Dynamic Characteristic Parameters Estimation Based EM-EKF Algorithm For MIMO Channel

With the development of science and technology, the wireless network communication technology plays an increasingly important strategic role in national socio-economic development and penetrates into all areas of society. Cognitive radio technology as the key enabling technology of next-generation communications networks can effectively solve the spectrum resources relatively “short” and “waste” in wireless network. Cognitive radio technology can be aware of its surrounding environment and adapts the transmission parameters in real-time. One of its key technologies is that the user have in-depth understanding capacity of the multi-dimensional and multi-scale characteristics of wave propagation in dynamic wireless environment. The research about the double-directional multiple-input multiple-output(MIMO) channel has become a hot topic in cognitive radio network. But most researches about the MIMO channel focuses on modeling the channel and describing the radio waves propagation characteristics in static scene with the propagation environments fixed, which will results that the user terminals are lack of the ability for perceiving dynamic radio wave propagation characteristics.In this paper, we propose a dynamic multi-dimensional parameter joint estimation method for estimation and tracking of radio wave propagation environment parameters in the dynamic wireless MIMO channel propagation environment. Thus, it can further enhance perceptive ability of the cognitive terminal for local radio propagation environment and complete adaptive communication services. In this paper we firstly analyze the features of dynamic wireless channel, and then characterize the parameters of dynamic radio channel. Then we build an observable MIMO channel model and simplify the model using two-Dimensional Discrete Fourier Transform(2-D DFT) in order to reduce the computational complexity of the derivation of the model. Based on the simplified model, this paper evolves the formulas of traditional EM algorithm in time domain to the ones in frequency domain. Moreover, we propose joint EM in frequency domain and EKF algorithm(EM-EKF) to estimate the propagation path parameters in dynamic wireless MIMO channel environment and achieve dynamic tracking of propagation path parameters.In order to assess the performance of the proposed algorithm in this paper, we build two typical simulation scenarios using the IlmProp simulation platform and respectively analysis the estimation performance of the proposed algorithm in the two simulation scenarios. The performance of the algorithm is further evaluated by analyzing the bias between estimates and true value of the propagation path parameter in delay-angle domain. Compared with the conventional SAGE algorithm, the proposed EM-EKF approach significantly outperforms the SAGE in both accuracy and computational cost from the simulation results.