METHOD AND DEVICE FOR MINIMAL PHASE GRADIENT DEMODULATION OF A SIGNAL REPRESENTATIVE OF DATA.

Abstract

Upon receiving a signal that has undergone data spectrum spreading under minimum phase gradient modulation, the receiving device generates local signals similar to the received carrier waves, demodulates the received waves by the local signals to obtain baseband signals, generates polar replicas including at least one non-zero level step to approximate the shape of a half-sinusoid, demodulates the baseband signals by correlation with the polar replicas, performs the algebraic summation of the baseband signals thus demodulated and samples the signal thus obtained in order to reconstruct the transmitted data.

Description

The invention relates to data transmission. More particularly, it applies to minimum phase gradient demodulation of a representative data signal. In French Patent Application FR-88 15855 (Main Patent) filed on December 2, 1988, the Applicants proposed a demodulation system for a signal representing data transmitted under a minimum phase gradient modulation called MSK (from the Anglo-Saxon terminology "Minimum Shift Keying"), in which the first and second carrier waves, originally in quadrature, are modulated by phase inversion combined with amplitude modulation, respectively, according to first and second polar signals with sinusoidal conformation, the polarity of which depends on successive bits of the data to be transmitted. The first and second bipolar signals consist of consecutive segments, taken alternately, of at least one pseudo-random sequence, said segments being rounded after polarity modulation by said successive bits of the data to be transmitted. According to the main patent, the reception of a signal that has undergone data spectrum spreading under minimum phase gradient modulation is carried out by a process that includes the following operations: - generating first and second local quadrature signals similar to the first and second quadrature carrier waves as received; - separately demodulating the waves received by said first and second local signals to obtain first and second baseband signals; - generating first and second polar replicas, each consisting of a periodic waveform whose pattern or "segment" includes a pulse and a plateau; - demodulating the first and second baseband signals by correlation with the second and first replicas, respectively; - algebraically summing the first and second baseband signals thus demodulated by correlation; - progressively shifting the first and second replicas.until the temporal coincidence of these replicas with the first and second baseband signals is achieved; and - sample the signal thus obtained in order to reconstruct the transmitted data. The Applicants have observed that the operation of aligning the replicas with the baseband signals is slowed down by the degradation of the signal / noise ratio existing between these replicas and the sinusoidal waves as received. More generally, the degradation of the signal-to-noise ratio reduces the overall performance achieved. The invention improves the situation. One aim of the invention is to provide improvements to the demodulation system of the main patent and which overcome the aforementioned drawback. In this regard, the invention aims to provide a demodulation system for an MSK signal which allows for rapid alignment of replicas with baseband signals. The invention relates to a data transmission method, according to any one of the claims of the main patent, in which first and second carrier waves, originally in quadrature, are modulated by phase inversion associated with amplitude modulation, respectively, according to first and second bipolar signals with a substantially sinusoidal conformation, the polarity of which depends on successive bits of the data to be transmitted, said first and second bipolar signals being composed of consecutive segments, taken alternately, of at least one pseudo-random sequence, said segments being rounded after polarity modulation by successive bits of the data to be transmitted, the method comprising, at the receiver,The following steps: a) generate first and second local quadrature signals similar to the first and second carrier waves as received; b) separately demodulate the waves received by said first and second local signals to obtain first and second baseband signals; c) generate first and second polar replicas, each consisting of a periodic waveform whose pattern or "segment" includes a pulse and a plateau; d) demodulate by correlation the first and second baseband signals with the second and first replicas, respectively; e) perform the algebraic summation of the first and second baseband signals thus demodulated by correlation; and f) sample the signal thus obtained in order to reconstruct the transmitted data. According to the general definition of the method according to the invention, the pattern or "segment" of the first and second polar replicas includes at least one non-zero level step to approximate the shape of a half-sinusoid. In practice, step c) includes - cl) the attenuation of the amplitude of the baseband signals during the duration of the non-zero