摘要
针对RFID标准中常用的数字调制方式,提出了一种新的低信噪比下测试识别方法。该方法采用自适应阈值小波消噪方法预处理识别信号,提高了在低信噪比下的调制识别能力,设计了一种基于遗传BP神经网络的识别分类器,进一步改善了低信噪比下的识别效果。仿真结果表明,该方法在信噪比为5dB时,识别正确率也能达到95%以上。
This paper proposes a new test and classification method at low SNR (signal-to-noise) on common digital modulation modes in RFID (radio frequency identification) standard. On one hand, it puts forward a method of pre-processing recognition signal by using self-study threshold wavelet denoise,which can improve the recognition ability at low SNR;on the other hand, it designs a recognition classifier of GA-BP (genetic algorithm-back propagation) nerual network,which can improve the classification effectiveness.Simulation result proves that this method is quite practical because the calssification accuracy is over 95% ever if the SNR is low to 5dB.
出处
《电子技术应用》
北大核心
2010年第4期111-114,118,共5页
Application of Electronic Technique
基金
国家863计划(No.2006AA04A104)
国家自然科学基金项目(NO.50677014
60876022)
高校博士点基金(20060532002)
湖南省科技计划项目(2008Gk2022)资助