JP3267657B2 - Demodulation method in digital communication - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a demodulation method in digital communication, and more particularly, to a method for synchronizing a demodulated baseband signal in digital communication with a clock signal sampling the demodulated baseband signal at the maximum opening of the eye. It relates to an improved demodulation system. In digital communication, a demodulated baseband signal is sampled at the maximum aperture of the eye, and digital data is reproduced by determining the polarity of the obtained sampling signal. Therefore, it is desirable that the maximum aperture of the eye of the demodulated baseband signal always coincides with the sampling phase of the clock signal.

[0002]

2. Description of the Related Art FIG. 7 is a diagram showing the configuration of a conventional demodulation system. In FIG. 7, reference numeral 31 denotes a demodulation unit for digital communication, 32 denotes an identification unit, 33 denotes a clock recovery unit, 34 denotes a non-linear processing unit,
35 is a limiter circuit, 36 is a differentiator circuit, 37 is a rectifier circuit, 38 is a narrow band filter formed by a PLL loop, 39
Is a phase comparator, 40 is a loop filter, 41 is a voltage controlled oscillator (VCO), and 42 is a timing adjustment unit.

[0003] The basic principle of the conventional clock recovery is that the reception I
A clock component is extracted from the F signal by a non-linear processing unit 34, and a high-Q narrow-band furnace wave unit 38 is interposed to reproduce a clock signal free of jitter and noise components. FIG. 8 is a diagram for explaining the operation of the conventional demodulation method. For example, a rectangular wave signal is obtained by non-linearly amplifying the demodulated baseband signal of the NRZ code system by the limiter circuit 35. Further, the square wave signal is differentiated and full-wave rectified by the differentiating circuit 36 and the rectifying circuit 37, thereby extracting a clock component synchronized with the code conversion point of the NRZ signal. Then, a phase error between the clock component and the clock signal of the voltage controlled oscillator 41 is detected by the phase comparator 39. If there is a phase error, the phase error is integrated by the loop filter 40, and the voltage control is performed by the obtained control voltage. Negative feedback is applied to the oscillator 41 so that the phase error becomes zero.

In this state, the voltage-controlled oscillator 41 operates in the NRZ
A clock signal having a half cycle of the minimum cycle of the signal is output, and the timing adjustment unit 42 performs a minute phase adjustment as needed. When the polarity of the demodulated baseband signal is identified by the identification unit 32, a so-called double sampling method is used to hit a code conversion point at the first timing and to detect an eye located at an intermediate point at the second timing. It hits the maximum opening.

However, in practice, there is linear distortion due to multipath fading, non-linear distortion due to a change in the operating point of the transmission power amplifier, etc., which changes the characteristics of the transmission path and the maximum aperture of the eye. . However, conventionally,
Since the eye opening is hit at the second timing after the elapse of a certain time from the first timing, the second timing cannot be adjusted even if the maximum opening of the eye fluctuates. For this reason, there has conventionally been a problem that the bit error rate is deteriorated.

Further, in order to realize such a clock reproduction by a digital PLL, it is necessary to configure a digital PLL using a clock source whose modulation speed is about 100 times. For this reason, in the conventional method, there was a limit to speeding up in digital processing.

[0007]

As described above, in the conventional demodulation method, even if the position of the maximum opening of the eye substantially fluctuates due to a change in the characteristics of the transmission path, the identification timing of the demodulated baseband signal is changed. I could not match this. An object of the present invention is to provide a demodulation method in digital communication that can always identify a demodulated baseband signal at the maximum opening of the eye even if the position of the maximum opening of the eye fluctuates due to a change in transmission path characteristics. It is in.

[0008]

The above-mentioned problem is solved by the structure shown in FIG. That is, the demodulation scheme in digital communication of the present invention, the tap coefficients h ₀ to h _m a FIR filter 1 for filtering the demodulated baseband signal
And a variance measuring unit 2 that measures the variance of the eye at each predetermined timing of the output of the FIR filter 1,
FIR filter 1 according to the variance value of the output of variance measurement unit 2
And a control unit 3 for controlling the tap coefficient of the FIR so that the variance value of the output of the variance measurement unit 2 is minimized.
The tap coefficient of the filter 1 is controlled.

