New subspace-based blind channel estimation for orthogonally coded MIMO-OFDM systems

Abstract

Two novel subspace-based blind channel estimation methods for orthogonally coded multiple-input multiple-output (MIMO) orthogonal frequency division multiplexing (OFDM) systems with more than two transmit antennas are proposed in this paper. The proposed methods exploit the null space induced by the orthogonal space-time block codes (OSTBC). They can be applied to systems with one or more receive antennas. Simulation results show that accurate estimates can be obtained. Index Terms—OFDM, MIMO, orthogonal space-time block code (OSTBC), blind channel estimation I. INTRODUCTION In recent years, space-time block coding (STBC) has emerged as a powerful approach to exploit spatial diversity and to combat fading in multiple-input multiple-output (MIMO) wireless communication systems. In 1998, Alamouti's code for two transmit antennas (1) has been incorporated as a simple and efficient transmit diversity technique. STBC with orthogonal designs for any number of transmit antennas was proposed in (2). The combination of orthogonal space-time block codes (OSTBC) and orthogonal frequency division multiplexing (OFDM), or simply OSTBC-OFDM, has drawn much attention because it attains the maximum transmit di- versity and has a simple maximum likelihood (ML) receiver structure. It is known that the performance of MIMO-OFDM systems critically depend on the accuracy of the estimated channel responses. Many channel estimation methods have been proposed in the past. Our paper focuses on blind channel estimation. Most of the existing blind MIMO channel estimation meth- ods can only deal with flat fading channels (3) (4). The OFDM system converts the frequency-selective MIMO channel into many flat-fading subcarriers in MIMO-OFDM systems. The subcarrier-wise approach may suffer from a prohibitively high computational complexity when the number of subcarriers is large. A coherent approach was proposed in (5) for blind channel estimation for orthogonally coded MIMO-OFDM systems with more than 2 transmit antennas. Because the approach is based on convex optimization and amounts to solving a semidefinite program (SDP), its computational cost may be rather high. In (6), the authors proposed a lower computational-complexity method. The method gives a good performance but it works only when there are more than one receive antenna. In this paper, we propose two new algorithms for blind chan- nel estimation. In the first part, we propose a new subspace- based method for blind channel estimation in MIMO-OFDM systems with more than 2 transmit antennas. The proposed method exploits the null space induced by the OSTBCs with rate less than one. The method works even when there is only one receive antenna. In the second part, we introduce a modification of the proposed method so that fewer received blocks are needed. The minimum number of received blocks needed can be as small as one for most practical systems. Numerical simulation is also given to verify the performance of the proposed methods.