Interference alignment (IA) has recently emerged as a promising interference mitigation technique for interference networks. In this letter, we focus on the IA non-iterative transceiver design problem in a multiple-input-multiple-output interfering b
roadcast channel (MIMO-IBC), and observed that there is previously unexploited flexibility in different permutations of user ordering. By choosing a good user ordering for a pre-determined IA inter-channel-interference allocation, an improved transceiver design can be accomplished. In order to achieve a more practical performance-complexity tradeoff, a suboptimal user ordering algorithm is proposed. Simulation shows the proposed suboptimal user ordering algorithm can achieve near-optimal performance compared to the optimal ordering while exhibiting only moderate computational complexity.
This paper presents an iterative geometric mean decomposition (IGMD) algorithm for multiple-input-multiple-output (MIMO) wireless communications. In contrast to the existing GMD algorithms, the proposed IGMD does not require the explicit computation
of the geometric mean of positive singular values of the channel matrix and hence is more suitable for hardware implementation. The proposed IGMD has a regular structure and can be easily adapted to solve problems with different dimensions. We show that the proposed IGMD is guaranteed to converge to the perfect GMD under certain sufficient condition. Three different constructions of the proposed algorithm are proposed and compared through computer simulations. Numerical results show that the proposed algorithm quickly attains comparable performance to that of the true GMD within only a few iterations.
