|Publication number||USRE38391 E1|
|Application number||US 09/772,290|
|Publication date||Jan 20, 2004|
|Filing date||Jan 29, 2001|
|Priority date||Dec 23, 1993|
|Publication number||09772290, 772290, US RE38391 E1, US RE38391E1, US-E1-RE38391, USRE38391 E1, USRE38391E1|
|Original Assignee||Stmicroelectronics S.A.|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (11), Non-Patent Citations (4), Classifications (11), Legal Events (1)|
|External Links: USPTO, USPTO Assignment, Espacenet|
This application is a continuation of application Ser. No. 09/516,565, filed Feb. 29, 2000, which in turn is a continuation of application Ser. No. 09/379,087, filed Aug. 23, 1999, which in turn is a reissue of U.S. Pat. No. 5,661,752, issued Aug. 26, 1997, entitled CIRCUIT FOR DETECTING WORD SEQUENCES IN A MODEM, filed as application Ser. No. 08/359,470 on Dec. 20, 1994, which applications are incorporated herein by reference.
1. Field of the Invention
The present invention relates to digital modems and more particularly to the detection of predetermined reference word sequences during the initial training procedure between two modems, or during intermediate phases of retraining or modification of the rate.
2. Discussion of the Related Art
In digital modems, the data to be transmitted are digitally encoded from a determined number of symbols and then transmitted as portions of sine wave signals whose phase and amplitude are modulated. Each symbol, for example 128 in accordance with the V32bis standard, corresponds to a signal having a predetermined phase and amplitude. To illustrate this conversion, the symbols can be disposed in a constellation such as illustrated in FIG. 1 where each symbol is disposed at a point corresponding to the phase and amplitude of the modulated signal representing the symbol. Thus, as represented in FIG. 2, each symbol is transmitted as a sine wave portion at a carrier frequency of, for example, 1800 Hz, during bit intervals corresponding to a baud rate of, for example, 2400 Hz. The sine wave is formed and decoded from a sampling frequency that is higher than the baud rate and is a multiple of the latter, for example 9600 Hz.
After decoding, each received Symbol is transformed into a digital word M including a first portion corresponding to the real value of the transmitted word and a second portion b corresponding to the imaginary value of this word. Thus, each word can be expressed by:
Due to the very high exchange rates between two modems, the quality of the telephone connection becomes increasingly important in order to provide reliable transfers of data between two modems. Possible impairments, such as attenuation of the high frequencies of the spectrum, near and far end echo with frequency offset, phase jitter and noise, influence the maximum bit rate possible for minimum errors. In the above V32bis standard, the demodulated complex signal received by a modem is coded on amplitude and phase and a constellation such as the one represented in FIG. 1 contains 128 possible values designed to be transmitted at a data rate of 14,400 bits/second at a baud rate of 2400 Hz. When the above mentioned impairments become more important, the difference between the received point and the ideal receive point, or receive error, becomes increasingly large until, eventually, the decision mechanism of the receiving modem confuses two or more adjacent points, thus causing intolerable errors in the reception.
In contrast, a connection with a slower data rate, such as 7,200 bits/second uses a constellation including only 16 values such as represented in FIG. 3 and permits a much larger tolerance to line impairments. These impairments may vary during the communication in which case the initial transmission rate has to be increased or decreased to find the optimal value. In general, a micro-controller is connected to the modem and can calculate the quality of the received signal, this value corresponding to the inverse of the average receive error. If this error becomes too important, the CCITT recommendations provide for a rate negotiation procedure which is quite short in duration to interrupt data transmission for a very short time of 288 bauds (120 ms) plus the round trip delay between the two modems.
The rate negotiation signal more particularly contains a preamble constituted by a sequence of predetermined signals, the sequence being repeated a determined number of times. Usually, this sequence corresponds to a succession of signals AA for the calling modem and to a succession of signals AC for the called modem. This sequence is repeated 56 times in the preamble. Considering the transmitted signals as a series of sine wave portions, a sequence of signals AA corresponds to an ideal sine wave at a 1800-Hz frequency, and a sequence of signals AC includes two components having a frequency of 600 and 3000 Hz, respectively.
A conventional process to detect the above successions of signals AA or AC uses two banks of digital filters at the front end of the receiver block of the modem. As represented in FIG. 4, the lower bank includes a prefilter F that allows only the frequencies ranging from 600 to 1800 Hz to pass, an energy calculator (providing the absolute square value of the signal) and a low-pass filter LPF1 of the first order. The upper bank includes a highly selective band-pass filter BPF centered on 600 or 1800 Hz followed by another energy calculator and by a low-pass filter LPF2. The outputs of the low-pass filters LPF1 and LPF2 are transmitted to a comparator 10. When sequences of signals AA or AC are present and when the band-pass filter BPF is centered on 600 or 1800 Hz, respectively, the energy in the upper bank is comparable to the energy in the lower bank. When a data signal is present, the upper bank provides a much lower level than the lower bank. This conventional process is satisfactory for the detection of sequences AA and AC during the initial handshake or retrain procedure where the duration of the analysis is relatively long. However, for operations to be carried out rapidly such as a rate negotiation operation, this procedure is less satisfactory since, in particular, it needs a relatively narrow band-pass filter BPF, which involves a relatively long response time. If it is desired to decrease this response time, selectivity is decreased, which results in a higher risk of false detection.
