Pub Date : 2012-12-24DOI: 10.1109/IROS.2012.6385784
C. Rossa, J. Lozada, A. Micaelli
A new concept of hybrid actuator for haptic devices is proposed. This system combines a controllable magnetorheological brake with a conventional DC motor. Both actuators are linked through an overrunning clutch. Thus, the motor is connected to the handle while the brake can exert a resistive force only in a defined direction. This configuration enables the brake and the motor to be engaged at the same time because the torque imposed by the motor is not canceled by the brake. The concept and its control laws have been investigated using a 1-DOF haptic device. The experimental results show that is possible to combine a powerful brake with a small DC motor. This approach reduces the power consumption, expand the range of forces, achieve global stability in the system providing thereby safety to the user. Besides, the proposed independent control laws enable the actuator to be adaptable in many different haptic applications.
{"title":"A new hybrid actuator approach for force-feedback devices","authors":"C. Rossa, J. Lozada, A. Micaelli","doi":"10.1109/IROS.2012.6385784","DOIUrl":"https://doi.org/10.1109/IROS.2012.6385784","url":null,"abstract":"A new concept of hybrid actuator for haptic devices is proposed. This system combines a controllable magnetorheological brake with a conventional DC motor. Both actuators are linked through an overrunning clutch. Thus, the motor is connected to the handle while the brake can exert a resistive force only in a defined direction. This configuration enables the brake and the motor to be engaged at the same time because the torque imposed by the motor is not canceled by the brake. The concept and its control laws have been investigated using a 1-DOF haptic device. The experimental results show that is possible to combine a powerful brake with a small DC motor. This approach reduces the power consumption, expand the range of forces, achieve global stability in the system providing thereby safety to the user. Besides, the proposed independent control laws enable the actuator to be adaptable in many different haptic applications.","PeriodicalId":6358,"journal":{"name":"2012 IEEE/RSJ International Conference on Intelligent Robots and Systems","volume":"101 1","pages":"4054-4059"},"PeriodicalIF":0.0,"publicationDate":"2012-12-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87489416","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 : 2012-12-24DOI: 10.1109/IROS.2012.6385671
S. Giannarou, Zhiqiang Zhang, Guang-Zhong Yang
Accurate recovery of the 3D structure of a deforming surgical environment during minimally invasive surgery is important for intra-operative guidance. One key component of reliable reconstruction is accurate camera pose estimation, which is challenging for monocular cameras due to the paucity of reliable salient features, coupled with narrow baseline during surgical navigation. With recent advances in miniaturized MEMS sensors, the combination of inertial and vision sensing can provide increased robustness for camera pose estimation particularly for scenes involving tissue deformation. The aim of this work is to propose a robust framework for intra-operative free-form deformation recovery based on structure-from-motion. A novel adaptive Unscented Kalman Filter (UKF) parameterization scheme is proposed to fuse vision information with data from an Inertial Measurement Unit (IMU). The method is built on a compact scene representation scheme suitable for both surgical episode identification and instrument-tissue motion modelling. Detailed validation with both synthetic and phantom data is performed and results derived justify the potential clinical value of the technique.
{"title":"Deformable structure from motion by fusing visual and inertial measurement data","authors":"S. Giannarou, Zhiqiang Zhang, Guang-Zhong Yang","doi":"10.1109/IROS.2012.6385671","DOIUrl":"https://doi.org/10.1109/IROS.2012.6385671","url":null,"abstract":"Accurate recovery of the 3D structure of a deforming surgical environment during minimally invasive surgery is important for intra-operative guidance. One key component of reliable reconstruction is accurate camera pose estimation, which is challenging for monocular cameras due to the paucity of reliable salient features, coupled with narrow baseline during surgical navigation. With recent advances in miniaturized MEMS sensors, the combination of inertial and vision sensing can provide increased robustness for camera pose estimation particularly for scenes involving tissue deformation. The aim of this work is to propose a robust framework for intra-operative free-form deformation recovery based on structure-from-motion. A novel adaptive Unscented Kalman Filter (UKF) parameterization scheme is proposed to fuse vision information with data from an Inertial Measurement Unit (IMU). The method is built on a compact scene representation scheme suitable for both surgical episode identification and instrument-tissue motion modelling. Detailed validation with both synthetic and phantom data is performed and results derived justify the potential clinical value of the technique.","PeriodicalId":6358,"journal":{"name":"2012 IEEE/RSJ International Conference on Intelligent Robots and Systems","volume":"13 1","pages":"4816-4821"},"PeriodicalIF":0.0,"publicationDate":"2012-12-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87865611","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 : 2012-12-24DOI: 10.1109/IROS.2012.6386048
Qian Zhao, K. Nakajima, H. Sumioka, Xiaoxiang Yu, R. Pfeifer
The biological hypothesis of spinal engine states that locomotion is mainly achieved by the spine, while the legs may serve as assistance. Inspired by this hypothesis, a compliant, multiple degree-of-freedom, biologically-inspired spine has been embedded into a quadruped robot, named Kitty, which has no actuation on the legs. In this paper, we demonstrate how versatile behaviors (bounding, trotting, and turning) can be generated exclusively by the spine's movements through dynamical interaction between the controller, the body, and the environment, known as embodiment. Moreover, we introduce information theoretic approach to quantitatively study the spine internal dynamics and its effect on the bounding gait based on three spinal morphologies. These three morphologies differ in the position of virtual spinal joint where the spine is easier to get bent. The experimental results reveal that locomotion can be enhanced by using the spine featuring a rear virtual spinal joint, which offers more freedom for the rear legs to move forward. In addition, the information theoretic analysis shows that, according to the morphological differences of the spine, the information structure changes. The relationship between the observed behavior of the robot and the corresponding information structure is discussed in detail.
