摘要
随着无线通信技术快速的发展,频谱环境变得日益复杂,传统通信信号盲检测与识别方法面临检测准确率低、识别速度慢及抗噪声能力弱等诸多问题。深度学习技术凭借其强大的特征自动提取能力和非线性映射能力,为解决上述问题提供了全新的技术途径。构建基于卷积神经网络的盲检测模型,实现复杂电磁环境下微弱信号的有效检测工作,设计多尺度特征融合的识别网络,完成多种调制方式的高精度分类任务。实验结果表明:所提方法在低信噪比条件下检测概率提升至90%以上,识别准确率达到95%,相比传统方法性能得到显著提高,验证了深度学习技术在通信信号处理领域的有效性。
With the rapid advancement of wireless communication technologies,spectrum environments have become increasingly complex.Traditional blind signal detection and recognition methods face challenges such as low detection accuracy,slow recognition speed,and weak noise resistance.Deep learning technology,leveraging its powerful feature extraction capabilities and nonlinear mapping abilities,provides a novel technical approach to address these issues.By constructing a convolutional neural network-based blind detection model,we achieved effective detection of weak signals in complex electromagnetic environments.Through designing a multi-scale feature fusion recognition network,we accomplished high-precision classification tasks for various modulation schemes.Experimental results demonstrate that the proposed method improves detection probability to over 90%and recognition accuracy to 95%under low signal-to-noise ratio conditions,signifi cantly outperforming traditional methods.This validates the eff ectiveness of deep learning technology in communication signal processing.
作者
高鹏
陈俊吉
Gao Peng;Chen Junji(Chongqing Yitong University,Chongqing 401520,China)
出处
《黑河学院学报》
2025年第12期182-185,共4页
Journal of Heihe University
基金
重庆市合川区科研项目“一体化形变智能北斗监测预警系统”(HCKJ-2024-113)
重庆市级教改项目“基于EIP-CDIO理念的‘12335’电子信息类专业实践教学体系的构建”(24JG101)。
关键词
深度学习
盲检测
信号识别
卷积神经网络
deep learning
blind detection
signal recognition
convolutional neural network
