Pub Date : 2021-07-19DOI: 10.1109/HUMANOIDS47582.2021.9555797
M. Schwarz, Sven Behnke
We present a method for real-time stereo scene capture and remote VR visualization that allows a human operator to freely move their head and thus intuitively control their perspective during teleoperation. The stereo camera is mounted on a 6D robotic arm, which follows the operator’s head pose. Existing VR teleoperation systems either induce high latencies on head movements, leading to motion sickness, or use scene reconstruction methods to allow re-rendering of the scene from different perspectives, which cannot handle dynamic scenes effectively. Instead, we present a decoupled approach which renders captured camera images as spheres, assuming constant distance. This allows very fast re-rendering on head pose changes while keeping the resulting temporary distortions during head translations small. We present qualitative examples, quantitative results in the form of lab experiments and a small user study, showing that our method outperforms other visualization methods.
{"title":"Low-Latency Immersive 6D Televisualization with Spherical Rendering","authors":"M. Schwarz, Sven Behnke","doi":"10.1109/HUMANOIDS47582.2021.9555797","DOIUrl":"https://doi.org/10.1109/HUMANOIDS47582.2021.9555797","url":null,"abstract":"We present a method for real-time stereo scene capture and remote VR visualization that allows a human operator to freely move their head and thus intuitively control their perspective during teleoperation. The stereo camera is mounted on a 6D robotic arm, which follows the operator’s head pose. Existing VR teleoperation systems either induce high latencies on head movements, leading to motion sickness, or use scene reconstruction methods to allow re-rendering of the scene from different perspectives, which cannot handle dynamic scenes effectively. Instead, we present a decoupled approach which renders captured camera images as spheres, assuming constant distance. This allows very fast re-rendering on head pose changes while keeping the resulting temporary distortions during head translations small. We present qualitative examples, quantitative results in the form of lab experiments and a small user study, showing that our method outperforms other visualization methods.","PeriodicalId":320510,"journal":{"name":"2020 IEEE-RAS 20th International Conference on Humanoid Robots (Humanoids)","volume":"11 24 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128377037","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-07-19DOI: 10.1109/HUMANOIDS47582.2021.9555801
T. Anzai, Yuta Kojio, Tasuku Makabe, K. Okada, M. Inaba
In this paper, we propose a novel flying humanoid robot platform with bi-copter flight unit. Humanoid robots have the ability to move by walking, but it is not sufficient for some tasks. To enhance the mobility of humanoid robots, we apply aerial robotics and develop a flying humanoid robot capable of walking and flying in the air. We describe the modeling and control of bi-copter and takeoff pose generation method for flying humanoid robot. We show the hardware implementations of a bi-copter flight unit, a humanoid robot and the whole system of flying humanoid robot. We perform several experiments to verify the effectiveness of the flight control, extended mobility and the implemented robot system including perception.
{"title":"Design and Development of a Flying Humanoid Robot Platform with Bi-copter Flight Unit","authors":"T. Anzai, Yuta Kojio, Tasuku Makabe, K. Okada, M. Inaba","doi":"10.1109/HUMANOIDS47582.2021.9555801","DOIUrl":"https://doi.org/10.1109/HUMANOIDS47582.2021.9555801","url":null,"abstract":"In this paper, we propose a novel flying humanoid robot platform with bi-copter flight unit. Humanoid robots have the ability to move by walking, but it is not sufficient for some tasks. To enhance the mobility of humanoid robots, we apply aerial robotics and develop a flying humanoid robot capable of walking and flying in the air. We describe the modeling and control of bi-copter and takeoff pose generation method for flying humanoid robot. We show the hardware implementations of a bi-copter flight unit, a humanoid robot and the whole system of flying humanoid robot. We perform several experiments to verify the effectiveness of the flight control, extended mobility and the implemented robot system including perception.","PeriodicalId":320510,"journal":{"name":"2020 IEEE-RAS 20th International Conference on Humanoid Robots (Humanoids)","volume":"60 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124720053","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-07-19DOI: 10.1109/HUMANOIDS47582.2021.9555682
Erik Berger, Alexander Uhlig
Safe and meaningful interaction with robotic systems during behavior execution requires accurate sensing capabilities. This can be achieved by the usage of force-torque sensors which are often heavy, expensive, and require an additional power supply. Consequently, providing accurate sensing capabilities to lightweight robots, with a limited amount of load, is a challenging task. Furthermore, such sensors are not able to distinguish between task-specific regular forces and external influences as induced by human co-workers. To solve this, robots often rely on a large number of manually generated rules which is a time-consuming procedure. This paper presents a data-driven machine learning approach that enhances robotic behavior with estimates of the expected proprioceptive forces (intrinsic) and unexpected forces (extrinsic) exerted by the environment. First, the robot’s common internal sensors are recorded together with ground truth measurements of the actual forces during regular and perturbed behavior executions. The resulting data is used to generate features that contain a compact representation of behavior-specific intrinsic and extrinsic fluctuations. Those features are then utilized for deep learning of proprioceptive models which enables a robot to accurately distinguish the amount of intrinsic and extrinsic forces. Experiments performed with the UR5 robot show a substantial improvement in accuracy over force values provided by previous research.