Thus, an object of the present invention is to provide a detection circuit for detecting a sequence of predetermined signals having a relatively short response time.
A further object of the present invention is to provide such a detection circuit that does not use any filtering circuitry.
These objects are achieved in one illustrative embodiment of the present invention, wherein a method for detecting a determined repetitive sequence of predetermined signals arriving on a modem, each signal being digitized as a word having a first portion corresponding to a real value and a second portion corresponding to an imaginary value, the method including the steps of delaying words of a sequence so that all the words are simultaneously present; linearly combining said words to provide a combined word with substantially null real and imaginary values; determining the modulus of each combined word; comparing this modulus with a threshold; when the modulus of the combined word is lower than the threshold, counting clock pulses corresponding to the rate at which the words arrive; and providing an identification signal when a predetermined number of clock signals is counted.
According to another embodiment of the present invention, each sequence includes two words, one of which is the inverse of the other, the linear combination being an addition.
According to a further embodiment of the present invention, each sequence includes two identical words, the linear combination being a subtraction.
According to an additional embodiment of the present invention, all the words of the sequence have a predetermined modulus, and the process further includes the steps of determining the modulus of each incoming word and resetting counting when the modulus deviates from its predetermined value by more than a predetermined threshold.
Another illustrative embodiment of the present invention provides a device for identifying a repetitive predetermined sequence of determined signals arriving on a modem, each of these signals being digitized as a word including a first portion corresponding to a real value and a second portion corresponding to an imaginary value. The device includes means for suitably delaying words of a sequence in order that all the words are present simultaneously; means for linearly combining the words to provide a combined word having substantially null real and imaginary values; means for calculating the modulus of each combined word; means for comparing this modulus with a threshold; means for counting clock pulses corresponding to the rate of arrival of the words; means for resetting the counting means when the modulus of the combined word is lower than the threshold; and means for providing an identification signal when a predetermined number of clock signals has been counted.
The foregoing and other objects, features, aspects and advantages of the invention will become apparent from the following detailed description of the present invention when taken in conjunction with the accompanying drawings.
FIG. 1 represents a constellation of 128 symbols in accordance with the V.32bis standard;
FIG. 2 represents by way of example the way in which each symbol is transmitted on a telephone line;
FIG. 3 represents a constellation of 16 symbols in accordance with the V.32bis standard;
FIG. 4 represents a circuit for identifying a sequence of signals AA or AC according to the prior art; and
FIG. 5 represents an exemplary embodiment of the invention for detecting a sequence of signals AA or AC.
The circuit according to the invention represented in FIG. 5 is part of a modem that during a receive phase receives digitized signals at its input 11. These signals arrive at the rate of a baud clock or symbol clock and are, for example, constituted by a word formed by two successive 16-bit portions corresponding to the real portion and to the imaginary portion of a received symbol, respectively. The input words are transmitted to a combinatory logic circuit 12 which, in the given example, includes a memory 13 of one word (real part and imaginary part) having an output provided to a complex adder 14. The other input of adder 14 receives the input signal 11 after passage through a multiplier 15. Thus, a linear combination of two successive words is carried out. When the sequence to be detected is constituted by a succession of identical words, for example AA . . . , multiplier 15 is a multiplier by −1 and, if this sequence is received, the output of adder 14 is normally zero. If the expected sequence is sequence ACAC . . . (refer FIGS. 1 and 3), that is, a sequence, wherein one signal is the complement of the next one (in the present description, “complement” is to be construed as “negative”), multiplier 15 is a multiplier by +1 and the output of adder 14 is normally zero if sequence ACAC . . . is received. More generally, it will be noted that it is merely necessary to provide a combinatory circuit 12 for combining successive signals having predetermined values in order to supply a zero value at the output when these signals are received.
In practice, since the lines are always noisy, the output of adder 14 is not strictly zero when the incoming signals are the expected signals. Accordingly, adder 14 is followed by an energy calculation circuit including a multiplier 17 receiving, on the one hand, directly the output of adder 14 and, on the other hand, this output through a circuit 18 providing the complex conjugate. The output signal 20 of multiplier 17 is provided to a first input of a comparator 21 that compares this Signal with a threshold TH1. Comparator 21 has an output at a high level when the signal at input 20 is higher than the threshold, and has an output at a low level when the signal is lower than the threshold. The output of comparator 21 is provided through an OR gate 22 to the reset input RS of a counter 23 that receives at its input 24 signals arriving at the same rate as data are introduced at input 11, that is, the signal at input 24 is normally the baud clock. Thus, as long as the output of comparator 21 is at a low level, counter 23 counts and as soon as output 21 reaches a high level, the counter is reset and the next counting begins at zero. The output of counter 23 is applied to an input 25 of a comparator 26 that compares this output with a predetermined value N, for example equal to 32. Thus, when a counting equal to 32 has occurred, a signal appears at the output terminal 28 of comparator 26. This signal at terminal 28 is an identification signal since it indicates that the output of comparator 21 has been 32 successive times at a low level, that is, its input was 32 successive times lower than threshold TH1. This means that signals corresponding to a sequence ACAC . . . were received 32 successive times (when multiplier 15 is a multiplier by +1) or a sequence AAA . . . (when multiplier 15 is a multiplier by −1).
