{"title":"Research on Human Features Semantic Segmentation Based on Laser Point Cloud","authors":"Tianyi Ma, Bokai Xuan, Jian Li, Yuexuan Xu, Minghe Liu, Qingsong Ding, Jianwen Wang, Hao Sun","doi":"10.1109/CYBER55403.2022.9907384","DOIUrl":"https://doi.org/10.1109/CYBER55403.2022.9907384","url":null,"abstract":"In view of the problems of health care for the semi-disabled elderly, this paper studies the semantic segmentation of human features in a bathing environment with a scrubbing device. Firstly three-dimensional point cloud data of different types to construct a human model is collected by the lidar. Secondly, overcome the influence of the water fog environment on the modeling by the hybrid filtering algorithm, and the human point cloud area is extracted. Finally, the human semantic segmentation model fusing the spatial feature extraction module and the channel attention module is proposed based on PointNet improvement. After training and testing on the target data set, the results show that the algorithm can accurately identify feature information for 3D human models of different types. The segmentation rate reaches 95.7%, which is 4.5% higher than the PointNet network, significantly improves the segmentation of human features, and has high engineering application value.","PeriodicalId":34110,"journal":{"name":"IET Cybersystems and Robotics","volume":"10 1","pages":"876-881"},"PeriodicalIF":0.0,"publicationDate":"2022-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"88450395","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}
{"title":"Evaluation and Comparison of Gmapping and Karto SLAM Systems","authors":"Shengshu Liu, Y. Lei, Xin Dong","doi":"10.1109/CYBER55403.2022.9907154","DOIUrl":"https://doi.org/10.1109/CYBER55403.2022.9907154","url":null,"abstract":"Gmapping and Karto are two classic laser-based SLAM algorithms widely used in various applications. This paper evaluated and compared the performances of these two algorithms. A series of experiments were conducted within the self-built outdoor environments. The parameters of algorithms were tuned, the performances of different parameter settings were evaluated and compared, and the pros and cons regarding mapping and localization accuracy and computational cost of two algorithms were discussed.","PeriodicalId":34110,"journal":{"name":"IET Cybersystems and Robotics","volume":"1 1","pages":"295-300"},"PeriodicalIF":0.0,"publicationDate":"2022-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"90840915","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}
Jianjun Yu, Ruiqi Li, Daoxiong Gong, Yixin Liu, Peng Liu
{"title":"Gait tracking control of biped robot based on adaptive gait switching algorithm","authors":"Jianjun Yu, Ruiqi Li, Daoxiong Gong, Yixin Liu, Peng Liu","doi":"10.1109/CYBER55403.2022.9907560","DOIUrl":"https://doi.org/10.1109/CYBER55403.2022.9907560","url":null,"abstract":"In order to make the walking gait of biped robot more human like, this paper takes the human walking data as the expected gait of robot, and uses the periodic characteristics of gait, proposes a gait tracking control strategy of Biped Robot Based on adaptive gait switching algorithm. Firstly, this paper establishes the complete dynamic models of left leg support phase (LSP) and right leg support phase (RSP) based on Lagrange method, then designs the corresponding LQR gait tracking control strategy, and uses the adaptive weighted particle swarm algorithm (A WPSO) to obtain the optimal controller parameters. Finally, the threshold range of plantar contact force in two periods are estimated based on the adaptive mechanism, and the occurrence of gait switching is detected according to the defined decision rules, thus trigger the control strategy in the next stage to realize the walking tracking control of biped robot. The experimental results show that only two LQR controllers to realize the accurate tracking of the desired gait of the biped robot, and the maximum gait speed reaches two steps/s, which is close to the human gait speed. Compared with other methods, the gait is more human like.","PeriodicalId":34110,"journal":{"name":"IET Cybersystems and Robotics","volume":"42 4 1","pages":"105-110"},"PeriodicalIF":0.0,"publicationDate":"2022-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"90424781","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}
{"title":"An Efficient Color and Geometric Feature Fusion Module for 6D Object Pose Estiamtion","authors":"Jiangeng Li, Hong Liu, Gao Huang, Guoyu Zuo","doi":"10.1109/CYBER55403.2022.9907032","DOIUrl":"https://doi.org/10.1109/CYBER55403.2022.9907032","url":null,"abstract":"6D pose estimation is widely used in robot tasks such as sorting and grasping. RGB-D-based methods have recently attained brilliant success, but they are still susceptible to heavy occlusion. Our critical insight is that color and geometry information in RGBD images are two complementary data, and the crux of the pose estimation problem under occlusion is fully leveraging them. Towards this end, we propose a new color and geometry feature fusion module that can efficiently leverage two complementary data sources from RGB-D images. Unlike prior fusion methods, we conduct a two-stage fusion strategy to do color-depth fusion and local-global fusion successively. Specifically, we fuse the color features extracted from RGB images into the point cloud in the first stage. In the second stage, we extract local and global features from the fused point cloud using an ASSANet-like network and splice them together to obtain the final fusion features. We conducted experiments on the widely used LineMod and YCB-Video datasets, which shows that our method improves the prediction accuracy while reducing the training time.","PeriodicalId":34110,"journal":{"name":"IET Cybersystems and Robotics","volume":"23 1","pages":"83-88"},"PeriodicalIF":0.0,"publicationDate":"2022-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74599402","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}