{"title":"Feature-based Deep Learning of Proprioceptive Models for Robotic Force Estimation","authors":"Erik Berger, Alexander Uhlig","doi":"10.1109/HUMANOIDS47582.2021.9555682","DOIUrl":"https://doi.org/10.1109/HUMANOIDS47582.2021.9555682","url":null,"abstract":"Safe and meaningful interaction with robotic systems during behavior execution requires accurate sensing capabilities. This can be achieved by the usage of force-torque sensors which are often heavy, expensive, and require an additional power supply. Consequently, providing accurate sensing capabilities to lightweight robots, with a limited amount of load, is a challenging task. Furthermore, such sensors are not able to distinguish between task-specific regular forces and external influences as induced by human co-workers. To solve this, robots often rely on a large number of manually generated rules which is a time-consuming procedure. This paper presents a data-driven machine learning approach that enhances robotic behavior with estimates of the expected proprioceptive forces (intrinsic) and unexpected forces (extrinsic) exerted by the environment. First, the robot’s common internal sensors are recorded together with ground truth measurements of the actual forces during regular and perturbed behavior executions. The resulting data is used to generate features that contain a compact representation of behavior-specific intrinsic and extrinsic fluctuations. Those features are then utilized for deep learning of proprioceptive models which enables a robot to accurately distinguish the amount of intrinsic and extrinsic forces. Experiments performed with the UR5 robot show a substantial improvement in accuracy over force values provided by previous research.","PeriodicalId":320510,"journal":{"name":"2020 IEEE-RAS 20th International Conference on Humanoid Robots (Humanoids)","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123826325","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-07-19DOI: 10.1109/HUMANOIDS47582.2021.9555773
Soya Shimizu, K. Ayusawa, G. Venture
This paper presents a method to estimate cost weights of cost functions and multiple joint motion time-series values of humanoid robots easily, using functional principal component analysis (FPCA) instead of direct optimal control (DOC) and inverse optimal control (IOC). Each given object’s cost weight exemplar can be converted into a point in the FPC space by applying FPCA. Cost weight and the FPC space enable to synthesize the motion model data and the cost function factor and therefore versatile motion data conveniently. The proposed method surpasses classic DOC and IOC methods in terms of calculation time and efficiency, in novel data analysis. Furthermore, proposed method is applied to the humanoid robot HRP4, to generate arm motions, as an experimental proof of concept with some cost functions. The accuracy of the motion generation is experimentally confirmed.
{"title":"Pseudo Direct and Inverse Optimal Control based on Motion Synthesis using FPCA","authors":"Soya Shimizu, K. Ayusawa, G. Venture","doi":"10.1109/HUMANOIDS47582.2021.9555773","DOIUrl":"https://doi.org/10.1109/HUMANOIDS47582.2021.9555773","url":null,"abstract":"This paper presents a method to estimate cost weights of cost functions and multiple joint motion time-series values of humanoid robots easily, using functional principal component analysis (FPCA) instead of direct optimal control (DOC) and inverse optimal control (IOC). Each given object’s cost weight exemplar can be converted into a point in the FPC space by applying FPCA. Cost weight and the FPC space enable to synthesize the motion model data and the cost function factor and therefore versatile motion data conveniently. The proposed method surpasses classic DOC and IOC methods in terms of calculation time and efficiency, in novel data analysis. Furthermore, proposed method is applied to the humanoid robot HRP4, to generate arm motions, as an experimental proof of concept with some cost functions. The accuracy of the motion generation is experimentally confirmed.","PeriodicalId":320510,"journal":{"name":"2020 IEEE-RAS 20th International Conference on Humanoid Robots (Humanoids)","volume":"124 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128147390","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-07-19DOI: 10.1109/HUMANOIDS47582.2021.9555804
Johannes Englsberger, A. Giordano, Achraf Hiddane, R. Schuller, F. Loeffl, George Mesesan, C. Ott
This paper introduces the Simplest Articulated Free-Floating (SAFF) model, a low-dimensional model facilitating the examination of controllers, which are designed for free-floating robots that are subject to gravity. Two different state-of-the-art control approaches, namely absolute CoM control accompanied by an assumption about the foot acceleration, and a controller combining absolute CoM and foot control objectives, are shown to yield exponential stability in the nominal case, while becoming unstable if the foot contact is lost. As an improvement over the state of the art, the so-called Relative-CoM-to-Foot (RCF) controller is introduced, which again yields exponential stability nominally, while preserving a BIBO stable behavior even in case of a complete contact loss. The controller performance is validated in various simulations.
