Automatic neural network structure searching method for communication signal modulation identification

• Main Authors: HUO YURONG, MA ZHIHAO, FANG YUQIANG, SONG WANJUN, CHEN WEIGAO
• Format: Patent
• Language: Chinese; English
• Published: 15.06.2021
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; PHYSICS

The invention discloses an automatic neural network structure search method for communication signal modulation identification, which reduces the consumption of computing resources by the network structure search method, gets rid of the dependence on large-scale high-performance computing equipment, and realizes the generation of a lightweight network structure model and the deployment of portable equipment; The method can be used for deployment and development of communication signal identification application software of a portable equipment end, solves the problem that a network structure search model is large and is difficult to deploy in mobile or portable equipment, and explores the application of network structure search in edge equipment from the algorithm level. 一种通信信号调制识别的自动神经网络结构搜索方法，减少网络结构搜索方法对计算资源的消耗，摆脱对大规模高性能计算设备的依赖，实现轻量化的网络结构模型生成以及在便携式设备的部署，可用于便携式设装备端通信信号识别应用软件的部署与开发，服网络结构搜索模型大、在移动或便携式设备部署难的问题，从算法层面探索网络结构搜索在边缘设备中的应用。