The above problem is solved by the structure shown in FIG. That is, the demodulation method in the digital communication according to the present invention includes a clock generation unit 5 having a variable clock frequency or clock phase, and a clock generation unit
, A variance measuring unit 7 for measuring the variance of the eye at the clock timing of the output of the quantizing unit 6, and a clock generating unit 5 according to the variance value of the output of the variance measuring unit 7. And a controller 8 for controlling the clock frequency or clock phase of the clock generator 5 so that the variance of the output of the variance measuring unit 7 is minimized. is there.

[0010]

In FIG. 1, the input / output of the FIR filter 1 is generally shown.
There is a relationship of Y (z) = H (z) X (z) between the forces, and the transfer function H (z) is expressed as H (z) = h₀+ H₁z^-1+ H_Twoz^-2+ ‥‥ + h_mz^-m It is represented by Now, for the sake of explanation, as an example, an FIR filter
Let each tap coefficient of data 1 be h ₀= -3, h₁= 12, h_Two=
17, h_Three= 12, h_Four= -3, otherwise 0
Then, the transfer function H₁(Z) is H₁(Z) =-3 + 12z^-1+ 17z^-2+ 12z^-3-3z^-Four Can be expressed as This is z^-1= E^-jwTAnd substitute the frequency characteristic H
₁(Ω), H₁(Ω) = e^-j2wT(−6cos2ωT + 24cosωT + 17). Where e^-j2wTThe size of is always 1
Then, if you plot inside the parentheses on the right side at angular frequency ω,
This is the frequency characteristic of the low-pass filter
I understand.

Next, each tap coefficient is shifted to the right by one,
h₁= -3, h_Two= 12, h_Three= 17, h_Four= 12, h
_Five= -3, otherwise 0, a new transfer function
Number H _Two(Z) is H_Two(Z) =-3z^-1+ 12z^-2+ 17z^-3+ 12z^-Four-3z^-Five = Z^-1(-3 + 12z^-1+ 17z^-2+ 12z^-3-3z^-Four) = Z^-1H₁(Z). Then, this frequency characteristic H_TwoFind (ω)
And H_Two(Ω) = e^-jwTH₁(Ω), but e^-jwTIs always 1, so H
_Two(Ω) = H₁(Ω). That is, the frequency characteristics change
No. On the other hand, the output Y of the FIR filter 1 in this case is
(Z) is: Y (z) = z^-1H₁(Z) X (z), where z^-1Is one unit time on the time axis
Output Y in this case,
(Z) is delayed by one unit time.

Accordingly, the control unit 3 controls the tap coefficients h ₀ to h _m of the FIR filter 1, without changing its frequency characteristic H (omega), back and forth only delay time of the demodulated baseband signal You can change it freely.
Therefore, the variance measurement unit 2 measures the variance of the eye at each predetermined timing of the output of the FIR filter 1, while the control unit 3 controls the tap coefficient of the FIR filter 1 so that the variance value is minimized. . Thus, even if the input demodulated baseband signal is distorted due to a change in transmission path characteristics, the eye is output in the maximum aperture state after passing through the FIR filter 1, so that the code can always be identified in an optimal state. .

Preferably, the control section 3 includes a comparing section 4 for comparing each variance value of the output of the variance measuring section 2 in a time series, so that the variance value is minimized based on the comparison output of the comparing section 4. The tap coefficient of the FIR filter 1 is controlled. In FIG. 2, the clock generator 5 has a variable clock frequency or clock phase, and the quantizer 6 quantizes the input demodulated baseband signal at the clock timing of the clock generator 5. On the other hand, the variance measuring unit 7 measures the variance of the eye at the clock timing of the output of the quantizing unit 6, and the control unit 8 controls the clock generating unit 5 so that the variance of the output of the variance measuring unit 7 is minimized. Clock frequency or clock phase. Therefore, even if the input demodulated baseband signal is distorted due to a change in the transmission path characteristics, the data quantized by the quantization unit 6 is always the data quantized at the maximum aperture of the eye. The code can be identified.

