Lan Zhao, Guoyin Tang, Yijun Du, Yuan She, Xiulan Wen
Point cloud registration attempts to register a pair of point cloud through rigid transformation to form a complete point cloud. Traditional methods rely on searching closest neighbors in the feature space and filtering the outliers through RANSAC to find the point-to-point correspondence. Recently, learning-based algorithms incorporate learning to local feature descriptors and perform better than the traditional methods. However, they continue to adopt the feature-based matching and point-level correspondence approach for pose estimation. In this work, attention mechanism is incorporated to learning transformation-invariant features and an end-to-end solution is used to predict the correspondences. A two-stream feature extraction architecture consisting primarily of feature extractor and transformer learns the rotation-invariant features through transformer layer. The network learns to predict the point coordinates mapped to the other point cloud and the probability between the correspondences in the overlap region. The rigid relative pose transformation can be solved from the predicted correspondences in a closed form. The registration method is applied to pose estimation in the robotic grasping and sorting task and achieves favorable result.
{"title":"Registration of 3D Point Cloud with Rotation Invariant Features Learning for Partial Overlapping","authors":"Lan Zhao, Guoyin Tang, Yijun Du, Yuan She, Xiulan Wen","doi":"10.1145/3598151.3598190","DOIUrl":"https://doi.org/10.1145/3598151.3598190","url":null,"abstract":"Point cloud registration attempts to register a pair of point cloud through rigid transformation to form a complete point cloud. Traditional methods rely on searching closest neighbors in the feature space and filtering the outliers through RANSAC to find the point-to-point correspondence. Recently, learning-based algorithms incorporate learning to local feature descriptors and perform better than the traditional methods. However, they continue to adopt the feature-based matching and point-level correspondence approach for pose estimation. In this work, attention mechanism is incorporated to learning transformation-invariant features and an end-to-end solution is used to predict the correspondences. A two-stream feature extraction architecture consisting primarily of feature extractor and transformer learns the rotation-invariant features through transformer layer. The network learns to predict the point coordinates mapped to the other point cloud and the probability between the correspondences in the overlap region. The rigid relative pose transformation can be solved from the predicted correspondences in a closed form. The registration method is applied to pose estimation in the robotic grasping and sorting task and achieves favorable result.","PeriodicalId":398644,"journal":{"name":"Proceedings of the 2023 3rd International Conference on Robotics and Control Engineering","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122528275","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}
Jie Li, Zongmin Liu, Jirui Wang, Zhenjie Gu, Yabo Shi
The initial weld position of large complex components has the characteristics of irregular shape and random spatial distribution. Traditional welding robots have technical bottlenecks such as low teaching programming efficiency and difficulty in offline programming in this application scenario. Therefore, at present, the method of "workers piling up" is used to weld large complex components. To overcome these problems, it is necessary and urgent to develop the intelligent initial weld position identification method for large components based on machine vision. Firstly, based on the theoretical background of YOLOv5 algorithm, an initial weld position recognition model for large complex components is established. Secondly, the model is optimized by changing the up-sampling method and fusing the Transformer Self-Attention mechanism. Finally, through the initial weld position detection experiment on the coco dataset, the experimental results show that the detection effect of this model is better than others and the original YOLOv5 model, and the detection speed of a single weld image is 11ms, and the average detection accuracy reaches 92.4%. Under the premise of ensuring the detection speed, the proposed method greatly improves the precision of initial weld position detection, and has better interference immunity and robustness.
