摘要
设计了一套航天密封继电器信号识别方法,基于微粒碰撞噪声检测(PIND)法,通过对信号的时域及频域的特征分析选取了多个具有代表性的特征。运用了机器学习方法,采用多层感知机并对其进行超参数寻优,找到了最优的超参数组合。对信号进行处理分析来判断密封继电器内是否存在多余物。提高了分类的准确率,同时也提高了模型的整体性能。
A set of aerospace sealed relay signal detection method is designed based on the Particle Impact Noise Detection (PIND) method and several representative features are selected by analyzing the time domain and frequency domain of the signal. By applying the machine learning method and using multi-layer perceptron as well as super-parameter optimization,the optimal hyper-parameter combination is found. The signal is processed and analyzed to determine if there is any excess in the sealed relay. The accuracy of the classification is improved,and the overall performance of the model is also improved.
作者
李响
蒋爱平
杨世华
薛永越
王国涛
LI Xiang;JIANG Ai⁃ping;YANG Shi⁃hua;XUE Yong⁃yue;WANG Guo⁃tao(Electronic Engineering College of Heilongjiang University of Technology,Harbin 150008,China;Military Apparatus Research Institute of Harbin Institute of Technology,Harbin 150001,China;Shanghai Aerospace Equipments Manufacturer,Shanghai 200245,China)
出处
《宇航计测技术》
CSCD
2020年第2期30-35,共6页
Journal of Astronautic Metrology and Measurement
基金
“国家自然科学基金”(51607059)
“黑龙江省自然科学基金”(QC2017059)
“黑龙江省博士后基金”(LBH-Z16169)
“黑龙江省高校基本科研业务费”(HDRCCX-201604)
“黑龙江省普通高校重点实验室开放课题”(2012TD007)
“黑龙江省教育厅科技成果培育”(TSTAUC2018016)资助项目。