plateau of the aftershocks, as a function of the height of said plateau. The invention also relates to a receiving device for implementing the method according to the invention, comprising: - receiving means for receiving the emitted signal; - a voltage-controlled local oscillator capable of generating first and second local quadrature signals similar to the first and second quadrature carrier waves as received, said carrier waves being modulated by phase inversion associated with amplitude modulation respectively according to first and second bipolar signals consisting of consecutive segments, taken alternately, of at least one pseudo-random sequence.said segments being rounded after polarity modulation by successive bits of the data to be transmitted; a demodulator suitable for separately demodulating the waves received by said first and second local signals to obtain first and second baseband signals; a pseudo-random sequence generator suitable for generating first and second polar replicas, each consisting of a periodic waveform whose pattern or "segment" includes a pulse and a plateau; a demodulator-correlator suitable for demodulating by correlation the first and second baseband signals with the second and first polar replicas, respectively; summing means suitable for performing the algebraic sum of the first and second baseband signals thus demodulated by correlation; and sampling means for the signal thus obtained in order to reconstruct the transmitted data. According to the general definition of the receiving device according to the invention, the pseudo-random sequence generator produces first and second polar replicas whose pattern or "segment" includes at least one non-zero step to approximate the shape of a half-sinusoid. According to a preferred embodiment of the device according to the invention, the device further comprises attenuating means connected to the output of the demodulator of the received waves in baseband signals and capable, on control of the pseudo-random sequence generator, of attenuating the amplitude of the baseband signals during the duration of the non-zero plateau of the aftershocks, depending on the height of said plateau. In practice, the attenuating means consist of two modules, each associated with a baseband signal, and each comprising: - at least two inputs, each receiving a control signal from the pseudo-random generator; - a gain amplifier having an input receiving the associated baseband signal and an output; - an inverter having an input connected to the output of the amplifier and an output; - switching means controlled by the control signals, comprising at least a first switch having an input connected to the output of the gain amplifier through a first resistor and an output; - a second switch having an input connected to ground through a second resistor and an output.the first and second resistors being connected in series through a third resistor; a third switch having an input connected to the output of the inverter through a fourth resistor and an output; and a fourth switch having an input connected to ground through a fifth resistor and an output, the fourth and fifth resistors being connected in series through a sixth resistor. Alternatively, each attenuating means module further includes a fifth switch having four inputs connected in series to ground through a seventh resistor and four outputs connected in series. Other features and advantages of the invention will become apparent from the examination of the detailed description below, and the accompanying drawings, in which - Figure 1 is a schematic view of the demodulator according to the invention - Figure 2 represents timing diagrams illustrating the replicas - Figure 3 represents a schematic view of the attenuating means for replicas consisting of a polar rectangular slot and a non-zero plateau; and - Figure 4 is a schematic view of the attenuating means for replicas consisting of a polar rectangular slot, a non-zero plateau and a zero plateau. The attached drawings contain numerous elements of a definite nature. They can therefore not only serve to clarify the description below, but also contribute to the definition of the invention, if necessary. Figure 1 shows the receiving side of the data transmission system according to the invention bearing general reference 14. The receiving device 14 is virtually identical to the receiving device 14 of the main patent, differing only in the addition of attenuating means which will be described in more detail below. The description of the receiving device is intentionally brief. Those skilled in the art may refer to the main patent for a more detailed description of the constituent elements of the receiving device 14. The receiving device 14 includes receiving means 18 for receiving the emitted MSK type signal 16. A separator 20 separates the carrier waves received on two channels. A voltage-controlled local oscillator 22 (VCXO) generates, via an amplifier 23 and a phase shifter 21, first and second local quadrature signals similar to the first and second quadrature carrier waves as received. A demodulator 24 separately demodulates the waves received by the first and