Patents Assigned to INSTITUTE OF AUTOMATION CHINESE ACADEMY OF SCIENCES
  • Publication number: 20210119908
    Abstract: The present invention discloses a data forwarding unit based on a Handle identifier, comprising a dynamic configuration module, a Handle identifier data identification module and a matching-forwarding module. The system of the present invention is applied to network devices such as switches and routers, and supports dynamic configuration of data packet analysis, matching and forwarding rules through data interaction with network systems such as SDN managers, so that the network devices can identify data packets based on the Handle identifier and perform the specified operation on the designated data packets with the Handle identifier according to the rules of dynamic configuration.
    Type: Application
    Filed: December 19, 2019
    Publication date: April 22, 2021
    Applicant: SHENYANG INSTITUTE OF AUTOMATION, CHINESE ACADEMY OF SCIENCES
    Inventors: Haibin YU, Peng ZENG, Dong LI, Zhibo LI, Jindi LIU, Xueting YU, Ming YANG
  • Patent number: 10970617
    Abstract: An acceleration and compression method for a deep convolutional neural network based on quantization of a parameter provided by the present application comprises: quantizing the parameter of the deep convolutional neural network to obtain a plurality of subcode books and respective corresponding index values of the plurality of subcode books; acquiring an output feature map of the deep convolutional neural network according to the plurality of subcode books and respective corresponding index values of the plurality of subcode books. The present application may implement the acceleration and compression for a deep convolutional neural network.
    Type: Grant
    Filed: August 21, 2015
    Date of Patent: April 6, 2021
    Assignee: INSTITUTE OF AUTOMATION CHINESE ACADEMY OF SCIENCES
    Inventors: Jian Cheng, Jiaxiang Wu, Cong Leng, Hanqing Lu
  • Patent number: 10964026
    Abstract: A refined segmentation system, method and device of an image shadow area are provided. The system of the present invention includes: a feature extraction network, a reverse fusion network, and a weighted fusion network. The feature extraction network includes a plurality of sampling layers which are arranged sequentially, a plurality of segmentation features of the shadow areas in the input images are obtained through the sampling layers sequentially. The reverse fusion network includes a plurality of layered reverse fusion branches, each of which includes a plurality of feature fusion layers arranged in sequence, and two input features are fused in sequence through each feature fusion layer. The weighted fusion network is configured to perform weighted fusion on outputs of the plurality of reverse fusion branches to obtain a final segmentation result of the shadow area of the input image.
    Type: Grant
    Filed: April 19, 2019
    Date of Patent: March 30, 2021
    Assignee: INSTITUTE OF AUTOMATION, CHINESE ACADEMY OF SCIENCES
    Inventors: Xin Zhao, Kaiqi Huang, Yupei Wang
  • Patent number: 10962976
    Abstract: A motion control method and system for a biomimetic robotic fish based on an adversarial structured control, includes: taking the accuracy and speed of motion to the target point as a reward term, and taking a power sum of servomotors as a loss term to construct an optimization objective function; optimizing parameters of a central pattern generator model that generates a global control quantity of a servomotor, after curing its parameters, optimizing the parameters of the servomotor compensation control model; iteratively optimizing the parameters of the model; obtaining the global control signal and compensation control signal of the biomimetic robotic fish through the trained model, and using the linear combination of the two sets of output signals as the control signal of the servomotor of the robotic fish to realize the motion control of the fish.
    Type: Grant
    Filed: November 11, 2020
    Date of Patent: March 30, 2021
    Assignee: INSTITUTE OF AUTOMATION, CHINESE ACADEMY OF SCIENCES
    Inventors: Zhengxing Wu, Junzhi Yu, Shuaizheng Yan, Jian Wang, Min Tan
  • Patent number: 10939845
    Abstract: A FFL-based magnetic particle imaging three-dimensional reconstruction method includes: acquiring current signal data of an induction coil during FFL-based three-dimensional scanning process of a scanned object; based on the current signal data, performing deconvolution through a preset kernel function to acquire a two-dimensional image data set, wherein the kernel function is a step function with L2 regularized constraint; based on the two-dimensional image data set, acquiring an initial three-dimensional image by using a Wiener filtering deconvolution algorithm; and based on the initial three-dimensional image, performing deconvolution through a Langevin function, and acquiring a final three-dimensional image by Radon transformation.
    Type: Grant
    Filed: June 22, 2020
    Date of Patent: March 9, 2021
    Assignee: INSTITUTE OF AUTOMATION, CHINESE ACADEMY OF SCIENCES
    Inventors: Jie Tian, Peng Zhang, Hui Hui, Kun Wang, Xin Yang
  • Publication number: 20210065021
    Abstract: The present invention relates to a working condition state modeling and model correcting method, comprising collecting data, and arranging the data in a chronological order to form a time sequence data set; preprocessing the time sequence data set; clustering the preprocessed time sequence data set, computing a central point data set of the duster, and generating a working condition data set and a working condition process data set; counting a working condition transition probability for the working condition process data set to form a working condition transition probability model data set; collecting the data, and detecting and processing the data; computing a working condition state transition mode phase by phase and processing.
