JPH0983596A - Digital quadrature detector - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce calculation amount in a reception FIR low-pass filter for the input IF signal of an arbitrary frequency and to highly accurately secure a base band signal. SOLUTION: Quadrature detection means 3 and 4 which detect a sample value signal which is A/D converted by a local frequency fs/4 and equivalently generate in-phase output and output, which have frequency offset Δf equivalent to the minimum value of |fc-(fs/4)| and whose sampling frequencies are fs/2, band-pass filters 5 and 6 which band-restrict the in-phase output and output and phase compensation means 7 and 19 compensating Δf included in the outputs of the band pass filters 5 and 6 are provided. When the carrier frequency of a signal after A/D conversion is considered to be fs/4 and it is detected, '0' is inserted into in-phase output and output at every other sample and the sampling frequency is equivalently reduced to 1/2. Thus, Δf by the consideration is compensated by the phase compensation means 7 and 19.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a digital quadrature detection device used in digital radio communication, and more particularly to reducing the amount of calculation of a digital filter constituting the device.

[0002]

2. Description of the Related Art A conventional digital quadrature detector is shown in FIG.
, The IF signal input terminal 1 to which the reception signal S (t) converted to the intermediate frequency (IF) is input, and the reception IF
An A / D converter 2 for sampling the signal S (t) at a sampling frequency fs and converting it into a digital signal;
Output sample value sequence S (nTs) of the D converter 2 (where T
(s = 1 / fs) is band-limited to the multipliers 3 and 4 for multiplying the local signal for quadrature detection at the local frequency fs / 4, and the outputs Xi (nTs) and Xq (nTs) of the multipliers 3 and 4. Reception FIR low-pass filters 17, 18 and reception F
It has an in-phase output terminal 8 and a quadrature output terminal 9 for outputting the outputs of the IR low-pass filters 17 and 18 as baseband signals.

In this digital quadrature detector, the reception F
In order to reduce the calculation amount of the IR low pass filters 17 and 18, the center frequency f _{IF of} the input IF signal S (t) and the A / D converter 2
The sampling frequency fs of the A / D converter 2 is set so that the relationship of f _IF = (4k + 1) fs / 4 (k = 0,1,2, ...) (1) holds with the sampling frequency fs of In addition, the local frequency of the quadrature detection unit is set to fs / 4.

In this case, the input IF signal S (t) (= I
(T) cos2πf _IF t + Q (t) sin2πf _IF t)
Is converted into an output sample value sequence S (nTs) represented by the following expression (2) by the A / D converter 2.

S (nTs) = I (nTs) cos (πn / 2) + Q (nTs) sin (πn / 2) (2) Then, this signal is distributed with the frequency fc of the following expression (3) as the center frequency. .

Fc = mfs ± (fs / 4) (m = 0, ± 1, ± 2, ...) (3) As an example, f _IF = 5fs / 4 (when k = 1 in the equation (1)) The input and output frequency spectra of the A / D converter 2 at this time are shown in FIGS.

On the other hand, in the quadrature detector, the multiplier 3 is used to quadrature detect S (nTs) at the local frequency fs / 4.
At, multiplying S (nTs) by cos (πn / 2),
In the multiplier 4, sin (πn / 2) is added to S (nTs)
Multiply (n = 0, ± 1, ± 2, ...). The local signal cos (πn / 2), s of the local frequency fs / 4,
in (πn / 2) is (1, 0, -1, 0), (0, 1,
Considering that it is a periodic sequence having 0, −1) as one period, outputs Xi (nTs), Xq (nT) of the quadrature detection unit are taken into consideration.
s) is expressed by the following equations (4) and (5), and has a form in which 0 is interpolated every other sample.

Xi (nTs) = S (nTs) cos (πn / 2) = I (nTs) (1 + (− 1) ⁿ ) / 2 (4) Xq (nTs) = S (nTs) sin (πn / 2) ) = Q (nTs) (1-(-1) ⁿ ) / 2 (5) As a result, the output of the quadrature detection unit equivalently has a sampling frequency of fs / 2. Further, the center frequency of the output of the quadrature detection unit is fc ± (fs
/ 4), and mfs and mfs ± (fs / 2) (m =
0, ± 1, ± 2, ...). The frequency spectrum of the output of this quadrature detector is shown in FIG. 7 (c).

