JP3111680B2 - Demodulator - 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 apparatus for improving the characteristics of an accumulation batch demodulation method for demodulating a digital phase modulation received signal.

[0002]

2. Description of the Related Art For example, in the literature (Honda et al .: Study on computational demodulation of PSK signal, IEICE Technical Report, CS87-10)
9, 1987) is a conventional burst demodulator.
It was configured as follows. The quasi-synchronous detector 1 performs quasi-synchronous detection on the digital phase-modulated burst reception signal 11 with an asynchronous reference carrier having a carrier approximation frequency. analog/
Digital converter 2, the high-speed clock (e.g., 16 or more about samples per symbol period) to the burst mode modulation wave after detection wave from the quasi synchronous detector 1 performs sampling and quantization operation in. The buffer memory 3 receives the sample point data from the analog / digital converter 2 using a reception filter.
After waveform shaping , all sample point data is temporarily stored for each burst. Digital signal processing demodulator 10
Is the sample point data signal 12 from the buffer memory 3.
The demodulation process is performed on the
And outputs it as a result.

The demodulator in the digital signal processing type demodulator 10
In the key processing, the optimal position to judge the data
While estimating the phase, the conventional burst demodulator device, this
In order to reduce the phase estimation error, a collective demodulation method for extracting an optimal phase point by high-speed sampling is adopted.

As shown in FIG. 10, a digital signal processing type demodulator 10 first estimates a carrier wave from a data sequence of a sample point data signal 12 from a buffer memory by a carrier wave estimating means 4 and corrects the frequency and phase of the data. I do. Next, from the data signal whose frequency and phase have been corrected, only the data of the optimum phase point (the sample point closest to the Nyquist point) estimated by the timing estimating means 7a is extracted by the sampler 7b. Further, the data is judged by the data judgment means 8 and output as a demodulated data signal 19. The timing estimating means 7a estimates the optimum phase point as follows. First, in the case of two-phase shift keying (hereinafter referred to as BPSK), FIG.
When the BPSK reception signal is sampled for one symbol as shown in the following expression (the solid line indicates the waveform of the reception signal having the amplitude a and the symbol period T. d indicates transmission data (-1 or 1), and here, 1, 1 , .y showing an example of a series of transmission data consisting of 1,1 represents the sampled signal, here shown by y _1, y _2), the sample points from the original Nyquist point τ (0 ≦ τ <T), the sampled signal y is represented by the sum of the signal component at that time and the intersymbol interference component. If there is an influence, the impulse response of the receiving filter can be expressed as follows: h (τ) = sin (πτ / T) · cos (πατ / T) / {(πτ / T) · (1− (2ατ / T) ^2)} α: roll-off rate for example y ₁ of the receive filter is y ₁ = d ₁ ah (τ _{+ D 2 ah (τ + T} ) + ... + d k + 1 ah (τ + kT) + d 0 ah (τ-T) + ... + d -k + 1 ah (τ-kT) becomes, m-th in the same sampled signal y _m is generally expressed as _{y m = Σ l d m +} 1 ah (τ + l T) -k ≦ l ≦ k. Further, as shown in FIG. 12, when the BPSK received signal is sampled X per symbol, if the initial phase difference is set to 0 for simplification , τ = (n−1) T / X.
N (1 ≦ 1) at the m-th symbol
The n ≦ X) th received signal y _m generally expressed as _{y m = Σ l d m +} 1 ah {(n-1) T / X + l T} -k ≦ l ≦ k. Now the number of symbols of one burst is L, the accumulated value Pn of the n-th obtained by accumulating the absolute square value of the amplitude of the n th sample, respectively, in each symbol _{Pn = Σ m | Σ l d} m + 1 ah (τ + l T ^{) | 2 1 ≦ m ≦ L} , the -k ≦ l ≦ k. Here, d _j (j ≦ 0, L + 1 ≦ j) does not exist, but _dj (j ≦ 0, L + 1 ≦
j) is a pseudo noise (PN) pattern. Next, in the case of four-phase phase shift keying (hereinafter referred to as QPSK), there are two sequences of in-phase and quadrature components, which are independent of each other. Therefore, Pn obtained in the same manner as in the case of BPSK is Pn = Σ _m [｛Σ _l d _{m + 1} ah (τ + l T)｝ ² + ｛Σ _l d _{m + 1} ah (τ + l T)｝ ² ] 1 ≦ m ≦ L, −k ≦ l ≦ k. Here, d and d represent a data sequence of an in-phase component and a quadrature component, respectively. And amplitude a = 1, when L is sufficiently large data is random, the above equation simply ^{_{Pn = 2LΣ l h 2 (τ}} + l T) = 2LΣ l h 2 {(n-1) T / X + l T} -k ≦ l ≦ k. In the above equation, Pn becomes maximum when τ = 0 (n = 1). Therefore, when the Nyquist point is sampled, Pn becomes maximum. However, since the initial phase difference is not always 0 in actual operation, Find n that maximizes Pn, and use that point as Nike
It is estimated as a strike point and is set as an optimal phase point .

