Pub Date : 2017-11-01DOI: 10.1109/HUMANOIDS.2017.8239546
D. Kanoulas, Chengxu Zhou, Anh Nguyen, Georgios Kanoulas, D. Caldwell, N. Tsagarakis
Reasoning about contacts between a legged robot's foot and the ground is a critical aspect of locomotion in natural terrains. This interaction becomes even more critical when the robot must move on rough surfaces. This paper presents a new visual contact analysis, based on curved patches that model local contact surfaces both on the sole of the robot's foot and in the terrain. The focus is on rigid, flat feet that represent the majority of the designs for current humanoids, but we also show how the introduced framework could be extended to other foot profiles, such as spherical feet. The footholds are localized visually in the environment's point cloud through a fast patch fitting process and a contact analysis between patches on the sole of the foot and in the surrounding environment. These patches aim to compose a spatial patch map for contact reasoning. We experimentally validate the introduced visionbased framework, using range data for rough terrain stepping demonstrations on the COMAN and WALK-MAN humanoids.
{"title":"Vision-based foothold contact reasoning using curved surface patches","authors":"D. Kanoulas, Chengxu Zhou, Anh Nguyen, Georgios Kanoulas, D. Caldwell, N. Tsagarakis","doi":"10.1109/HUMANOIDS.2017.8239546","DOIUrl":"https://doi.org/10.1109/HUMANOIDS.2017.8239546","url":null,"abstract":"Reasoning about contacts between a legged robot's foot and the ground is a critical aspect of locomotion in natural terrains. This interaction becomes even more critical when the robot must move on rough surfaces. This paper presents a new visual contact analysis, based on curved patches that model local contact surfaces both on the sole of the robot's foot and in the terrain. The focus is on rigid, flat feet that represent the majority of the designs for current humanoids, but we also show how the introduced framework could be extended to other foot profiles, such as spherical feet. The footholds are localized visually in the environment's point cloud through a fast patch fitting process and a contact analysis between patches on the sole of the foot and in the surrounding environment. These patches aim to compose a spatial patch map for contact reasoning. We experimentally validate the introduced visionbased framework, using range data for rough terrain stepping demonstrations on the COMAN and WALK-MAN humanoids.","PeriodicalId":143992,"journal":{"name":"2017 IEEE-RAS 17th International Conference on Humanoid Robotics (Humanoids)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129002936","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 : 2017-11-01DOI: 10.1109/HUMANOIDS.2017.8239543
A. Albini, S. Denei, G. Cannata
This paper describes a procedure aimed to integrate tactile sensors into a real robot in order to create a platform suitable for human-robot physical interaction experiments. Furthermore, a framework for human-robot physical interaction based on tactile feedback and prioritized tasks control is presented. The framework has been successfully tested by defining and executing three physical interaction tasks. A further experiment has been performed, simulating a human intervention during a task execution.
{"title":"Enabling natural human-robot physical interaction using a robotic skin feedback and a prioritized tasks robot control architecture","authors":"A. Albini, S. Denei, G. Cannata","doi":"10.1109/HUMANOIDS.2017.8239543","DOIUrl":"https://doi.org/10.1109/HUMANOIDS.2017.8239543","url":null,"abstract":"This paper describes a procedure aimed to integrate tactile sensors into a real robot in order to create a platform suitable for human-robot physical interaction experiments. Furthermore, a framework for human-robot physical interaction based on tactile feedback and prioritized tasks control is presented. The framework has been successfully tested by defining and executing three physical interaction tasks. A further experiment has been performed, simulating a human intervention during a task execution.","PeriodicalId":143992,"journal":{"name":"2017 IEEE-RAS 17th International Conference on Humanoid Robotics (Humanoids)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129657850","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}
This paper proposes an innovative method to employ a ducted-fan propulsion system to drive the legs of a bipedal robot, named Jet-HR1 (Jet Humanoid Robot ver.1), over broad ditches. two-dimensional gaits are chosen along with trajectory planning to calculate the thrust torque offered by the ducted fan for balancing the gravitational moment. The effectiveness and practicality are validated by Jet-HR1 stepping over a wide ditch with 370 mm, using almost 80% of its leg, both in simulation and experiment.
