Introductionmentioning

confidence: 99%

A Pinching Strategy for Fabrics Using Wiping Deformation

Shibata

Hirai

2016

Robotics

View full text Add to dashboard Cite

This paper discusses a strategy by which a robotic hand can use the physical properties of a fabric to pinch the fabric. Pinching may be accomplished by using a wiping motion, during which the movement and deformation of a deformable object occur simultaneously. The wiping motion differs from the displacement of a deformable object. During the wiping motion, there is contact, but no relative movement, between the manipulator and the object, whereas, during displacement, there is both contact and relative movement between the object and the floor. This paper first describes wiping motion and distinguishes wiping slide from wiping deformation by displacement of the internal points of an object. Wiping motion is also shown to be an extended scheme of pushing and sliding of rigid objects. Our strategy for pinching a fabric is accomplished with a combination of wiping deformation and residual deformation of the fabric under unloaded conditions. Using this strategy, a single-armed robotic hand can pinch both surfaces of the fabric without handover motion.

show abstract