Adaptive Multicarrier CDMA Space-Time Receivers - Adaptive Signal Processing in Wireless Communications - page 301

channel with timing and frequency mismatch. Based on this model, we proposed a new
adaptive multicarrier interference subspace rejection receiver. We incorporated the least
complex and more practical ISR interference rejection mode to simultaneously suppress
MAI, ISI, and ICI at the signal combining step. We also proposed a realistic implemen-
tation of the new MC-ISR receiver that includes an efficient strategy for carrier offset
recovery in a multicarrier and multiuser detection scheme. MC-ISR supports both the
MT-CDMA- and MC-DS-CDMA-air interfaces. Furthermore, the assessment of the
new MC-ISR receiver was oriented toward an implementation in a future, real-world
wireless system. Indeed, we analyzed the performance of MC-ISR in an unknown time-
varying Rayleigh channel with multipath, carrier offset, and cross-correlation between
subcarrier channels and took into account all channel estimation errors. As another con-
tribution in this work, we derived a link/system-level performance analysis of MC-ISR
based on the Gaussian assumption (GA) and validated it by simulations. Under realistic
propagation conditions and in the presence of channel estimation errors, simulation
results validated the performance analysis and confirmed the net advantage of the full
interference suppression capabilities of MC-ISR. With two receiving antennas and nine
MT-CDMA subcarriers in 5 MHz bandwidth, MC-ISR provides about 4,320 kbps at low
mobility for DBPSK, i.e., an increase of 115% in throughput over current 3G DS-CDMA
with MRC.

Appendix

Derivation of the Interference Variance after MC-ISR Combining

Our goal is to estimate the variances:



∑ ≠

C

∑

K

∑ n

n + 1

H

    = ˆ

u

d




Var δ

d

+ δ

d

+ δ

d

Var ξ

u

λ

u

WY

MAIkn ,, ICIkn ,, ISIkn ,, k ′,, ′ ′ kn , kn , nkn ′′ ,,

u
ud

1



kK ′=− nn ′= − 1

∑

K

∑

n + 1 H

Y nkn

d

W

ˆ

d d

+ ξλ ( 1 0 . 5 3 )

kn ′′ ′ , kn , kn , ′′ ,,

kK
kk

′=−
′≠

nn ′= − 1



∑ ξλ

n + 1

H

d

d

ˆ

d d



+ WY

 .

kn , ′ kn , kn , nkn ′ ,,

nn
nn

′= −
′≠

1



Let us consider the general problem of deriving the variance of the sum of random
complex variables. We first introduce the variables x α , α ∈{1, …, N t } and ξ α , α ∈{1, …, N t },
with the following properties: E[ξ α ξ ∗ α′ ] = M ξ , ∀α ≠ α′, E[ξ α ξ ∗ α ] = V ξ , E[ x α ] = 0, and
Va r[ Σ N α=1 x α ] = 0. hen we assume that ξ α and x α are independent. Thus, we derive the
variance as follows: