Method for slicing K-best detection in multiple-input multiple-output wireless communications system
AssigneeGeorgia Tech Research Corporation
InventorsSangwook Suh, John R. Barry
Life of this patent
Estimating the term…
Abstract
In a MIMO detector, a slicing list detection scheme employs a complex plane to represent symbols in an alphabet, in which all decision regions are bounded by either vertical lines or horizontal lines. A K-best scheme accesses the complex plane and an offline-generated lookup table to detect elements of a received vector. At each level, the system prunes all but the K-best candidates from each surviving node through the slicing list detector.
Claims18
A method of detecting a MIMO signal transmitted from N antennas, the MIMO signal representing a transmitted vector of N elements, each of the N elements being transmitted by a different one of the N antennas and each of the N elements including a selected symbol from an alphabet ofM symbols, the MIMO signal represented as a received vector of NR elements, in which each of the NR elements corresponds to a different receiver antenna of an array of receiver antennas, the method comprising the steps of:
- (a) defining a finite number of decision regions within a complex plane , each decision region defined to encompass a different combination of K symbols in the alphabet, in which each decision region is bounded by threshold lines consisting of horizontal threshold lines that are parallel to the real axis and vertical threshold lines that are parallel to the imaginary axis, and in which K is a selected number taken from a set of integers;
- (b) modifying the received vector to be represented as a complex point p having a real component and an imaginary component;
- (c) nulling a reconstructed transmitted vector and then performing a loop of steps including:
- (i) mapping the complex point p onto the complex plane;
- (ii) then comparing the real component of the complex point p to the vertical threshold lines and imaginary component of complex point p to the horizontal threshold lines to determine which decision region encompasses the complex point p;
- (iii) then accessing a lookup table, in which the lookup table has been compiled offline, to determine which K symbols associated with the decision region that encompasses the complex point p are closest to the complex point p and adding the closest K symbols to a candidate set of detected symbol vectors;
- (iv) then subtracting the detected symbol from the complex decision point and returning to the mapping step;
- and
- (v) then when N detected symbols have been detected in the reconstructed transmitted vector, exiting the loop;
- and
- (d) after exiting the loop, returning the reconstructed transmitted vector.
A detecting method for detecting a MIMO signal transmitted from N antennas, the MIMO signal representing a transmitted vector of N elements, each of the N elements being transmitted by a different one of the N antennas and each of the N elements including a selected symbol from an alphabet of M symbols, the MIMO signal represented as a received vector of NR elements, in which each of the NR elements corresponds to a different receiver antenna of an array of receiver antennas, the method comprising the steps of:
- (a) defining a finite number of decision regions within a complex plane, each decision region defined to encompass a different combination of K symbols in the alphabet, in which each decision region is bounded by threshold lines consisting of horizontal threshold lines that are parallel to the real axis and vertical threshold lines that are parallel to the imaginary axis, and in which K is a selected number taken from a set of integers, wherein the horizontal threshold lines are chosen so that each of the horizontal threshold lines intersects the imaginary axis at integers ranging from −√{square root over (M)} to √{square root over (M)} and wherein the vertical threshold lines are chosen so that each of the vertical threshold lines intersects the real axis at integers ranging from −√{square root over (M)} to √{square root over (M)};
- (b) wherein each symbol includes a real component and an imaginary component in which the real components and imaginary components consist of odd integers from −√{square root over (M)}+1 to √{square root over (M)}−1, wherein the step of defining a finite number of decision regions comprises the steps of dividing a complex plane into a total of 4 (√{square root over (M)}+1)2 sub-regions, in which 4M of the sub-regions are defined to be square in shape are grouped together in a center portion of the complex plane, while 4(1 +2√{square root over (M)}) of the sub-regions are placed at edges and the corners of the center portion and are of a semi-infinite rectangular in shape and selecting boundaries for each selected decision region so that each of the K symbols associated with the selected decision region is one of K best selections for a complex point p mapped into the selected decision region;
- (c) modifying the received vector to be represented as a complex point p having a real component and an imaginary component;
- (d) nulling a reconstructed transmitted vector and then performing a loop of steps including:
- (i) mapping the complex point p onto the complex plane;
- (ii) then comparing the real component of the complex point p to the vertical threshold lines and imaginary component of complex point p to the horizontal threshold lines to determine which decision region encompasses the complex point p;
- (iii) then accessing a lookup table, in which the lookup table has been compiled offline, to determine which K symbols associated with the decision region that encompasses the complex point p are closest to the complex point p and adding the closest K symbols to the reconstructed transmitted vector;
- (iv) then subtracting the detected symbol from the complex decision point and returning to the mapping step;
- and
- (v) then when N detected symbols have been detected in the reconstructed transmitted vector, exiting the loop;
- and
- (e) after exiting the loop, returning the reconstructed transmitted vector.
A receiver for receiving a MIMO signal transmitted from N antennas, the MIMO signal representing a transmitted vector of N elements, each of the N elements being transmitted by a different one of the N antennas and each of the N elements including a selected symbol from an alphabet of M symbols, the MIMO signal represented as a received vector of NR elements, in which each of the NR elements corresponds to a different receiver antenna of an array of receiver antennas, the receiver comprising:
- (a) an array of NR antennas;
- (b) a memory on which is stored a lookup table that has been compiled offline and that relates values of the received vector to symbols from the alphabet;
- (c) a signal processing circuit, responsive to the antennas and in data communication with the memory, that is programmed to execute steps, including:
- (i) define a finite number of decision regions within a complex plane, each decision region defined to encompass a different combination of K symbols in the alphabet, in which each decision region is bounded by threshold lines consisting of horizontal threshold lines that are parallel to the real axis and vertical threshold lines that are parallel to the imaginary axis, and in which K is a selected number taken from a set of integers;
- (ii) modify the received vector to be represented as a complex point p having a real component and an imaginary component;
- (iii) null a reconstructed transmitted vector and then execute a loop of steps including:
- (1) map the complex point p onto the complex plane;
- (2) then compare the real component of the complex point p to the vertical threshold lines and imaginary component of complex point p to the horizontal threshold lines to determine which decision region encompasses the complex point p;
- (3) then access the lookup table to determine which K symbols associated with the decision region that encompasses the complex point p are closest to the complex point p and adding the closest one of the K symbols to the reconstructed symbol vector;
- (4) then subtract the detected symbol from the complex decision point and returning to the mapping step;
- and
- (5) then when N detected symbols have been detected in the reconstructed transmitted vector, exit the loop;
- and
- (iv) after exiting the loop, return the reconstructed transmitted vector.
Dependencies are inferred from the claim text.