{"title":"A novel initial weld position identification method for large complex components based on improved YOLOv5","authors":"Jie Li, Zongmin Liu, Jirui Wang, Zhenjie Gu, Yabo Shi","doi":"10.1145/3598151.3598173","DOIUrl":"https://doi.org/10.1145/3598151.3598173","url":null,"abstract":"The initial weld position of large complex components has the characteristics of irregular shape and random spatial distribution. Traditional welding robots have technical bottlenecks such as low teaching programming efficiency and difficulty in offline programming in this application scenario. Therefore, at present, the method of \"workers piling up\" is used to weld large complex components. To overcome these problems, it is necessary and urgent to develop the intelligent initial weld position identification method for large components based on machine vision. Firstly, based on the theoretical background of YOLOv5 algorithm, an initial weld position recognition model for large complex components is established. Secondly, the model is optimized by changing the up-sampling method and fusing the Transformer Self-Attention mechanism. Finally, through the initial weld position detection experiment on the coco dataset, the experimental results show that the detection effect of this model is better than others and the original YOLOv5 model, and the detection speed of a single weld image is 11ms, and the average detection accuracy reaches 92.4%. Under the premise of ensuring the detection speed, the proposed method greatly improves the precision of initial weld position detection, and has better interference immunity and robustness.","PeriodicalId":398644,"journal":{"name":"Proceedings of the 2023 3rd International Conference on Robotics and Control Engineering","volume":"85 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127885018","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}
Four-rotor UAV has the problem of being susceptible to wind disturbance and coupling of its own channel in the flight process, which will have a negative impact on its trajectory tracking. This paper focus on the trajectory tracking in the flight process of the four-rotor UAV, a super twisting sliding mode(STSM) composite controller based on the extended state observer(ESO) is designed for the position loop of the UAV, attitude loop adopts the traditional cascade PID controller. First, the ground coordinate system is established as a reference for trajectory tracking, and then the kinematics modeling of the UAV system is carried out by using Newton-Euler method, and the desired trajectory is entered into the position controller. At the same time, an ESO is established based on a general UAV model, and disturbances and internal channel coupling of the UAV are observed and fedback to the corresponding controller. Finally, the composite proposed controller is simulated to track a desired trajectory. Based on the simulation results,draw the conclusion of the composite controller this paper proposed improved the robustness of the four-rotor UAV in the trajectory tracking scenario.
{"title":"A Super-Twisting-Sliding-Mode Controller for four-rotor UAV","authors":"Song Hu, Zhenxing Sun, Ting Wang","doi":"10.1145/3598151.3598154","DOIUrl":"https://doi.org/10.1145/3598151.3598154","url":null,"abstract":"Four-rotor UAV has the problem of being susceptible to wind disturbance and coupling of its own channel in the flight process, which will have a negative impact on its trajectory tracking. This paper focus on the trajectory tracking in the flight process of the four-rotor UAV, a super twisting sliding mode(STSM) composite controller based on the extended state observer(ESO) is designed for the position loop of the UAV, attitude loop adopts the traditional cascade PID controller. First, the ground coordinate system is established as a reference for trajectory tracking, and then the kinematics modeling of the UAV system is carried out by using Newton-Euler method, and the desired trajectory is entered into the position controller. At the same time, an ESO is established based on a general UAV model, and disturbances and internal channel coupling of the UAV are observed and fedback to the corresponding controller. Finally, the composite proposed controller is simulated to track a desired trajectory. Based on the simulation results,draw the conclusion of the composite controller this paper proposed improved the robustness of the four-rotor UAV in the trajectory tracking scenario.","PeriodicalId":398644,"journal":{"name":"Proceedings of the 2023 3rd International Conference on Robotics and Control Engineering","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130698335","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}
Qian Chen, Lizhe Qi, Zhongwei Hua, Z. Yang, Yunquan Sun
In order to solve the problem of low exploration efficiency caused by repeatedly backtracking the frontiers in large and complex environments, a hierarchical autonomous exploration algorithm is proposed. Firstly, based on the robot's static model to judge the environment's traversability, which allows robots to explore complex 3D environments. Following that, a hybrid strategy is used to filter out frontiers within a local planning horizon, thus complete exploration of the local area is achieved by solving the traveling salesman problem (TSP). Finally, the sparse global topology map generates transfer paths between sub-areas, transferring the robot to another sub-area to resume exploration. Compared to the RRT autonomous exploration algorithm and the GBP autonomous exploration algorithm, the method in this work reduces the exploration path by more than 13.8% and the exploration time by more than 23.7%. The results show that the proposed algorithm significantly improves the autonomous exploration efficiency of mobile robots in large and complex environments.