Preferably, the control unit 8 includes a comparing unit 9 for comparing each variance value of the output of the variance measuring unit 7 in a time series, so that the variance value is minimized based on the comparison output of the comparing unit 9. The clock frequency or the clock phase of the clock generator 5 is controlled.

[0015]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments according to the present invention will be described below in detail with reference to the accompanying drawings. Note that the same reference numerals indicate the same or corresponding parts throughout the drawings. FIG. 3 is a diagram showing the configuration of the demodulation system of the first embodiment. In FIG.
The SK quadrature demodulator, 22 _I and 22 _Q are A / D converters (A
_{/ D), 1 I, 1} Q are each tap coefficient h ₀ to h _m a variable digital FIR (Finite Impulse Response) filter, Z ^-1 is the unit time delay element, × multiplication circuit, + an addition circuit, 23 identification of the code, 24 digital rectifier, 2 distributed measurement unit, 3 a control unit, 3 ₁ the inverter circuit, 3 ₂ J-K type flip-flop (FF),
3 ₃ ROM up-down counter (CTR), the 3 ₄ stores the tap coefficients h ₀ to h _m in different phases, 4
Comparing unit 4 ₁ memory (MEM), 4 ₂ is a comparator (CMP).

[0016] Such a transmission digital FIR filter 1 function H (z) is represented by _{H (z) = h 0 +} h 1 z -1 + h 2 z -2 + ‥‥ + h m z -m, each tap coefficients h ₀ to h _m is previously Elect value such as to have a without causing waveform distortion, and the optimum roll-off filter in the noise elimination frequency characteristic H (ω). FIG. 4 is a diagram for explaining the operation of the demodulation method of the first embodiment.

The frequency of the clock signal CLK _{2 in} FIG.
According to Nyquist's sampling theorem, the frequency is selected to be at least twice the frequency of the baseband signal band. Then, for example, when the digital demodulated baseband signal I sampled by the A / D converter 22 _I is imaged in an analog manner, it can be imaged as an eye pattern as shown in FIG. Data for each code point is E ₀ or -E _0. The digital rectifier 24 forms a full-wave rectified rectified output by removing the polarity bit from the digital demodulated baseband signal I. The distributed measurement unit 2 is sequentially measured variance at each predetermined section N on the basis of the data E ₀ for each demodulated data E _i and code point sampled by the clock signal CLK _1. The variance Ψ is obtained by, for example, Ψ = Σ _{i = 0} ^N (E ₀ −E _i ) ² . Here, if the clock signal CLK ₁ is sampled digital demodulated base band signals I at the maximum opening of the eye, the value of its variance Ψ is supposed that always minimized. However, actually, since the waveform distortion occurs due to the influence of the transmission line distortion, the value of the dispersion Ψ also changes every moment.

In FIG. 3, the comparator 4 of the comparing section 4
₂ compares the current variance Ψ with the previous variance Ψ _-1 ; if Ψ> Ψ- ₁ , outputs a logic 1 level; otherwise, outputs a logic 0 level I do. This output is connected to the J-K input terminal of flip-flop 3 _2, the flip-flop 3 ₂ When the output of the comparator 4 ₂ is a logic 0 level does not reverse its output Q, the comparator 4 _second output Is a logical 1 level,
The output Q is inverted by the clock signal CLK ₁ ′ generated when the shared data Ψ becomes available.
Then, this output Q is the up-down counter 3 ₃ up / down mode is controlled, the up-down counter 3 ₃ simultaneously is one count-up or a count-down with the clock signal CLK ₁ '. Then, according to the count value M at that time, tap coefficients h ₀ to h _m of the various phases is read from the ROM 3 _4.