{"title":"Jet-HR1: Two-dimensional bipedal robot step over large obstacle based on a ducted-fan propulsion system","authors":"Zhifeng Huang, Biao Liu, Jiapeng Wei, Qingsheng Lin, J. Ota, Yun Zhang","doi":"10.1109/HUMANOIDS.2017.8246905","DOIUrl":"https://doi.org/10.1109/HUMANOIDS.2017.8246905","url":null,"abstract":"This paper proposes an innovative method to employ a ducted-fan propulsion system to drive the legs of a bipedal robot, named Jet-HR1 (Jet Humanoid Robot ver.1), over broad ditches. two-dimensional gaits are chosen along with trajectory planning to calculate the thrust torque offered by the ducted fan for balancing the gravitational moment. The effectiveness and practicality are validated by Jet-HR1 stepping over a wide ditch with 370 mm, using almost 80% of its leg, both in simulation and experiment.","PeriodicalId":143992,"journal":{"name":"2017 IEEE-RAS 17th International Conference on Humanoid Robotics (Humanoids)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125454842","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 : 2017-11-01DOI: 10.1109/HUMANOIDS.2017.8239551
Anna Lena Kleesattel, W. Potthast, K. Mombaur
Amputee sprinting motions are of high interest due to the remarkable performances of individual athletes and the resulting question what the share of their running-specific prostheses is. The goal of our study was to compare twelve ablebodied and amputee sprinting motions which were synthesized by combinations of ten optimality criteria. We created rigid multi-body system models with 13 degrees of freedom in the sagittal plane for both an able-bodied and an amputee sprinter. The joints are powered by torque actuators, except for the prosthetic joint which is equipped with a passive linear springdamper system. The sprinting motions are the solutions of an optimal control problem with periodicity constraints and objective functions which combine different optimality criteria. For both athletes, we found realistic human-like sprinting motions. The analysis of the motions suggested that the amputee athlete applies less active joint torque in the hip and knee of his affected leg compared to the able-bodied athlete.
{"title":"Towards a better understanding of human sprinting motions with and without prostheses","authors":"Anna Lena Kleesattel, W. Potthast, K. Mombaur","doi":"10.1109/HUMANOIDS.2017.8239551","DOIUrl":"https://doi.org/10.1109/HUMANOIDS.2017.8239551","url":null,"abstract":"Amputee sprinting motions are of high interest due to the remarkable performances of individual athletes and the resulting question what the share of their running-specific prostheses is. The goal of our study was to compare twelve ablebodied and amputee sprinting motions which were synthesized by combinations of ten optimality criteria. We created rigid multi-body system models with 13 degrees of freedom in the sagittal plane for both an able-bodied and an amputee sprinter. The joints are powered by torque actuators, except for the prosthetic joint which is equipped with a passive linear springdamper system. The sprinting motions are the solutions of an optimal control problem with periodicity constraints and objective functions which combine different optimality criteria. For both athletes, we found realistic human-like sprinting motions. The analysis of the motions suggested that the amputee athlete applies less active joint torque in the hip and knee of his affected leg compared to the able-bodied athlete.","PeriodicalId":143992,"journal":{"name":"2017 IEEE-RAS 17th International Conference on Humanoid Robotics (Humanoids)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121799177","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 : 2017-11-01DOI: 10.1109/HUMANOIDS.2017.8239547
Masahiro Bando, Masaki Murooka, Iori Yanokura, Shunichi Nozawa, K. Okada, M. Inaba
For robots to act in outdoor environments, the ability for locomotion is an important factor because it influences the range of activities. Humanoid robots are acquiring the ability for locomotion in horizontal directions so far, and the ability for vertical locomotion is now required in order to extend their range of activities further. Therefore, we propose rappelling as a method for vertical locomotion. Rappelling is a method to descend from a high place using a rope and a belay device. With this method, humanoid robots are expected to get the ability for vertical locomotion without additional actuators. In this paper, we propose a full-body motion for rappelling based on the sequential transition of the centroid position and the contact states. We also propose the reliable manipulation of a rope grasped by both hands. With the proposed methods, we conducted an experiment of rappelling with HRP2, and it could successfully descend from a height of 2m50cm in 277 seconds.
{"title":"Rappelling by a humanoid robot based on transition motion generation and reliable rope manipulation","authors":"Masahiro Bando, Masaki Murooka, Iori Yanokura, Shunichi Nozawa, K. Okada, M. Inaba","doi":"10.1109/HUMANOIDS.2017.8239547","DOIUrl":"https://doi.org/10.1109/HUMANOIDS.2017.8239547","url":null,"abstract":"For robots to act in outdoor environments, the ability for locomotion is an important factor because it influences the range of activities. Humanoid robots are acquiring the ability for locomotion in horizontal directions so far, and the ability for vertical locomotion is now required in order to extend their range of activities further. Therefore, we propose rappelling as a method for vertical locomotion. Rappelling is a method to descend from a high place using a rope and a belay device. With this method, humanoid robots are expected to get the ability for vertical locomotion without additional actuators. In this paper, we propose a full-body motion for rappelling based on the sequential transition of the centroid position and the contact states. We also propose the reliable manipulation of a rope grasped by both hands. With the proposed methods, we conducted an experiment of rappelling with HRP2, and it could successfully descend from a height of 2m50cm in 277 seconds.","PeriodicalId":143992,"journal":{"name":"2017 IEEE-RAS 17th International Conference on Humanoid Robotics (Humanoids)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132584477","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 : 2017-11-01DOI: 10.1109/HUMANOIDS.2017.8246928
Brandon J. DeHart, D. Kulić
The ability of a legged system to balance depends on both the control strategy used and the system's physical design. To quantify a system's inherent balance capabilities, we define momentum gains for general 2D and 3D models. We provide two methods for calculating these gains, and relate both velocity and momentum gains to the centroidal momentum of a system, a commonly used measure of aggregate system behavior. Finally, we compare velocity and momentum gains as criteria for the design of simple balancing systems using a parameterized optimization framework.