{"title":"A Hierarchical Autonomous Exploration Algorithm for Large-scale and Complex Environments with Mobile Robot","authors":"Qian Chen, Lizhe Qi, Zhongwei Hua, Z. Yang, Yunquan Sun","doi":"10.1145/3598151.3598191","DOIUrl":"https://doi.org/10.1145/3598151.3598191","url":null,"abstract":"In order to solve the problem of low exploration efficiency caused by repeatedly backtracking the frontiers in large and complex environments, a hierarchical autonomous exploration algorithm is proposed. Firstly, based on the robot's static model to judge the environment's traversability, which allows robots to explore complex 3D environments. Following that, a hybrid strategy is used to filter out frontiers within a local planning horizon, thus complete exploration of the local area is achieved by solving the traveling salesman problem (TSP). Finally, the sparse global topology map generates transfer paths between sub-areas, transferring the robot to another sub-area to resume exploration. Compared to the RRT autonomous exploration algorithm and the GBP autonomous exploration algorithm, the method in this work reduces the exploration path by more than 13.8% and the exploration time by more than 23.7%. The results show that the proposed algorithm significantly improves the autonomous exploration efficiency of mobile robots in large and complex environments.","PeriodicalId":398644,"journal":{"name":"Proceedings of the 2023 3rd International Conference on Robotics and Control Engineering","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132006044","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}
As the technology of virtual reality (VR) and augmented reality (AR) continue to advance at a rapid pace, there has been an increasing focus among researchers on the perception of haptic feedback in virtual environments. In recent years, vibration-based texture haptic modeling research has gradually received attention. Using vibration signals to analyze the texture features of objects and describing them using mathematical models, object textures can be reproduced. This technique is not only widely used in VR and AR, but also has important research value and application prospects in haptic assistive technology for the blind. This article proposes a virtual texture haptic reconstruction model based on the Informer model, aiming to address issues such as poor accuracy, weak generalization ability, and poor real-time performance of texture haptic reconstruction models. The algorithm is based on the encoder-decoder framework of Informer model, and uses interaction information as input to extract features from the user-texture interaction parameters (sliding speed and interaction pressure) to generate dynamic vibration acceleration signals in real time. The algorithm is experimentally validated by performance evaluation of the model to verify the accuracy as well as the generalization of the algorithm.
{"title":"Research on texture haptic reconstruction method based on Informer model","authors":"Dapeng Chen, Dongliang Zhu, Jia Liu, Geng Chen, Ying-Sang Fang, Yunjie Zhang","doi":"10.1145/3598151.3598273","DOIUrl":"https://doi.org/10.1145/3598151.3598273","url":null,"abstract":"As the technology of virtual reality (VR) and augmented reality (AR) continue to advance at a rapid pace, there has been an increasing focus among researchers on the perception of haptic feedback in virtual environments. In recent years, vibration-based texture haptic modeling research has gradually received attention. Using vibration signals to analyze the texture features of objects and describing them using mathematical models, object textures can be reproduced. This technique is not only widely used in VR and AR, but also has important research value and application prospects in haptic assistive technology for the blind. This article proposes a virtual texture haptic reconstruction model based on the Informer model, aiming to address issues such as poor accuracy, weak generalization ability, and poor real-time performance of texture haptic reconstruction models. The algorithm is based on the encoder-decoder framework of Informer model, and uses interaction information as input to extract features from the user-texture interaction parameters (sliding speed and interaction pressure) to generate dynamic vibration acceleration signals in real time. The algorithm is experimentally validated by performance evaluation of the model to verify the accuracy as well as the generalization of the algorithm.","PeriodicalId":398644,"journal":{"name":"Proceedings of the 2023 3rd International Conference on Robotics and Control Engineering","volume":"61 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124113264","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}
Haolei Ru, Kuo Su, Jiangjian Xiao, Gen Xu, Kangkang Song, Gang Xu, Jianfei Ge, Weicheng Huang, M. Sheng
Magnetic adsorption wall climbing robots are extensively used for surface inspection, maintenance, and repair of ships, large chemical storage tanks, and steel structures, thereby avoiding high-risk work and lowering the occurrence of accidents. The research status of magnetic adsorption wall climbing robots at home and abroad is summarized using the three main adsorption methods of magnetic adsorption wall climbing robots: permanent magnetic adsorption, electromagnetic adsorption, and electric permanent magnetic adsorption, and the main development trend of future magnetic adsorption wall climbing robots is given using the magnetic adsorption methods and application scenarios to provide a reference basis.
