Pub Date : 2017-12-01DOI: 10.1109/ROBIO.2017.8324716
Xueqian Wang, Bo Xia, Gang Li, Houde Liu, Bin Liang
This paper focuses on the problem of dynamic control of planar schematic dual-arm free-floating space robot. When the arms are approaching the target, one manipulator is designed for the task planning and the other is used to keep the posture of the space base unchanged. The whole system is virtually decomposed into several subsystems and kinematic and dynamic of each system is firstly analyzed. Then, the virtual decomposition controller of each subsystem is designed based on the built subsystems, using the virtual decomposition control (VDC) method. And all subsystem controller consists of the whole system controller. Since the space robot is free floating, there is no exerted force from the environment to the space base, which leads to 3 non-holonomic constraints on the system dynamic and is also taken into consideration in the control algorithm. Finally, simulations are performed on Matlab / Simulink platform and the simulation results show that the algorithm is effective.
{"title":"Modelling and control of dual-arm free-floating space robot using virtual decomposition control for capturing target","authors":"Xueqian Wang, Bo Xia, Gang Li, Houde Liu, Bin Liang","doi":"10.1109/ROBIO.2017.8324716","DOIUrl":"https://doi.org/10.1109/ROBIO.2017.8324716","url":null,"abstract":"This paper focuses on the problem of dynamic control of planar schematic dual-arm free-floating space robot. When the arms are approaching the target, one manipulator is designed for the task planning and the other is used to keep the posture of the space base unchanged. The whole system is virtually decomposed into several subsystems and kinematic and dynamic of each system is firstly analyzed. Then, the virtual decomposition controller of each subsystem is designed based on the built subsystems, using the virtual decomposition control (VDC) method. And all subsystem controller consists of the whole system controller. Since the space robot is free floating, there is no exerted force from the environment to the space base, which leads to 3 non-holonomic constraints on the system dynamic and is also taken into consideration in the control algorithm. Finally, simulations are performed on Matlab / Simulink platform and the simulation results show that the algorithm is effective.","PeriodicalId":197159,"journal":{"name":"2017 IEEE International Conference on Robotics and Biomimetics (ROBIO)","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123784751","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}
In view of the insufficient of flexibility and safety on the existing rigid robot, and the soft-bodied robot has the defect in the accuracy. In this paper, a robot compliant joint based on MRF (magneto-rheological fluid) is designed, the driving mechanism of the joint is analyzed and a kinematic model is established. The compliant joint is embedded with a magneto-rheological fluid actuator in a conventional robot joint, in order to increase the active compliance control function of the joint. And the safety performance is increased on the basis of the control accuracy. We process a soft joint prototype. Test of the torque output of the MRF transmission, the corner output of the passive rod and the force output were carried out. Experimental results show, Magneto-rheological fluid compliant joints can transfer the force of 0.2N and the response time can be within 200ms.
{"title":"Design and experimental study of compliant joints of robot based on magneto-rheological fluid","authors":"Shibo Cai, Pengfei Wang, Liubin Tian, Fang Xu, Libin Zhang","doi":"10.1109/ROBIO.2017.8324711","DOIUrl":"https://doi.org/10.1109/ROBIO.2017.8324711","url":null,"abstract":"In view of the insufficient of flexibility and safety on the existing rigid robot, and the soft-bodied robot has the defect in the accuracy. In this paper, a robot compliant joint based on MRF (magneto-rheological fluid) is designed, the driving mechanism of the joint is analyzed and a kinematic model is established. The compliant joint is embedded with a magneto-rheological fluid actuator in a conventional robot joint, in order to increase the active compliance control function of the joint. And the safety performance is increased on the basis of the control accuracy. We process a soft joint prototype. Test of the torque output of the MRF transmission, the corner output of the passive rod and the force output were carried out. Experimental results show, Magneto-rheological fluid compliant joints can transfer the force of 0.2N and the response time can be within 200ms.","PeriodicalId":197159,"journal":{"name":"2017 IEEE International Conference on Robotics and Biomimetics (ROBIO)","volume":"39 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126968329","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 : 2017-12-01DOI: 10.1109/ROBIO.2017.8324748
Yujin Wakita, Hideyuki Tanaka, Y. Matsumoto
A projection function is applied to information sharing between a robot and the user controlling it when they coexist within the same environment. Projection information can be projected in multiple areas and modes simultaneously. We introduce a dual-use projection function as a projection user interface and robot hand pointer for easy control of a mobile manipulator for application in service robots in daily living environments. Dual projection in a physical environment can allow the user to understand the robot and task environment and complete the task. Application of the proposed function to a basic service robot and the results of preliminary experiments are described.
