A Gesture-Based Natural Human–Robot Interaction Interface With Unrestricted Force Feedback

Liang, Yinhao; Du, Guanglong; Li, Chunquan; Chen, Chuxin; Wang, Xueqian; Liu, Peter

doi:10.1109/tim.2022.3149109

Cited by 10 publications

(6 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In view of the sensors used, the MoCap-based LfD methods can also be classified into marker-based methods and marker-less methods [17]. The marker-less methods, such as the ones using RGB-D cameras [5,9,31] or Leap-motions [13], are portable and more friendly to the human workers' intuitive body movements than the marker-based methods [9]. However, data accuracy of the marker-based demonstration methods tends to be much higher than the marker-less methods [17] as long as the occlusion problem is eliminated.…”

Section: Rgb Cameramentioning

confidence: 99%

“…In the goal update laws ( 6) and ( 10), the hyper-parameters are β p = 0.01 and β q = 0.01. In the press movement primitive, hyper-parameters of the controller (13) and goal update law (14) are K τ = I, L τ = 2I, and β f = 0.001. For each object, the operator conducted 10 times of assembly demonstration.…”

Section: Robotmentioning

confidence: 99%

“…Learning from demonstration is a paradigm for robots to learn target skills by imitating the human opera-tors' demonstration [1,13,22]. With LfD methods, human workers, who are even non-expert in robotics, are capable of programming robots with complex skills, which makes LfD a promising way to program robots in small-batch, high-mix, and customized manufacturing [9,36].…”

Section: Introductionmentioning

confidence: 99%

“…There is no technical threshold for the workers to demonstrate the assembly tasks with the MoCap sensors. What they do is nothing other than what they used to do in the assembly lines [13]. In contrast to the widely-applied teleoperation [21] and kinesthetic teaching [10] methods, the MoCap-based methods are applicable without the absence of the robots.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Robot Learning from Demonstration for Assembly with Camera-Supplemented Optical Motion Capture Sensors

Hu,

Yan,

Yang

et al. 2023

Preprint

View full text Add to dashboard Cite

Motion capture (MoCap) sensors hold considerable potential in various industrial fields. One of their applications is serving as demonstration data collectors in robot learning from demonstration (LfD) systems. Within these systems, human workers transfer their expertise to robots through demonstrating targeted tasks, thereby providing significant value within the realm of industrial manufacturing. By supplementing the optical MoCap sensors with an RGB camera, a novel assembly demonstration platform and the corresponding policy learning method is proposed here. The 6-dimensional (6D) pose of the parts and video of the human workers' assembly demonstration are simultaneously obtained and temporally aligned. By exploiting both the accuracy and abundance of the demonstration data, the action recognition and parameter estimation method are proposed to learn the policy. The robotic assembly experiments are conducted to validate the proposed demonstration platform and policy learning method. As the experiment results indicate, with the proposed demonstration platform and the policy learning method, the robot can efficiently accomplish precise assembly tasks with a high success rate.

show abstract

Section: Rgb Cameramentioning

confidence: 99%

Section: Robotmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Robot Learning from Demonstration for Assembly with Camera-Supplemented Optical Motion Capture Sensors

Hu,

Yan,

Yang

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…On the other side, to provide the user kinaesthetic feedback, it is necessary to generate, for example, force fields that often involve bulky devices, in which the haptic feedback is bound to limited workspaces and bandwidth [ 10 ], especially for what concerns teleoperation in both industrial [ 11 ] and surgical [ 12 ] realms.…”

Section: Introductionmentioning

confidence: 99%

Rendering Immersive Haptic Force Feedback via Neuromuscular Electrical Stimulation

Galofaro

D’Antonio

Lotti

et al. 2022

Sensors

View full text Add to dashboard Cite

Haptic feedback is the sensory modality to enhance the so-called “immersion”, meant as the extent to which senses are engaged by the mediated environment during virtual reality applications. However, it can be challenging to meet this requirement using conventional robotic design approaches that rely on rigid mechanical systems with limited workspace and bandwidth. An alternative solution can be seen in the adoption of lightweight wearable systems equipped with Neuromuscular Electrical Stimulation (NMES): in fact, NMES offers a wide range of different forces and qualities of haptic feedback. In this study, we present an experimental setup able to enrich the virtual reality experience by employing NMES to create in the antagonists’ muscles the haptic sensation of being loaded. We developed a subject-specific biomechanical model that estimated elbow torque during object lifting to deliver suitable electrical muscle stimulations. We experimentally tested our system by exploring the differences between the implemented NMES-based haptic feedback (NMES condition), a physical lifted object (Physical condition), and a condition without haptic feedback (Visual condition) in terms of kinematic response, metabolic effort, and participants’ perception of fatigue. Our results showed that both in terms of metabolic consumption and user fatigue perception, the condition with electrical stimulation and the condition with the real weight differed significantly from the condition without any load: the implemented feedback was able to faithfully reproduce interactions with objects, suggesting its possible application in different areas such as gaming, work risk assessment simulation, and education.

show abstract

Contact classification for human–robot interaction with densely connected convolutional neural network and convolutional block attention module

Aydin,

Avaroğlu

2024

SIViP

View full text Add to dashboard Cite

A Gesture-Based Natural Human–Robot Interaction Interface With Unrestricted Force Feedback

Cited by 10 publications

References 45 publications

Robot Learning from Demonstration for Assembly with Camera-Supplemented Optical Motion Capture Sensors

Robot Learning from Demonstration for Assembly with Camera-Supplemented Optical Motion Capture Sensors

Rendering Immersive Haptic Force Feedback via Neuromuscular Electrical Stimulation

Contact classification for human–robot interaction with densely connected convolutional neural network and convolutional block attention module

Contact Info

Product

Resources

About