高斯信道下极化码构造方法研究

Study on construction methods for polar codes in AWGN channels

极化码是基于信道极化(Channel Polarization)现象,首次以构造性的方法逼近信道容量的码。极化码的构造可以归结为子信道的选择问题,它直接决定了连续删除译码算法的性能和极化码的渐进性,Arikan给出了在BEC信道下极化码的详细构造方法,然而在高斯信道下这种方法的复杂度会随着码长的增加而指数增加,文中分别详细给出了扎起高斯信道下利用蒙特卡罗方法、密度进化、高斯近似进行极化码构造的步骤和计算方法,并从极化效果、帧错误率和复杂度上分析和比较了这三种方法在性能以及复杂度上的区别,仿真结果表明基于高斯近似的构造方法具有低复杂度和低错误概率的性能。

Polar code is based on the polarization channel （Channel Polarization） phenomenon, for the first time to construct the method of approximation of channel capacity. The construction of the polar codes can be attributed to the selection of channel, which directly determines the performance of the successive cancellation decoding algorithm and the asymptotic behavior of polar codes, Arikan gave out under the BEC channel polarization code construction method in detail, but the complexity of this method under gaussian channel will increase with the yard and index increased, this paper gives out a detailed erected under gaussian channel using monte carlo method, the density evolution, the Gaussian approximation to steps and calculation methods of polarization code structure, and from the polarization effect, frame error rate and the complexity of analysis and comparison of these three methods on the performance and complexity of difference. The simulation results show that based on the structure of the Gaussian approximation method has low complexity and low performance of the error probability.