[0005]

In the conventional burst demodulation apparatus as described above, the collective demodulation for collective demodulation for extracting an optimum phase point by high-speed sampling (for example, about 16 samples or more per symbol period) in order to reduce a phase estimation error. Since the method is adopted, when the transmission speed becomes high, the analog / digital converter requires a high-speed operation, the buffer memory has a huge number of stored data, and the digital signal processing type demodulator has a problem that the amount of operation is large. .

An object of the present invention is to provide an accumulation batch demodulation method for extracting an optimum phase point by low-speed sampling so that a burst demodulator reduces a load and reduces a phase estimation error. .

[0007]

A demodulation device according to the present invention comprises:
In order to solve the above problems, an analog / digital converter for sampling a quasi-coherently detected digital phase modulation reception signal and performing analog / digital conversion, and an analog / digital converter
Interpolation timing generating means for performing an operation on the sampling data from the digital converter to estimate the Nyquist point and generating a plurality of pieces of interpolation timing information; and the sampling data signal with the plurality of pieces of interpolation timing information from the interpolation timing generation means. A data correcting means for estimating Nyquist point data by performing an interpolation operation and generating them as a plurality of temporary demodulated data sequences, and selecting a demodulated data sequence by comparing and judging the plurality of temporary demodulated data sequences from the data correcting means; And a data determining means for determining the demodulated data sequence from the timing determining means and outputting it as a demodulated data signal.

The data is corrected by the timing determining means.
For each of the plurality of temporary demodulated data sequences from the
The absolute value of the data amplitude of
A comparison decision is made to select a demodulated data sequence .

Alternatively, the data is supplemented by the timing determining means.
For each of the plurality of temporary demodulated data sequences from the correct means,
Compare and judge the absolute variance value and select the demodulated data sequence
You .

Further , the data correction means is used by the timing determining means.
For each of the plurality of temporary demodulated data sequences from the stage,
Compare the correlation value between the known pattern part and the predetermined known pattern.
A comparison decision is made to select a demodulated data sequence.

Further , the data correction means is used by the timing determining means.
For each of the plurality of temporary demodulated data sequences from the stage,
Soft-decision data series and hard-decision data
The demodulation data sequence is selected by comparing and determining the correlation value with the data sequence.
Select.

Further, the carrier power to noise power ratio measuring means for measuring the calculated carrier power to noise power ratio the mean value and variance values of said sampled data signal from the analog / digital converter is provided, the interpolation timing generator The means performs an operation on the accumulated value in accordance with the carrier power to noise power ratio information to estimate the Nyquist point and generate interpolation timing information.

