基于基展开模型的时变信道阶数和径数盲估计

Blind Estimation of the Order and Path Number of Time-Varying Channel under Basis Expansion Model

针对基展开模型的时变信道阶数和径数盲估计问题,采用了一种子空间投影算法来进行估计。该算法充分利用输入子空间和输出子空间具有的同构关系,将当前的观测数据投影到由过去和将来的观测数据所张成的子空间,其投影误差矩阵包含了时变信道的阶数和径数信息,进而可通过求投影误差矩阵的秩和范数来估计信道阶数和径数。仿真表明,与MDL、AIC和Liavas准则相比,该算法可在较低的信噪比下实现时变信道的阶数估计。

For blind estimation of the order and path number of time-varying channel under basis expansion model, this paper presents a subspace projection algorithm. After recognizing the isomorphic relations between the input and output subspaces, by projecting current observation data into a subspace spanned by past and future data, the obtained projection error matrix contains the information of the order and path number of the time-varying channel, so the order and path number can be esti- mated based on the norm and rank of the projection error matrix. Simulation results indicate that compared with the MDL, AIC and Liavas criteria, this algorithm could realize the order estimation of time-varying channel under low SNR conditions.