Conventional Minimum-Variance (MV) channel estimation is affected by two sources of error, namely the finite number of samples used to estimate the covariance matrix and the asymptotic bias due to interference and additive noise. On the other hand, widely linear (WL) filtering has been shown to improve the estimation of improper complex signals. Researchers have recently demonstrated that the application of WL processing principles can significantly improve the performance of subspace-based channel estimation algorithms. However, in contrast to MV estimation algorithms, subspace-based algorithms assume knowledge of the total number of users in the system, and must be coupled with sophisticated user enumeration algorithm at the expense of increased complexity. In this work, in an effort to combine the practical advantages of MV channel estimation algorithms with the performance of WL filters we propose a widely linear version of the MV channel estimator in the context of multicarrier(MC) CDMA systems employing real modulation. We use numerical simulations to demonstrate that the widely linear minimum-variance algorithm yields more accurate channel estimates compared to the conventional MV algorithm. By considering two simplified transmission/reception models, we also show analytically that the widely linear estimator on average reduces both types of error.
Identifer | oai:union.ndltd.org:LACETR/oai:collectionscanada.gc.ca:QMM.112550 |
Date | January 2007 |
Creators | Abdallah, Saeed. |
Publisher | McGill University |
Source Sets | Library and Archives Canada ETDs Repository / Centre d'archives des thèses électroniques de Bibliothèque et Archives Canada |
Language | English |
Detected Language | English |
Type | Electronic Thesis or Dissertation |
Format | application/pdf |
Coverage | Master of Engineering (Department of Electrical and Computer Engineering.) |
Rights | All items in eScholarship@McGill are protected by copyright with all rights reserved unless otherwise indicated. |
Relation | alephsysno: 002713644, proquestno: AAIMR51440, Theses scanned by UMI/ProQuest. |
Page generated in 0.0461 seconds