{"title":"Research status and development trend of magnetic adsorption wall climbing robot","authors":"Haolei Ru, Kuo Su, Jiangjian Xiao, Gen Xu, Kangkang Song, Gang Xu, Jianfei Ge, Weicheng Huang, M. Sheng","doi":"10.1145/3598151.3598166","DOIUrl":"https://doi.org/10.1145/3598151.3598166","url":null,"abstract":"Magnetic adsorption wall climbing robots are extensively used for surface inspection, maintenance, and repair of ships, large chemical storage tanks, and steel structures, thereby avoiding high-risk work and lowering the occurrence of accidents. The research status of magnetic adsorption wall climbing robots at home and abroad is summarized using the three main adsorption methods of magnetic adsorption wall climbing robots: permanent magnetic adsorption, electromagnetic adsorption, and electric permanent magnetic adsorption, and the main development trend of future magnetic adsorption wall climbing robots is given using the magnetic adsorption methods and application scenarios to provide a reference basis.","PeriodicalId":398644,"journal":{"name":"Proceedings of the 2023 3rd International Conference on Robotics and Control Engineering","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117013419","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}
Abstract-Nowadays, robot industry is booming, and the control requirements are higher with more user-friendly features. Therefore, it is significant to study the mechanism of human upper limb movement for robot control optimization. In this study, a musculoskeletal model of human upper limb was established, and the motion data of the upper limb was collected by a stereo photogrammetry system, which was used for model scaling to match the measurement and physiological characteristics of the experimental subjects and generated the generalized coordinate values of the elbow and shoulder joints. Then, Opensim simulates the neuromuscular control mechanism during the pull and trigger action of the subject and analyzes the upper limb dynamics. The RMS and MAX marker errors of the dynamic simulation were within the tolerable range, and the experimental EMG signals could match well with the muscle activation of the simulation results, which verified the correctness of the simulation. It provides a theoretical basis for upper limb kinetic analysis, and also provides thoughts for robot control optimization.
{"title":"Opensim-Based Dynamics Simulation of Upper Limb","authors":"Pengfei Liu, Jin Hua, Li Wang, Yue Zhao","doi":"10.1145/3598151.3598158","DOIUrl":"https://doi.org/10.1145/3598151.3598158","url":null,"abstract":"Abstract-Nowadays, robot industry is booming, and the control requirements are higher with more user-friendly features. Therefore, it is significant to study the mechanism of human upper limb movement for robot control optimization. In this study, a musculoskeletal model of human upper limb was established, and the motion data of the upper limb was collected by a stereo photogrammetry system, which was used for model scaling to match the measurement and physiological characteristics of the experimental subjects and generated the generalized coordinate values of the elbow and shoulder joints. Then, Opensim simulates the neuromuscular control mechanism during the pull and trigger action of the subject and analyzes the upper limb dynamics. The RMS and MAX marker errors of the dynamic simulation were within the tolerable range, and the experimental EMG signals could match well with the muscle activation of the simulation results, which verified the correctness of the simulation. It provides a theoretical basis for upper limb kinetic analysis, and also provides thoughts for robot control optimization.","PeriodicalId":398644,"journal":{"name":"Proceedings of the 2023 3rd International Conference on Robotics and Control Engineering","volume":"28 9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132059942","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}
Surgical instruments play a vital role in the medical field. Surgical forceps can help improve the accuracy of operations, thereby reducing the risk of surgery and increasing the success rate. Currently, most surgical forceps only provide force feedback information. This paper proposes a new type of fiber Bragg grating force/position sensing surgical forceps, where the upper forceps are connected by three arc-shaped cylindrical surfaces and two circular rings for force/position detection, the lower forceps are employed for temperature compensation and contact sensing, whereas the structural optimization of the surgical forceps is accomplished via finite element analysis. After optimization, the measurement range of the sensor is 0-10N, and the maximum sensitivity is 91.3μɛ/N. harmonic response analysis is conducted to establish that the operational frequency range of the surgical forceps lies between 0 to 1802.8 Hz.
{"title":"Surgical Forceps with Force/Position Perception Based on FBG","authors":"Songtao Ouyang, Lekai Cheng, Wenxing Xiao, Jianyong Long, Q. Liang","doi":"10.1145/3598151.3598168","DOIUrl":"https://doi.org/10.1145/3598151.3598168","url":null,"abstract":"Surgical instruments play a vital role in the medical field. Surgical forceps can help improve the accuracy of operations, thereby reducing the risk of surgery and increasing the success rate. Currently, most surgical forceps only provide force feedback information. This paper proposes a new type of fiber Bragg grating force/position sensing surgical forceps, where the upper forceps are connected by three arc-shaped cylindrical surfaces and two circular rings for force/position detection, the lower forceps are employed for temperature compensation and contact sensing, whereas the structural optimization of the surgical forceps is accomplished via finite element analysis. After optimization, the measurement range of the sensor is 0-10N, and the maximum sensitivity is 91.3μɛ/N. harmonic response analysis is conducted to establish that the operational frequency range of the surgical forceps lies between 0 to 1802.8 Hz.","PeriodicalId":398644,"journal":{"name":"Proceedings of the 2023 3rd International Conference on Robotics and Control Engineering","volume":"42 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125559277","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}
Virtual reality motion sickness (VRMS) is a key obstacle in user experience and development of virtual reality technology. Given the existed condition of technology, the adjusts on the virtual reality scene and content is convenient and low-cost way to reduce the influence of VRMS. This paper aims to investigate whether the additional task in a virtual reality content can reduce the VRMS level. We collect the EEG data and simulator sickness questionnaire from the subjects during their riding a virtual reality roll coaster with and without an additional shooting task. In final, the statistic method is used to analyze the rhythm energies and differential entropies of the EEG. The SSQ results show that the shooting task can reduce the VRMS severity. The difference of the EEG features between the shooting task and no-task is also more significant than those between different field of view, that is a further proof of the reduction effect of the additional shooting task on VRMS. This study reveals the effect of additional shooting task on the VRMS and provides a novel beginning point of research on the use of additional task to prevent VRMS.