{"title":"Design of Programmable Droplet Manipulation Platform Based on Magnetic Control","authors":"Xianmiao Zhang, Jie Liu, Jiying Liu, Yu-zhou Wang, Mian Zhang, Hongbiao Xiang","doi":"10.1109/CYBER55403.2022.9907625","DOIUrl":"https://doi.org/10.1109/CYBER55403.2022.9907625","url":null,"abstract":"Droplet actuation simplifies the handling of various reagents or samples and can be applied to a wide range of fields, including chemistry, biology, biomedical, and others. This paper presents a programmable droplet control system based on a magnetoelastic membrane and electromagnetic pillar array. Different magnetic blocks with different magnetization directions were designed on the silicone rubber membrane, and the magnetoelastic membrane deformed under the magnetic field generated by the array of electromagnetic pillars. By combining the gravitational forces of the droplet and the deformation of the magnetic membranes, the motion of the droplet can be controlled. Furthermore, the surface of membranes was ablated with a laser machine to impart superhydrophobic properties. The simulation results show that with the different magnetic fields, the droplet can move lengthwise, widthwise, and diagonally in the horizontal plane, and multiple droplets can be merged and mixed. In contrast to the traditional droplet control method, the droplet programmable movement control system utilizing superhydrophobic magnetoelastic membranes and an electromagnetic pillar array has better stationarity, flexibility and does not affect the basic properties of the droplets.","PeriodicalId":34110,"journal":{"name":"IET Cybersystems and Robotics","volume":"21 1","pages":"1224-1229"},"PeriodicalIF":0.0,"publicationDate":"2022-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74911572","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}
Yunfeng Bai, Fengming Li, Man Zhao, Wei Wang, Yibin Li, R. Song
{"title":"Robot Manipulation Skill Learning Based on Dynamic Movement Primitive*","authors":"Yunfeng Bai, Fengming Li, Man Zhao, Wei Wang, Yibin Li, R. Song","doi":"10.1109/CYBER55403.2022.9907262","DOIUrl":"https://doi.org/10.1109/CYBER55403.2022.9907262","url":null,"abstract":"This paper proposes a robot automatic valve turning control strategy based on teaching learning, which consists of teaching, model learning and task repetition. The first stage is the teaching and learning stage. The robot learns motor skills by observing the human performing tasks. In order to accurately learn motor skills from demonstrations, data alignment is performed on the teaching data through Dynamic Time Warping (DTW). The second stage is the model construction and learning stage. The high-level learning strategy aims to learn motor skills from demonstrations through Dynamic Movement Primitives (DMP), using the statistical approach Gaussian Mixture Model and Gaussian Mixture Regression (GMM-GMR) to analyze the data from demonstrations. And the valve turning is repetition. To verify the effectiveness of the proposed control strategy, the experiment of the butterfly valve closing is performed. The results show that the robot is able to learn and reproduce the valve reaching and turning tasks. It completes the valve closing action by turning the valve for 7 turns.","PeriodicalId":34110,"journal":{"name":"IET Cybersystems and Robotics","volume":"67 1","pages":"579-584"},"PeriodicalIF":0.0,"publicationDate":"2022-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"76529737","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}
{"title":"A Novel Lightweight Architecture of Deep Convolutional Neural Networks","authors":"Baicheng Liu, Xi’ai Chen, Zhi Han, Huidi Jia, Yandong Tang","doi":"10.1109/CYBER55403.2022.9907319","DOIUrl":"https://doi.org/10.1109/CYBER55403.2022.9907319","url":null,"abstract":"Deep convolutional neural networks have achieved much success in many computer vision tasks. However, a network has millions of parameters which limit its inference speed and usage for some situations with limited storage space. Low-rank based methods and pruning methods are verified effective to compress the number of parameters and accelerate inference speed of deep convolutional neural networks. As the price, the performance of the networks decreases. To overcome this problem, in this paper, we design a novel low-rank and sparse architecture of convolutional neural networks. Besides accelerating inference speed and reducing parameters, our approach achieves better performance than baseline networks.","PeriodicalId":34110,"journal":{"name":"IET Cybersystems and Robotics","volume":"70 1","pages":"230-235"},"PeriodicalIF":0.0,"publicationDate":"2022-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"76724050","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}
