Robust object grasping in clutter via singulation

Kiatos, Marios; Malassiotis, Sotiris

doi:10.1109/icra.2019.8793972

Cited by 37 publications

(32 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Finally, to produce more optimal policies, we take consideration of the workspace limits and their nature (open or constraining walls), which are ignored in [21]. Kiatos et al [22] employ a pushing policy with discrete actions that is learned via Q-learning in order to singulate a target object from each surrounding clutter using depth features that approximate the topography of the scene. However, they made restricted assumptions for the height of the obstacles and the properties of the target object.…”

Section: Related Workmentioning

confidence: 99%

“…However, they made restricted assumptions for the height of the obstacles and the properties of the target object. Sarantopoulos and Kiatos [23] extended the work of [22] by modelling the Qfunction with two different networks, one for each primitive action.…”

Section: Related Workmentioning

confidence: 99%

“…In contrast to [22], [23], which use discrete actions to singulate the target object, we use continuous actions which increases the agent's pushing options. Furthermore, [23] uses only a single policy (argmax between the Q-values of the discrete directions), while in the proposed method we have multiple policies: primitive policies and a high level policy which selects between those.…”

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

Total Singulation With Modular Reinforcement Learning

Sarantopoulos

Kiatos

Doulgeri

et al. 2021

IEEE Robot. Autom. Lett.

Self Cite

View full text Add to dashboard Cite

Prehensile robotic grasping of a target object in clutter is challenging because, in such conditions, the target touches other objects, resulting to the lack of collision free grasp affordances. To address this problem, we propose a modular reinforcement learning method which uses continuous actions to totally singulate the target object from its surrounding clutter. A high level policy selects between pushing primitives, which are learned separately. Prior knowledge is effectively incorporated into learning, through action primitives and feature selection, increasing sample efficiency. Experiments demonstrate that the proposed method considerably outperforms the state-of-theart methods in the singulation task. Furthermore, although training is performed in simulation the learned policy is robustly transferred to a real environment without a significant drop in success rate. Finally, singulation tasks in different environments are addressed by easily adding a new primitive and by retraining only the high level policy.The research leading to these results has received funding from the European Community's Framework Programme Horizon 2020 under grant agreement No 871704, project BACCHUS.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Total Singulation With Modular Reinforcement Learning

Sarantopoulos

Kiatos

Doulgeri

et al. 2021

IEEE Robot. Autom. Lett.

Self Cite

View full text Add to dashboard Cite

show abstract

“…However, they evaluated a large number of potential pushes to find the optimal action. Kiatos et al [32] trained a deep Q-network to select push actions in order to singulate a target object from its surrounding clutter with the minimum number of pushes using depth features to approximate the topography of the scene. Sarantopoulos and Kiatos [33] extended the work of [32] by modelling the Q-function with two different networks, one for each primitive action, leading to higher success rates and faster network convergence.…”

Section: B Learning-based Approachesmentioning

confidence: 99%

Bridging the gap between learning and heuristic based pushing policies

Kiatos¹,

Sarantopoulos²,

Malassiotis³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Non-prehensile pushing actions have the potential to singulate a target object from its surrounding clutter in order to facilitate the robotic grasping of the target. To address this problem we utilize a heuristic rule that moves the target object towards the workspace's empty space and demonstrate that this simple heuristic rule achieves singulation. Furthermore, we incorporate this heuristic rule to the reward in order to train more efficiently reinforcement learning (RL) agents for singulation. Simulation experiments demonstrate that this insight increases performance. Finally, our results show that the RL-based policy implicitly learns something similar to one of the used heuristics in terms of decision making.

show abstract

“…RL-based, training-optimal push policies were presented in [124] for given depth observations of a scene. These policies facilitate grasping objects in cluttered scenes where the target can be invisible by using a deep neural network algorithm with Qlearning.…”

Section: B Suction and Multifunctional Graspingmentioning

confidence: 99%

Review of Deep Reinforcement Learning-Based Object Grasping: Techniques, Open Challenges, and Recommendations

2020

View full text Add to dashboard Cite

The motivation behind our work is to review and analyze the most relevant studies on deep reinforcement learning-based object manipulation. Various studies are examined through a survey of existing literature and investigation of various aspects, namely, the intended applications, techniques applied, challenges faced by researchers and recommendations for minimizing obstacles. This review refers to all relevant articles on deep reinforcement learning-based object manipulation and solutions. The object grasping issue is a major manipulation challenge. Object grasping requires detection systems, methods and tools to facilitate efficient and fast agent training. Several studies have proposed that object grasping and its subtypes are the main elements in dealing with the environment and agent. Unlike other review articles, this review article provides different observations on deep reinforcement learning-based manipulation. The results of this comprehensive review of deep reinforcement learning in the manipulation field may be valuable for researchers and practitioners because they can expedite the establishment of important guidelines.

show abstract

Robust object grasping in clutter via singulation

Cited by 37 publications

References 13 publications

Total Singulation With Modular Reinforcement Learning

Total Singulation With Modular Reinforcement Learning

Bridging the gap between learning and heuristic based pushing policies

Review of Deep Reinforcement Learning-Based Object Grasping: Techniques, Open Challenges, and Recommendations

Contact Info

Product

Resources

About