{"title":"Embodiment enables the spinal engine in quadruped robot locomotion","authors":"Qian Zhao, K. Nakajima, H. Sumioka, Xiaoxiang Yu, R. Pfeifer","doi":"10.1109/IROS.2012.6386048","DOIUrl":"https://doi.org/10.1109/IROS.2012.6386048","url":null,"abstract":"The biological hypothesis of spinal engine states that locomotion is mainly achieved by the spine, while the legs may serve as assistance. Inspired by this hypothesis, a compliant, multiple degree-of-freedom, biologically-inspired spine has been embedded into a quadruped robot, named Kitty, which has no actuation on the legs. In this paper, we demonstrate how versatile behaviors (bounding, trotting, and turning) can be generated exclusively by the spine's movements through dynamical interaction between the controller, the body, and the environment, known as embodiment. Moreover, we introduce information theoretic approach to quantitatively study the spine internal dynamics and its effect on the bounding gait based on three spinal morphologies. These three morphologies differ in the position of virtual spinal joint where the spine is easier to get bent. The experimental results reveal that locomotion can be enhanced by using the spine featuring a rear virtual spinal joint, which offers more freedom for the rear legs to move forward. In addition, the information theoretic analysis shows that, according to the morphological differences of the spine, the information structure changes. The relationship between the observed behavior of the robot and the corresponding information structure is discussed in detail.","PeriodicalId":6358,"journal":{"name":"2012 IEEE/RSJ International Conference on Intelligent Robots and Systems","volume":"11 1","pages":"2449-2456"},"PeriodicalIF":0.0,"publicationDate":"2012-12-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"86850999","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 : 2012-12-24DOI: 10.1109/IROS.2012.6386105
C. Datta, C. Jayawardena, I. Kuo, B. MacDonald
Service robots for personal and domestic use are increasingly gaining momentum. Easy and efficient programming of such robots is an enormous research and commercial space that is beginning to be explored. In this paper, we present RoboStudio, a Visual Programming Environment (VPE) to program the interactive behavior of personal service robots. RoboStudio lies at the intersection of VPEs which aid in authoring the robot user interface and control logic. A novel contribution of this work is that it advances the research in authoring service applications on robotic platforms, specially for researchers who do development in decentralized multidisciplinary teams and validate their research goals through field trials. Furthermore, service robot programming environments is a novel area of research, particularly when it comes to expressing what the robot does in a declarative syntax.
