Abstract:

Different from traditional multi-antenna systems, cooperative communication systems are inherently time-asynchronous due to the nature of relay distribution. In this paper, we investigate cooperative communication systems based on distributed linear convolutional space-time codes. An equivalent asynchronous system model is obtained by converting the relays’ transmission delays to a zero vector of the corresponding
generator polynomials. With the banded Toeplitz property of the equivalent channel matrix, we propose a block minimum mean square error equalization method using the Trench algorithm to solve the Toeplitz linear systems. In the proposed equalization method, computation of the high order banded Toeplitz linear system is simplified to that of a low order system. The method avoids high order matrix inversion operations of the
traditional minimum mean square error equalization method, leading to substantial reduction of computational complexity. Simulation results show that satisfactory system performance can be achieved with non-minimumorder non-trace-orthogonal generator polynomials, proper data frame length and large delay.

Key words: asynchronous cooperative communication, distributed linear convolutional space-time code, Toeplitz, minimum mean square error, equalization