DocumentCode
3412734
Title
The asymptotic multiuser efficiency of two-stage detection in mismatched AWGN channels
Author
Gray, Steven D. ; Brady, David
Author_Institution
Dept. of Electr. & Comput. Eng., Northeastern Univ., Boston, MA, USA
Volume
3
fYear
1993
fDate
11-14 Oct 1993
Firstpage
758
Abstract
This paper investigates the loss of near-far resistance for two multiuser detectors due to imperfect channel estimation in the additive white Gaussian noise multiple access channel. The symbol error rates and the Asymptotic Multiuser Efficiencies (AMEs) of the two-stage hard-limiting detector are determined for imperfect channel knowledge in the two-user asynchronous channel. The degradation of the AME due to channel mismatch is described for a variety of mismatch conditions, including errors in estimation of timing and complex amplitude. It is shown that when perfect knowledge of the channel does not exist, the two-stage detector becomes near-far limited. Also presented is a comparison of the exponentially complex maximum likelihood sequence detector and the hard-decision two-stage detector with imperfect knowledge of the channel
Keywords
Gaussian channels; Gaussian noise; amplitude estimation; communication complexity; error statistics; military communication; signal detection; timing; white noise; asymptotic multiuser efficiency; exponentially complex maximum likelihood sequence detector; hard-decision two-stage detector; imperfect channel estimation; mismatched AWGN channels; multiuser detectors; symbol error rates; two-user asynchronous channel; Additive white noise; Amplitude estimation; Channel estimation; Degradation; Detectors; Error analysis; Estimation error; Maximum likelihood detection; Maximum likelihood estimation; Timing;
fLanguage
English
Publisher
ieee
Conference_Titel
Military Communications Conference, 1993. MILCOM '93. Conference record. Communications on the Move., IEEE
Conference_Location
Boston, MA
Print_ISBN
0-7803-0953-7
Type
conf
DOI
10.1109/MILCOM.1993.408717
Filename
408717
Link To Document