{"title":"RoboStudio: A visual programming environment for rapid authoring and customization of complex services on a personal service robot","authors":"C. Datta, C. Jayawardena, I. Kuo, B. MacDonald","doi":"10.1109/IROS.2012.6386105","DOIUrl":"https://doi.org/10.1109/IROS.2012.6386105","url":null,"abstract":"Service robots for personal and domestic use are increasingly gaining momentum. Easy and efficient programming of such robots is an enormous research and commercial space that is beginning to be explored. In this paper, we present RoboStudio, a Visual Programming Environment (VPE) to program the interactive behavior of personal service robots. RoboStudio lies at the intersection of VPEs which aid in authoring the robot user interface and control logic. A novel contribution of this work is that it advances the research in authoring service applications on robotic platforms, specially for researchers who do development in decentralized multidisciplinary teams and validate their research goals through field trials. Furthermore, service robot programming environments is a novel area of research, particularly when it comes to expressing what the robot does in a declarative syntax.","PeriodicalId":6358,"journal":{"name":"2012 IEEE/RSJ International Conference on Intelligent Robots and Systems","volume":"79 1","pages":"2352-2357"},"PeriodicalIF":0.0,"publicationDate":"2012-12-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"86888835","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 : 2012-12-24DOI: 10.1109/IROS.2012.6385730
D. Soltero, M. Schwager, D. Rus
We present an online algorithm for a robot to shape its path to a locally optimal configuration for collecting information in an unknown dynamic environment. As the robot travels along its path, it identifies both where the environment is changing, and how fast it is changing. The algorithm then morphs the robot's path online to concentrate on the dynamic areas in the environment in proportion to their rate of change. A Lyapunov-like stability proof is used to show that, under our proposed path shaping algorithm, the path converges to a locally optimal configuration according to a Voronoi-based coverage criterion. The path shaping algorithm is then combined with a previously introduced speed controller to produce guaranteed persistent monitoring trajectories for a robot in an unknown dynamic environment. Simulation and experimental results with a quadrotor robot support the proposed approach.
{"title":"Generating informative paths for persistent sensing in unknown environments","authors":"D. Soltero, M. Schwager, D. Rus","doi":"10.1109/IROS.2012.6385730","DOIUrl":"https://doi.org/10.1109/IROS.2012.6385730","url":null,"abstract":"We present an online algorithm for a robot to shape its path to a locally optimal configuration for collecting information in an unknown dynamic environment. As the robot travels along its path, it identifies both where the environment is changing, and how fast it is changing. The algorithm then morphs the robot's path online to concentrate on the dynamic areas in the environment in proportion to their rate of change. A Lyapunov-like stability proof is used to show that, under our proposed path shaping algorithm, the path converges to a locally optimal configuration according to a Voronoi-based coverage criterion. The path shaping algorithm is then combined with a previously introduced speed controller to produce guaranteed persistent monitoring trajectories for a robot in an unknown dynamic environment. Simulation and experimental results with a quadrotor robot support the proposed approach.","PeriodicalId":6358,"journal":{"name":"2012 IEEE/RSJ International Conference on Intelligent Robots and Systems","volume":"46 1","pages":"2172-2179"},"PeriodicalIF":0.0,"publicationDate":"2012-12-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"86926049","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 : 2012-12-24DOI: 10.1109/IROS.2012.6385854
Seung-Joon Yi, Byoung-Tak Zhang, D. Hong, Daniel D. Lee
During heavy work, humans utilize whole body motions in order to generate large forces. In extreme cases, exaggerated weight shifts are used to impart large impact forces. There have been approaches to design stable whole body impact motions based on precise dynamic models of the robot and the target object, but they have practical limitations as the uncertainty in the ensuing reaction forces can lead to instability. In the current work, we describe a motion controller for a humanoid robot that generates impacts at an end effector while keeping the robot body balanced before and after the impact. Instead of relying on the accuracy of the impact dynamics model, we use a simplified model of the robot and biomechanically motivated push recovery controllers to reactively stabilize the robot against unknown perturbations from the impact. We demonstrate our approach in physically realistic simulations, as well as experimentally on a small humanoid robot platform.
{"title":"Active stabilization of a humanoid robot for impact motions with unknown reaction forces","authors":"Seung-Joon Yi, Byoung-Tak Zhang, D. Hong, Daniel D. Lee","doi":"10.1109/IROS.2012.6385854","DOIUrl":"https://doi.org/10.1109/IROS.2012.6385854","url":null,"abstract":"During heavy work, humans utilize whole body motions in order to generate large forces. In extreme cases, exaggerated weight shifts are used to impart large impact forces. There have been approaches to design stable whole body impact motions based on precise dynamic models of the robot and the target object, but they have practical limitations as the uncertainty in the ensuing reaction forces can lead to instability. In the current work, we describe a motion controller for a humanoid robot that generates impacts at an end effector while keeping the robot body balanced before and after the impact. Instead of relying on the accuracy of the impact dynamics model, we use a simplified model of the robot and biomechanically motivated push recovery controllers to reactively stabilize the robot against unknown perturbations from the impact. We demonstrate our approach in physically realistic simulations, as well as experimentally on a small humanoid robot platform.","PeriodicalId":6358,"journal":{"name":"2012 IEEE/RSJ International Conference on Intelligent Robots and Systems","volume":"58 1","pages":"4034-4039"},"PeriodicalIF":0.0,"publicationDate":"2012-12-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87125883","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 : 2012-12-24DOI: 10.1109/IROS.2012.6386029
Z. Kira, R. Hadsell, G. Salgian, S. Samarasekera
In this paper, we present a system for detecting pedestrians at long ranges using a combination of stereo-based detection, classification using deep learning, and a cascade of specialized classifiers that can reduce false positives and computational load. Specifically, we use stereo to perform detection of vertical structures which are further filtered based on edge responses. A convolutional neural network was then designed to support the classification of pedestrians using both appearance and stereo disparity-based features. A second convolutional network classifier was trained specifically for the case of long-range detections using appearance only. We further speed up the classifier using a cascade approach and multi-threading. The system was deployed on two robots, one using a high resolution stereo pair with 180 degree fisheye lenses and the other using 80 degree FOV lenses. Results are demonstrated on a large dataset captured in a variety of environments.