{"title":"From Space to Earth – Relative-CoM-to-Foot (RCF) control yields high contact robustness","authors":"Johannes Englsberger, A. Giordano, Achraf Hiddane, R. Schuller, F. Loeffl, George Mesesan, C. Ott","doi":"10.1109/HUMANOIDS47582.2021.9555804","DOIUrl":"https://doi.org/10.1109/HUMANOIDS47582.2021.9555804","url":null,"abstract":"This paper introduces the Simplest Articulated Free-Floating (SAFF) model, a low-dimensional model facilitating the examination of controllers, which are designed for free-floating robots that are subject to gravity. Two different state-of-the-art control approaches, namely absolute CoM control accompanied by an assumption about the foot acceleration, and a controller combining absolute CoM and foot control objectives, are shown to yield exponential stability in the nominal case, while becoming unstable if the foot contact is lost. As an improvement over the state of the art, the so-called Relative-CoM-to-Foot (RCF) controller is introduced, which again yields exponential stability nominally, while preserving a BIBO stable behavior even in case of a complete contact loss. The controller performance is validated in various simulations.","PeriodicalId":320510,"journal":{"name":"2020 IEEE-RAS 20th International Conference on Humanoid Robots (Humanoids)","volume":"69 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127578012","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-07-19DOI: 10.1109/HUMANOIDS47582.2021.9555684
Dominik Urbaniak, Alejandro Agostini, Dongheui Lee
Task and motion planning deals with complex tasks that require a robot to automatically define and execute multi-step sequences of actions in cluttered scenarios. In this context, a linear motion is often not sufficient to approach a target object since collisions of the gripper with other objects or the target object might occur. Thus, motion planners should be able to generate collision-free trajectories for every particular configuration of obstacles for grounding the symbolic actions of the task plan. Current approaches either search for feasible motions offline using computationally expensive trial-and-error processes on physically realistic simulations or learn a set of motion parameters for particular object configuration spaces with little generalization. This work proposes an appealing alternative by efficiently generating trajectories for the collision-free execution of symbolic actions in variable scenarios without the need of intensive offline simulations. Our approach combines the benefit of learning from demonstration, to quickly generate an initial set of motion parameters for each symbolic action, with policy improvement with path integrals, to diversify this initial set of parameters to cope with different obstacle configurations. We show how the improved flexibility is achieved after a few minutes of training and successfully solves tasks requiring different sequences of picking and placing actions in variable configurations of obstacles.
{"title":"Combining Task and Motion Planning using Policy Improvement with Path Integrals","authors":"Dominik Urbaniak, Alejandro Agostini, Dongheui Lee","doi":"10.1109/HUMANOIDS47582.2021.9555684","DOIUrl":"https://doi.org/10.1109/HUMANOIDS47582.2021.9555684","url":null,"abstract":"Task and motion planning deals with complex tasks that require a robot to automatically define and execute multi-step sequences of actions in cluttered scenarios. In this context, a linear motion is often not sufficient to approach a target object since collisions of the gripper with other objects or the target object might occur. Thus, motion planners should be able to generate collision-free trajectories for every particular configuration of obstacles for grounding the symbolic actions of the task plan. Current approaches either search for feasible motions offline using computationally expensive trial-and-error processes on physically realistic simulations or learn a set of motion parameters for particular object configuration spaces with little generalization. This work proposes an appealing alternative by efficiently generating trajectories for the collision-free execution of symbolic actions in variable scenarios without the need of intensive offline simulations. Our approach combines the benefit of learning from demonstration, to quickly generate an initial set of motion parameters for each symbolic action, with policy improvement with path integrals, to diversify this initial set of parameters to cope with different obstacle configurations. We show how the improved flexibility is achieved after a few minutes of training and successfully solves tasks requiring different sequences of picking and placing actions in variable configurations of obstacles.","PeriodicalId":320510,"journal":{"name":"2020 IEEE-RAS 20th International Conference on Humanoid Robots (Humanoids)","volume":"77 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132450363","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-07-19DOI: 10.1109/HUMANOIDS47582.2021.9555671
Tasuku Makabe, T. Anzai, Youhei Kakiuchi, K. Okada, M. Inaba
Humanoid research aimed at verifying human movements and substituting for work deals with movement acquisition in various environments to which humans can adapt. On the other hand, water exists as an environment in which humanoids cannot adapt, even though humans adapt and demand work alternatives. Therefore, there is room to construct a platform that can operate underwater and the ground and verify the method of acquiring operation underwater. This study constructs the humanoid that can operate in both land and water environments using modular components that can easily change the body structure. While changing the environment’s force, such as frictional force, water resistance, and buoyancy, the humanoid performed moving movements such as swimming and walking in multiple lands and water settings. Walking motion experiments have shown that the underwater environment’s viscosity effectively reduces the speed of falls and prevents damage in humanoid experiments. We investigated a solution to the problem that humanoids are vulnerable to disturbance in an environment where friction is brutal to obtain through swimming.