Next, mfs ± (fs / 2) (m = 0, ±) corresponding to the second harmonic component contained in the output of the quadrature detector.
In order to suppress the components (1, ± 2, ...), band limitation is performed by using the reception FIR low pass filters 17 and 18. As a result, baseband I and Q signals are obtained as shown in FIG.

At this time, the reception FIR low-pass filter 1
When the number of taps of 7 and 18 is L taps, the number of multiply-accumulation operations per output sample is reduced to L / 2 because the sampling frequency of the input signal is fs / 2.

As described above, in the conventional digital quadrature detector, the baseband signal is detected from the input IF signal while reducing the calculation amount in the reception FIR low pass filter.

[0012]

However, in the conventional digital quadrature detector, the center frequency f _{IF of the} input IF signal is
And the sampling frequency fs of the A / D converter are required to be satisfied by the equation (1). If this condition is not satisfied, a frequency offset may occur or the quadrature detection unit There is a problem that the 0 interpolation effect cannot be obtained and the amount of calculation in the reception FIR low-pass filter increases.

The present invention solves the above-mentioned conventional problems, and reduces the calculation amount in the reception FIR low-pass filter with respect to the input IF signal of an arbitrary center frequency while reducing the baseband signal. It is an object of the present invention to provide a digital quadrature detection device that can be obtained with high accuracy.

[0014]

Therefore, according to the present invention, the input IF signal is sampled at the sampling frequency fs, A / D-converted into a sampled signal having a center frequency fc, and then quadrature detected to obtain a baseband signal. In the digital quadrature detector, the A / D-converted sample value signal is quadrature-detected at the local frequency fs / 4 to obtain fc and fs.
Quadrature detecting means for generating an in-phase output and a quadrature output equivalently having a sampling frequency of fs / 2 with a frequency offset corresponding to the minimum value of the difference from / 4, and band-limiting the in-phase output and the quadrature output respectively. First and second
And a phase compensating means for compensating for the frequency offset contained in the outputs of the first and second band pass filters.

Further, the A / D-converted sample value signal is subjected to quadrature detection at a local frequency fs / 4 to obtain fc and fs /
Has a frequency offset corresponding to the minimum value of the difference from 4,
Quadrature detection means equivalently generating an in-phase output and a quadrature output having a sampling frequency of fs / 2, first and second band-pass filters for band limiting the in-phase output and the quadrature output, respectively, and first and second A symbol discrimination point detecting means for detecting a symbol discrimination point from the output sample value sequence of the second band pass filter, and a baseband for performing baseband differential detection using the value of the symbol discrimination point detected by the symbol discrimination point detecting means. The differential detection means and the phase compensation means included in the output of the baseband differential detection means for compensating the phase error caused by the frequency offset are provided.

Further, the sample value obtained by A / D conversion is represented by | f
Quadrature detection means for performing quadrature detection at a local frequency corresponding to the minimum value of c | to generate an in-phase output and a quadrature output, and a frequency mfs + (fs included in the in-phase output and the quadrature output.
/ 2) (m = 0, ± 1, ± 2, ...) The first and second low-pass filters that suppress the components of the interfering wave and the output sample value string of the first and second low-pass filters are set to 1 First and second thinning processing means for thinning out every sample and converting the sampling frequency to fs / 2, and first and second
First and second receiving low-pass filters for band limiting the output of the thinning-out processing means are provided.

[0017]

When the input IF signal with the center frequency f _IF is A / D converted at the sampling frequency fs, the center frequency of the signal after A / D conversion is fc = ((mD ± N) / D) fs (however,
D and N are positive integers and m = 0, ± 1, ± 2 ,. When the carrier frequency of this signal is regarded as fs / 4 and quadrature detection is performed, the in-phase output and the quadrature output are zero every other sample.
Is interpolated, and the sampling frequency is equivalently reduced to 1/2. Therefore, when the in-phase output and the quadrature output are input to the bandpass filter for filter processing, the number of product-sum operations of the L-tap bandpass filter is L.
It is reduced to / 2.