[0013]

In the burst demodulator of the present invention, the analog
For sampling data from log / digital converter
The operation is performed by the interpolation timing
The exact point is estimated, and the
Information is generated. This interpolation timing generation means
A data correction unit based on the plurality of pieces of interpolation timing information;
Performs an interpolation operation on the sampling data signal in
Nyquist point data is estimated by
A sequence is generated. The plurality from the data correction means
Of the temporarily demodulated data sequence
The demodulation data sequence is selected by comparing the
Data sequence is determined by the data determination means and the demodulated data signal is
Output as a signal.

[0014]

DESCRIPTION OF THE PREFERRED EMBODIMENTS A burst demodulator according to one embodiment of the present invention.
Is shown in FIG. The quasi-synchronous detector 1, the analog / digital converter 2, the buffer memory 3, the carrier estimating means 4, and the data judging means 8 are the same as those shown in FIGS.
And corresponding. The interpolation timing generation means 5 performs an operation on the correction data signal 13 from the carrier wave estimation means 4 to estimate a Nyquist point, and generates interpolation timing information 14 and 15. The data correction means 6 includes the interpolation timing generation means 5
The Nyquist point data is estimated by performing an interpolation operation on the correction data signal 13 from the carrier wave estimating means 4 with the interpolation timing information 14 and 15 from, and is generated as temporary demodulated data sequences 16 and 17. The timing deciding means 7 applies the provisional demodulated data sequences 16 and 17 from the data
The value obtained by cumulatively adding the absolute value or absolute square value of the data amplitude for each burst is compared and determined, and the demodulated data sequence 18 is determined.
And outputs it to the data determination means 8.

The burst demodulator of the above embodiment employs an accumulation batch demodulation method for extracting an optimum phase point by low-speed sampling.

The interpolation timing generating means 5 includes, for example, a QP
At the time of SK, as shown in FIG. 2, the cumulative addition value Pn obtained by performing an operation on the correction data signal 13 from the carrier wave estimating means 4 is calculated from the function of the deviation τ (−T / 2 ≦ τ ≦ T / 2) from the Nyquist point. The Nyquist point is estimated for the thought-normalized cumulative addition value P (τ), and interpolation timing information 14 and 15 are generated. _{P (τ) = 2LΣ L h} 2 (τ + L T) / P (0) -k ≦ L ≦ k 1 a 2 samples per symbol period T in FIG. 2 for example
The ratio K of the second and second cumulative addition values P1 and P2 is P (3
T / 16) and P (11T / 16) and P (13T /
16) and P (21T / 16) are indistinguishable. (1) Outside of P1 and P2, outside of P2 by 5T / 16 from timing point of P2 (2) Inside of P1 and P2 The Nyquist point is estimated as 3T / 16 inside of P2 from the timing point of P1, and as interpolation timing information δ ₁ 14 and δ ₂ 15,
δ ₁ = 5/8 and δ ₂ = 3/8 are generated.

As shown in FIG. 3, for example, in the case of BPSK, the data correction means 6 performs primary interpolation (for example, internal division) or a correction data signal from the carrier wave estimation means 4 with δ ₁ and δ ₂ from the interpolation timing generation means 5. The Nyquist point data is estimated by performing a quadratic interpolation (for example, Lagrange's method) operation, and is generated as temporary demodulated data sequences 16 and 17. For example, in FIG. 3, the second of the (m-1) th symbol and the first of the mth symbol
The second sample point data y ₂ (m−1) and y ₁ (m)
Is divided internally by δ ₁ : (1−δ ₁ ), and the temporarily demodulated data u ₁
(M) is obtained. Similarly, the first and second sample point data y ₁ (m−1) and y ₂ (m
-1) is internally divided by δ ₂ : (1−δ ₂ ) to obtain provisionally demodulated data u ₂ (m). A data sequence composed of u ₁ (m) and u ₂ (m) is generated as temporary demodulated data sequences 16 and 17, respectively. u ₁ (m) = (1−δ ₁ ) × y ₂ (m−1) + δ ₁ × y ₁ (m) u ₂ (m) = (1−δ ₂ ) × y ₁ (m−1) + δ ₂ × y ₂ (m-1)