{"title":"Development of Amphibious Humanoid for Behavior Acquisition on Land and Underwater","authors":"Tasuku Makabe, T. Anzai, Youhei Kakiuchi, K. Okada, M. Inaba","doi":"10.1109/HUMANOIDS47582.2021.9555671","DOIUrl":"https://doi.org/10.1109/HUMANOIDS47582.2021.9555671","url":null,"abstract":"Humanoid research aimed at verifying human movements and substituting for work deals with movement acquisition in various environments to which humans can adapt. On the other hand, water exists as an environment in which humanoids cannot adapt, even though humans adapt and demand work alternatives. Therefore, there is room to construct a platform that can operate underwater and the ground and verify the method of acquiring operation underwater. This study constructs the humanoid that can operate in both land and water environments using modular components that can easily change the body structure. While changing the environment’s force, such as frictional force, water resistance, and buoyancy, the humanoid performed moving movements such as swimming and walking in multiple lands and water settings. Walking motion experiments have shown that the underwater environment’s viscosity effectively reduces the speed of falls and prevents damage in humanoid experiments. We investigated a solution to the problem that humanoids are vulnerable to disturbance in an environment where friction is brutal to obtain through swimming.","PeriodicalId":320510,"journal":{"name":"2020 IEEE-RAS 20th International Conference on Humanoid Robots (Humanoids)","volume":"99 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134641307","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-07-19DOI: 10.1109/HUMANOIDS47582.2021.9555772
Kento Kawaharazuka, Yuya Koga, Manabu Nishiura, Yusuke Omura, Yuki Asano, K. Okada, Koji Kawasaki, M. Inaba
While musculoskeletal humanoids have the advantages of various biomimetic structures, it is difficult to accurately control the body, which is challenging to model. Although various learning-based control methods have been developed so far, they cannot completely absorb model errors, and recognition errors are also bound to occur. In this paper, we describe a method to modify the movement of the musculoskeletal humanoid by applying external force during the movement, taking advantage of its flexible body. Considering the fact that the joint angles cannot be measured, and that the external force greatly affects the nonlinear elastic element and not the actuator, the modified motion is reproduced by the proposed muscle-based compensation control. This method is applied to a musculoskeletal humanoid, Musashi, and its effectiveness is confirmed.
{"title":"Motion Modification Method of Musculoskeletal Humanoids by Human Teaching Using Muscle-Based Compensation Control","authors":"Kento Kawaharazuka, Yuya Koga, Manabu Nishiura, Yusuke Omura, Yuki Asano, K. Okada, Koji Kawasaki, M. Inaba","doi":"10.1109/HUMANOIDS47582.2021.9555772","DOIUrl":"https://doi.org/10.1109/HUMANOIDS47582.2021.9555772","url":null,"abstract":"While musculoskeletal humanoids have the advantages of various biomimetic structures, it is difficult to accurately control the body, which is challenging to model. Although various learning-based control methods have been developed so far, they cannot completely absorb model errors, and recognition errors are also bound to occur. In this paper, we describe a method to modify the movement of the musculoskeletal humanoid by applying external force during the movement, taking advantage of its flexible body. Considering the fact that the joint angles cannot be measured, and that the external force greatly affects the nonlinear elastic element and not the actuator, the modified motion is reproduced by the proposed muscle-based compensation control. This method is applied to a musculoskeletal humanoid, Musashi, and its effectiveness is confirmed.","PeriodicalId":320510,"journal":{"name":"2020 IEEE-RAS 20th International Conference on Humanoid Robots (Humanoids)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127197912","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-07-19DOI: 10.1109/HUMANOIDS47582.2021.9555776
Juan D. Gamba, A. C. Leite, R. Featherstone
This paper presents a simulation study of the balancing problem for a monopod robot in which the lower body (the leg) has been modified to include a passively spring-loaded prismatic joint. Such a mechanism can move by hopping but can also stand and balance on a single point. We aim to investigate the extent to which a balance controller can deal with the large values and rapid changes in the spring-damper forces, while controlling the absolute positions and orientations of its parts and balancing on one leg. It can be shown that a good performance is achieved if the spring-loaded joint is instrumented and calibrated so that its position and velocity, as well as the stiffness and damping coefficients, are considered when calculating the controller state variables. We also demonstrate the effectiveness of the balance controller by adding a high-order sliding mode (HOSM) observer based on the finite-time algorithm for robust parameter estimation of the stiffness and damping coefficients. The stability analysis and convergence proofs are presented based on the Lyapunov stability theory. Numerical simulations are included to illustrate the performance and feasibility of the proposed methodology.
