Patents by Inventor Shengfei ZHANG

Shengfei ZHANG has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).

  • Patent number: 12064600
    Abstract: Disclosed is an acoustic method for precisely controlling and metering manual injection. The method includes: rotating a main housing to allow ratchet grooves to collide with ratchets on a ratchet wheel; obtaining collision voice and vibration signals by a switch sensor and a perception sensor; performing a Mel-frequency cepstral coefficient (MFCC) algorithm and classification with deep neural networks (DNN) on signals detected by a voice sensor, to output processed and recognized voice features; performing an empirical mode decomposition (EMD) algorithm and classification with cyclic delay diversity (CDD) on signals detected by a vibration sensor, to output processed and recognized vibration signals; and inputting the output voice features, the vibration signals, and reception time differences between the voice signals and the vibration signals into a support vector machine (SVM) classifier jointly for final recognition and classification, to recognize whether manual injection is correctly performed.
    Type: Grant
    Filed: March 15, 2023
    Date of Patent: August 20, 2024
    Assignee: Phray Technology Co., Ltd.
    Inventors: Yongxin Wang, Leijie Wang, Jucai Fang, Yanling Liu, Weidong Zhang, Wuyi Ming, Shengfei Zhang
  • Publication number: 20230347052
    Abstract: Disclosed is an acoustic method for precisely controlling and metering manual injection. The method includes: rotating a main housing to allow ratchet grooves to collide with ratchets on a ratchet wheel; obtaining collision voice and vibration signals by a switch sensor and a perception sensor; performing a Mel-frequency cepstral coefficient (MFCC) algorithm and classification with deep neural networks (DNN) on signals detected by a voice sensor, to output processed and recognized voice features; performing an empirical mode decomposition (EMD) algorithm and classification with cyclic delay diversity (CDD) on signals detected by a vibration sensor, to output processed and recognized vibration signals; and inputting the output voice features, the vibration signals, and reception time differences between the voice signals and the vibration signals into a support vector machine (SVM) classifier jointly for final recognition and classification, to recognize whether manual injection is correctly performed.
    Type: Application
    Filed: March 15, 2023
    Publication date: November 2, 2023
    Inventors: Yongxin WANG, Leijie WANG, Jucai FANG, Yanling LIU, Weidong ZHANG, Wuyi MING, Shengfei ZHANG