IP Library Granted Patent US 8,023,581
Granted Patent B2
US 8,023,581 · App. 11/857,269 · Granted Sep 20, 2011

Max-log receiver for multiple-input multiple-output (MIMO) systems

View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 8,023,581
App. No.
11/857,269
Granted
Sep 20, 2011
Kind
B2
Abstract

A method includes the steps of i) listing out all possibilities for a first symbol of a two stream signal; ii) determining a second symbol of the two stream signal for each of the first symbol listed out, iii) evaluating a metric for each of the first symbol and second symbol pair, iv) listing out all possibilities for second symbol, v) determining a first symbol for each choice of the second symbol listed out, vi) evaluating a metric for each of the second symbol and first symbol pair, vii) determining an exact maximum log likelihood ratio for all bits using the metrics, and viii) decoding codeword(s) in the two stream signal using the determined exact maximum log likelihood ratio for all bits.

Claims (22)

1. A method comprising the steps of:

i) listing out all possibilities for a first symbol of a two stream signal;

ii) determining a second symbol of the two stream signal for each said first symbol listed out,

iii) evaluating a metric for each said first symbol and second symbol pair,

iv) listing out all possibilities for second symbol,

v) determining a first symbol for each choice of said second symbol listed out,

vi) evaluating a metric for each said second symbol and first symbol pair,

vii) determining an exact max-log likelihood ratio for each bit using the metrics, and

viii) decoding codeword(s) in the two stream signal using the determined exact max-log likelihood ratios for all bits.

2. The method of claim 1 , wherein said first symbol is designated symbol- 1 and said second symbol is designated symbol- 2 .

3. The method of claim 2 , wherein maximum max-log likelihood ratios for 2 Log(M) bits in said two stream signal is obtained with O(M) complexity when symbol- 1 and symbol- 2 belong to a constellation of cardinality M, said complexity related to evaluating the metrics of 2M hypotheses O(M) denoting that the said complexity scales linearly in M.

4. The method of claim 3 , wherein further reduction in said complexity can be achieved by avoiding the redundant computation in the two modified QR decompositions_H=∥h 2 ∥ 2 UL,H=∥h 1 ∥ 2 VR, where the matrix H represents the channel with h 1 and h 2 denoting its first and second columns, respectively, U and V are scaled semi-unitary matrices and L, R are square lower and upper triangular matrices, respectively.

5. The method of claim 2 wherein the metric for x 1,j which is the j th possibility for symbol- 1 is defined by Q(x i,j )=|z 1 −L 11 x 1,j | 2 +min x 2 |z 2 −L 21 x 1,j −x 2 | 2 wherein z 1 and z 2 are obtained by linearly transforming the vector of received observations, L 11 and L 21 are elements of a lower triangular matrix obtained using a modified QR decomposition; defining q 1j =z 2 −L 21 x 1,j and expressing Q(x 1,j ) as Q(x 1,j )=|z 1 −L 11 x 1,j | 2 +min x 2 R |q 1j R −x 2 R | 2 +min x 2 1 |q 1j I −x 2 I | 2 , with q 1j R ,q 1j I denoting the real and imaginary parts of q 1j , respectively, and both x 2 R ,x 2 I belonging to a common real-valued constellation, the two minimizations for computing Q(x 1,j ) being done in parallel using simple slicing (rounding) operations each with a complexity independent of the cardinality of the said real-valued constellation.

6. The method of claim 1 , wherein said metric for x 2,j which is the j th possibility for symbol- 2 is based on Q(x 2,j )=|w 2 −R 22 x 2,j | 2 +min x 1 |W 1 −R 12 x 2,j −x 1 | 2 ; wherein w 1 and w 2 are obtained by linearly transforming the vector of received observations, R 12 and R 22 are elements of an upper triangular matrix obtained using a modified QR decomposition; defining q 2j =w 1 −R 12 x 2,j and expressing Q(x 2,j ) as Q(x 2,j )=|w 2 −R 22 x 2,j | 2 +min x 1 R |q 2j R −x 1 R |+min x 1 I |q 2j I −x 1 I | 2 , with q 2j R ,q 2j I denoting the real and imaginary parts of q 2j , respectively and both x 1 R , x 1 I belonging to a real-valued constellation, the two minimizations for computing Q(x 2,j ) being done in parallel using simple slicing operations, thereby efficiently determining all {Q(x 2,j )} j=1 M .

7. A method comprising the steps of:

i) using the method in claim 1 to decode the two codewords associated with the two streams

ii) conducting a redundancy check on the two decoded codewords

iii) In case the cyclic redundancy check of only one codeword is true, re-encoding, modulating and subtracting that codeword from the received signal to obtain a single stream signal,

iv) listing out all possibilities for the remaining symbol in the single stream signal,

v) evaluating a metric for each possibility of the remaining symbol,

vi) determining the max-log likelihood ratios for all bits using the metrics,

vii) decoding the remaining codeword in the single stream signal using the determined max-log likelihood ratios for all bits.

Assignments (2)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Feb 27, 2012
From: NEC LABORATORIES AMERICA, INC.
To: NEC CORPORATION
Reel/Frame 027767/0918 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Nov 8, 2007
From: PRASAD, NARAYAN; WANG, XIAODONG; MADIHIAN, MOHAMMAD
To: NEC LABORATORIES AMERICA, INC.
Reel/Frame 020086/0005 →