New blind equalisation for non-constant modulus signals using a segment cost function Online publication date: Thu, 30-May-2019
by Wei Rao; Huijun Xu; Jianqiu Zhang
International Journal of Wireless and Mobile Computing (IJWMC), Vol. 16, No. 4, 2019
Abstract: It is well known that for constant modulus (i.e., magnitude) signals the famous constant modulus (CM) blind equalisation algorithm implemented in a fractionally spaced equaliser can present a zero steady-state mean square error (MSE), which means completely eliminating the distortions introduced in transmitting signals through channels. But for non-constant modulus signals it suffers a large steady-state MSE. In order to overcome this defect, a segment cost function according to the CM criterion is suggested. The distinctive feature of the segment cost function is that the equalised signals (of the non-constant modulus signals) are divided into three segments to form a ⊓ shape where the ideal signals have a constant modulus instead of non-constant. And then a new blind equalisation algorithm seeks to minimise this segment cost function by applying a stochastic gradient method is proposed. When employing the proposed algorithm to equalise the 4-PAM or 16-QAM non-constant modulus signals, just as using the CM blind equalisation algorithm to equalise the 4-QAM constant modulus signal, a zero steady-state MSE can be obtained, which is derived. Compared to the classical bind equalisation algorithms, such as CMA, MMA, MCMA, or CMA+SDD, the proposed algorithm yields improved performance, especially for higher SNR.
Online publication date: Thu, 30-May-2019
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Wireless and Mobile Computing (IJWMC):
Login with your Inderscience username and password:
Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.
If you still need assistance, please email firstname.lastname@example.org