    Type: Application
    Filed: February 21, 2019
    Publication date: March 4, 2021
    Applicant: SHENYANG INSTITUTE OF AUTOMATION, CHINESE ACADEMY OF SCIENCES
    Inventors: Wenli SHANG, Peng ZENG, Xianda LIU, Jianming ZHAO, Long YIN, Chunyu CHEN, Jiansong AO, Guoyu TONG
  • Publication number: 20210065371
    Abstract: A refined segmentation system, method and device of an image shadow area are provided. The system of the present invention includes: a feature extraction network, a reverse fusion network, and a weighted fusion network. The feature extraction network includes a plurality of sampling layers which are arranged sequentially, a plurality of segmentation features of the shadow areas in the input images are obtained through the sampling layers sequentially. The reverse fusion network includes a plurality of layered reverse fusion branches, each of which includes a plurality of feature fusion layers arranged in sequence, and two input features are fused in sequence through each feature fusion layer. The weighted fusion network is configured to perform weighted fusion on outputs of the plurality of reverse fusion branches to obtain a final segmentation result of the shadow area of the input image.
    Type: Application
    Filed: April 19, 2019
    Publication date: March 4, 2021
    Applicant: INSTITUTE OF AUTOMATION, CHINESE ACADEMY OF SCIENCES
    Inventors: Xin ZHAO, Kaiqi HUANG, Yupei WANG
  • Patent number: 10935986
    Abstract: A gliding depth control method for a biomimetic gliding robotic dolphin includes: obtaining a preset gliding depth and a preset yaw angle; obtaining an estimated velocity by a sliding mode observer based on depth information and inertial navigation information, and obtaining a control quantity of pectoral fins on both sides of the biomimetic gliding robotic dolphin by a yaw controller in combination with the preset yaw angle; obtaining a segmented diving velocity reference trajectory by constructing and segmenting a B├ęzier curve; obtaining a diving control quantity by a model predictive control method in combination with the estimated velocity; obtaining a target position of a piston through a buoyancy principle, and obtaining a control quantity of the piston according to a current position of the piston; and controlling the biomimetic gliding robotic dolphin to glide based on the control quantity of the piston and the control quantity of the pectoral fins.
    Type: Grant
    Filed: October 14, 2020
    Date of Patent: March 2, 2021
    Assignee: INSTITUTE OF AUTOMATION, CHINESE ACADEMY OF SCIENCES
    Inventors: Junzhi Yu, Zhengxing Wu, Jian Wang, Shuaizheng Yan, Min Tan
  • Publication number: 20210056415
    Abstract: An information processing method based on contextual signals and a prefrontal cortex-like network includes: selecting a feature vector extractor based on obtained information to perform feature extraction to obtain an information feature vector; inputting the information feature vector into the prefrontal cortex-like network, and performing dimensional matching between the information feature vector and each contextual signal in an input contextual signal set to obtain contextual feature vectors to constitute a contextual feature vector set; and classifying each feature vector in the contextual feature vector set by a feature vector classifier to obtain classification information of the each feature vector to constitute a classification information set. An information processing system based on contextual signals and a prefrontal cortex-like network includes an acquisition module, a feature extraction module, a dimensional matching module, a classification module and an output module.
    Type: Application
    Filed: April 19, 2019
    Publication date: February 25, 2021
    Applicant: INSTITUTE OF AUTOMATION, CHINESE ACADEMY OF SCIENCES
    Inventors: Guanxiong ZENG, Yang CHEN, Shan YU
  • Patent number: 10923136
    Abstract: A speech extraction method based on the supervised learning auditory attention includes: converting an original overlapping speech signal into a two-dimensional time-frequency signal representation by a short-time Fourier transform to obtain a first overlapping speech signal; performing a first sparsification on the first overlapping speech signal, mapping intensity information of a time-frequency unit of the first overlapping speech signal to preset D intensity levels, and performing a second sparsification on the first overlapping speech signal based on information of the preset D intensity levels to obtain a second overlapping speech signal; converting the second overlapping speech signal into a pulse signal by a time coding method; extracting a target pulse from the pulse signal by a trained target pulse extraction network; converting the target pulse into a time-frequency representation of the target speech to obtain the target speech by an inverse short-time Fourier transform.
