Channel quantization for multiuser diversity
A method and system for channel quantization for multiuser diversity are provided. A quantizer may utilize quantization levels determined based on an objective function J that corresponds to a performance metric for a finite-rate feedback multiuser downlink communication system. The quantization levels may be obtained based on the results of maximizing the objective function J by means of an optimization algorithm. The optimization algorithm may be a multivariate optimization algorithm that may be implemented utilizing an iterative approach. The quantizer may utilize the quantization levels to quantize channel state information (CSI) to achieve multiuser diversity in the finite-rate feedback multiuser system. The performance metric may be based on signal-to-noise ratio (SNR), bit error rate (BER), and/or system capacity. The quantization levels may be utilized in a multiuser system where distributed quantization is needed for optimal selection.
1. A method for processing signals in a communication system, the method comprising:
maximizing a function J corresponding to a metric in a multiuser system with finite-rate feedback, wherein said function J generates a sum over a range n=1 to N of a product of μ n and p n , where n corresponds to a current quantization level, μ n is a conditional mean corresponding to said metric and p n is a probability that one of a plurality of receivers in said multiuser system comprises a signal-to-noise ratio (SNR) that is in an n th quantization interval [x n−1 , x n ], where x n−1 is a first partitioning threshold for said n th quantization interval, and x n−1 is a second partitioning threshold for said n th quantization interval; and
determining quantization levels to be utilized by a quantizer in said multiuser system with finite-rate feedback based on said maximized function J.
2. The method according to claim 1 , wherein said probability p n is defined by the following expression:
p
n
=
(
F
γ
k
K
(
x
n
)
-
F
γ
k
K
(
x
n
-
1
)
)
where K is a total number of said plurality of receivers, k is an index that indicates a current of said plurality of receivers, γ k is a signal-to-noise ratio (SNR) corresponding to each of said plurality of receivers, and F γ k K (x) is a cumulative distribution function of the largest of said SNR corresponding to each of said plurality of receivers.
3. The method according to claim 1 , comprising maximizing said function J based on an algorithm that converges to a specified threshold.
4. The method according to claim 3 , wherein said algorithm is a multivariate optimization algorithm.
5. The method according to claim 3 , wherein said algorithm is an iterative algorithm.
6. The method according to claim 1 , wherein said metric is a signal-to-noise ratio (SNR).
7. The method according to claim 1 , wherein said metric is a bit-error-rate (BER).
8. The method according to claim 1 , wherein said metric is a system capacity.
9. The method according to claim 1 , comprising storing said determined quantization levels in said plurality of receivers in said multiuser system.
10. The method according to claim 9 , further comprising dynamically selecting at least a portion of said stored determined quantization levels in said plurality of receivers based on said metric.
11. A system for processing signals in a communication system, the system comprising:
a multiuser system with finite-rate feedback comprising a plurality of receivers;
each of said plurality of receivers comprises a quantizer; and
said quantizer is operable to determine quantization levels based on maximizing a function J corresponding to a metric in said multiuser system with finite-rate feedback, wherein said function J generates a sum over a range n=1 to N of a product of μ n and p n , where n corresponds to a current quantization level, μ n is a conditional mean corresponding to said metric and p n is a probability that one of the plurality of receivers in said multiuser system comprises a signal-to-noise ratio (SNR) that is in an n th quantization interval [x n−1 , x n ], where x n−1 is a first partitioning threshold for said n th quantization interval, and x n is a second partitioning threshold for said n th quantization interval.
12. The system according to claim 11 , wherein said probability p n is defined by the following expression:
p
n
=
(
F
γ
k
K
(
x
n
)
-
F
γ
k
K
(
x
n
-
1
)
)
where K is a total number of said plurality of receivers, k is an index that indicates a current of said plurality of receivers, γ k is a signal-to-noise ratio (SNR) corresponding to each of said plurality of receivers, and
F
γ
k
K
(
x
)
is a cumulative distribution function of the largest of said SNR corresponding to each of said plurality of receivers.
13. The system according to claim 11 , wherein said function J is maximized based on an algorithm that converges to a specified threshold.
14. The system according to claim 13 , wherein said algorithm is a multivariate optimization algorithm.
15. The system according to claim 13 , wherein said algorithm is an iterative algorithm.
16. The system according to claim 11 , wherein said metric is a signal-to-noise ratio (SNR).
17. The system according to claim 11 , wherein said metric is a bit-error-rate (BER).
18. The system according to claim 11 , wherein said metric is a system capacity.
19. The system according to claim 11 , wherein each of said plurality of receivers comprises a memory that stores said determined quantization levels.
20. The system according to claim 19 , wherein each of said plurality of receivers comprises a processor that dynamically selects at least a portion of said stored determined quantization levels.