The reading of the tap coefficients from the ROM is disclosed in Japanese Patent Application Laid-Open No. 61-108236, and this portion can be configured in the same manner. FIG.
In has increased the read phase of the tap coefficients h ₀ to h _m at a point in time result which is delayed from A to B, the dispersion Ψ in the direction of arrow a. Therefore, when the read phase of the tap coefficients h ₀ to h _m is inverted from B to A by [psi> [psi _-1, dispersion [psi is reduced in the direction of arrow b. Furthermore, when advancing the read phase of [psi <[psi tap coefficients by _-1 h ₀ to h _m from A and C, the dispersion [psi is further reduced in the direction of arrow c. In addition, Ψ
<When the read phase of the tap coefficients h ₀ to h _m by [psi _-1 advances from C to D, the dispersion [psi increases in the direction of arrow d. Therefore, when the read phase of the tap coefficients h ₀ to h _m is inverted from D to C by [psi> [psi _-1, dispersion [psi is reduced in the direction of arrow e. Thus, the read phase of the tap coefficients h ₀ to h _m is controlled to always distributed Ψ is minimized.

[0020] Although said the case of controlling the tap coefficients h ₀ to h _m of the FIR filter 1 _I and 1 _Q based on the variance of the demodulated baseband signals I in this embodiment, dispersion of the demodulated baseband signal Q May be controlled based on Alternatively, the tap coefficients of the FIR filters 1 _I and 1 _Q may be independently controlled based on the respective variances of the demodulated baseband signals I and Q. Or,
The tap coefficients of the FIR filters 1 _I and 1 _Q may be controlled based on the average value of each variance of the demodulated baseband signals I and Q.

When a quadrature signal such as QPSK is demodulated, a phase error Δθ of a leading phase or a lagging phase from a code point can be detected according to a fluctuation of a reception IF frequency. In such a case, even if the comparison unit 4 that performs the time-series comparison of the variances is not provided as described above, for example, by detecting that the variance value at a certain time is larger than a predetermined value, of the direction of canceling the phase error [Delta] [theta], it may control the up-down counter 3 ₃ directly.

FIG. 5 is a diagram showing the configuration of a demodulation system according to a second embodiment. In FIG.
25 _I and 25 _Q are roll-off filters (F), 6 _I and 6
_Q is A / D converter (A / D), 24 is a digital rectifier, 7 distributed measurement unit, 8 the control unit, 8 ₁ inverter circuit, 8 ₂ J-K type flip-flop (FF),
8 ₃ is an up / down counter (CTR), 8 ₄ is D / A
Converter (D / A), 9 a comparison unit, 9 ₁ memory (ME
M), 9 ₂ a comparator (CMP), 5 is a voltage controlled oscillator (VCO).

FIG. 6 is a diagram for explaining the operation of the demodulation system of the second embodiment. The digital rectifier 24 forms a full-wave rectified output by, for example, removing the polarity bit from the digital demodulated baseband signal I. Then, the dispersion measuring unit 7 outputs the clock signal C of the voltage controlled oscillator 5.
The variance in each predetermined section N is sequentially measured based on each demodulated data E _i sampled by the LK and the data E _{0 of the} code point. The variance Ψ is obtained by, for example, Ψ = Σ _{i = 0} ^N (E ₀ −E _i ) ² . Here, if the clock signal CLK performs A / D conversion of the demodulated baseband signal I at the maximum opening of the eye, the value of the variance Ψ should be minimum. However, in practice, the value of the variance Ψ changes every moment due to the fluctuation of the reception IF frequency f, the distortion of the demodulated baseband signal I, and the like.