    Type: Grant
    Filed: April 19, 2019
    Date of Patent: February 16, 2021
    Assignee: INSTITUTE OF AUTOMATION, CHINESE ACADEMY OF SCIENCES
    Inventors: Jiaming Xu, Yating Huang, Bo Xu
  • Publication number: 20210042929
    Abstract: A three-dimensional object detection method includes: extracting a target in a two-dimensional image by a pre-trained deep convolutional neural network to obtain a plurality of target objects; determining a point cloud frustum in a corresponding three-dimensional point cloud space based on each target object; segmenting the point cloud in the frustum based on a point cloud segmentation network to obtain a point cloud of interest; and estimating parameters of a 3D box in the point cloud of interest based on a network with the weighted channel features to obtain the parameters of the 3D box for three-dimensional object detection. According to the present invention, the features of the image can be learned more accurately by the deep convolutional neural network and the parameters of the 3D box in the point cloud of interest are estimated based on the network with the weighted channel features.
    Type: Application
    Filed: April 19, 2019
    Publication date: February 11, 2021
    Applicant: INSTITUTE OF AUTOMATION, CHINESE ACADEMY OF SCIENCES
    Inventors: Xin ZHAO, Kaiqi HUANG, Zhe LIU
  • Publication number: 20210040846
    Abstract: The present invention relates to a rotating self-drilling device for extraterrestrial objects.
    Type: Application
    Filed: December 28, 2018
    Publication date: February 11, 2021
    Applicant: ShenYang Institute of Automation, Chinese Academy of Sciences
    Inventors: Jinguo LIU, Feiyu ZHANG, Mangkuan WANG, Yuwang LIU
  • Publication number: 20210039726
    Abstract: A reconfigurable joint track compound mobile robot has a main vehicle body, yaw joints and an auxiliary track module. The main vehicle body has a main track, and a clutch brake and a first wheel joint arranged in a main track driving wheel. A second wheel joint is arranged in a main track driven wheel. The main vehicle body is provided with main track driving mechanisms and a wheel joint driving mechanism. The main track driving wheel is driven to rotate by the main track driving mechanisms, which are connected with the clutch brake. The second wheel joint is driven to rotate by the wheel joint driving mechanism. Each wheel joint is correspondingly connected with the yaw joints, which are rotatably connected with the auxiliary track module. A yaw driving mechanism that drives the auxiliary track module to swing is arranged in each yaw joint.
    Type: Application
    Filed: December 31, 2018
    Publication date: February 11, 2021
    Applicant: ShenYang Institute of AutoMation, Chinese Academy of Sciences
    Inventors: Jinguo LIU, Xing LI, Jian DING, Yuwang LIU
  • Patent number: 10915815
    Abstract: An information processing method based on contextual signals and a prefrontal cortex-like network includes: selecting a feature vector extractor based on obtained information to perform feature extraction to obtain an information feature vector; inputting the information feature vector into the prefrontal cortex-like network, and performing dimensional matching between the information feature vector and each contextual signal in an input contextual signal set to obtain contextual feature vectors to constitute a contextual feature vector set; and classifying each feature vector in the contextual feature vector set by a feature vector classifier to obtain classification information of the each feature vector to constitute a classification information set. An information processing system based on contextual signals and a prefrontal cortex-like network includes an acquisition module, a feature extraction module, a dimensional matching module, a classification module and an output module.
    Type: Grant
    Filed: April 19, 2019
    Date of Patent: February 9, 2021
    Assignee: INSTITUTE OF AUTOMATION, CHINESE ACADEMY OF SCIENCES
    Inventors: Guanxiong Zeng, Yang Chen, Shan Yu
  • Publication number: 20210015395
    Abstract: A FFL-based magnetic particle imaging three-dimensional reconstruction method includes: acquiring current signal data of an induction coil during FFL-based three-dimensional scanning process of a scanned object; based on the current signal data, performing deconvolution through a preset kernel function to acquire a two-dimensional image data set, wherein the kernel function is a step function with L2 regularized constraint; based on the two-dimensional image data set, acquiring an initial three-dimensional image by using a Wiener filtering deconvolution algorithm; and based on the initial three-dimensional image, performing deconvolution through a Langevin function, and acquiring a final three-dimensional image by Radon transformation.
    Type: Application
    Filed: June 22, 2020
    Publication date: January 21, 2021
    Applicant: INSTITUTE OF AUTOMATION, CHINESE ACADEMY OF SCIENCES
    Inventors: Jie TIAN, Peng ZHANG, Hui HUI, Kun WANG, Xin YANG
  • Publication number: 20210011457
    Abstract: An online monitoring device of 3D printing equipment includes a signal collection module, a signal processing module, a feature extraction module, a monitoring module and a knowledge base module. A vibration signal of a preset component of the 3D printing equipment is collected by a vibration sensor. The collected vibration signal of each preset component is converted from an analog signal to a digital signal and the spectrum characteristics are extracted. Based on the spectrum characteristics of each preset component, the operation state type of the preset component is obtained by a comparative analysis model. The knowledge base module is configured to store newly added samples and initial samples of the 3D printing equipment. The initial samples include spectrum characteristic information and corresponding fault category of known faults, and the newly added samples include spectrum characteristic information and corresponding fault category of new faults.