{"title":"Projection function and hand pointer for user-interface of daily service robot","authors":"Yujin Wakita, Hideyuki Tanaka, Y. Matsumoto","doi":"10.1109/ROBIO.2017.8324748","DOIUrl":"https://doi.org/10.1109/ROBIO.2017.8324748","url":null,"abstract":"A projection function is applied to information sharing between a robot and the user controlling it when they coexist within the same environment. Projection information can be projected in multiple areas and modes simultaneously. We introduce a dual-use projection function as a projection user interface and robot hand pointer for easy control of a mobile manipulator for application in service robots in daily living environments. Dual projection in a physical environment can allow the user to understand the robot and task environment and complete the task. Application of the proposed function to a basic service robot and the results of preliminary experiments are described.","PeriodicalId":197159,"journal":{"name":"2017 IEEE International Conference on Robotics and Biomimetics (ROBIO)","volume":"39 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115566278","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 : 2017-12-01DOI: 10.1109/ROBIO.2017.8324573
Yun Qi, Jinkai Xu, Zhanjiang Yu, Huadong Yu
In monitoring high-speed micro-milling, acoustic emission is used to explore the relationship between the machining parameters and the acoustic emission signal under different processing parameters. The acquired acoustic emission signal is denoised by singular value decomposition based on Hankel matrix, and the characteristic values of the denoising signal is calculated by ensemble empirical mode decomposition and the Hilbert-Huang transform. Results show that the characteristic values of the acoustic emission signal can represent the change in machining parameters, such as the spindle speed, and the acoustic emission signal is suitable for monitoring the micro-milling process.
{"title":"Acoustic emission monitoring in high-speed micro end-milling based on SVD-EEMD method","authors":"Yun Qi, Jinkai Xu, Zhanjiang Yu, Huadong Yu","doi":"10.1109/ROBIO.2017.8324573","DOIUrl":"https://doi.org/10.1109/ROBIO.2017.8324573","url":null,"abstract":"In monitoring high-speed micro-milling, acoustic emission is used to explore the relationship between the machining parameters and the acoustic emission signal under different processing parameters. The acquired acoustic emission signal is denoised by singular value decomposition based on Hankel matrix, and the characteristic values of the denoising signal is calculated by ensemble empirical mode decomposition and the Hilbert-Huang transform. Results show that the characteristic values of the acoustic emission signal can represent the change in machining parameters, such as the spindle speed, and the acoustic emission signal is suitable for monitoring the micro-milling process.","PeriodicalId":197159,"journal":{"name":"2017 IEEE International Conference on Robotics and Biomimetics (ROBIO)","volume":"107 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116377448","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 : 2017-12-01DOI: 10.1109/ROBIO.2017.8324646
Yuan Luo, Yu Liu, Yi Zhang, Boyu Wang, Zhou Ye
Recently long short-term memory (LSTM) recurrent neural networks (RNN) have achieved greater success in acoustic models for the large vocabulary continuous speech recognition system. In this paper, we propose an improved hybrid acoustic model based on deep bidirectional long short-term memory (DBLSTM) RNN. In this new acoustic model, maxout neurons are used in the fully-connected part of DBLSTM to solve the problems of vanishing and exploding gradient. At the same time, the dropout regularization algorithm is used to avoid the over-fitting during the training process of neural network. In addition, in order to adapt the bidirectional dependence of DBLSTM at each time step, a context-sensitive-chunk (CSC) back-propagation through time (BPTT) algorithm is proposed to train DBLSTM neural network. Simulation experiments have been made on Switchboard benchmark task. The results show that the WER of the improved hybrid acoustic model is 14.5%, and the optimal network structures and CSC configurations are given.