Indeed, it is necessary to wait for a determined number of passages of signal 20 at a low level to, detect with certainty a sequence because the combinatory circuit 12 provides a low output each time successive signals exhibiting a predetermined (for example identical or reverse) relation are provided thereto; this can occur during a signal transmission for specific signals, as is apparent by studying the constellations of FIGS. 1 and 3.
However, there is a case when an erroneous detection of a sequence such as ACAC . . . may happen, that is, when no signal arrives at terminal 11, or more precisely when only noise arrives due to a temporary misfunction of the telephone link. The lower branch of the circuit of FIG. 5 solves this problem and avoids other ambiguities.
This lower branch includes a circuit for determining the energy of the incoming signal. The circuit includes a multiplier 31, whose first input directly receives the signal of terminal 11, and whose second input receives its complex conjugate through a circuit 32. The average energy A2 of signal A or C is subtracted from this energy (it should be noted that these signals have the safe modulus). This operation is carried out in an adder 33 whose output is provided to an absolute value calculator. 34. The output of circuit 34 is provided to the first, input of a comparator 35, whose second input receives a threshold signal TH2. The output of comparator 35 is provided to a second input of the OR gate 22. Thus, if the signal at terminal 11 is a signal having the same modulus as a signal A or C, the output of circuit 34 will be substantially zero and the output of comparator 35 will be zero. Accordingly, this signal will not affect the output of the OR gate 22. In contrast, if the signal at terminal 11 has a modulus different from the modulus of A, for example because it corresponds to a symbol having a different value or because it corresponds only to noise the output of comparator 35 will be at a high level and counter 23 will be reset.
Various other safety circuits can be devised by those skilled in the art.
Of course, although for the sake of simplification, the above description has been made using terms sometimes corresponding to analog systems, it will clearly appear to those skilled in the art that all the elements of the circuit according to the invention process digital signals and that the components of the circuit illustrated as hardware are often in practice embodied as software operating on a processor.
Having thus described one particular embodiment of the invention, various alterations, modifications, and improvements will readily occur to those skilled in the art. Such alterations, modifications, and improvements are intended to be part of this disclosure, and are intended to be within the spirit and scope of the invention. Accordingly, the foregoing description is by way of example only and is not intended to be limiting. The invention is limited only as defined in the following claims and the equivalents thereto.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US3872430||Nov 23, 1973||Mar 18, 1975||Boudreau Paul Emile||Method and apparatus of error detection for variable length words using a polynomial code|
|US4035581||Apr 8, 1976||Jul 12, 1977||Licentia Patent-Verwaltungs-G.M.B.H.||Code word detecting method|
|US4606045||Mar 16, 1984||Aug 12, 1986||Trw Inc.||Method and apparatus for detecting an equalizer training period in a receiving-end modem|
|US4829462||Jun 13, 1986||May 9, 1989||International Business Machines Corporation||Communication bit pattern detection circuit|
|US4847877||Nov 28, 1986||Jul 11, 1989||International Business Machines Corporation||Method and apparatus for detecting a predetermined bit pattern within a serial bit stream|
|US5127027||Nov 7, 1990||Jun 30, 1992||Canon Kabushiki Kaisha||Training signal detecting apparatus|
|US5202905||Nov 16, 1990||Apr 13, 1993||Fujitsu Limited||Moden signal detecting system for adaptive differential PCM CODEC|
|US5537549||Apr 28, 1993||Jul 16, 1996||Allen-Bradley Company, Inc.||Communication network with time coordinated station activity by time slot and periodic interval number|
|US5793809||May 31, 1995||Aug 11, 1998||Paradyne Corporation||Transparent technique for Mu-law modems to detect an all-digital circuit connection|
|EP0238100A2||Sep 21, 1983||Sep 23, 1987||Trw Inc.||Improved modem signal acquisition technique|
|JPH0537511A||Title not available|
|1||French Search Report from French patent application No. 9315942 dated Sep. 13, 1994.|
|2||IEEE Transactions on Communications, vol. 27, No. 9, Sep. 1979, New York US, pp. 1296-1301.|
|3||Patent Abstracts of Japan vol. 17, No. 328 (E-1385) Jun. 22, 1993 & JP-A-05037511 (NEC).|
|4||R. Wilson et al. "Generation and Performance of Quadraphase Welti Codes for Radar and Synchronization of Coherent and Differentially Coherent PSK".|
|U.S. Classification||375/222, 375/368, 375/340|
|International Classification||H04L27/38, H04L7/04, H04L7/02, H04L7/10|
|Cooperative Classification||H04L27/38, H04L7/046|
|European Classification||H04L7/04B10, H04L27/38|