The output of this band-pass filter contains a frequency offset of Δf (= │fc- (fs / 4) │ minimum value) due to the above-mentioned "deem", and this frequency offset is phase-compensated. And the baseband signal with high precision is output.

Further, when a symbol identification point is detected from the output of the bandpass filter and baseband differential detection is performed using this symbol identification point, it is caused by the frequency offset Δf included in the output of the differential detection. The phase error θe = 2πΔfT is removed by the phase compensating means.

When the A / D-converted signal is orthogonally transformed at the local frequency having the minimum value of | fc |, the sampled sequence after the orthogonal transformation is passed through a simple low-pass filter and then thinned out. By doing so, the sampling frequency is converted to fs / 2, and the number of product-sum operations in subsequent filter processing is reduced to 1/2. This simple low-pass filter has a function of suppressing an interference wave component which may possibly overlap with the signal component after the thinning-out process when the interference wave is mixed in the input IF signal. Therefore, even if the IF signal includes an interference wave, a highly accurate baseband signal can be obtained with a small number of calculations in the reception filter.

[0021]

【Example】

(First Embodiment) As shown in FIG. 1, a digital quadrature detector according to a first embodiment outputs an IF signal input terminal 1 to which a received IF signal S (t) having a center frequency _fIF is inputted and the IF signal. An A / D converter 2 for sampling at a sampling frequency fs and converting it into a sample value signal, and a quadrature detection of an A / D converted sample value sequence S (nTs) (Ts = 1 / fs) at a local frequency fs / 4. Cos (πn
/ 2), sin (πn / 2) (n = 0, ± 1, ± 2,
, And receiving FIR band pass filters 5 and 6 for band limiting the outputs of the multipliers 3 and 4, and
A phase compensator 7 for compensating the frequency offset of the outputs of the reception FIR band pass filters 5, 6, a compensation function memory 19 for supplying a compensation function to the phase compensator 7, and a phase compensator 7.
The in-phase output terminal 8 outputs the in-phase output and the quadrature output as a baseband signal.

I input to this digital quadrature detector
The center frequency f _IF of the F signal S (t) is generally expressed by the following equation.

F _IF = (N / D) fs (N, D: positive integer) (6) At this time, the output sample value sequence S (nT of the A / D converter 2
s) is distributed with the frequency fc of the following equation (7) as the center frequency.

Fc = mfs ± f _IF = ((mD ± N) / D) fs (m = 0, ± 1, ± 2, ...) (7) As an example, operation when N = 9 and D = 8 Will be described. At this time, equations (6) and (7) are given by the following equations (8) and (9).
The frequency spectrum of the input and output of the A / D converter 2 is as shown in FIGS. 3 (a) and 3 (b).

F _IF = (9/8) fs (8) fc = (m ± (9/8)) fs (m = 0, ± 1, ± 2, ...) (9) At this time, S (nTs) Is expressed by the following equation (10).

S (nTs) = I (nTs) cos (2π (fs / 8) nTs) + Q (nTs) sin (2π (fs / 8) nTs) (10) In the quadrature detector, this S (nTs) is calculated. Local frequency f
In order to perform the quadrature detection at s / 4, the multiplier 3 causes the local signal cos (2π (fs / 4) nTs) = cos (πn /
2) is multiplied by S (nTs), and the multiplier 4 local signal sin (2π (fs / 4) nTs) = sin (πn /
2) is multiplied by S (nTs). This cos (πn /
2), sin (πn / 2) is (1, 0, -1, 0),
Considering that it is a periodic sequence having (0, 1, 0, -1) as one period, outputs Xi (nTs), X of the quadrature detection unit
q (nTs) has a form in which 0 is interpolated every other sample, and is expressed by the following equation.

