Efficient optimal ML detector
View Patent ↗An efficient optimal maximum-likelihood output detector reducing the complexity of demodulation/decoding computations in multiple-input multiple-output communication systems. A plurality of received signals may be combined into a plurality of combined received signals by multiplication of a matrix representing the plurality of received signals with another matrix that meets certain conditions. The plurality of combined received signals may then allow for slicing operations as well as calculation of distance metrics with significantly reduced complexity.
1. A method of detecting multiple-input multiple-output (MiMO) transmissions, comprising:
performing by one or more processors and/or circuits integrated within a single chip:
receiving a plurality of signals;
combining the plurality of signals to produce a plurality of combined signals, the plurality of combined signals including: a first spatial stream, representing a series of first transmitted constellation points, each transmitted constellation point being a member of a first set of candidate constellation points, and each candidate constellation point representing a contribution to the plurality of combined signals, and a second spatial stream, representing a series of second transmitted constellation points, each second transmitted constellation point being a member of a second set of candidate constellation points;
and for each candidate constellation point of the first set of candidate constellation points: subtracting the contribution to the plurality of combined signals, and calculating a distance metric using a slicer to determine a closest candidate constellation point among the second set of candidate constellation points.
2. The method of claim 1 , wherein the plurality of signals are combined by multiplying the plurality of signals with a matrix W, wherein the product of a Hermitian transposition of the matrix W and the matrix W equals a diagonal or nearly diagonal matrix.
3. The method of claim 2 , wherein the product of the matrix W and a channel matrix H is triangular or nearly triangular.
4. The method of claim 1 , wherein each of the candidate constellation points of the first set of candidate constellation points represents bits, comprising calculating a soft output for each of the bits using the distance metrics.
5. The method of claim 4 , comprising: for each bit B
represented in candidate constellation points of the first set of candidate constellation points:
finding a first minimum distance metric among the distance metrics, wherein a value of bit B of the candidate constellation point associated with the first minimum distance metric is zero;
finding a second minimum distance metric among the distance metrics wherein a value of bit B of the candidate constellation point associated with the second minimum distance metric is one; and
producing the soft output by calculating a difference between the first minimum distance metric and the second minimum distance metric.
6. The method of claim 1 , comprising finding one candidate constellation point among the first set of candidate constellation points, wherein the distance metric calculated for the one candidate constellation point is a minimum distance metric among the distance metrics calculated for the first set of candidate constellation points, to produce a hard output.
7. The method of claim 3 , wherein each of the candidate constellation points of the first set of candidate constellation points represents bits, comprising:
for each bit B represented in candidate constellation points of the first set of candidate constellation points:
finding a first minimum distance metric among the distance metrics, wherein a value of bit B of the candidate constellation point associated with the first minimum distance metric is zero;
finding a second minimum distance metric among the distance metrics, wherein a value of bit B of the candidate constellation point associated with the second minimum distance metric is one; and
producing a soft output by calculating a difference between the first minimum distance metric and the second minimum distance metric.
8. A method of detecting multiple-input multiple-output transmissions, comprising:
performing by one or more processors and/or circuits integrated within a single chip:
receiving a plurality of signals;
combining the plurality of signals to produce a plurality of combined signals, the plurality of combined signals being usable to calculate distance metrics using a slicing operation;
calculating a plurality of distance metrics using the plurality of combined signals; and
calculating log-likelihood ratios for a plurality of bits represented in one or more transmitted symbols using the plurality of distance metrics, the transmitted symbols having been transmitted using the plurality of signals.
9. A detector, configured to detect multiple-input multiple-output transmissions, comprising:
a sigal receiver, configured to receive a plurality of signals;
a signal combiner, configured to combine the plurality of signals to produce a plurality of combined signals, the plurality of combined signals being usable to calculate distance metrics using a slicing operation;
a distance metrics calculator, configured to calculate a plurality of distance metrics using the plurality of combined signals; and
a log-likelihood calculator, configured to calculate log-likelihood ratios for a plurality of bits represented in one or more transmitted symbols using the plurality of distance metrics, the transmitted symbols having been transmitted using the plurality of signals.