Identification and Equalization Algorithm for MIMO System Based on Higher Order Statistics in Frequency Domain

Combining higher order statistics that can identify non-minimum phase system with shift block method of blind separation (SHIBBS) algorithm which is good in separation and moderate amount of computation, an algorithm was proposed to identify and separate mixed sources for muti-input and muti-output (MIMO) system in frequency domain. However, the received signals have some disadvantages such as the uncertain permutation and gain ambiguities. Under the condition that the sources are independent from each other, the permutation problem was solved by using cross fourth-order statistics, then we get the separated signals; and constant modulus algorithm, employing the LMS for adaptation, is used in the gain ambiguities. Simulations show that the proposed algorithm is feasible and practical.