Xi (nTs) = S (nTs) cos (πn / 2) = I (nTs) (cos (2π (fs / 8) nTs) + cos (2π (3fs / 8) nTs)) / 2 + Q (nTs) ) (Sin (2π (3fs / 8) nTs) -sin (2π (fs / 8) nTs)) / 2 (11) Xq (nTs) = S (nTs) sin (πn / 2) = I (nTs) ( sin (2π (3fs / 8) nTs) + sin (2π (fs / 8) nTs)) / 2 + Q (nTs) (cos (2π (fs / 8) nTs) -cos (2π (3fs / 8) nTs)) / 2 (12) The center frequency of the output of the quadrature detection unit is fc ± (fs / 4) after multiplication by the multipliers 3 and 4, and mfs + (11fs /
8) (= mfs + (3fs / 8)), mfs + (7fs
/ 8), mfs- (7fs / 8), mfs- (11fs
/ 8) (= mfs + (5fs / 8)) (m = 0, ± 1,
. ±., And the frequency spectrum of the output of the quadrature detector becomes as shown in FIG. 3 (c). Further, the sampling frequency is equivalently fs / 2. As shown in the figure, the output of the quadrature detector has a frequency offset Δf represented by the following equation.

The minimum value of Δf = | fc− (fs / 4) | = fs / 8 (13) Next, mfs + (3fs / 8) corresponding to the second harmonic component included in the output of the quadrature detection unit, mfs + (5fs / 8)
In order to suppress the component of (m = 0, ± 1, ± 2, ...),
Reception FIR band pass filters 5 and 6 having a center frequency Δf
To limit the bandwidth.

As a result, the frequency spectrum shown in FIG. 3 (d) is obtained. At this time, assuming that the number of taps of the reception FIR bandpass filters 5 and 6 is L taps, the number of multiply-accumulation operations per output sample is reduced to L / 2 because the sampling frequency of the input signal is fs / 2. To be done.

On the other hand, this filtering process is performed using the equation (1
1) and (12) are equivalent to removing the 3fs / 8 component, and filter outputs Yi (nTs), Yq (nT
s) is represented by the following equation.

Yi (nTs) = R (nTs) cos (2π (fs / 8) nTs + φ (nTs)) / 2 (14) Yq (nTs) = R (nTs) sin (2π (fs / 8) nTs + φ (NTs)) / 2 (15) R (nTs) = (I (nTs) ² + Q (nTs) ² ) ^1/2 (16) φ (nTs) = tan ⁻¹ (Q (nTs) / I (nTs) (17) In the phase compensator 7, the reception FIR bandpass filters 5 and 6 can be obtained by performing the operations of the following equations (18) and (19).
The frequency offset Δf = fs / 8 that occurs in the output of is compensated. Note that cos (nπ / 4), s in the equation
in (nπ / 4) is supplied from the compensation function memory 19.
Therefore, in the compensation function memory 19, a plurality of compensation function values cos (nπ / 4), sin (nπ) depending on the time n.
/ 4) is stored.

Ie (nTs) = Yi (nTs) cos (nπ / 4) + Yq (nTs) sin (nπ / 4) (18) Qe (nTs) = − Yi (nTs) sin (nπ / 4) + Yq (nTs) ) Cos (nπ / 4) (19) As a result, a baseband signal is obtained as shown in the following equations (20) and (21).

Ie (nTs) = R (nTs) cos (φ (nTs)) / 2 = I (nTs) / 2 (20) Qe (nTs) = R (nTs) sin (φ (nTs)) / 2 = Q (nTs) / 2 (21 ) Thus, the digital quadrature detection device of the first embodiment, the input IF signal, regardless of the value of the center frequency f _IF, quadrature detection at a local frequency fs / 4 , Equivalently converting the sampling frequency to fs / 2, and receiving FIR
The calculation amount of the bandpass filter is reduced. And
A highly accurate baseband signal is obtained by compensating the frequency offset of the filter output by the phase compensator.

(Second Embodiment) The digital quadrature detection apparatus of the second embodiment compensates for the phase error contained in the filter output after performing the baseband delay detection.