{"title":"Robust Balancing Control of a Spring-legged Robot based on a High-order Sliding Mode Observer","authors":"Juan D. Gamba, A. C. Leite, R. Featherstone","doi":"10.1109/HUMANOIDS47582.2021.9555776","DOIUrl":"https://doi.org/10.1109/HUMANOIDS47582.2021.9555776","url":null,"abstract":"This paper presents a simulation study of the balancing problem for a monopod robot in which the lower body (the leg) has been modified to include a passively spring-loaded prismatic joint. Such a mechanism can move by hopping but can also stand and balance on a single point. We aim to investigate the extent to which a balance controller can deal with the large values and rapid changes in the spring-damper forces, while controlling the absolute positions and orientations of its parts and balancing on one leg. It can be shown that a good performance is achieved if the spring-loaded joint is instrumented and calibrated so that its position and velocity, as well as the stiffness and damping coefficients, are considered when calculating the controller state variables. We also demonstrate the effectiveness of the balance controller by adding a high-order sliding mode (HOSM) observer based on the finite-time algorithm for robust parameter estimation of the stiffness and damping coefficients. The stability analysis and convergence proofs are presented based on the Lyapunov stability theory. Numerical simulations are included to illustrate the performance and feasibility of the proposed methodology.","PeriodicalId":320510,"journal":{"name":"2020 IEEE-RAS 20th International Conference on Humanoid Robots (Humanoids)","volume":"110 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115450553","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-07-19DOI: 10.1109/HUMANOIDS47582.2021.9555681
Thomas Eiband, Dongheui Lee
Learning from Demonstration (LfD) can significantly speed up the knowledge transfer from human to robot, which has been proven for relatively unconstrained actions such as pick and place. However, transferring contact or force-based skills (contact skills) to a robot is noticeably harder since force and position constraints need to be considered simultaneously. We propose a set of contact skills, which differ in the force and kinematic constraints. In a first user study, several subjects were asked to term a variety of force-based interactions, from which skill names were derived. In a second and third user study, the identified skill names are used to let a test group of subjects classify the shown interactions. To evaluate the skill recognition from the robot perspective, we propose a feature-based classification scheme to recognize such skills with a robotic system in a LfD setting. Our findings prove that humans are able to understand the meaning of the different skills and, using the classification pipeline, the robot is able to recognize the different skills from human demonstrations.
{"title":"Identification of Common Force-based Robot Skills from the Human and Robot Perspective","authors":"Thomas Eiband, Dongheui Lee","doi":"10.1109/HUMANOIDS47582.2021.9555681","DOIUrl":"https://doi.org/10.1109/HUMANOIDS47582.2021.9555681","url":null,"abstract":"Learning from Demonstration (LfD) can significantly speed up the knowledge transfer from human to robot, which has been proven for relatively unconstrained actions such as pick and place. However, transferring contact or force-based skills (contact skills) to a robot is noticeably harder since force and position constraints need to be considered simultaneously. We propose a set of contact skills, which differ in the force and kinematic constraints. In a first user study, several subjects were asked to term a variety of force-based interactions, from which skill names were derived. In a second and third user study, the identified skill names are used to let a test group of subjects classify the shown interactions. To evaluate the skill recognition from the robot perspective, we propose a feature-based classification scheme to recognize such skills with a robotic system in a LfD setting. Our findings prove that humans are able to understand the meaning of the different skills and, using the classification pipeline, the robot is able to recognize the different skills from human demonstrations.","PeriodicalId":320510,"journal":{"name":"2020 IEEE-RAS 20th International Conference on Humanoid Robots (Humanoids)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-07-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114635104","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：