{"title":"Maxout neurons based deep bidirectional LSTM for acoustic modeling","authors":"Yuan Luo, Yu Liu, Yi Zhang, Boyu Wang, Zhou Ye","doi":"10.1109/ROBIO.2017.8324646","DOIUrl":"https://doi.org/10.1109/ROBIO.2017.8324646","url":null,"abstract":"Recently long short-term memory (LSTM) recurrent neural networks (RNN) have achieved greater success in acoustic models for the large vocabulary continuous speech recognition system. In this paper, we propose an improved hybrid acoustic model based on deep bidirectional long short-term memory (DBLSTM) RNN. In this new acoustic model, maxout neurons are used in the fully-connected part of DBLSTM to solve the problems of vanishing and exploding gradient. At the same time, the dropout regularization algorithm is used to avoid the over-fitting during the training process of neural network. In addition, in order to adapt the bidirectional dependence of DBLSTM at each time step, a context-sensitive-chunk (CSC) back-propagation through time (BPTT) algorithm is proposed to train DBLSTM neural network. Simulation experiments have been made on Switchboard benchmark task. The results show that the WER of the improved hybrid acoustic model is 14.5%, and the optimal network structures and CSC configurations are given.","PeriodicalId":197159,"journal":{"name":"2017 IEEE International Conference on Robotics and Biomimetics (ROBIO)","volume":"454 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116410189","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 : 2017-12-01DOI: 10.1109/ROBIO.2017.8324571
Xi Zhang, Jian Zhang
Line-structured light sensor (LLS) can provide the capability of three-dimensional point acquisition for robotics. As one type of 3D scanners, LLS has been widely used in many filed of robotics for its strong anti-interference, fast scanning speed and high measuring accuracy. Researchers have been studying the methods to improve the accuracy, and the operability of the LLS for many years. In this paper, calibration methods for LLS are reviewed, which covers the calibration target and the light plane calibration methods. What's more, some potential improvements are discussed and analyzed. This review paper can be used as a reference guide for researchers to choose suitable calibration method and calibration target for developing an applicable LLS.
{"title":"Summary on calibration method of line-structured light sensor","authors":"Xi Zhang, Jian Zhang","doi":"10.1109/ROBIO.2017.8324571","DOIUrl":"https://doi.org/10.1109/ROBIO.2017.8324571","url":null,"abstract":"Line-structured light sensor (LLS) can provide the capability of three-dimensional point acquisition for robotics. As one type of 3D scanners, LLS has been widely used in many filed of robotics for its strong anti-interference, fast scanning speed and high measuring accuracy. Researchers have been studying the methods to improve the accuracy, and the operability of the LLS for many years. In this paper, calibration methods for LLS are reviewed, which covers the calibration target and the light plane calibration methods. What's more, some potential improvements are discussed and analyzed. This review paper can be used as a reference guide for researchers to choose suitable calibration method and calibration target for developing an applicable LLS.","PeriodicalId":197159,"journal":{"name":"2017 IEEE International Conference on Robotics and Biomimetics (ROBIO)","volume":"48 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122706433","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 : 2017-12-01DOI: 10.1109/ROBIO.2017.8324511
U. Scarcia, R. Meattini, C. Melchiorri
The mapping of the human intention to a dexterous anthropomorphic robotic hand is still an open issue among researchers. The complexity behind this problems comes mainly from three factors: the kinematics differences between the users and the robotic hand(s); the differences in size and motion capabilities among different users hands; and the high number of degrees of freedom present in an anthropomorphic hand. In this work, we present a procedure for the determination of a linear transformation capable to interface the user and the robot kinematics and therefore to allow a precise and natural control of the mechanical device. The main assumption that we make is that different human hand kinematics differ-with a good approximation-for a scaling factor only, whereas the proportions between the phalanges lengths and the relative orientation of the fingers are kept almost constant in healthy people [1]. We also assume that, being the considered robotic hand highly anthropomorphic, this condition holds also between the user and the robotic hand. In addition, while for a robotic hand the definition of a reference frame fixed to the palm is a free choice, for the human hand tracked with some external system it is completely software dependent. Therefore additional rotational and translational corrective terms have to be introduced to compensate for the different placement of the palm reference frame with respect to the fingers. We have applied this approach to control the UB-Hand IV using a commercial device called Leap Motion, able to track with a good accuracy the pose of the palm and the positions of the key points of the human hand, i.e. the end points of the hand bones [2].