This device, as shown in FIG.
Output sample value string Yi of the R bandpass filters 5 and 6
(NTs), Yq (nTs) (Ts = 1 / fs), the symbol identification point detection unit 10 for detecting the values Yi (nT), Yq (nT) (T: symbol period) of the symbol identification points, and the symbol identification check. Cosine ID of the modulation phase difference from the output of the output unit 10
(NT) and sine QD (nT) generating baseband differential detection section 11 and output ID of baseband differential detection section 11
(NT), frequency offset included in QD (nT) Δ
The phase compensating unit 12 for compensating the phase error θe (= 2πΔfT) caused by f and the compensation coefficient memory 20 for supplying the compensation coefficient to the phase compensating unit 12 are provided. Other configurations are first
There is no difference from the device of the embodiment.

This device performs quadrature detection by the operation procedure described in the first embodiment, and the filters 5 and 6 have the formula (14).
~ Yi (nTs) and Yq (nTs) represented by (17)
Is output as a filter output.

Yi (nTs) = R (nTs) cos (2π (fs / 8) nTs + φ (nTs)) / 2 (14) Yq (nTs) = R (nTs) sin (2π (fs / 8) nTs + φ (NTs)) / 2 (15) R (nTs) = (I (nTs) ² + Q (nTs) ² ) ^1/2 (16) φ (nTs) = tan ⁻¹ (Q (nTs) / I (nTs) (17) The symbol identification point detection unit 10 uses Yi (nTs), Yq (n
The value Yi at the symbol identification point expressed by the following equation from Ts)
(NT) and Yq (nT) (T: symbol period) are detected.

Yi (nT) = Acos (2π (fs / 8) nT + φ (nT)) / 2 (22) Yq (nT) = Asin (2π (fs / 8) nT + φ (nT)) / 2 (23) A = R (nT) (24) φ (nT) = tan ⁻¹ (Q (nT) / I (nT)) (25) The baseband differential detection unit 11 calculates the following equations (26) and (27). Performs baseband differential detection.

ID (nT) = Yi (nT) Yi ((n-1) T) + Yq (nT) Yq ((n-1) T) (26) QD (nT) = Yi ((n-1) T ) Yq (nT) -Yi (nT) Yq ((n-1) T) (27) As a result, the cosine and sine of the modulation phase difference are generated as shown in the following equation.

ID (nT) = A ² cos (Δφ (nT) + θe) (28) QD (nT) = A ² sin (Δφ (nT) + θe) (29) Δφ (nT) = φ (nT) ) −φ ((n−1) T) (30) θe = 2πΔfT (Δf = fs / 8) (31) In the phase compensator 12, the calculation as in the following equations (32) and (33) is performed. By doing so, the phase error θe due to the frequency offset Δf generated in the output of the baseband delay detection unit 11 is compensated. Note that cos θe and si in the same equation
nθe is supplied from the compensation coefficient memory 20.

Ie (nT) = ID (nT) cosθe + QD (nT) sinθe (32) Qe (nT) = QD (nT) cosθe−ID (nT) sinθe (33) As a result, the baseband as shown in the following formula The signal is obtained.

Ie (nT) = A ² cos (Δφ (nT)) (34) Qe (nT) = A ² sin (Δφ (nT)) (35) Thus, the digital quadrature detection of the second embodiment is performed. In the device,
Quadrature detection is performed on the input IF signal at the local frequency fs / 4, regardless of the value of the center frequency f _IF of the input IF signal,
Equivalently, the sampling frequency is converted to fs / 2 to reduce the calculation amount of the reception FIR band pass filter.
A high-accuracy baseband signal is obtained by phase-compensating the frequency offset Δf of the filter output in the form of compensating the constant phase error θe after baseband differential detection. In this case, the compensation coefficient memory 20 only needs to store only two data cos θe and sin θe that do not depend on the time n as the compensation function value. Therefore, a large memory capacity is not required.

(Third Embodiment) The digital quadrature detection apparatus according to the third embodiment reduces the sampling frequency by thinning out the sample value sequence after the quadrature detection processing.