{"title":"Quantifying balance capabilities using momentum gain","authors":"Brandon J. DeHart, D. Kulić","doi":"10.1109/HUMANOIDS.2017.8246928","DOIUrl":"https://doi.org/10.1109/HUMANOIDS.2017.8246928","url":null,"abstract":"The ability of a legged system to balance depends on both the control strategy used and the system's physical design. To quantify a system's inherent balance capabilities, we define momentum gains for general 2D and 3D models. We provide two methods for calculating these gains, and relate both velocity and momentum gains to the centroidal momentum of a system, a commonly used measure of aggregate system behavior. Finally, we compare velocity and momentum gains as criteria for the design of simple balancing systems using a parameterized optimization framework.","PeriodicalId":143992,"journal":{"name":"2017 IEEE-RAS 17th International Conference on Humanoid Robotics (Humanoids)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132761346","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 : 2017-11-01DOI: 10.1109/HUMANOIDS.2017.8246943
Barry Ridge, Timotej Gaspar, A. Ude
As robotic systems have become more and more complex and difficult to manage, various software architectures, libraries and programming paradigms have been introduced aimed at high-level control and integration of their constituent parts. The Robot Operating System (ROS) has, for many, become the de facto software framework for communication standardisation and hardware interface abstraction, and various packages within its ecosystem have come to the fore as being reliable design choices for dictating control flow. ROSbased software packages that use state machines as their core methodology to bridge the gap between low-level imperative task scripting and higher-level task planning have proven particularly popular. However, while they provide much in terms of power and flexibility, their overall task-level simplicity can often be obfuscated at the script-level by boilerplate code, intricate structure and lack of code reuse between state machine prototypes. In this paper, we aim to address this deficit by proposing a code generation, templating and metascripting methodology for state machine assembly, as well as an accompanying application programming interface (API), for the rapid, modular development of robot control programs. The API has been developed to function effectively as either a frontend for concise scripting or a back-end for code generation for visual programming systems. Its capabilities are demonstrated in an experiment using a simulated humanoid robot platform.
{"title":"Rapid state machine assembly for modular robot control using meta-scripting, templating and code generation","authors":"Barry Ridge, Timotej Gaspar, A. Ude","doi":"10.1109/HUMANOIDS.2017.8246943","DOIUrl":"https://doi.org/10.1109/HUMANOIDS.2017.8246943","url":null,"abstract":"As robotic systems have become more and more complex and difficult to manage, various software architectures, libraries and programming paradigms have been introduced aimed at high-level control and integration of their constituent parts. The Robot Operating System (ROS) has, for many, become the de facto software framework for communication standardisation and hardware interface abstraction, and various packages within its ecosystem have come to the fore as being reliable design choices for dictating control flow. ROSbased software packages that use state machines as their core methodology to bridge the gap between low-level imperative task scripting and higher-level task planning have proven particularly popular. However, while they provide much in terms of power and flexibility, their overall task-level simplicity can often be obfuscated at the script-level by boilerplate code, intricate structure and lack of code reuse between state machine prototypes. In this paper, we aim to address this deficit by proposing a code generation, templating and metascripting methodology for state machine assembly, as well as an accompanying application programming interface (API), for the rapid, modular development of robot control programs. The API has been developed to function effectively as either a frontend for concise scripting or a back-end for code generation for visual programming systems. Its capabilities are demonstrated in an experiment using a simulated humanoid robot platform.","PeriodicalId":143992,"journal":{"name":"2017 IEEE-RAS 17th International Conference on Humanoid Robotics (Humanoids)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117084194","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 : 2017-11-01DOI: 10.1109/HUMANOIDS.2017.8246906
Z. Aftab, Asad Ali
The article presents design and preliminary simulation results for a lower-body exoskeleton system. The device has 5DoF per leg and a common hip rotation degree between both legs. The mathematical model of the combined human-exoskeleton system is developed using the Robotics Toolbox in Matlab. Simulations are carried out to observe joint torque and power profiles for straight-line on-ground walking in the sagittal plane. The results indicate the validity of developed model and show promise for the design of links and actuation systems in future.