{"title":"The effect of additional shooting task on alleviating virtual reality motion sickness","authors":"C. Hua, Jia-Chao Tao, Lining Chai, Zhanfeng Zhou","doi":"10.1145/3598151.3598180","DOIUrl":"https://doi.org/10.1145/3598151.3598180","url":null,"abstract":"Virtual reality motion sickness (VRMS) is a key obstacle in user experience and development of virtual reality technology. Given the existed condition of technology, the adjusts on the virtual reality scene and content is convenient and low-cost way to reduce the influence of VRMS. This paper aims to investigate whether the additional task in a virtual reality content can reduce the VRMS level. We collect the EEG data and simulator sickness questionnaire from the subjects during their riding a virtual reality roll coaster with and without an additional shooting task. In final, the statistic method is used to analyze the rhythm energies and differential entropies of the EEG. The SSQ results show that the shooting task can reduce the VRMS severity. The difference of the EEG features between the shooting task and no-task is also more significant than those between different field of view, that is a further proof of the reduction effect of the additional shooting task on VRMS. This study reveals the effect of additional shooting task on the VRMS and provides a novel beginning point of research on the use of additional task to prevent VRMS.","PeriodicalId":398644,"journal":{"name":"Proceedings of the 2023 3rd International Conference on Robotics and Control Engineering","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134603843","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}
Gesture recognition is still actively researched in non-contact human-computer interaction (HCI), where the Dynamic Time Warping (DTW) algorithm is commonly employed. However, the computation load of traditional DTW algorithms during the matching phase presents a challenge for gesture recognition, especially when the number of reference gestures in the template library increases. In order to solve this problem, a Constraints-based Dynamic Time Warping (CDTW) method is proposed in this paper, including Global-Path Constraint, First-Frame Constraint, and Feature-Vector Constraint. These three constraints are expected to limit the region of the warping path, exclude several reference gestures from the template library, and reduce computation load directly with reduced elements of the feature vector. To verify the proposed CDTW method, Microsoft Kinect-based comparative experiments with the traditional DTW method have been carried out. Experimental results show that our CDTW method boosts the efficiency of gesture recognition with a 17% decrease in recognition time and a 3% increase in average recognition accuracy, compared to the traditional DTW algorithm.
{"title":"A Constraints-based Dynamic Time Warping Method for Gesture Recognition with Kinect","authors":"Jiaojiao Liu, Xiong Lu, Hongbin Yin, Xianglin Tao, Chuanlong Zhang, Aaron Quigley","doi":"10.1145/3598151.3598181","DOIUrl":"https://doi.org/10.1145/3598151.3598181","url":null,"abstract":"Gesture recognition is still actively researched in non-contact human-computer interaction (HCI), where the Dynamic Time Warping (DTW) algorithm is commonly employed. However, the computation load of traditional DTW algorithms during the matching phase presents a challenge for gesture recognition, especially when the number of reference gestures in the template library increases. In order to solve this problem, a Constraints-based Dynamic Time Warping (CDTW) method is proposed in this paper, including Global-Path Constraint, First-Frame Constraint, and Feature-Vector Constraint. These three constraints are expected to limit the region of the warping path, exclude several reference gestures from the template library, and reduce computation load directly with reduced elements of the feature vector. To verify the proposed CDTW method, Microsoft Kinect-based comparative experiments with the traditional DTW method have been carried out. Experimental results show that our CDTW method boosts the efficiency of gesture recognition with a 17% decrease in recognition time and a 3% increase in average recognition accuracy, compared to the traditional DTW algorithm.","PeriodicalId":398644,"journal":{"name":"Proceedings of the 2023 3rd International Conference on Robotics and Control Engineering","volume":"60 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115076236","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：