What this result shows us is that the ML antenna-combining technique discussed herein is equivalent
to the adaptive antenna array using the MSINR cost function-based weights, .
This tells us that the MLE has interference suppression capabilities provided the metric value is correctly
computed. In other words, the interference statistics are included in the metric calculation.
wMSINR R1
In # h
??“ | |2
| |2
| |2
+ VA ??“
??“
.
.
. .
.
.
.
.
. .
.
.
r1(k)
M(k)
X
X
X
sh(k)
rj(k)
rD(k)
h1(k) ?†
hj(k) ?†
hD(k) ?†
Let us now discuss methods to include interference statistics in the ML metric calculation. In the
method presented above the interference is considered in the interference plus noise covariance
matrix. This is usually obtained by subtracting the estimated desired signal from the received signal.
Hence the accuracy of the RIn matrix depends on how accurate the desired signal can be estimated.
Alternatively, the RIn matrix inversion can be avoided if the M interfering signals with I multipaths
are estimated and also removed from the received signal. The very significant drawback of this second
approach is that the number of states in the VA increases and the metric now also becomes dependent
on the accuracy of estimating the interference.
Pages:
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524