{"title":"Long-Range Pedestrian Detection using stereo and a cascade of convolutional network classifiers","authors":"Z. Kira, R. Hadsell, G. Salgian, S. Samarasekera","doi":"10.1109/IROS.2012.6386029","DOIUrl":"https://doi.org/10.1109/IROS.2012.6386029","url":null,"abstract":"In this paper, we present a system for detecting pedestrians at long ranges using a combination of stereo-based detection, classification using deep learning, and a cascade of specialized classifiers that can reduce false positives and computational load. Specifically, we use stereo to perform detection of vertical structures which are further filtered based on edge responses. A convolutional neural network was then designed to support the classification of pedestrians using both appearance and stereo disparity-based features. A second convolutional network classifier was trained specifically for the case of long-range detections using appearance only. We further speed up the classifier using a cascade approach and multi-threading. The system was deployed on two robots, one using a high resolution stereo pair with 180 degree fisheye lenses and the other using 80 degree FOV lenses. Results are demonstrated on a large dataset captured in a variety of environments.","PeriodicalId":6358,"journal":{"name":"2012 IEEE/RSJ International Conference on Intelligent Robots and Systems","volume":"36 1","pages":"2396-2403"},"PeriodicalIF":0.0,"publicationDate":"2012-12-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87536884","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 : 2012-12-24DOI: 10.1109/IROS.2012.6385496
B. Munske, J. Kotlarski, T. Ortmaier
This paper presents a new university course combining theoretical lectures with a robot competition. The main intention is to offer a practical course including hands-on experiences being close to current research topics in the field of mobile robotics. For this purpose, a commercial mobile robot is equipped with state of the art sensors, allowing autonomous execution of manipulation tasks. The course consists of three thematical sections. At the beginning of each, lectures provide the theoretical background. On this basis, the participants are addressed to develop algorithms and to solve specific tasks, delivered in homework packages, self-reliantly or in a team. Finally, the developed software components need to be merged to solve a predefined scenario, e.g. autonomous part handling. At the end of a thematic section, students demonstrate their developed solutions within a challenge and explain their approaches in a presentation. Starting with teleoperation and object recognition, the RobotChallenge ends up with navigation in unknown terrain. Besides others, the participants acquire soft-skills, such as project and team management. Being carried out for the first time in winter term 2011/2012, the RobotChallenge successfully promotes profound understanding of mobile robotics that is applied during practical experiences. It turns out, that aspiring to win competions lead to a high motivation of the students w.r.t. development of appropriate solutions.
{"title":"The RobotChallenge - A research inspired practical lecture","authors":"B. Munske, J. Kotlarski, T. Ortmaier","doi":"10.1109/IROS.2012.6385496","DOIUrl":"https://doi.org/10.1109/IROS.2012.6385496","url":null,"abstract":"This paper presents a new university course combining theoretical lectures with a robot competition. The main intention is to offer a practical course including hands-on experiences being close to current research topics in the field of mobile robotics. For this purpose, a commercial mobile robot is equipped with state of the art sensors, allowing autonomous execution of manipulation tasks. The course consists of three thematical sections. At the beginning of each, lectures provide the theoretical background. On this basis, the participants are addressed to develop algorithms and to solve specific tasks, delivered in homework packages, self-reliantly or in a team. Finally, the developed software components need to be merged to solve a predefined scenario, e.g. autonomous part handling. At the end of a thematic section, students demonstrate their developed solutions within a challenge and explain their approaches in a presentation. Starting with teleoperation and object recognition, the RobotChallenge ends up with navigation in unknown terrain. Besides others, the participants acquire soft-skills, such as project and team management. Being carried out for the first time in winter term 2011/2012, the RobotChallenge successfully promotes profound understanding of mobile robotics that is applied during practical experiences. It turns out, that aspiring to win competions lead to a high motivation of the students w.r.t. development of appropriate solutions.","PeriodicalId":6358,"journal":{"name":"2012 IEEE/RSJ International Conference on Intelligent Robots and Systems","volume":"11 1","pages":"1072-1077"},"PeriodicalIF":0.0,"publicationDate":"2012-12-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"86512154","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 : 2012-12-24DOI: 10.1109/IROS.2012.6385834
R. Vertechy, A. Frisoli, M. Solazzi, D. Pellegrinetti, M. Bergamasco
This paper presents the development of an interaction-torque control architecture for a rehabilitation exoskeleton (RehabExos) with flexible joints equipped with internal torque sensors. The architecture consists of an outer control loop, which is based on the kineto-statics of the RehabExos, and on four identical-independent inner full-state joint-torque controllers (IJTFC). The considered IJTFC demonstrate good stability, responsiveness and accuracy in tracking the desired torques generated by the outer controller. Comparison with other control strategies available from the literature is shown via analytical and experimental results, which highlight the improved performances of the proposed IJTFC method in controlling the RehabExos robot.
{"title":"An interaction-torque controller for robotic exoskeletons with flexible joints: Preliminary experimental results","authors":"R. Vertechy, A. Frisoli, M. Solazzi, D. Pellegrinetti, M. Bergamasco","doi":"10.1109/IROS.2012.6385834","DOIUrl":"https://doi.org/10.1109/IROS.2012.6385834","url":null,"abstract":"This paper presents the development of an interaction-torque control architecture for a rehabilitation exoskeleton (RehabExos) with flexible joints equipped with internal torque sensors. The architecture consists of an outer control loop, which is based on the kineto-statics of the RehabExos, and on four identical-independent inner full-state joint-torque controllers (IJTFC). The considered IJTFC demonstrate good stability, responsiveness and accuracy in tracking the desired torques generated by the outer controller. Comparison with other control strategies available from the literature is shown via analytical and experimental results, which highlight the improved performances of the proposed IJTFC method in controlling the RehabExos robot.","PeriodicalId":6358,"journal":{"name":"2012 IEEE/RSJ International Conference on Intelligent Robots and Systems","volume":"24 1","pages":"335-340"},"PeriodicalIF":0.0,"publicationDate":"2012-12-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"86095635","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 : 2012-12-24DOI: 10.1109/IROS.2012.6385863
E. Weng, L. Fu
In this paper, we present a boosting approach by combining the pose estimation and the upper body tracking to on-line recognize human actions. Instead of using a predefined pose to initialize the human skeleton, we construct a key poses database with depth HOG features as searching indexes. When user enters the camera view, we automatically search the database to get the initial skeleton. Then we use the particle filter to track human upper body parts. At the same time, we feed the tracking joints into the hidden Markov models to on-line spot and recognize the performed action. In order to rectify tracking errors, we apply the action recognition results and reuse our key poses database to reinforce the tracking process. Our contributions of the proposed approach are three-fold. First, our method can recognize human poses and actions at the same time. Second, with the key poses database and action recognition results as the feedback, the tracking process becomes more efficient and accurate. Third, we propose a spotting method based on the gradient of HMM probabilities, which thus enables our method to achieve on-line spotting and recognition. Experimental results demonstrate the effectiveness of the proposed approach.
{"title":"On-line human action recognition by combining joint tracking and key pose recognition","authors":"E. Weng, L. Fu","doi":"10.1109/IROS.2012.6385863","DOIUrl":"https://doi.org/10.1109/IROS.2012.6385863","url":null,"abstract":"In this paper, we present a boosting approach by combining the pose estimation and the upper body tracking to on-line recognize human actions. Instead of using a predefined pose to initialize the human skeleton, we construct a key poses database with depth HOG features as searching indexes. When user enters the camera view, we automatically search the database to get the initial skeleton. Then we use the particle filter to track human upper body parts. At the same time, we feed the tracking joints into the hidden Markov models to on-line spot and recognize the performed action. In order to rectify tracking errors, we apply the action recognition results and reuse our key poses database to reinforce the tracking process. Our contributions of the proposed approach are three-fold. First, our method can recognize human poses and actions at the same time. Second, with the key poses database and action recognition results as the feedback, the tracking process becomes more efficient and accurate. Third, we propose a spotting method based on the gradient of HMM probabilities, which thus enables our method to achieve on-line spotting and recognition. Experimental results demonstrate the effectiveness of the proposed approach.","PeriodicalId":6358,"journal":{"name":"2012 IEEE/RSJ International Conference on Intelligent Robots and Systems","volume":"66 1","pages":"4112-4117"},"PeriodicalIF":0.0,"publicationDate":"2012-12-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"86104662","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：