    Type: Application
    Filed: July 7, 2020
    Publication date: January 14, 2021
    Applicants: INSTITUTE OF AUTOMATION, CHINESE ACADEMY OF SCIENCES, CLOUD COMPUTING CENTER CHINESE ACADEMY OF SCIENCES, DongGuan, Guangdong (CN)
    Inventors: Gang XIONG, Jiawei LIAO, Zhen SHEN, Xiuqin SHANG, Chao GUO, Jun YAN, Can LUO, Xiao WANG, Feiyue WANG
  • Publication number: 20200402526
    Abstract: A speech extraction method based on the supervised learning auditory attention includes: converting an original overlapping speech signal into a two-dimensional time-frequency signal representation by a short-time Fourier transform to obtain a first overlapping speech signal; performing a first sparsification on the first overlapping speech signal, mapping intensity information of a time-frequency unit of the first overlapping speech signal to preset D intensity levels, and performing a second sparsification on the first overlapping speech signal based on information of the preset D intensity levels to obtain a second overlapping speech signal; converting the second overlapping speech signal into a pulse signal by a time coding method; extracting a target pulse from the pulse signal by a trained target pulse extraction network; converting the target pulse into a time-frequency representation of the target speech to obtain the target speech by an inverse short-time Fourier transform.
    Type: Application
    Filed: April 19, 2019
    Publication date: December 24, 2020
    Applicant: INSTITUTE OF AUTOMATION, CHINESE ACADEMY OF SCIENCES
    Inventors: Jiaming XU, Yating HUANG, Bo XU
  • Publication number: 20200352491
    Abstract: A non-contact brain blood oxygen detecting system includes a mobile terminal device. The mobile terminal device includes a control module, a transmitting module, a receiving module and a display module. The control module is connected to the transmitting module, the receiving module and the display module, respectively. The transmitting module in the mobile terminal device is configured to emit dual-wavelength near-infrared light to a detected subject. The receiving module is configured to receive a light signal after propagation fed back by the detected subject, and to perform data conversion on the received light signal to obtain a digital signal containing blood oxygen information. The control module is configured to obtain the blood oxygen information of the detected subject according to the digital signal obtained by the receiving module. The display module is configured to display the blood oxygen information obtained by the control module.
    Type: Application
    Filed: December 28, 2017
    Publication date: November 12, 2020
    Applicant: INSTITUTE OF AUTOMATION, CHINESE ACADEMY OF SCIENCES
    Inventors: Tianzi JIANG, Xin ZHANG, Nianming ZUO
  • Patent number: 10818311
    Abstract: An auditory selection method based on a memory and attention model, including: step S1, encoding an original speech signal into a time-frequency matrix; step S2, encoding and transforming the time-frequency matrix to convert the matrix into a speech vector; step S3, using a long-term memory unit to store a speaker and a speech vector corresponding to the speaker; step S4, obtaining a speech vector corresponding to a target speaker, and separating a target speech from the original speech signal through an attention selection model. A storage device includes a plurality of programs stored in the storage device. The plurality of programs are configured to be loaded by a processor and execute the auditory selection method based on the memory and attention model. A processing unit includes the processor and the storage device.
    Type: Grant
    Filed: November 14, 2018
    Date of Patent: October 27, 2020
    Assignee: INSTITUTE OF AUTOMATION, CHINESE ACADEMY OF SCIENCES
    Inventors: Jiaming Xu, Jing Shi, Bo Xu
  • Patent number: 10813130
    Abstract: The present invention relates to cognitive wireless sensor network technologies, and in particular to a medium access control method for a cognitive sensor network based on broadcast preamble sampling. A cognitive node in the cognitive sensor network adopts a periodical dormancy-awakening mechanism. Firstly, a state of a primary user is judged by using a spectrum sensing technology. If the primary user is not active, a data sending node establishes a communication link by using a broadcast preamble code. Each neighbor node simultaneously considers information of hops from a gateway according to an awakening order, and independently determines to serve as a relay node for forwarding data. Furthermore, a transmission conflict between the cognitive node and the primary user may be caused due to return or missed alarm of the primary user. The present invention adopts a retransmission mechanism based on confirmation to ensure transmission reliability of data packets.
    Type: Grant
    Filed: September 7, 2017
    Date of Patent: October 20, 2020
    Assignee: SHENYANG INSTITUTE OF AUTOMATION, CHINESE ACADEMY OF SCIENCES
    Inventors: Wei Liang, Haibin Yu, Meng Zheng, Manyi Du, Shuai Liu, Yutuo Yang