低复杂度的Morse电报自动译码算法  

Low-Complexity Automatic Decoding Algorithm for Morse Telegraphy

作　　者：高长江 王俊波[1] 吴立强 毛钦曙 彭志新 梁中凯 GAO Changjiang;WANG Junbo;WU Liqiang;MAO Qinshu;PENG Zhixin;LIANG Zhongkai(Southeast University,Nanjing Jiangsu 211189,China;Nanjing Panda Handa Technology Co.,Ltd.,Nanjing Jiangsu 210001,China;Unit 69246 of PLA,Urumqi Xinjiang 830017,China;Unit 92211 of PLA,Haikou Hainan 570000,China)

机构地区：[1]东南大学,江苏南京211189 [2]南京熊猫汉达科技有限公司,江苏南京210001 [3]中国人民解放军69246部队,新疆乌鲁木齐830017 [4]中国人民解放军92211部队,海南海口570000

出　　处：《通信技术》2024年第12期1254-1259,共6页Communications Technology

基　　金：慧眼行动——复杂电磁环境下的短波手键报的自动译码技术。

摘　　要：通过短波信道传输的手键报(Morse电报),在应急通信领域具有不可替代的作用。基于传统机器学习聚类的自动译码方法非常契合于计算资源有限的情况,但这类算法在低信噪比条件下的识别准确率有待改善。提出了一种面向资源受限环境的Morse电报自动译码方法,利用短时傅里叶变换从Morse信号中提取出一维能量特征,使用Savitzky-Golay滤波器进行特征的平滑降噪,基于统计特性计算Morse信号判决门限,并提取出信号与间隔的时间长度特征,通过K均值聚类方法将信号分类为点和划,将间隔分类为码间隔、字间隔和组间隔,最终得到完整的Morse译码报文。实验结果显示,所提算法在信噪比高于-10 dB的仿真数据下能够达到95.9%以上的码准确率,明显优于同类算法。The manual Morse code transmission over HF(High Frequency)channels,also known as Morse telegraphy,plays an irreplaceable role in emergency communications.Automatic decoding methods based on conventional machine learning clustering are well-suited for scenarios with limited computing resources.However,the recognition accuracy of such algorithms under low signal-to-noise ratio conditions needs improvement.This paper proposes an automatic decoding method for Morse telegraphy tailored for resource-constrained environments.It utilizes short-time Fourier transform to extract one-dimensional energy features from Morse signals,employs Savitzky-Golay filter for feature smoothing and noise reduction,calculates Morse signal decision thresholds based on statistical properties,and extracts signal and spacing time length features.Then,through K-means clustering method,it classifies the signals into dots and dashes,and categorizes the spacing into inter-character spacing,inter-word spacing and inter-group spacing,which finally obtains the complete decoded Morse message.Experimental results demonstrate that the proposed algorithm achieves a code accuracy of over 95.9%on simulated data with signal-to-noise ratios higher than-10 dB,outperforming similar algorithms significantly.

关 键 词：短波通信 Morse电报 自动译码 资源受限 K均值聚类 

分 类 号：TN922[电子电信—通信与信息系统]