second local signals generated by the oscillator 22 to obtain at the output of the filtering means 26 baseband SQ and SI signals. A pseudo-random sequence generator 28 generates first and second polar replicas P1M and P2M. In the main patent, the replicas P1M and P2M each consist of a periodic waveform whose pattern or "segment" S" comprises a polar rectangular pulse (Figure 2b) or a polar rectangular pulse followed by a plateau at 0° (Figure 2c). The S" segments have a duration Tseg equal to the duration of the S' segments of the main patent. In part c of Figure 2, the polar rectangular pulses have a duration aTseg, where a is the shape ratio, equal here to 0.7. In parts b to f of figure 2, only the replica P1M is shown. In practice, the replicas P1M and P2M delivered by the generator 28 are identical and offset from each other by the duration of half a segment S". We refer again to Figure 1. A demodulator-correlator 30 demodulates by correlation the first and second baseband signals SQ and SI with the second and first polar replicas P2M and P1M, respectively. Summing means 32 perform the algebraic summation of the first and second baseband signals thus demodulated by correlation. The demodulated signal, representative of the data thus obtained at the output of the means 32, is integrated and sampled by integration and sampling means 34 in order to reconstruct the transmitted data. A priori, the two replicas P1M and P2M are not in temporal coincidence with the baseband signals SI and SQ; their correlation will therefore give a result of zero. It is therefore necessary to provide in the receiving device of the invention means suitable for seeking temporal coincidence between the first and second replicas P1M and P2M with the first and second baseband signals SQ and SI. Functionally, the search means provide an algebraic sum 32, 40 of the baseband signals thus demodulated by correlation with the replicas 30, 38, appropriate filtering 35, 42, squaring 36, 44, and an algebraic sum 46 of the signals thus processed. From these search means, information is obtained that represents the correlation between said replicas and the baseband signals, detected by the threshold detector 49. In this coincidence search operation, we gradually shift the first and second replicas 50, 51, 53 until we approach the temporal coincidence of these replicas with the baseband signals. After this coincidence search operation, a coincidence tracking operation is performed during which the first and second replicas are aligned as precisely as possible with the first and second baseband signals. This tracking is done using phase deviation analysis. The coincidence search operation can also consist of shifting (block 99) the replicas shifted in advance or behind by the duration of half a segment Tseg, until one of the two temporal coincidences of these replicas thus shifted is obtained with the baseband signals, detected by the threshold detector 47. The Applicants observed that the calibration operation of the 50, 51, 53 replicas (Figures 2b and 2c) with the baseband signals is slowed down by a degradation of the signal-to-noise ratio between these replicas and the received sinusoidal waves (Figure 2a). For example, they measured that the degradation of this signal-to-noise ratio is on the order of -0.9 dB in the case of replicas consisting of a periodic waveform whose pattern includes a polar rectangular pulse (Figure 2b) on the one hand; and on the order of -0.4 dB in the case of replicas consisting of a periodic waveform whose pattern includes a polar rectangular pulse followed by a plateau at 0 with a shape ratio of 0.7 (Figure 2c) on the other hand. To further reduce degradation, the Applicants improved the receiver device 14 of the main Patent by making replicas approaching the shape of a half-sinusoid according to the invention. For this purpose, attenuating means 90 are interposed between the filtering means 26 and the d-modulator-correlator 30. These attenuating means which receive the baseband SQ and SI signals are controlled by control signals 101 and 102 delivered by the pseudo-random sequence generator 28. In part d of figure 2, a chronogam is shown illustrating a replica whose pattern includes a polar notch followed by a non-zero plateau to approximate the shape of a half-sinusoid. More precisely, the pattern includes a first straight line segment D4, with a value or height + 0.36, having a width equal to 20% of its period Tseg, a second straight line segment D5, with a value + 1, having a width equal to 60% of its period Tseg, and a third straight line segment D6, with a value + 0.36, having a width equal to 20% of its triode Tseg. The Applicants noted that with a replica exhibiting such a shape pattern, the signal-to-noise ratio is degraded by -0.2 dB. We now refer to figure 3 which schematically represents mitigating means allowing the generation of pseudo-random replicas described in reference to part d of figure 2. The attenuating means 90 consist of two modules, each associated with a baseband signal. Each module has two inputs, 105A and 105B, each receiving a control signal, 101A and 101B, from the pseudo-random generator 28. A gain amplifier