Abstract
In this paper, we propose a new low-complexity optimal transmitting antenna selection algorithm for multiple-input multiple-output (MIMO) systems. Different from the conventional optimal antenna selection methods, our proposed algorithm can approximate the inverse matrix subject to an error tolerance. However, the antenna selection optimality can often be obtained as the same outcome from the exact matrix inverse. Furthermore, our proposed matrix inverse approximation algorithm can be greatly expedited by using parallel computing. With significantly reduced complexity using many microprocessors, our proposed method is very appealing to the future MIMO technologies. © 2013 IEEE.