In FIG. 5, the comparator 9 of the comparing section 9
₂ compares the current variance Ψ with the previous variance Ψ _-1 ; if Ψ> Ψ- ₁ , outputs a logic 1 level; otherwise, outputs a logic 0 level I do. This output is connected to the J-K input terminal of flip-flop _82, the flip-flop ₈₂ is when the output of the comparator 9 ₂ is a logic 0 level does not reverse its output Q, the comparator 9 _second output Is a logical 1 level,
The output Q is inverted by the clock signal CLK 'generated when the distributed data becomes available. Then, the the output Q is the up-down counter 8 ₃ up / down mode is controlled, the up-down counter 8 ₃ simultaneously is one count-up or a count-down with the clock signal CLK '. By D / A conversion of the count value M at that time by the D / A converter 8 ₄ controls the oscillation frequency of the voltage controlled oscillator 5.

[0025] In FIG. 6, changing the count value of the up-down counter 8 ₃ at a certain time point from M-1 to M-2, the phase of the clock signal CLK from the CLK-A C
LK-B, so that the variance Ψ increases in the direction of arrow a. Therefore, the count value is changed to M-2 by Ψ> Ψ _-1.
, The phase of the clock signal CLK advances from CLK-B to CLK-A, and the dispersion Ψ decreases in the direction of arrow b. Furthermore, changing [psi <[psi count value by _-1 from M-1 to M, the phase of the clock signal CLK goes from CLK-B in CLK-C, the dispersion [psi is further reduced in the direction of arrow c. Further, when the count value is changed from M to M + 1 according to Ψ <Ψ _−1, the phase of the clock signal CLK advances from CLK-C to CLK-D, and the dispersion Ψ becomes an arrow d.
In the direction of. Therefore, when the count value is inverted from M + 1 to M by Ψ> Ψ- ₁ , the clock signal CL
The phase of K is delayed from CLK-D to CLK-C, and the dispersion Ψ decreases in the direction of arrow e. Thus, the control voltage of the voltage controlled oscillator 5 is controlled so that the dispersion 常 に is always minimized.

In this embodiment, the case where the control voltage of the voltage controlled oscillator 5 is controlled based on the dispersion of the digital demodulated baseband signal I has been described. However, the control is performed based on the dispersion of the demodulated baseband signal Q. May be. Alternatively, the control voltages of the plurality of voltage controlled oscillators may be independently controlled based on the respective variances of the demodulated baseband signals I and Q. Alternatively, the control voltage of the voltage controlled oscillator 5 may be controlled based on the average value of each variance of the demodulated baseband signals I and Q, and the like.

In the case where a quadrature signal such as QPSK is demodulated, the phase error Δθ of the leading phase or the lagging phase from the code point can be detected according to the fluctuation of the reception IF frequency. In such a case, even if the comparison unit 4 that performs the time-series comparison of the variances is not provided as described above, for example, by detecting that the variance value at a certain time is larger than a predetermined value, in the direction of canceling the phase error [Delta] [theta], it may be configured to control the up-down counter 8 ₃ directly.

Although the clock frequency of the voltage controlled oscillator 5 is controlled in this embodiment, the clock phase of the voltage controlled oscillator 5 may be controlled. In the above embodiment, the case of QPSK has been described.
The present invention can be applied to M and the like.

[0029]

As described above, according to the present invention, an FIR filter 1 for filtering a demodulated baseband signal is provided.
As variable tap coefficients h ₀ to h _m a is, FIR
A dispersion measuring unit for measuring a variance of an eye at each predetermined timing of an output of the filter; a control unit for controlling a tap coefficient of the FIR filter according to a variance value of an output of the dispersion measuring unit; Controls the tap coefficient of the FIR filter 1 so that the variance of the output of the variance measuring unit 2 is minimized. Therefore, even if the input demodulated baseband signal is distorted due to a change in the transmission path characteristics, the signal passes through the FIR filter 1. After that, since the maximum opening of the eye always matches the sampling phase of the clock signal CLK, the code can always be identified in an optimal state. Moreover, according to the present invention, a PLL and a VCO associated with clock recovery are not required, and the speed of clock synchronization can be increased. In addition, since the calculation can be performed at the data rate, it can be applied to transmission of high-speed data.

