Pulse or digital communications – Systems using alternating or pulsating current – Plural channels for transmission of a single pulse train
Reexamination Certificate
1998-12-31
2002-07-16
Pham, Chi (Department: 2631)
Pulse or digital communications
Systems using alternating or pulsating current
Plural channels for transmission of a single pulse train
C375S341000
Reexamination Certificate
active
06421394
ABSTRACT:
BACKGROUND OF THE INVENTION
A modem is an electronic device that incorporates both a modulator and a demodulator into a single piece of signal conversion equipment. Interfacing directly to a communication channel, modems establish communication links between various computer systems and terminal equipment. The International Telegraph and Telephone Consultative Committee (CCITT), which determines protocols and standards for telephone and telegraph equipment, has authored a number of recommendations describing modem operation. See CCITT Data Communication Over the Telephone Network Series V Recommendations, BLUE BOOK Volume VIII-Fascicle VIII.1 Melbourne, 1988. In most cases, the communications channel is the general switched telephone network (GSTN) or a two-or four-wire leased circuit. Originally, these channels were assigned for voiceband transmission, so they are bandlimited from 300 hz to 3400 hz. The CCITT V.32. bis recommendation specifies the symbol rate of 2400 bauds per second. See CCITT Recommendation V32.bis Geneva, 1991. To achieve data transmission rates of 7200, 9600, 12000, and 14400 bits per second, three, four, five or six bits must be transmitted in one symbol interval. By adding a single bit to a binary symbol with K bits, the number of waveforms to be produced by the modulator is increased from 2
k
+2
k+1
. To achieve the same value of the probability of error, an increase in alphabet size within the same bandwidth requires a 3 dB increase in the signal to noise ratio (SNR).
Traditionally, a modem was implemented using analog discrete components. Today, digital circuits centered around one or two high performance digital signal processors (DSPs) can meet the demands of modem methods without the difficulties associated with analog circuitry. A digital modem implementation offers programmability, realizability of sophisticated methods, temperature insensitivity, ease of design, and often reduced cost when compared with analog implementations.
Modem communication technology has advanced to enable an increase in data communication rates. In order to achieve higher communication rates, communication bandwidth usually expands. But, because phone lines are bandlimited, new techniques for effective data communication rate improvements have emerged.
Another of the CCITT telecommunication standards is the V.34 modem 33.6 kbps. See ITU Recommendation V.34, “A modem operating at data signalling rates of up to 33,600 bit/s for use on the general switched telephone network and on leased point-to-point 2-wire telephone-type circuits”, Geneva, Oct. 9-18, 1996. Such variable data rates are made possible by using advanced signal processing techniques such as channel coding using 4D convolutional codes (see L. F. Wei, “Trellis-Coded Modulation with Multi-dimensional Constellations,” IEEE Transactions on Information Theory, vol. IT-33, no. 4, July 1987.), shell mapping (see R. Laroia, N. Farvardin, S. A. Tretter, “On Optimal Shaping of Multidimensional Constellations”, IEEE Transactions on Information Theory, vol. IT-40, no. 4, July 1994.), preceding (see R. Laroia, “Coding for Intersymbol Interference Channels—Combined Coding and Precoding,” IEEE Transactions on Information Theory, vol. IT-4, no. 4, July 1996.), nonlinear encoding (see W. L. Betts, E. S. Zuranski, “Method and Apparatus for Adaptively Providing Precoding and Preemphasis conditioning to Signal Data for Transfer over a Communication Channel,” U.S. Pat. No. 5,396,519, March 1995.) and pre-emphasis filtering at the transmitter (see W. L. Betts, prior).
The 4D convolutional codes yield a coding gain above 4.2 db, and an associated Viterbi subset decoder has reduced complexity due to the lower number of state transitions between the states. Shell mapping is a technique for non-equiprobably signaling that reduces the transmitted signal power and thereby yields a coding gain of 0.7-0.9 dB. The precoder implements decision feedback equalization at the transmitter in a way that creates a valid signal for the Viterbi subset decoder at the receiver. Finally, an intelligent choice of the pre-emphasis filter reduces the transmitted signal power at the low frequencies and hence, the level of nonlinear harmonics in the network echo is reduced for a higher level of cancellation.
In the receiver, after demodulation, the baseband signal is subjected to timing recovery, carrier recovery, fractionally-spaced equalization, followed by the Viterbi decoder which decodes the 4D codes of Wei. The Viterbi subset decoder is the most computationally expensive block in the receiver, and the complexity depends on the 16, 32, or 64 state codes being used.
The Viterbi subset decoder (and the equalizer in the self-learning mode) require the knowledge of the closest hard point on a constellation. When precoding is enabled there is constellation expansion, and the point slicing method, which yields the closest hard point on the constellation, assumes that the hard points lie on the infinite lattice. See M. V. Eyuboglu, “Flexible Precoding for V. FAST,” Int. Conf. on Data Transmission—Advances in Modem and ISDN Technology and Applications, 1992, Vol. 356, Ch.26, No.356, pp.13-18, and also G. D. Forney, Jr., “Advances in Modem Technology since V.32/ V.32bis,” Int. Conf. on Data Transmission —Advances in Modem and ISDN Technology and Applications, 1992, Vol.356, Ch.26, No.356, pp.1-6. A common method for determining the hard points is the Infinite Point Slicing (IPS) method which has a simple implementation. When the constellation expansion is high compared to the constellation size, the distance between the constellation points is effectively reduced in order to maintain the constant transmitted signal power. Therefore, the precoding is disabled in this case, which can result in inter-symbol interference (ISI) caused by severely distorted channels. Consequently, the point slicing method must yield hard points from the particular finite size constellation used in the transmitter of the remote modem and be capable of distinguishing the symbols from one another to yield the correct hard point from among the hard points in the finite size constellation.
Presently, there are no computationally efficient methods for yielding hard points in a 4-quadrant finite constellation that is generalized for all constellation configurations in the V.34 modem standard. The IPS is a computationally efficient method, however, it generates erroneous hard points when it is necessary to consider the constellation boundary. The necessity for a method that does not generate erroneous hard points when it is necessary to consider the constellation boundary is stated in Wei.
In at least one previous work, Martinez, K., Mack, G., “Viterbi Decoder for Wireline Modems,” U.S. Pat. No. 4,709,377, (1985), a point slicing method was presented. In that work, the QAM constellation that was assumed is not general in the sense that it has a rectangular shape in the first quadrant. The received soft points are therefore translated into the first quadrant, the nearest hard point is found in the first quadrant, then the hard point is reflected into the quadrant of its corresponding soft point. Due to the rectangular nature of the constellation boundary, the bounds of the perimeter in the first quadrant are straight lines, and the soft points outside this perimeter are shifted inside the perimeter by truncating the real or imaginary parts.
SUMMARY OF THE INVENTION
The present invention addresses and solves the problems of selecting erroneous hard points found in the prior art by determining a valid hard point constellation and projecting received external soft points into the constellation boundary to locate the closest valid hard point. The present invention provides a point slicing method for digital transmission based on QAM (Quadrature Amplitude Modulation) constellations. Upon receiving noisy constellation points at the equalizer output, such as shown in
FIG. 3
for the V.34 modem, the proposed method obtains the closest valid constellation hard point in each subset. The com
Pham Chi
Tran Khai
LandOfFree
Computer method and apparatus for modem point slicing does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Computer method and apparatus for modem point slicing, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Computer method and apparatus for modem point slicing will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2886493