As shown in FIG. 4, the timing determining means 7 accumulates the absolute value or absolute square value of the data amplitude for each burst in the accumulator 71 for each of the provisionally demodulated data sequences 16 and 17 from the data correcting means 6. Value S ₁ cumulatively added to
And it determines that the larger the correct series of values in a comparator 72 to S _2, and select one of the tentatively demodulated data series 16 and 17 as the demodulated data stream 18 by selector 73, and outputs the data determination means 8. _{S 1 = Σ m | u 1} (m) | or _{Σ m | u 1 (m)} | 2 1 ≦ m ≦ L S 2 = Σ m | u 2 (m) | or _{Σ m | u 2 (m)} | ² 1 ≦ m ≦ L

In the above embodiment, the timing determining means 7 is used.
It may be configured instead of cumulative adder 71 in distributed calculator 71 or pattern correlator 71. Place of the dispersion calculator 71
The absolute variances V ₁ and V ₂ of the data amplitude and the pattern
Correlation value D ₁ and D with known pattern portion and a predetermined known pattern of the case of the correlator 71 and tentatively demodulated data sequence 16 17
₂ is calculated respectively. Then, the operation result is output to the comparator 72
And the smaller or larger value may be determined to be a correct series.

In the above embodiment, the timing determining means 7 may be constituted by a soft decision unit 74, a hard decision unit 75 and a correlator 76 instead of the accumulator 71 as shown in FIG. This place
In this case, soft decision data series {u ₁ } s and {u ₂ } obtained by performing soft decision and hard decision by soft decision unit 74 and hard decision unit 75, respectively .
s and the hard decision data sequence {u _1} h and {u _2} a h phase
The correlation value C ₁ and C ₂ input to the correlator 76 and calculated by the correlator 76 are compared by the comparator 72 , and the larger value is determined to be a correct sequence . Here, the soft decision data sequence is
This is a data sequence obtained by quantizing the amplitude of the provisional demodulated data sequence with a plurality of bits, and the hard decision data sequence is a data sequence in the case of 1-bit quantization.

In the above embodiment , as shown in FIG. 6, a carrier power to noise power ratio measuring means 9 is provided. Here, the carrier power (C), the variance value and the average value of the correction data signal 13 from the carrier estimating means 4 are obtained. as determined noise power (N), respectively, of these the measured carrier power to noise power ratio (C / N) information 20, the interpolation timing generator 5 this C / N
The Nyquist point may be estimated for the cumulative addition value P (τ) corresponding to the information 20, and the interpolation timing information 14 and 15 may be generated. For example, at the time of BPSK, as shown in FIG.
Changes according to the value of C / N, estimating the Nyquist point from P (τ) corresponding to C / N has the effect of reducing the phase estimation error with respect to the noise superimposed signal.

In the above embodiment, the interpolation timing generating means 5 has been described as operating on the correction data signal 13 from the carrier wave estimating means 4. However, as shown in FIG. An operation may be performed. Since the calculation uses the sample point data on the complex plane, the accumulated addition value P (τ) becomes the same even if there is a residual frequency by quasi-synchronous detection.

In the above embodiment, the waveform of the output of the analog / digital converter 2 is described as being shaped.
It goes without saying that analog-to-digital conversion may be performed after waveform shaping as long as the reception filter takes account of aliasing and the like.

[0024]

According to the burst demodulator of the present invention as described above, low-speed samples (for example, 2
Since the batch collective demodulation method for extracting the optimum phase point by sampling is adopted, the load can be reduced compared to the conventional method using high-speed sampling, and high-speed transmission can be dealt with. Further, there is an effect that the phase estimation error can be reduced by reducing the number of samples due to the introduction of the interpolation timing generation means.

[Brief description of the drawings]

FIG. 1 is a functional block diagram of a burst demodulator according to an embodiment of the present invention.