Further, according to the present invention, a clock generator 5 that can change the clock frequency or clock phase, a quantizer 6 that quantizes the demodulated baseband signal at the clock timing of the clock generator 5, and a quantizer 6 A variance measuring unit 7 for measuring the variance of the eye at the clock timing of the output of the control unit, and a control unit 8 for controlling the clock frequency or the clock phase of the clock generating unit 5 according to the variance value of the output of the variance measuring unit 7. The unit 8 controls the clock frequency or the clock phase of the clock generation unit 5 so that the dispersion value of the output of the dispersion measurement unit 7 is minimized. Therefore, even if the input demodulated baseband signal is distorted due to a change in the transmission path characteristics. ,
Since the data quantized by the quantization unit 6 is always data quantized at the maximum opening of the eye, the code can always be identified in an optimal state.

[Brief description of the drawings]

FIG. 1 is a diagram showing the basic configuration of the present invention.

FIG. 2 is a diagram showing the basic configuration of the present invention.

FIG. 3 is a diagram illustrating a configuration of a demodulation method according to a first embodiment.

FIG. 4 is a diagram for explaining the operation of the demodulation method according to the first embodiment;

FIG. 5 is a diagram illustrating a configuration of a demodulation method according to a second embodiment.

FIG. 6 is a diagram illustrating the operation of the demodulation method according to the second embodiment.

FIG. 7 is a diagram showing a configuration of a conventional demodulation system.

FIG. 8 is a diagram for explaining the operation of the conventional demodulation method.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 FIR filter 2 dispersion measurement unit 3 control unit 5 clock generation unit 6 quantization unit 7 dispersion measurement unit 8 control unit

Continuation of the front page (56) References JP-A-2-46045 (JP, A) JP-A-4-104542 (JP, A) JP-A-4-160843 (JP, A) JP-A-59-15352 (JP) JP-A-59-171230 (JP, A) JP-A-4-10728 (JP, A) JP-A-3-62753 (JP, A) JP-A-4-364637 (JP, A) 59-208949 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) H04L 27/00-27/38

Claims (4)

(57) [Claims] 1. An FIR filter (1) for filtering a demodulated baseband signal, comprising: a tap coefficient (h ₀ )
h _m) and the variable ones, distributed measurement unit that measures the variance of the eye at each predetermined timing of the output of the FIR filter (1) and (2), the FIR filter in accordance with the variance value of the output of the distributed measurement unit (2) ( A control unit (3) for controlling the tap coefficient of (1), wherein the control unit (3) controls the tap coefficient of the FIR filter (1) so that the variance of the output of the variance measuring unit (2) is minimized. A demodulation method in digital communication. 2. The control section (3) includes a comparison section (4) for comparing each variance value of the output of the variance measurement section (2) in a time series.
2. A demodulation method in digital communication according to claim 1, wherein the tap coefficient of the FIR filter is controlled based on the comparison output of the comparison unit so that the variance value is minimized. 3. A clock generation unit (5) capable of changing a clock frequency or a clock phase, a quantization unit (6) for quantizing a demodulated baseband signal at a clock timing of the clock generation unit (5), and a quantization unit. A variance measuring unit (7) for measuring the variance of the eye at the clock timing of the output of (6), and controlling the clock frequency or clock phase of the clock generating unit (5) according to the variance of the output of the variance measuring unit (7) The control unit (8) controls the clock frequency or the clock phase of the clock generation unit (5) so that the variance value of the output of the variance measurement unit (7) is minimized. Characteristic demodulation method in digital communication. 4. The control unit (8) includes a comparison unit (9) that compares each variance value of the output of the variance measurement unit (7) in a time series.
4. The demodulation method in digital communication according to claim 3, wherein the clock frequency or the clock phase of the clock generator is controlled based on the comparison output of the comparator to minimize the variance.