{"title":"Mapping human hand fingertips motion to an anthropomorphic robotic hand","authors":"U. Scarcia, R. Meattini, C. Melchiorri","doi":"10.1109/ROBIO.2017.8324511","DOIUrl":"https://doi.org/10.1109/ROBIO.2017.8324511","url":null,"abstract":"The mapping of the human intention to a dexterous anthropomorphic robotic hand is still an open issue among researchers. The complexity behind this problems comes mainly from three factors: the kinematics differences between the users and the robotic hand(s); the differences in size and motion capabilities among different users hands; and the high number of degrees of freedom present in an anthropomorphic hand. In this work, we present a procedure for the determination of a linear transformation capable to interface the user and the robot kinematics and therefore to allow a precise and natural control of the mechanical device. The main assumption that we make is that different human hand kinematics differ-with a good approximation-for a scaling factor only, whereas the proportions between the phalanges lengths and the relative orientation of the fingers are kept almost constant in healthy people [1]. We also assume that, being the considered robotic hand highly anthropomorphic, this condition holds also between the user and the robotic hand. In addition, while for a robotic hand the definition of a reference frame fixed to the palm is a free choice, for the human hand tracked with some external system it is completely software dependent. Therefore additional rotational and translational corrective terms have to be introduced to compensate for the different placement of the palm reference frame with respect to the fingers. We have applied this approach to control the UB-Hand IV using a commercial device called Leap Motion, able to track with a good accuracy the pose of the palm and the positions of the key points of the human hand, i.e. the end points of the hand bones [2].","PeriodicalId":197159,"journal":{"name":"2017 IEEE International Conference on Robotics and Biomimetics (ROBIO)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129785080","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 : 2017-12-01DOI: 10.1109/ROBIO.2017.8324386
Yuangen Wei, Wenzeng Zhang
Inspired from flexible bending of octopus' tentacles and outside-driving kind of traditional hand exoskeletons, this paper proposes a novel self-adaptive underactuated multi-fingered hand (OS Hand), which has four flexible tentacles. Each tentacle is similar to an octopus' tentacle, and consists of an artificial muscle which goes through all joints, eight serial-hinged joints, and force-changeable assembly. The force-changeable assembly is mainly composed of a spring and elastic rubber membrane, which is coordinated for stable grasping by a layer of rubber material in the surface of the tentacle. Each tentacle can execute different grasping modes depending on the shapes and dimensions of the objects grasped and grip objects in a gentle and form-fitting manner. The OS Hand combines good qualities of both powerful grasp of traditional grippers and form-fitting grasp of flexible hands. Kinematic analysis and mathematic model disclose the distribution of contact forces and compare the tentacles in the OS Hand with traditional rigid tentacles. Experimental results show that the OS Hand is valid for precise pinching, self-adaptive powerful encompassing, and grasping forces are freely changeable in a wide range. With the advantages of high self-adaptation, various grasp configurations and large range of grasping forces, the OS Hand has a wide range of applications in the area of service robotics which requires a lot of flexible operations of general grasping, moving, and releasing.
{"title":"OS hand: Octopus-inspired self-adaptive underactuated hand with fluid-driven tentacles and force-changeable artificial muscles","authors":"Yuangen Wei, Wenzeng Zhang","doi":"10.1109/ROBIO.2017.8324386","DOIUrl":"https://doi.org/10.1109/ROBIO.2017.8324386","url":null,"abstract":"Inspired from flexible bending of octopus' tentacles and outside-driving kind of traditional hand exoskeletons, this paper proposes a novel self-adaptive underactuated multi-fingered hand (OS Hand), which has four flexible tentacles. Each tentacle is similar to an octopus' tentacle, and consists of an artificial muscle which goes through all joints, eight serial-hinged joints, and force-changeable assembly. The force-changeable assembly is mainly composed of a spring and elastic rubber membrane, which is coordinated for stable grasping by a layer of rubber material in the surface of the tentacle. Each tentacle can execute different grasping modes depending on the shapes and dimensions of the objects grasped and grip objects in a gentle and form-fitting manner. The OS Hand combines good qualities of both powerful grasp of traditional grippers and form-fitting grasp of flexible hands. Kinematic analysis and mathematic model disclose the distribution of contact forces and compare the tentacles in the OS Hand with traditional rigid tentacles. Experimental results show that the OS Hand is valid for precise pinching, self-adaptive powerful encompassing, and grasping forces are freely changeable in a wide range. With the advantages of high self-adaptation, various grasp configurations and large range of grasping forces, the OS Hand has a wide range of applications in the area of service robotics which requires a lot of flexible operations of general grasping, moving, and releasing.","PeriodicalId":197159,"journal":{"name":"2017 IEEE International Conference on Robotics and Biomimetics (ROBIO)","volume":"24 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128453390","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 : 2017-12-01DOI: 10.1109/ROBIO.2017.8324687
Cheng Li, H. Gu
Leveling error is critical for final alignment accuracy in accurate assembly tasks, which needs to be sensed and further compensated. In this paper, we propose an easy and cost-saving leveling measurement method. The parameter sensitivity is studied, and the leveling compensation approach in robotic systems is stated. We conduct experiments to examine its accuracy performance, and show its effectiveness in a demonstration assembly case.