As shown in FIG. 4, this apparatus samples an IF signal input terminal 1 to which an IF signal is input and the input IF signal at a sampling frequency fs, and outputs a center frequency fc = mfs ± f _IF (m = 0. , ± 1, ± 2, ...), and an A / D converter 2 for converting the sampled value signal S (nTs) into a local frequency f.
_In order to perform quadrature detection with ₀ (= minimum value of | fc |), cos
Multipliers 3 and 4 for multiplying (2πf ₀ nTs) and sin (2πf ₀ nTs), and a simple FIR low-pass filter 13 for suppressing an interfering wave component mixed in the received IF signal,
14, thinning-out processing units 15 and 16 for thinning out the output sample value strings of the FIR low-pass filters 13 and 14 every other sample, reception FIR low-pass filters 17 and 18 for band-limiting the outputs of the thinning-out processing units 15 and 16, and the base. An in-phase output terminal 8 and a quadrature output terminal 9 for outputting a band signal are provided.

In this device, by providing the simple FIR low-pass filters 13 and 14, even if the interfering wave is mixed in the input IF signal, the baseband signal can be generated by removing it.

[0046] represented as the center frequency f _IF 1 of the center frequency f _IF 0 and disturbance IF signal of the desired wave IF signal input to the device generally following equation.

F _IF 0 = (N0 / D0) fs (N0, D0: positive integer) (36) f _IF 1 = (N1 / D1) fs (N1, D1: positive integer) (37) At this time, A / The desired wave component and the disturbing wave component in the output sample value sequence of the D converter 2 are distributed with the frequencies fc0 and fc1 of the following equations (38) and (39) as center frequencies.

Fc0 = mfs ± f _IF 0 = ((mD0 ± N0) / D0) fs (m = 0, ± 1, ± 2, ...) (38) fc1 = mfs ± f _IF 1 = ((mD1 ± N1 ) / D1) fs (m = 0, ± 1, ± 2, ...) (39) As an example, N0 = 9, D0 = 8, N1 = 11, D
The operation when 1 = 8 will be described. At this time, the formula (3
6) to (39) are expressed by the following equations (40) to (42), and the frequency spectrums at the input and output of the A / D converter 2 are as shown in FIGS.

F _IF 0 = (9/8) fs (40) f _IF 1 = (11/8) fs (41) fc0 = (m ± (9/8)) fs (42) fc1 = (m ± ( 11/8)) fs (42) (m = 0, ± 1, ± 2, ...) At this time, the output S (nTs) of the A / D converter 2 is expressed by the following equation.

S (nTs) = SD (nTs) + SU (nTs) (43) SD (nTs) = I (nTs) cos (2π (fs / 8) nTs) + Q (nTs) sin (2π (fs / 8) nTs) (44) SU (nTs) = Wi (nTs) cos (2π (3fs / 8) nTs) + Wq (nTs) sin (2π (3fs / 8) nTs) (45) However, SD (nTs) is a desired wave The component, SU (nTs), represents the interfering wave component.

On the other hand, the quadrature detector performs quadrature detection at the local frequency f ₀ expressed by the following equation. f ₀ = | fc ₀ | min = fs / 8 (46) Therefore, cos is calculated by the multiplier 3 for S (nTs).
(2π f ₀ nTs) is multiplied, and the multiplier 4 calculates sin (2π
f ₀ nTs). As a result, the outputs Xi (nTs) and Xq (nTs) of the quadrature detection unit are expressed as follows.

