Multidimensional Deformable Object Manipulation Based on DN-Transporter Networks

Teng, Yadong; Lu, Huimin; Li, Yujie; Kamiya, Tohru; Nakatoh, Yoshihisa; Serikawa, Seiichi; Gao, Peng

doi:10.1109/tits.2022.3168303

Cited by 12 publications

(17 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Applications of DRL and DIL methods on cloth-shaping have been explored since 2018. The adopted methods are BC (Section 2.1) [328,327,386,362], MFRL (Section 3.1) [244,392,171,126,146,234,146,224] and MBRL (Section 3.3) [402,155,237,10,234,233]. So far, MBRL and BC have achieved better performance than MFRL methods [327,155,386,233].…”

Section: Data-driven Control In Cloth-shapingmentioning

confidence: 99%

“…The canonical task of bag manipulation is item loading (ILD) [362] that involves opening the bag (OB), inserting items (II) [327] and lifting the bag (LB). Lifting a bag with an item in it is also known as an item containing (IC) [330].…”

Section: Bag Manipulationmentioning

confidence: 99%

“…II can be more complex if there are other bags for distraction [327]. The next common task is bag moving (BM) [362], which includes lifting the bag, creating displacement and placing the bag at the target position. The third standard task is item unloading (IU), which requires opening the bag and picking the items out.…”

Section: Bag Manipulationmentioning

confidence: 99%

“…The complex dynamic between the rigid/deformable items with the bag [387] gives another layer of challenge for the accomplishment of ILD. During II, the perception needs to reason whether the objects are within the region of open contour [330,362] of the bag. While conducting IC, the system also needs to reason if items will remain inside the bag under their interaction with the bag and gravity [330].…”

Section: Challenges In Bag Manipulationmentioning

confidence: 99%

“…Seita et al (2021a) [327] and Teng et al (2022) [362] leverage supervised learning to fit a pixel-based goal-condition value heatmap for P&P action primitives to perform a set of skills for bag manipulations. Weng et al (2021) [387] investigate graph-based key point dynamic learning for predicting the interaction between rigid items and a bag.…”

Section: Perception In Bag Manipulationmentioning

confidence: 99%

See 4 more Smart Citations

Data-Driven Robotic Manipulation of Cloth-like Deformable Objects: Present, Challenges, and Future Prospects

Kadi¹,

Terzic²

2022

Preprint

View full text Add to dashboard Cite

Manipulating cloth-like deformable objects (CDOs) is a long-standing problem in the robotics community. CDOs are flexible (non-rigid) objects that do not show a detectable level of com- pression strength while two points on the article are pushed towards each otherand include objects such as ropes (1D), fabrics (2D) and bags (3D). In general, CDOs’ many degrees of freedom (DoF) introduce severe self-occlusion and complex state-action dynamics as significant obstacles for perception and manipulation systems. These challenges exacerbate existing issues of modern robotic control methods such as imitation learning (IL) and reinforcement learning (RL). This review focuses on the application details of data-driven control methods on four major task families in this domain: cloth-shaping, rope manipulation, dressing and bag manip- ulation. Furthermore, we identify specific inductive biases in these four domains that present challenges for more general IL and RL algorithms, and summarise the future direction for the development of the field.

show abstract