Goal-aware generative adversarial imitation learning from imperfect demonstration for robotic cloth manipulation

Tsurumine, Yoshihisa; Matsubara, Takamitsu

doi:10.1016/j.robot.2022.104264

Cited by 8 publications

(6 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In order to robustly learn desirable policies even from subpar demonstrations, Tsurumine et al [ 122 ] propose goal-aware generative adversarial imitation learning (GA-GAIL), which uses a second discriminator to distinguish the goal state in parallel with the first discriminator that indicates the demonstration data. To accomplish stable policy learning from two discriminators, GA-GAIL uses the entropy-maximizing deep P-network (EDPN) as a generator.…”

Section: Deep Rl For Robotic Manipulationmentioning

confidence: 99%

A Survey on Deep Reinforcement Learning Algorithms for Robotic Manipulation

Han

Mulyana

Stankovic³

et al. 2023

Sensors

View full text Add to dashboard Cite

Robotic manipulation challenges, such as grasping and object manipulation, have been tackled successfully with the help of deep reinforcement learning systems. We give an overview of the recent advances in deep reinforcement learning algorithms for robotic manipulation tasks in this review. We begin by outlining the fundamental ideas of reinforcement learning and the parts of a reinforcement learning system. The many deep reinforcement learning algorithms, such as value-based methods, policy-based methods, and actor–critic approaches, that have been suggested for robotic manipulation tasks are then covered. We also examine the numerous issues that have arisen when applying these algorithms to robotics tasks, as well as the various solutions that have been put forth to deal with these issues. Finally, we highlight several unsolved research issues and talk about possible future directions for the subject.

show abstract

Section: Deep Rl For Robotic Manipulationmentioning

confidence: 99%

A Survey on Deep Reinforcement Learning Algorithms for Robotic Manipulation

Han

Mulyana

Stankovic³

et al. 2023

Sensors

View full text Add to dashboard Cite

show abstract

“…However, the number of bins needed to approximate a continuous action space grows exponentially with increasing dimensionality. Generative adversarial networks [27,76,16], variational autoencoders [47] and combined Categorical and Gaussian distributions [66,23,14] have been used to learn from multimodal demonstrations. Nevertheless, these models tend to be sensitive to hyperparameters, such as the number of clusters used [66].…”

Section: Learning Robot Manipulation Policies From Demonstrationsmentioning

confidence: 99%

“…Indeed, human demonstrations often contain diverse ways that a task can be accomplished. A natural choice is then to treat policy learning as a distribution learning problem: instead of representing a policy as a deterministic map π θ (x), learn the entire distribution of actions conditioned on the current robot state p(y|x) [27,76,25,66]. Recent works use diffusion models for learning such state-conditioned action distributions for robot manipulation policies from demonstrations [54,11,60] and show they outperform deterministic or other alternatives, such as variational autoencoders [47], mixture of Gaussians † Equal contribution Fig.…”

Section: Introductionmentioning

confidence: 99%

3D Scene Localization and Mapping Based on Omnidirectional SLAM

Xi-lin

Huang

Zhang

et al. 2021

IOP Conf. Ser.: Earth Environ. Sci.

View full text Add to dashboard Cite

The vision SLAM technology is the key technology of robot navigation and unmanned driving. In order to further study the accuracy of multi-view SLAM positioning and mapping applications, this paper combines omnidirectional imaging and SLAM technology, improves the existing technical model in the field of omnidirectional imaging, studies the pose map optimization and SLAM overall adjustment technology based on object constraints and global consistency, and carries out omnidirectional SLAM point cloud reconstruction and 3D mapping. Through the time synchronization experiment of Ladybug 5 Plus and Femtomes Mini2-D-INS, the two positioning results of the system and the GNSS measurement results are compared and analysed.

show abstract

“…Till now network-based imitation learning methods involve sampling human-operated skills and then training neural networks using the acquired sample data to achieve skill learning. As of now, neural network-based imitation learning methods primarily include BC methods (Li et al, 2022), where neural networks are directly trained, and Generative Adversarial Imitation Learning (GAIL) methods (Tsurumine and Matsubara, 2022), which approximate strategies through generative adversarial techniques. Neural network-based imitation learning methods excel in strategy approximation.…”

mentioning

confidence: 99%

Peg-in-hole assembly skill imitation learning method based on ProMPs under task geometric representation

Zang,

Wang,

Zha

et al. 2023

Front. Neurorobot.

View full text Add to dashboard Cite

IntroductionBehavioral Cloning (BC) is a common imitation learning method which utilizes neural networks to approximate the demonstration action samples for task manipulation skill learning. However, in the real world, the demonstration trajectories from human are often sparse and imperfect, which makes it challenging to comprehensively learn directly from the demonstration action samples. Therefore, in this paper, we proposes a streamlined imitation learning method under the terse geometric representation to take good advantage of the demonstration data, and then realize the manipulation skill learning of assembly tasks.MethodsWe map the demonstration trajectories into the geometric feature space. Then we align the demonstration trajectories by Dynamic Time Warping (DTW) method to get the unified data sequence so we can segment them into several time stages. The Probability Movement Primitives (ProMPs) of the demonstration trajectories are then extracted, so we can generate a lot of task trajectories to be the global strategy action samples for training the neural networks. Notalby, we regard the current state of the assembly task as the via point of the ProMPs model to get the generated trajectories, while the time point of the via point is calculated according to the probability model of the different time stages. And we get the action of the current state according to the target position of the next time state. Finally, we train the neural network to obtain the global assembly strategy by Behavioral Cloning.ResultsWe applied the proposed method to the peg-in-hole assembly task in the simulation environment based on Pybullet + Gym to test its task skill learning performance. And the learned assembly strategy was also executed on a real robotic platform to verify the feasibility of the method further.DiscussionAccording to the result of the experiment, the proposed method achieves higher success rates compared to traditional imitation learning methods while exhibiting reasonable generalization capabilities. It shows that the ProMPs under geometric representation can help the BC method make better use of the demonstration trajectory and thus better learn the task skills.

show abstract

Goal-aware generative adversarial imitation learning from imperfect demonstration for robotic cloth manipulation

Cited by 8 publications

References 26 publications

A Survey on Deep Reinforcement Learning Algorithms for Robotic Manipulation

A Survey on Deep Reinforcement Learning Algorithms for Robotic Manipulation

3D Scene Localization and Mapping Based on Omnidirectional SLAM

Peg-in-hole assembly skill imitation learning method based on ProMPs under task geometric representation

Contact Info

Product

Resources

About