G, having an input GE receiving the baseband signal and an output GS, amplifies the signal SQ. An inverter I, having an input IE connected to the output of the amplifier GS and an output IS, reverses the polarity of the signal SQ. Switching means C, controlled according to control signals 101A and 101B, are capable of attenuating the gain of the amplifier G and the inverter I during the duration of the non-zero plateau of the replicas as a function of the height (or value) of said plateau. The C switch means include the individual switches in C1 to C4, each of whose outputs CîS to C4S is connected in series to the demodulator-correlator 30. Structurally, switch C1 has an input ClE connected to the output GS of the gain amplifier G through a resistor R1. Switch C2 has an input C2E connected to ground through a resistor R2; resistors R1 and R2 are connected in series via a resistor R3. Resistors R1, R2 and R3 are used to attenuate the signal from amplifier G according to an attenuation ratio U = 0.36, that is, the value of the non-zero plateau of replica P1M described with reference to part d of figure 2. Similarly, resistors R4, R5 and R6 are used to attenuate the gain of the inverter I according to a loss ratio -U. The attenuation occurs during the non-zero plateaus, i.e., during the straight sections D4 and D6 (Figure 2d). The control signals 101A and 101B from generator 28 regulate this attenuation by switching the switching means C. It will be noted that to generate the aftershocks whose pattern includes a rectangular niche (figure 2b) the attenuating means 90 can be reduced to just a gain amplifier G, an inverter I, two switches C1 and C3 controlled by a control signal from the generator 28. Thus arranged the attenuating means 90 no longer perform their attenuation function but only the control of the switching means C. The replicas whose shape pattern includes a polar rectangular slot followed by a 0 plateau (figure 2c), require attenuating means equipped in addition to the elements listed above for a replica with a polar rectangular slot, with the two switches C2 and C4 connected directly to ground and controlled according to the two control signals from generator 28. We now refer to part e of Figure 2, which shows a chronogram illustrating an aftershock whose shape pattern includes a polar rectangular slit followed by a non-zero plateau and a zero plateau. More precisely, the aftershock comprises a first straight line segment D7, with a value of 0 and a width equal to 10% of its period Tseg; a second straight line segment D8, with a value of +0.618 and a width equal to 20% of its period Tseg; a third straight line segment D9, with a value of +1 and a width equal to 40% of its period Tseg; a fourth straight line segment D10, with a value of ±0.618 and a width equal to 20% of its period Tseg; and a fifth straight line segment D11, with a value of 0 and a width equal to 10% of its period Tseg. With a replica exhibiting such a shape pattern, the Applicants measured that the signal-to-noise ratio is degraded by -0.144 dB. We now refer to part f of figure 2, in which a chronogram has been shown illustrating a replica whose shape pattern also includes five levels. More specifically, the replica P1M includes a first straight segment D12, with a value of 0, having a width equal to 5% of its period Tseg, a second straight segment D13, with a value of +0.36, having a width equal to 15% of its period Tseg, a third straight segment D14, with a value of +1, having a width equal to 60% of its period Tseg, a fourth straight segment D15, with a value of +0.36, having a width equal to 15% of its period Tseg and a fifth straight segment D16, with a value of 0, having a width equal to 5% of its period Tseg. With a replica exhibiting such a shape pattern, the Applicants measured that the signal-to-noise ratio is degraded by -0.2 dB. We now refer to figure 4 which represents an embodiment of the attenuating means for the aftershocks described with reference to parts e and f of figure 2. The attenuating means are almost identical to those described with reference to figure 3, only four additional switches C5, C6, C7 and C8 have been added, each input C5E, C6E, C7E and C8E being connected to ground through a resistor R7. Switches C5, C6, C7 and C8 are switched in response to a third control signal 101C from generator 28 and received on the third input 105C of the means switches C. The switching of switches C5 to C8 takes place during the duration of the straight sections D7 and D11 (figure 2e) or the straight sections D12 and D16 (figure 2f) which correspond to zero steps. Replicas whose shape pattern approaches a rounded conformation to reduce signal-to-noise ratio degradation can be used in modulation systems other than the MSK system, for example, systems with or without étalem of the don-hete-s spectrum, involving phase changes such as the so-called BPSK modulation from the Anglo-Saxon terminology "Binary phase Shift Keying". This reduction in signal-to-noise ratio degradation can be achieved by replicas whose shape pattern approximates a rounded conformation along a stepped contour such that a curve representing a cardinal sine sin xx

