论文部分内容阅读
提出并实现了一种基于Madalines人工神经元网络的通信网信号识别系统的研制方案．该系统可以在复杂的线路及干扰情况下，实时完成传真、计算机Modem通信、BP机自动寻呼、拨号音、双音频、电子合成语、热线音乐、噪声和话音等信号的自动识别工作，系统最大监测容量可达16384线，检出率超过30％，虚警率低于万分之一．
Proposed and implemented a Madalines artificial neural network based communication network signal recognition system development program. The system can automatically recognize the signals such as fax, computer Modem communication, BP automatic paging, dial tone, dual audio, electronic compound language, hotline music, noise and voice under complex line and interference conditions. The system The maximum monitoring capacity of up to 16,384 lines, the detection rate of over 30%, false alarm rate is less than one ten thousandth.