Probabilistic 3D multilabel real-time mapping for multi-object manipulation

Wada, Kazumi; Okada, Kei; Inaba, Masayuki

doi:10.1109/iros.2017.8206394

Cited by 10 publications

(6 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…[120] was then extended in ref. [122] by considering a multi-object manipulation. Even though the 3D object segmentation improves grasping efficiency, notably in cluttered place, there are still failures due to picking motion.…”

Section: Grasping In Cluttered Environmentmentioning

confidence: 99%

Comprehensive Review on Reaching and Grasping of Objects in Robotics

Mohammed¹,

Chua²,

Kwek³

2021

Robotica

View full text Add to dashboard Cite

SUMMARY Interaction between a robot and its environment requires perception about the environment, which helps the robot in making a clear decision about the object type and its location. After that, the end effector will be brought to the object’s location for grasping. There are many research studies on the reaching and grasping of objects using different techniques and mechanisms for increasing accuracy and robustness during grasping and reaching tasks. Thus, this paper presents an extensive review of research directions and topics of different approaches such as sensing, learning and gripping, which have been implemented within the current five years.

show abstract

Section: Grasping In Cluttered Environmentmentioning

confidence: 99%

Comprehensive Review on Reaching and Grasping of Objects in Robotics

Mohammed¹,

Chua²,

Kwek³

2021

Robotica

View full text Add to dashboard Cite

show abstract

“…representations use them for navigation [16,17,18]; however, some also use them for manipulation. Wada et al [19] used a voxel grid to store occupancy and semantics of objects in a cluttered scene to select the next target object and grasp point. MoreFusion [20] is a system that uses voxels to perform multi-object reasoning to improve 6D pose estimation; the system gives accurate object poses and can perform precise pick-and-place in cluttered scenes.…”

Section: Voxel Representation For Manipulationmentioning

confidence: 99%

Coarse-to-Fine Q-attention: Efficient Learning for Visual Robotic Manipulation via Discretisation

James¹,

Wada²,

Laidlow³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Reflecting on the last few years, the biggest breakthroughs in deep reinforcement learning (RL) have been in the discrete action domain. Robotic manipulation, however, is inherently a continuous control environment, but these continuous control reinforcement learning algorithms often depend on actor-critic methods that are sample-inefficient and inherently difficult to train, due to the joint optimisation of the actor and critic. To that end, we explore how we can bring the stability of discrete action RL algorithms to the robot manipulation domain. We extend the recently released ARM algorithm, by replacing the continuous next-best pose agent with a discrete next-best pose agent. Discretisation of rotation is trivial given its bounded nature, while translation is inherently unbounded, making discretisation difficult. We formulate the translation prediction as the voxel prediction problem by discretising the 3D space; however, voxelisation of a large workspace is memory intensive and would not work with a high density of voxels, crucial to obtaining the resolution needed for robotic manipulation. We therefore propose to apply this voxel prediction in a coarse-to-fine manner by gradually increasing the resolution. In each step, we extract the highest valued voxel as the predicted location, which is then used as the centre of the higher-resolution voxelisation in the next step. This coarse-to-fine prediction is applied over several steps, giving a near-lossless prediction of the translation. We show that our new coarse-to-fine algorithm is able to accomplish RLBench tasks much more efficiently than the continuous control equivalent, and even train some real-world tasks, tabular rasa, in less than 7 minutes, with only 3 demonstrations. Moreover, we show that by moving to a voxel representation, we are able to easily incorporate observations from multiple cameras. Videos and code found at: https://sites.google.com/ view/c2f-q-attention.

show abstract

“…Many studies have concentrated on grasping or grasp pose detection, such as segmenting the target object [78] [79], combining red-green-blue color image with its corresponding depth image (i.e. RGB-D) based-multimodal data [80], templatebased approach with convex hull [81] [82], use of a bounding box [83], [84] and detection of an object in a cluttered scene by using two-stream CNNs [85].…”

Section: A 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

Probabilistic 3D multilabel real-time mapping for multi-object manipulation

Cited by 10 publications

References 21 publications

Comprehensive Review on Reaching and Grasping of Objects in Robotics

Comprehensive Review on Reaching and Grasping of Objects in Robotics

Coarse-to-Fine Q-attention: Efficient Learning for Visual Robotic Manipulation via Discretisation

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

Contact Info

Product

Resources

About