Claims

Claims 1. A data transmission method according to any one of the claims of the main patent, wherein the first and second carrier waves, originally in quadrature, are modulated by phase inversion associated with amplitude modulation respectively according to first and second bipolar signals with substantially sinusoidal conformation, the polarity of which depends on successive bits of the data to be transmitted, said first and second bipolar signals being composed of consecutive segments, taken alternately, of at least one pseudo-random sequence, said segments being rounded after modulation, in polarity by said successive bits of the data to be transmitted, said method comprising, at the receiver,the following steps - a) generate first and second local quadrature signals similar to the first and second carrier waves as received - b) separately demodulate the waves received by said first and second local signals to obtain first and second baseband signals (SQ and SI) - c) generate first and second polar replications (P1M and P2M), each consisting of a periodic waveform whose pattern or "segment" includes a pulse and a plateau - d) demodulate by correlation the first and second baseband signals (SQ and SI) with the second and first replications (P2M and P1M),respectively - e) perform the algebraic summation of the first and second baseband signals (SQ and SI) thus demodulated by correlation - f) sample the signal thus obtained in order to reconstruct the transmitted data; characterized in that the pattern or "segment" (S") of the first and second polar replicas includes at least one non-zero level plateau to approximate the shape of a half-sinusoid.

2. Method according to claim 1, characterized in that step c) comprises - cl) the attenuation of the amplitude of the baseband signals during the duration of the non-zero plateau of the replicas (P1M and P2M), as a function of the height of said plateau.

3. Receiver device for implementing the method according to any one of claims 1 and 2 and any one of claims 13 to 27 of the main Patent, comprising - receiver means (18) for receiving the emitted signal (16) - a voltage-controlled local oscillator (22) capable of generating first and second local quadrature signals similar to the first and second quadrature carrier waves as received, said carrier waves being modulated by phase inversion associated with amplitude modulation respectively according to first and second bipolar signals consisting of consecutive segments, taken alternately from at least one pseudo-random sequence,said segments being rounded after polarity modulation by successive bits of the data to be transmitted - a demodulator (24) suitable for separately demodulating the waves received by said first and second local signals to obtain first and second baseband signals (SQ and SI) - a pseudo-random sequence generator (28) suitable for generating first and second polar replicas (P1M and P2), each consisting of a periodic waveform whose pattern or "segment" includes a pulse and a plateau - a demodulator-correlator (30) suitable for demodulating by correlation the first and second baseband signals (SQ and SI) with the second and first polar replicas (P2M and P1M),respectively - summing means (32) suitable for performing the algebraic summation of the first and second baseband signals thus demodulated by correlation; and - sampling means (34) of the signal thus obtained in order to reconstruct transmitted data, characterized in that the generator (28) generates first and second polar replicas whose pattern or "segment" includes at least one non-zero plateau to approximate the shape of a half-sinusoid.

4. Device according to claim 3, characterized in that it further comprises attenuating means (90) connected to the output of the demodulator (24) of the received waves in baseband signals and capable, on control of the pseudo-random sequence generator, of attenuating the amplitude of the baseband signals during the duration of the non-zero plateau of the aftershocks, depending on the height of said plateau.

5. Device according to claim 4, characterized in that the attenuating means (90) consist of two modules, each associated with a baseband signal, and each comprising: - at least two inputs (105A, 105B), each receiving a control signal (101A, 101B) from the pseudo-random generator (28); - a gain amplifier (G) having an input (GE) receiving the associated baseband signal (SQ) and an output (GS); - an inverter (I) having an input (IE) connected to the output of the amplifier (GS) and an output (IS); - switching means (C) controlled by the control signals (101A, 101B), comprising at least a first switch (C1) having an input (C1E) connected to the output of the gain amplifier (GS) through a first resistor (R1) and an output (C1S); - a second switch (C2) having an input (C2E) connected to ground through a second resistor (R2),the first and second resistors (R1 and R2) being connected in series through a third resistor (R3) - a third switch (C3) having an input (C3E) connected to the output of the inverter (IS) through a fourth resistor (R4) - a fourth switch (C4) having an input (C4E) connected to ground through a fifth resistor (R5), and an output (C4S), the fourth and fifth resistors (R4 and R5) being connected in series through a sixth resistor (R6).

6. Device according to claim 5, characterized in that each module of the attenuating means further comprises a fifth switch (C5) having four inputs connected in series through a seventh resistor (R7) and four outputs mounted in series.