Pulse or digital communications – Transmitters
Reexamination Certificate
2000-03-14
2003-05-20
Vo, Don N. (Department: 2631)
Pulse or digital communications
Transmitters
C332S103000, C375S298000
Reexamination Certificate
active
06567477
ABSTRACT:
BACKGROUND
The present invention relates generally to communication apparatus, and more particularly, to methods that provide for constellation pruning and partial Gray coding of the pruned constellation that may be used with a transmitter hardware architecture that implements two-dimensional, uncoded signaling for equiprobable signals suitable for multiple gigabit/second satellite communications.
At gigabit/second and higher data rates, hardware limitations can prevent implementation of relatively complex satellite communication schemes with features such as coded modulation, shaping codes, multi-dimensional constellations, and non-equiprobable signaling. The present invention addresses simpler and more easily realized communication scheme using two-dimensional, uncoded modulation with equiprobable signals.
Previous work on two-dimensional constellation design has generally not associated the constellation with a particular hardware architecture. Such work has therefore tended to emphasize power measures related to the transmitter output (RF) power, such as the ratio at the receiver of the bit energy to the noise power density, E
b
/N
0
. However, the transmitter prime power is a more useful criterion because it directly corresponds to the energy drain on the satellite's batteries. Unless the constellation design is assumed fixed, the transmitter prime power and the output power are not linearly related because the efficiency with which the transmitter converts the former to the latter depends on the particular constellation employed.
Constellation design methods are discussed in articles by: R. W. Lucky et al. entitled “On the optimum performance of N-ary systems having two degrees of freedom”, IRE Transactions on Communications Systems, 10:185-192. June 1962; S. Moskowitz, entitled “Signal design for coherent M-ary communication systems using stochastic gradients. In Proceedings of the 5th Hawaii International Conference on System Science, pages 171-173, Honolulu. Hi., January 1972; G. J. Foschini, et al. entitled Optimization of two-dimensional signal constellations in the presence of Gaussian noise”, IREE Trans. on Communications COM-22:28-38. January 1974; and S. Emami et al. entitled “Signal selection for non-symmetric sources in two dimensions”, in Southeastcon, pages 461-464, Birmingham, Ala. April 1992.
The constellation design methods discussed in articles attempt to minimize the probability of symbol error while constraining the average or peak value of E
b
/N
0
, usually by following an approximate gradient descent and renormalizing the symbol locations after each gradient update step in order to satisfy the power constraint. These methods are not applicable in the current context, because the hardware architecture restricts the symbols to a set of discrete locations that cannot be continuously adjusted, but only allowed or disallowed. In an article by J. Salz, et al. entitled “Data transmission by combined AM and PM”, in The Bell System Technical Journal, 50(7):7399-2419, September 1971, the symbol space was assumed to be quantized to N
a
equally spaced amplitude levels and N
p
equally spaced phases per amplitude level. The corresponding constellations, dense near the origin and increasingly sparse away from it, radically differ from those discussed herein. Furthermore, the topic of pruning was not discussed in the Salz, et al. article, as K was always assumed to equal N
a
N
p
, where K is the number of constellation symbols.
Coding for certain non-rectangular constellations was examined in an article by T. Yamazato, et al. entitled “An arrangement technique of Gray-code table for signal constellation of modified QAM and triangular-shaped signal set”, in IEICE Transactions, E 74(9): 2579-2585, September 1991. In that work, a one-to-one correspondence between the symbols of a rectangular, Gray-coded constellation and a non-rectangular constellation was established by “manually” warping and transferring the symbols to lie on top of each other. Although this unsystematic technique may sometimes provide a good initial coding that can then be improved by the “swapping” coding algorithm developed herein, the technique seems ill-suited to irregularly shaped constellations.
In the above cited Foschini, et al. article, it was shown that under a constraint on the average value of E
b
/N
0
, the hexagonal or “honeycomb” constellation is asymptotically optimal as the number of symbols goes to infinity. In a hexagonal constellation, the symbols are located at the centers of densely packed hexagons. For various finite constellation sizes, the superiority of the hexagonal constellation over other standard constellations (e.g., circular, cross, octagonal, pentagonal, rectangular, and triangular constellations) has also been demonstrated as discussed in articles by: K. Kawai, et al., entitled “Optimum combination of amplitude and phase modulation scheme and its application to data transmission modem”, in Proceedings of IEEE International conference on communications, pages 29.6-29.11, Phil., Pa., June 1972; M. K. Simon et al. entitled “Hexagonal multiple phase-and-amplitude-shift-keyed signal sets”, in IEEE Trans. on Communications, COM-71:1108-1115, October 1973; and G. D. Forney, Jr., et al., entitled “Efficient modulation for band-limited channels. IEEE Trans. on Selected Areas in Communications, 2(5):632-647, September 1984.
Accordingly, it is an objective of the present invention to provide for constellation pruning and partial Gray coding methods for use with a transmitter hardware architecture that implements multiple gigabit/second satellite communications.
SUMMARY OF THE INVENTION
To meet the above and other objectives, the present invention provides for improved constellation pruning and partial Gray coding methods. Such methods may be employed with a transmitter hardware architecture that implements multiple gigabit/second satellite communications. The transmitter hardware, which implements two-dimensional, uncoded signaling for equiprobable signals, comprises an array of interchangeable transmitter stages. Each stage of the transmitter includes a constant-signal phase shift keying (PSK) modulator and can generate multiple phase states. By way of example, stages with four or six states produce rectangular QAM or hexagonal constellations of symbols, respectively.
A more complex variant of the transmitter design also implements on-off keying, wherein any stage may be keyed into an additional “off” state, providing increased power efficiency. In the off state, the supply current to the power amplifier of a stage is interrupted and the power consumed by the stage is essentially zero. When several stages destructively interfere with each other, they are keyed into their off states to save power without altering the output of the transmitter.
Using this transmitter architecture, hexagonal constellations, for example, are produced when the transmitter stages have six phase states, and QAM (quadrature amplitude modulation or rectangular) constellations when there are four phase states.
Designing the transmitter requires determining the number of phase states per stage, the number of stages, whether to enable on-off keying, and which symbols to “prune” from excessively large constellations. Two design optimization criteria used in implementing the present invention are the transmitter prime power and the bit error rate. Although the transmitter output power is a more conventional criterion, the transmitter prime power is preferable because it directly relates to the amount of power consumed by the transmitter.
For binary communications, the pruned constellation is coded by assigning binary sequences to its symbols. Computationally efficient strategies for pruning and coding, provided by the present invention, yield coded hexagonal constellations, for example, that outperform traditional Gray-coded QAM constellations. In some cases, judicious selection of the design parameters can reduce the transmitter prime power by over 3 dB.
Signals generated using the transmitter co
Chethik Frank
Feldman David D.
Float Kenneth W.
Lockheed Martin Corporation
Nguyen Dung X.
Vo Don N.
LandOfFree
Two-dimensional constellation generation, pruning, and... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Two-dimensional constellation generation, pruning, and..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Two-dimensional constellation generation, pruning, and... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3028224