Training sequence-based distributed space-time block codes with frequency domain equalization

This paper proposes a new distributed space-time block code (DSTBC) for single-carrier communications systems based on training sequence (TS). The proposed DSTBC is devised for amplify-and-forward (AF) relay networks over frequency selective fading channels. The data is transmitted from the source to the destination by a block-wise manner, resembling the block level of Alamouti scheme. A TS is padded to the tail of each information-bearing data block. The use of TS in this paper is twofold. First, it plays as a cyclic extension, thereby allowing us to construct a low-complexity receiver in the frequency domain. Second, it can be used for channel estimation. In contrast to zero padding technique, the introduction of TS makes it difficult to decouple the detection of data blocks. To recover the orthogonality of the equivalent space-time channel, we carefully treat the interference induced by the training sequences. As the second contribution, a least square (LS) channel estimation approach is also proposed for the proposed DSTBC. The optimal training sequences are designed to minimize the effect of additive noise. Bit-error-rate (BER) performance comparison is carried out by computer simulation.