Pulse or digital communications – Receivers – Angle modulation
Reexamination Certificate
1999-02-04
2002-07-16
Chin, Stephen (Department: 2734)
Pulse or digital communications
Receivers
Angle modulation
C375S341000, C375S265000, C714S792000, C714S795000
Reexamination Certificate
active
06421400
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to communication systems that employ Phase Shift Keying (PSK) modulation. More particularly, the present invention relates to the use of improved techniques for quantization and distance metric determination in a demodulator for PSK-modulated communication system.
2. Description of the Related Art
M-phase PSK modulation is a technique commonly applied in communication systems.
FIG. 1
shows a typical model for a communication channel employing PSK modulation. The channel includes an encoder
102
, a PSK modulator
104
, a PSK demodulator
108
, and a soft-decision decoder
110
. At the transmitter end, encoder
102
encodes user data U
j
to provide protection against errors. The encoded data V
i
is modulated onto a carrier signal using phase modulation. In other words, the data symbol V
i
is coded as one of M phases (relative to a reference phase of a constant frequency signal), so that during time period i, the carrier signal has a phase angle ∠X
i
which corresponds to the data symbol V
i
. As the phase-modulated carrier signal travels from the transmitter to the receiver, it experiences various forms of interference. In the model of
FIG. 1
, the interference n
i
is modeled as an additive noise source
106
. The corrupted phase-modulated carrier signal Y
i
is received by demodulator
108
which determines a quantized signal Y
Qi
that most closely corresponds to the signal Y
i
. The quantized signal Y
Qi
is recovered through the use of a reference phase generator by comparing received symbols with the reference phase. Soft-decision decoder
110
then decodes the quantized encoded data signal Y
Qi
to determine an error-corrected estimate Û
j
of the original user data.
As part of the decoding process, the soft-decision decoder
110
calculates a set of distance metrics for each quantized encoded data signal Y
Qi
. A distance metric is a measurement of how “close” the quantized data signal Y
Qi
is to valid constellation signal points. Common distance metrics are discussed further below.
When encoder
102
is a trellis encoder, soft-decision decoder
110
is typically a Viterbi decoder. A Viterbi decoder tracks the most likely receive signal sequences through multiple trellis stages, and compiles sums of distance metrics to determine path weights indicating the relative likelihood of the most likely receive signal sequences. After a predetermined number of trellis stages, the Viterbi decoder will begin rendering decoding decisions. The decoding decisions Û
j
correspond to an allowable sequence of encoded symbols V
i
which are “closest to” the quantized signal sequence Y
Qi
. Accordingly, the quantization method and the distance measurement method both affect the performance of trellis code modulated (TCM) systems. Details on maximum-likelihood decoding and the Viterbi algorithm can be found in many standard textbooks, including J. G. Proakis,
Digital Communications
: 2
ed
, McGraw-Hill Book Company, New York, N.Y., (c) 1989, pp. 454-459 and 610-616.
As shown in
FIG. 2
, demodulator
108
typically mixes the received signal Y
i
with a local carrier to produce In-phase (“I”) and Quadrature-phase (“Q”) baseband signals. The I and Q signals are used for phase comparison with the local carrier to identify encoded data symbols in the M-signal constellation.
FIG. 2
shows the demodulator
108
and the decoder
110
at the receiver end. The phase comparison is typically done by first quantizing the I and Q coordinates into one of a number of discrete levels. Quantizer
202
maps the received I and Q signals to one of these discrete quantization levels. The quantized I and Q signals, I
Q
and Q
Q
, respectively, are then compared by the decoder
110
with the I and Q signals corresponding to the data symbols in the M-signal constellation.
Examples of PSK constellations are shown in FIG.
3
and FIG.
4
. The constellation signal points each represent one encoded data symbol. The phase angle of the constellation signal points is measured in degrees from the in-phase (I) axis.
FIG. 3
shows a 4-signal PSK constellation, and
FIG. 4
shows an 8-signal PSK constellation. These constellations provide 4 phases and 8 phases, respectively, for conveying data symbols from the transmitter to the receiver.
Some VLSI implementations of quantizers in M-PSK demodulator use Cartesian coordinates as a means of quantizing I and Q signals. The Cartesian quantization method is equivalent to dividing the signal constellation space into a grid pattern.
FIG. 5
shows a 15-level quantization of the 4-signal PSK constellation. The received I signal is quantized to one of 15 amplitude levels numbered from 0 to 14, and the received Q signal is similarly quantized. The quantization levels are spaced such that the 4 PSK signal points are at the corners of the grid.
FIG. 6
shows a 15-level quantization of the 8-signal PSK constellation. Again, the I and Q signals are each quantized to one of 15 amplitude levels numbered from 0 to 14. Here, the quantization levels are spaced so that the four signals on the I and Q axes are at the edges of the grid. After the I and Q signals have been mapped to one of the squares in the quantization grid, distance calculations to the signal points may be performed and used by the decoder to determine a most likely symbol sequence.
Metrics commonly used to calculate distance between two signal points in an M-PSK modulated system are squared Euclidean distance and absolute distance.
FIG. 7
shows a received signal point quantized at (11,13). The absolute distance between (11,13) and the nearest data symbol (14,14) is:
|11−14|+|13−14|=4
whereas the squared Euclidean distance is:
(11−14)
2
+(13−14)
2
=10.
In general, the absolute distance calculation is simpler to implement than the squared Euclidean distance calculation for digital implementations in VLSI. However, using the absolute distance in an 8-signal PSK constellation increases the error rate of the decoder.
FIG. 8
shows the bit error rate (BER) as a function of the signal-to-noise ratio (SNR) for an 8-signal PSK TCM system. The system which uses absolute distance metric has a higher BER than the system which uses the Euclidean distance metric. The difference in performance between the two metric methods can be measured in terms of the change in SNR necessary to produce an equivalent performance. From
FIG. 8
it is observed that to achieve the same BER, a system using the absolute distance metric requires a SNR that is at least 0.25 dB higher than the system using the squared Euclidean distance metric. In other words, for an 8-signal PSK constellation, use of the absolute distance metric translates into a 0.25 dB loss in SNR relative to use of the Euclidean distance metric in 15-level Cartesian coordinate quantization.
While the Euclidean distance metric may provide a lower Bit Error Rate, the mathematical computations associated with the calculation of the Euclidean distance make this metric expensive to implement for VLSI applications. Thus, what is needed is an apparatus and method for performing quantization and using computationally efficient distance metrics without incurring significant performance losses.
SUMMARY OF THE INVENTION
Generally speaking, the present invention fills these needs by providing improved quantization methods and improved distance metrics for a PSK demodulator/decoder. It should be appreciated that the present invention can be implemented in numerous ways, including implementation as a process, an apparatus, a system, a device, or a method. Several inventive embodiments of the present invention are described below.
In one embodiment, the present invention provides a digital communications receiver which includes a PSK demodulator and a soft-decision decoder. The PSK demodulator is configured to accept a receive signal and responsively produce quantized baseband signal components which include a include a quant
Mogre Advait
Rhee Do-jun
Chin Stephen
Conley Rose & Tayon
Kim Heechul
LSI Logic Corporation
LandOfFree
System and method using polar coordinate representation for... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and method using polar coordinate representation for..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method using polar coordinate representation for... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2834176