{"title":"Simulating a wearable lower-body exoskeleton device for torque and power estimation","authors":"Z. Aftab, Asad Ali","doi":"10.1109/HUMANOIDS.2017.8246906","DOIUrl":"https://doi.org/10.1109/HUMANOIDS.2017.8246906","url":null,"abstract":"The article presents design and preliminary simulation results for a lower-body exoskeleton system. The device has 5DoF per leg and a common hip rotation degree between both legs. The mathematical model of the combined human-exoskeleton system is developed using the Robotics Toolbox in Matlab. Simulations are carried out to observe joint torque and power profiles for straight-line on-ground walking in the sagittal plane. The results indicate the validity of developed model and show promise for the design of links and actuation systems in future.","PeriodicalId":143992,"journal":{"name":"2017 IEEE-RAS 17th International Conference on Humanoid Robotics (Humanoids)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124507916","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 : 2017-11-01DOI: 10.1109/HUMANOIDS.2017.8239538
Markus Giftthaler, J. Buchli
This paper presents a state and state-input constrained variant of the discrete-time iterative Linear Quadratic Regulator (iLQR) algorithm, with linear time-complexity in the number of time steps. The approach is based on a projection of the control input onto the nullspace of the linearized constraints. We derive a fully constraint-compliant feedforward-feedback control update rule, for which we can solve efficiently with Riccati-style difference equations. We assume that the relative degree of all constraints in the discrete-time system model is equal to one, which often holds for robotics problems employing rigid-body dynamic models. Simulation examples, including a 6 DoF robotic arm, are given to validate and illustrate the performance of the method.
{"title":"A projection approach to equality constrained iterative linear quadratic optimal control","authors":"Markus Giftthaler, J. Buchli","doi":"10.1109/HUMANOIDS.2017.8239538","DOIUrl":"https://doi.org/10.1109/HUMANOIDS.2017.8239538","url":null,"abstract":"This paper presents a state and state-input constrained variant of the discrete-time iterative Linear Quadratic Regulator (iLQR) algorithm, with linear time-complexity in the number of time steps. The approach is based on a projection of the control input onto the nullspace of the linearized constraints. We derive a fully constraint-compliant feedforward-feedback control update rule, for which we can solve efficiently with Riccati-style difference equations. We assume that the relative degree of all constraints in the discrete-time system model is equal to one, which often holds for robotics problems employing rigid-body dynamic models. Simulation examples, including a 6 DoF robotic arm, are given to validate and illustrate the performance of the method.","PeriodicalId":143992,"journal":{"name":"2017 IEEE-RAS 17th International Conference on Humanoid Robotics (Humanoids)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126193881","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 : 2017-11-01DOI: 10.1109/HUMANOIDS.2017.8246935
Manfred Eppe, Matthias Kerzel, Sascha S. Griffiths, Hwei Geok Ng, S. Wermter
We present a proof of concept to show how a deep network for end-to-end visuomotor learning to grasp is coupled with an attention focus mechanism for state-of-the-art object detection with convolutional neural networks. The cognitively motivated integration of both methods in a single robotic system allows us to realize a high-level interface to use the visuomotor network in environments with several objects, which otherwise would only be usable in environments with a single object. The resulting system is deployed on a humanoid robot, and we perform several real-world grasping experiments that demonstrate the feasibility of our approach.
{"title":"Combining deep learning for visuomotor coordination with object identification to realize a high-level interface for robot object-picking","authors":"Manfred Eppe, Matthias Kerzel, Sascha S. Griffiths, Hwei Geok Ng, S. Wermter","doi":"10.1109/HUMANOIDS.2017.8246935","DOIUrl":"https://doi.org/10.1109/HUMANOIDS.2017.8246935","url":null,"abstract":"We present a proof of concept to show how a deep network for end-to-end visuomotor learning to grasp is coupled with an attention focus mechanism for state-of-the-art object detection with convolutional neural networks. The cognitively motivated integration of both methods in a single robotic system allows us to realize a high-level interface to use the visuomotor network in environments with several objects, which otherwise would only be usable in environments with a single object. The resulting system is deployed on a humanoid robot, and we perform several real-world grasping experiments that demonstrate the feasibility of our approach.","PeriodicalId":143992,"journal":{"name":"2017 IEEE-RAS 17th International Conference on Humanoid Robotics (Humanoids)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130557642","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：