Persistent Homology for Effective Non-Prehensile Manipulation

Vieira, Ewerton R.; Nakhimovich, Daniel; Gao, Kai; Wang, Rui; Yu, Jingjin; Bekris, Kostas E.

doi:10.1109/icra46639.2022.9811848

Cited by 15 publications

(12 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The former is motivated by human behaviors of implicitly grouping objects before pushing them simultaneously. Topological tools, such as persistent homology, have been successfully employed to comprehensively identify manageable clusters of objects for recognizing the appropriated pushing actions [1]. The second direction employs Monte-Carlo Tree Search (MCTS) to explore the feasible identified pushing actions with a high reward level in terms of: the number of obstacles removed from the path to the target; and how dispersed the clusters are after the performed actions.…”

Section: The Framework Exploresmentioning

confidence: 99%

“…The proposed method takes advantage of persistent homology to inform the selection of efficient and robust pushing actions, which outperform, in terms of planning time and number of actions, prior efforts in the same domain [1], [23], [24]. Moreover, it provides high-level solutions that are robust under uncertainty (pose, arm movements, nonprehensile manipulation of objects).…”

Section: Related Workmentioning

confidence: 99%

“…Usually, c = √ 2 and, for limited number of iterations, c is chosen to be less than √ 2. The optimistic path is selected by choosing the action a maximizing the UCB U (X , a), given by (1).…”

Section: A Monte-carlo Tree Searchmentioning

confidence: 99%

“…3. When the target is too close to the walls S and N , it is possible to apply a planar rotation matrix R −ϕ to obtain Π(X ) with incidence angle ϕ as in [1].…”

Section: B Proposed Approach For Non-prehensile Manipulationmentioning

confidence: 99%

“…Having selected the closest cluster to the gripper CC(X , r) c , the pushing actions are based on the smallest rectangle that contains CC(X , r) c , denoted by circumscribed rectangle R(X , r). The arm will push the obstacles in the region R(X , r) either towards the north wall N or the south wall S. Differently from previous work [1], the two types of pushing action are also explored during the expansion stage of the MCTS. That is, for a selected R(X , r) two actions can be performed to remove the obstacles inside R(X , r) either by a sweeping movement from the bottom of R(X , r) to the top of R(X , r) or from top to bottom of R(X , r), denote the actions by up and down, respectively.…”

Section: B Proposed Approach For Non-prehensile Manipulationmentioning

confidence: 99%

See 4 more Smart Citations

Persistent Homology Guided Monte-Carlo Tree Search for Effective Non-Prehensile Manipulation

Vieira¹,

Gao²,

Nakhimovich³

et al. 2022

Preprint

View full text Add to dashboard Cite

Performing object retrieval tasks in messy realworld workspaces involves the challenges of uncertainty and clutter. One option is to solve retrieval problems via a sequence of prehensile pick-n-place operations, which can be computationally expensive to compute in highly-cluttered scenarios and also inefficient to execute. The proposed framework selects the option of performing non-prehensile actions, such as pushing, to clean a cluttered workspace to allow a robotic arm to retrieve a target object. Non-prehensile actions, allow to interact simultaneously with multiple objects, which can speed up execution. At the same time, they can significantly increase uncertainty as it is not easy to accurately estimate the outcome of a pushing operation in clutter. The proposed framework integrates topological tools and Monte-Carlo tree search to achieve effective and robust pushing for object retrieval problems. In particular, it proposes using persistent homology to automatically identify manageable clustering of blocking objects in the workspace without the need for manually adjusting hyper-parameters. Furthermore, MCTS uses this information to explore feasible actions to push groups of objects together, aiming to minimize the number of pushing actions needed to clear the path to the target object. Real-world experiments using a Baxter robot, which involves some noise in actuation, show that the proposed framework achieves a higher success rate in solving retrieval tasks in dense clutter compared to state-of-the-art alternatives. Moreover, it produces high-quality solutions with a small number of pushing actions improving the overall execution time. More critically, it is robust enough that it allows to plan the sequence of actions offline and then execute them reliably online with Baxter.

show abstract

Section: The Framework Exploresmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: A Monte-carlo Tree Searchmentioning

confidence: 99%

“…3. When the target is too close to the walls S and N , it is possible to apply a planar rotation matrix R −ϕ to obtain Π(X ) with incidence angle ϕ as in [1].…”

Section: B Proposed Approach For Non-prehensile Manipulationmentioning

confidence: 99%

Section: B Proposed Approach For Non-prehensile Manipulationmentioning

confidence: 99%

See 3 more Smart Citations