Xi (nTs) = S (nTs) cos (2πf ₀ nTs) = XDi (nTs) + XUi (nTs) (47) XDi (nTs) = SD (nTs) cos (2πf ₀ nTs) = I (nTs) / 2 + (I (nTs) cos (2π (fs / 4) nTs) + Q (nTs) sin (2π (fs / 4) nTs)) / 2 (48) XUi (nTs) = SU (nTs) cos (2πf) ₀ nTs) = (Wi (nTs) cos (2π (fs / 4) nTs) + Wq (nTs) sin (2π (fs / 4) nTs)) / 2+ (Wi (nTs) cos (2π (fs / 2)) nTs) + Wq (nTs) sin (2π (fs / 2) nTs)) / 2 (49) Xq (nTs) = S (nTs) sin (2πf 0 nTs) = XDq (nTs) + XUq (nTs) (50) X q (nTs) = SD (nTs ) sin (2πf 0 nTs) = Q (nTs) / 2 + (I (nTs) sin (2π (fs / 4) nTs) -Q (nTs) cos (2π (fs / 4 ) NTs)) / 2 (51) XUq (nTs) = SU (nTs) sin (2πf ₀ nTs) = (Wi (nTs) sin (2π (fs / 2) nTs) −Wq (nTs) cos (2π (fs) / 2) nTs)) / 2- (Wi (nTs) sin (2π (fs / 4) nTs) -Wq (nTs) cos (2π (fs / 4) nTs)) / 2 (52) where XDi (nTs) ), XUi (nTs) is the desired wave component and the interfering wave component in the in-phase output Xi (nTs), and XDq (nTs) and XUq (nTs) are the quadrature output X.
The desired wave component and the interfering wave component in q (nTs) are shown. Further, the frequency spectrum of the output of the quadrature detection unit is multiplied by the multipliers 3 and 4 to generate fc0, fc1 and f _0.
And are changed, as shown in FIG. 5 (c).

Next, when there is a possibility that the component of the interfering wave after the quadrature detection overlaps with the component of the desired wave after the thinning-out process, the interfering wave component is suppressed beforehand by the FIR low pass filters 13 and 14. . FIG. 5 (c) or formula (49)
As can be seen from (52), mfs + (fs / 2) (m
= 0, ± 1, ± 2, ...), an interfering wave component is generated, but this interfering wave component is generated when a thinning-out process described later for converting the sampling frequency to fs / 2 is executed. It is superimposed on the frequency of the desired wave component of the baseband after the thinning process (for example, fs, 2fs, 3fs, ...
The desired wave component corresponding to is the sampling frequency fs /
When converted to 2, fs / 2, fs, 3fs / 2, ...
These fs / 2 and 3fs / 2 overlap the interference wave component of mfs + (fs / 2). After the desired wave and the interfering wave are superposed, the interfering wave cannot be removed.

Therefore, this interference wave component is suppressed in advance by simple FIR low-pass filters (for example, about 7 taps) 13 and 14. After that, in the thinning processing units 15 and 16,
Set the output sample value string of the FIR low-pass filters 13 and 14 to 1.
Thinning out every sample, sampling frequency fs / 2
Convert to As a result, the frequency spectrum shown in FIG. 5D is obtained.

Next, as shown in FIG. 5D, for example, an unnecessary component other than the desired wave component distributed around the frequency mfs (m = 0, ± 1, ± 2, ...) Is received by the FIR low-pass filter 17, By suppressing at 18, it is possible to obtain a baseband signal as shown in FIG. At this time, the number of taps of the reception FIR low pass filters 17 and 18 is L
If it is a tap, since the sampling frequency of the signals input to the filters 17 and 18 is converted to fs / 2, the number of product-sum operations per output sample is reduced to L / 2.

As described above, in the digital quadrature detector of the third embodiment, the sampling frequency is converted to fs / 2 by performing the decimating process after the quadrature detecting process, and the number of multiplication and addition operations in the reception FIR low-pass filter in the subsequent stage. 1 /
It has been reduced to 2. Further, even when an interfering wave is mixed in the input IF signal, a part of this component that greatly affects the result is suppressed by a simple FIR low-pass filter,
After that, an excellent baseband signal can be obtained by performing the decimation process and the process of the reception FIR low pass filter.

[0057]

As is apparent from the above description of the embodiments, the digital quadrature detector according to the present invention calculates the amount of calculation of the reception FIR band pass filter regardless of the value of the center frequency f _IF of the input IF signal. It is possible to obtain a highly accurate baseband signal by reducing the frequency offset and compensating for the frequency offset.