FIG. 2 is a view for explaining the operation of an interpolation timing generation means shown in FIG. 1;

FIG. 3 is a view for explaining the operation of the data correction means shown in FIG. 1;

FIG. 4 is a functional block diagram of a timing determining unit shown in FIG. 1;

FIG. 5 is a functional block diagram of another embodiment of the timing determining means shown in FIG. 1;

FIG. 6 is a functional block diagram of another embodiment including a carrier power to noise power ratio measuring means according to the present invention.

FIG. 7 is a view for explaining the operation of the carrier power to noise power ratio measuring means shown in FIG. 6;

FIG. 8 is a functional block diagram of another embodiment showing the present invention.

FIG. 9 is a functional block diagram of a conventional burst demodulator.

FIG. 10 is a functional block diagram of the digital signal processing type demodulator shown in FIG. 9;

FIG. 11 is a view for explaining one sample operation per symbol of the timing estimating means shown in FIG. 10;

FIG. 12 is a view for explaining an X-sample operation per symbol of the timing estimating means shown in FIG. 10;

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 quasi-synchronous detector 2 analog / digital converter 3 buffer memory 4 carrier estimation means 5 interpolation timing generation means 6 data correction means 7 timing determination means 8 data determination means 9 carrier power to noise power ratio measurement means 11 reception signal 12 sampling points Data signal 13 Corrected data signal 14, 15 Interpolation timing information 16, 17 Temporary demodulated data sequence 18 Demodulated data sequence 19 Demodulated data signal

──────────────────────────────────────────────────続 き Continued on front page (58) Field surveyed (Int. Cl. ⁷ , DB name) H04L 27/00-27/38

Claims (6)

(57) [Claims] An analog-to-digital converter that samples a quasi-coherently detected digital phase-modulated reception signal and performs analog-to-digital conversion, and performs an operation on sampled data from the analog-to-digital converter to estimate a Nyquist point. Interpolation timing generating means for generating a plurality of pieces of interpolation timing information, and performing an interpolation operation on the sampling data signal with the plurality of pieces of interpolation timing information from the interpolation timing generation means to estimate Nyquist point data, and calculate the Nyquist point data. Data correcting means for generating a demodulated data sequence, timing determining means for comparing and judging the plurality of temporary demodulated data sequences from the data correcting means and selecting a demodulated data sequence, and demodulated data sequence from the timing determining means Data judgment to judge and output as demodulated data signal Demodulating apparatus characterized by providing a stage. 2. The timing determining means according to claim 1, wherein said timing correcting means is a data correcting means.
For each of these multiple provisional demodulated data sequences,
Compare the absolute value of the data amplitude or the cumulative addition value of the absolute square value.
And selecting a demodulated data sequence.
Item 7. The demodulation device according to Item 1. 3. The timing decision means is a data correction means.
For each of these multiple temporary demodulated data sequences, the absolute
Comparison of variance values and selection of demodulated data sequence
The demodulation device according to claim 1, wherein 4. A data deciding means for the timing deciding means.
For each of the multiple temporary demodulated data sequences,
Compare the correlation value between the pattern part and the predetermined
And selecting a demodulated data sequence.
Item 7. The demodulation device according to Item 1. 5. A data determining means for the timing determining means.
For each of these multiple temporary demodulated data sequences,
Soft-decision data series and hard-decision data with fixed and hard decisions
Compare and judge the correlation value with the sequence and select the demodulated data sequence
The demodulator according to claim 1, wherein 6. A carrier power to noise power ratio measuring means for calculating an average value and a variance value of the sampling data signal from the analog / digital converter and measuring a carrier power to noise power ratio, wherein the interpolation timing generation is performed. 2. The demodulator according to claim 1, wherein the means performs an operation on the accumulated value in accordance with the carrier power to noise power ratio information, estimates a Nyquist point, and generates interpolation timing information.