{"title":"A leveling measurement method and its application in robotic systems","authors":"Cheng Li, H. Gu","doi":"10.1109/ROBIO.2017.8324687","DOIUrl":"https://doi.org/10.1109/ROBIO.2017.8324687","url":null,"abstract":"Leveling error is critical for final alignment accuracy in accurate assembly tasks, which needs to be sensed and further compensated. In this paper, we propose an easy and cost-saving leveling measurement method. The parameter sensitivity is studied, and the leveling compensation approach in robotic systems is stated. We conduct experiments to examine its accuracy performance, and show its effectiveness in a demonstration assembly case.","PeriodicalId":197159,"journal":{"name":"2017 IEEE International Conference on Robotics and Biomimetics (ROBIO)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129924740","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 : 2017-12-01DOI: 10.1109/ROBIO.2017.8324409
Wataru Yamazaki, Ming Ding, J. Takamatsu, T. Ogasawara
Manipulation performed by humans contains a lot of information that helps robots to learn how to handle objects. Since hand poses and motions are related to manipulated objects, extracting this information is one of the important tasks for robotics community. This paper presents a framework to recognize human manipulations including hand motions, hand poses, and shapes of manipulated objects using egocentric RGB-D videos. Our framework is straightforward but powerful through the efficient use of depth information and egocentric vision. We estimate hand poses with an example-based method through the limited appearances of the hand in egocentric vision. First, from a sensed point cloud, our framework distinguishes hands, manipulated objects and an environment using skin color detection and limitation on the range of the moving hand. Next, we estimate a hand pose by aligning the extracted hand point cloud with a pre-recorded database of hand point clouds of different poses. The position and orientation of the head-mounted sensor are estimated to acquire the hand motion in the world coordinate system. Then, the type of hand motion is classified using Dynamic Programming matching between a series of velocity vectors of estimated and a database of wrist trajectories. Finally, we experiment the effectiveness of our framework for hand motion recognition to validate our work.
{"title":"Hand pose estimation and motion recognition using egocentric RGB-D video","authors":"Wataru Yamazaki, Ming Ding, J. Takamatsu, T. Ogasawara","doi":"10.1109/ROBIO.2017.8324409","DOIUrl":"https://doi.org/10.1109/ROBIO.2017.8324409","url":null,"abstract":"Manipulation performed by humans contains a lot of information that helps robots to learn how to handle objects. Since hand poses and motions are related to manipulated objects, extracting this information is one of the important tasks for robotics community. This paper presents a framework to recognize human manipulations including hand motions, hand poses, and shapes of manipulated objects using egocentric RGB-D videos. Our framework is straightforward but powerful through the efficient use of depth information and egocentric vision. We estimate hand poses with an example-based method through the limited appearances of the hand in egocentric vision. First, from a sensed point cloud, our framework distinguishes hands, manipulated objects and an environment using skin color detection and limitation on the range of the moving hand. Next, we estimate a hand pose by aligning the extracted hand point cloud with a pre-recorded database of hand point clouds of different poses. The position and orientation of the head-mounted sensor are estimated to acquire the hand motion in the world coordinate system. Then, the type of hand motion is classified using Dynamic Programming matching between a series of velocity vectors of estimated and a database of wrist trajectories. Finally, we experiment the effectiveness of our framework for hand motion recognition to validate our work.","PeriodicalId":197159,"journal":{"name":"2017 IEEE International Conference on Robotics and Biomimetics (ROBIO)","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126804372","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：