电子学报 ›› 2019, Vol. 47 ›› Issue (7): 1566-1574.DOI: 10.3969/j.issn.0372-2112.2019.07.023

• 学术论文 • 上一篇    下一篇

低信噪比下RSC码快速迭代寻优识别算法

吴昭军1, 张立民1, 钟兆根2, 孙雪丽2   

  1. 1. 海军航空大学信息融合研究所, 山东烟台 264001;
    2. 海军航空大学航空基础学院, 山东烟台 264001
  • 收稿日期:2018-04-04 修回日期:2018-09-03 出版日期:2019-07-25
    • 通讯作者:
    • 张立民
    • 作者简介:
    • 吴昭军 男,1992年9月出生于四川蓬溪.现为海军航空大学博士研究生.主要研究方向信道编码盲识别.E-mail:wuzhaojun1992@qq.com
    • 基金资助:
    • 国家自然科学基金重大研究计划 (No.91538201); 泰山学者工程专项经费 (No.Ts201511020)

Fast Iterative Recognition of RSC Encoder at Low SNR

WU Zhao-jun1, ZHANG Li-min1, ZHONG Zhao-gen2, SUN Xue-li2   

  1. 1. Department of Information Fusion, Naval Astronautical University, Yantai, Shandong 264001, China;
    2. School of Aviation Basis, Naval Astronautical University, Yantai, Shandong 264001, China
  • Received:2018-04-04 Revised:2018-09-03 Online:2019-07-25 Published:2019-07-25
    • Supported by:
    • National Natural Science Foundation of China Major Research Project (No.91538201); Taishan Scholars Special Funding for Construction Projects (No.Ts201511020)

摘要: 为了解决现有算法在RSC码多项式参数识别过程中,实时性不好和容错性差两大缺点,提出了具有低信噪比适应能力的RSC码快速迭代识别算法.首先根据RSC码元之间的线性约束关系,定义了双曲正切符合度概念,该概念能够表征在某一多项式参数下,截获码元之间的线性关系成立的可能性大小;其次将截获码元总的双曲正切符合度值作为代价函数,然后将待识别的多项式参数的概率值作为代价函数自变量,从而将RSC码参数识别问题转化为多元函数极大值求解问题;最后利用变步长梯度上升方法,在有限次的迭代下,完成在连续概率空间中代价函数极大值求解,最终完成RSC码参数识别.提出的算法收敛速度快且稳定,除了具有较强的低信噪比适应能力外,其计算量与编码器寄存器个数以及码元路数成平方倍数增长.仿真实验表明:提出的算法最多在第5次迭代时,就能完成参数的收敛,同时低信噪比的适应能力较强,即使在0dB条件下,RSC码多项式参数识别率能达到90%以上;与现有的相关算法相比,所提算法的低信噪比适应能力提高了近3dB,同时完成一次参数识别的时间大大降低.

关键词: RSC码, 双曲正切符合度, 变步长, 梯度上升法, 识别

Abstract: In order to solve the defects which are poor error tolerance and large amount of calculation in current algorithms in recognition of the RSC encoder, a fast iterative recognition algorithm which has excellent performance was proposed. Firstly, according to the linear constraint relation between RSC symbols, the concept of hyperbolic tangent conformation was defined; this can measure the possibility of the linear relationship between the symbols under a certain polynomial parameter. Secondly, the total hyperbolic tangent coincidence value of the intercepted symbol was used as a cost function, and then the probability value of the polynomial parameters were regarded as the cost function independent variable and the problem of RSC code identification was transformed into the maximum value of the multivariate function. Finally, the variable step gradient method was used to solve the maximum value of the cost function in the continuous probability space at the finite iteration. The proposed algorithm has a fast and stable convergence speed. In addition to the strong adaptive ability of low SNR, the computational complexity increases squarely with the number of encoder registers and the number of symbols. The simulation experiment showed that the proposed algorithm could achieve the convergence of the parameters at most fifth iterations, while have strong ability to suit to the low SNR. Even the SNR is 0dB, the correct identification rate of RSC code can reach more than 90%. Compared with the existing algorithm, the proposed algorithm improved the adaptive capacity of low SNR by nearly 3dB, at the same time,the time consuming is greatly reduced.

Key words: recursive systematic convolutional (RSC) codes, hyperbolic tangent conformation, variable step size, gradient ascent method, recognition

中图分类号: