一种低功耗架构下毫米波信道估计算法研究

Millimeter wave channel estimation algorithm under lowpower consumption architecture

针对低分辨率模数转换器量化误差大、信道估计准确性低等问题,提出一种两阶段信道估计算法。利用量化过程构造优化问题,采用梯度下降算法估计支撑集;根据估计的支撑集将信道估计问题降维;采用最大化期望算法降低量化误差并估计信道复增益。仿真结果表明,提出的算法比目前最先进的对比算法性能提高了6 dB,且时间复杂度由对数线性数量级降为线性数量级。

The low-resolution analog-to-digital converter results in increased quantization error,which leads to the decreasing accuracy of the channel estimation.This paper proposed a two-stage channel estimation algorithm.First,the optimization problem is modeled with the quantization process,and the support set is estimated via gradient descent.Then,the dimension of the channel matrix is reduced according to the estimated support set.Finally,the expectation maximization is employed to reduce the quantization error and hence estimate the channel complex gain.Simulation results demonstrated that the proposed method improves the performance of channel estimation by 6dB as compared to the state-of-the-art algorithm.Meanwhile,the computational complexity of the proposed method is reduced from logarithmic linear to linear.