Further, in the apparatus configured to compensate for the phase error θe after the baseband differential detection, it is sufficient to prepare two data independent of the time n as the compensation coefficient at the time of phase compensation, so the compensation coefficient is stored. It is also possible to reduce the amount of memory used.

Further, in the device for converting the sampling frequency into fs / 2 by performing the thinning-out process, even if the interfering wave is mixed in the input IF signal, this component is suppressed by the simple FIR low-pass filter and the An accurate baseband signal can be obtained with a small number of calculations.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a digital quadrature detection device according to a first embodiment,

FIG. 2 is a block diagram showing a configuration of a digital quadrature detection device according to a second embodiment,

FIG. 3 is a frequency spectrum diagram of each part in the digital quadrature detection device of the first and second embodiments,

FIG. 4 is a block diagram showing a configuration of a digital quadrature detection device according to a third embodiment,

FIG. 5 is a frequency spectrum diagram of each part in the digital quadrature detection device of the third embodiment,

FIG. 6 is a block diagram showing a configuration of a conventional digital quadrature detection device,

FIG. 7 is a frequency spectrum diagram of each unit in a conventional digital quadrature detection device.

[Explanation of symbols]

 1 IF signal input terminal 2 A / D converter 3, 4 Multiplier 5, 6, 17, 18 Receive FIR bandpass filter 7 Phase compensator 8 In-phase output terminal 9 Quadrature output terminal 10 Symbol discrimination point detector 11 Baseband delay Detection unit 12 Phase compensation unit 13, 14 FIR low-pass filter 15, 16 Decimation processing unit 19 Memory for compensation function 20 Memory for compensation coefficient

Claims (3)

[Claims] 1. An input IF signal is sampled at a sampling frequency fs.
In a digital quadrature detector that obtains a baseband signal by performing quadrature detection after sampling with, and performing A / D conversion into a sample value signal having a center frequency fc, the A / D converted sample value signal is converted into a local frequency fs.
Quadrature detection means for performing quadrature detection at / 4, having a frequency offset corresponding to the minimum value of the difference between fc and fs / 4, and equivalently generating an in-phase output and a quadrature output whose sampling frequency is fs / 2. A first for band limiting the in-phase output and the quadrature output respectively
And a second band pass filter, and a phase compensating means for compensating for the frequency offset included in the outputs of the first and second band pass filters. 2. An input IF signal is sampled at a sampling frequency fs.
In a digital quadrature detector that obtains a baseband signal by performing quadrature detection after sampling with, and performing A / D conversion into a sample value signal having a center frequency fc, the A / D converted sample value signal is converted into a local frequency fs.
Quadrature detection means for performing quadrature detection at / 4, having a frequency offset corresponding to the minimum value of the difference between fc and fs / 4, and equivalently generating an in-phase output and a quadrature output whose sampling frequency is fs / 2. A first for band limiting the in-phase output and the quadrature output respectively
And a second band-pass filter, a symbol identification point detection unit that detects a symbol identification point from the output sample value sequences of the first and second band-pass filters, and the symbol identification detected by the symbol identification point detection unit. A baseband differential detection means for performing baseband differential detection using a point value; and a phase compensation means included in the output of the baseband differential detection means for compensating for a phase error caused by the frequency offset. Characteristic digital quadrature detector. 3. An input IF signal is sampled at a sampling frequency fs.
In a digital quadrature detection device that obtains a baseband signal by performing quadrature detection after sampling with, and performing A / D conversion into a sample value signal of center frequency fc, the A / D converted sample value is set to the minimum value of | fc | Quadrature detection means for performing quadrature detection at a corresponding local frequency to generate an in-phase output and a quadrature output, and a frequency mfs + included in the in-phase output and the quadrature output.
First and second low-pass filters for suppressing the interference wave components near (fs / 2) (m = 0, ± 1, ± 2, ...), and output sample values of the first and second low-pass filters The column is thinned out every other sample, and the sampling frequency is f
It is provided with first and second decimation processing means for converting to s / 2, and first and second reception low-pass filters for band limiting the outputs of the first and second decimation processing means. Digital quadrature detector.