Xue Feng, Zhang Li-xun, Wang Chao, Wang Zhen-han, Fan Yu-he
{"title":"Research on Structure Design and Control of Plane 3-DOF Cable Driven Virtual Microgravity Training System","authors":"Xue Feng, Zhang Li-xun, Wang Chao, Wang Zhen-han, Fan Yu-he","doi":"10.1109/CYBER55403.2022.9907691","DOIUrl":"https://doi.org/10.1109/CYBER55403.2022.9907691","url":null,"abstract":"Astronauts' microgravity environment simulation training on the ground is an important preparation for space operation tasks. In view of the problems of high cost, short single training time and low simulation accuracy of the existing microgravity training equipment, a virtual microgravity training system driven by plane 3 degrees of freedom (hereinafter referred to as “DOF”) cable is proposed. The system controls the motion of the virtual object pulled by the cable by sampling the astronauts' operating force on the working object; so that the virtual object conforms to the motion law in the microgravity environment. The structure of the system is designed. Aiming at the problems of insufficient workspace and high requirements for the performance of the driving unit caused by the unreasonable distribution of cable tension in the previous cable drive system, a control strategy of optimizing cable tension using genetic algorithm is proposed. The simulation results show that the motion of virtual mass under the action of operating force conforms to the motion law in microgravity environment, and has high simulation accuracy; The cable tension changes smoothly and the system has good stability. It can realize the simulated operation training of moving objects with different masses in microgravity environment.","PeriodicalId":34110,"journal":{"name":"IET Cybersystems and Robotics","volume":"30 4","pages":"1218-1223"},"PeriodicalIF":0.0,"publicationDate":"2022-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"72616493","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}
Q. Bu, P. Lv, Kexin Zhang, Xiaobo Dou, Fei Luo, Xufeng Zhou
{"title":"Photovoltaic and energy storage control of partially observable distribution network based on deep reinforcement learning","authors":"Q. Bu, P. Lv, Kexin Zhang, Xiaobo Dou, Fei Luo, Xufeng Zhou","doi":"10.1109/CYBER55403.2022.9907595","DOIUrl":"https://doi.org/10.1109/CYBER55403.2022.9907595","url":null,"abstract":"After a large number of distributed power sources are connected to the distribution network, the volatility and uncertainty brought by them may lead to the over-limit of the distribution network voltage and the increase of network losses; at the same time, the distribution network itself is also in a partially observable state. In view of these problems, photovoltaic and energy storage are selected as the control objects. In this paper, a photovoltaic energy storage linkage control technology based on deep reinforcement learning is designed, and an example is used to verify the feasibility and effectiveness of the method proposed in this paper.","PeriodicalId":34110,"journal":{"name":"IET Cybersystems and Robotics","volume":"72 1","pages":"871-875"},"PeriodicalIF":0.0,"publicationDate":"2022-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"72732260","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}
Yankui Wang, Wenhao Yao, Min Dong, Yixuan Li, Longxing Zhu, Sheng Bi
{"title":"Prediction of Battery Capacity Based on Deep Residual Network","authors":"Yankui Wang, Wenhao Yao, Min Dong, Yixuan Li, Longxing Zhu, Sheng Bi","doi":"10.1109/CYBER55403.2022.9907034","DOIUrl":"https://doi.org/10.1109/CYBER55403.2022.9907034","url":null,"abstract":"Consistency is essential to the life of battery packs. Therefore, there is a special process to determine the capacity of lithium batteries in their production process (aka grading). However, this process takes a very long time. We propose a new method based on deep learning, which uses data collected by sensors before the grading process to predict the battery capacity, hoping to reduce the time consumed in the whole process. We propose an end-to-end battery capacity prediction model. In our processing steps, complex feature extraction steps are not needed. On the contrary, we use a residual network to complete it automatically. We modified the original ResNet to suit our task. Convolution1D and global pooling layers are used to extract the time series feature. To improve the model's accuracy, we design a fusion model to deal with the time series of multi-step processes. Transfer learning is applied to help us train the model faster. The results on the test set show that the root mean square error of the predicted capacity of our fusion model is 4mAh, which is a 45% decline compared with the benchmark model. We visualize the extracted features, interpret the model and explain the possible mechanism of our model. Furthermore, based on our analysis, suggestions for improving prediction performance are put forward.","PeriodicalId":34110,"journal":{"name":"IET Cybersystems and Robotics","volume":"226 1 1","pages":"462-467"},"PeriodicalIF":0.0,"publicationDate":"2022-07-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"72934177","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}