JPS5831060B2 - Single band frequency division multiplex signal demodulator - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention provides a single sideband frequency division multiplexing signal demodulator that converts between a single sideband frequency division multiplexed (hereinafter abbreviated as SSB-FDM) signal and a baseband signal by digital signal processing. Regarding.

Conventionally, analog balanced modulators and analog bandpass filters are used to convert a predetermined number of baseband signals into SSB-FDM signals or, conversely, to demodulate a predetermined number of baseband signals from SSIE3-FDM signals. is being carried out.

In contrast, in recent years, digital signal processing has enabled 88B
- Attempts are beginning to be made to realize modulation and demodulation of FDM signals.

The following are typical known methods for implementing modulation and demodulation of SSB-FDM signals through this digital signal processing.

(1) Publication “■E” published in the United States in December 1971
EETRANS ACTIONS ONCOMMUNIC
ATIONTECHNOLOGY, VOL, C0M-1
The paper "Systems Analysis of a TD" published on pages 1050 to 1059 of "9, 6"
M-FDM Translator/Digital
A-Type Channel Bank” (Reference 1)
.

(2) Publication rIEEEET published in the United States in September 1974
RANSACTNON8ONCOMMUNICATIO
N.S.

VOL, COM-22, Fu 9” page 1199-No. 12
The paper on page 05 “TDM-F DM Transmu
ltiplexer: Digital Polyp
hase and FFT” (Reference 2).

By modulating and demodulating SSB-FDM signals using such digital signal processing, it is possible to improve and make the transmission characteristics uniform, to make the equipment more compact and economical, and to simplify the manufacturing and maintenance of the equipment. 88B-FD
It is expected that mutual conversion between M signals and time division multiplexed (abbreviated as TDM) signals will become possible with high quality and economically.

By the way, in the modulation and demodulation method of SSB-FDM signals using digital signal processing, the equipment size and ultimately the equipment price are almost determined by the number of multiplications required per unit time, so in order to make the above expectations a reality, Therefore, it is required to obtain a modulation/demodulation method that requires essentially fewer multiplications per unit time.

SUMMARY OF THE INVENTION An object of the present invention is to provide a single sideband frequency division multiplex signal demodulation device using novel digital signal processing that requires fewer multiplications per unit time than conventional systems.

Another object of the present invention is to provide a single sideband frequency division multiplex signal demodulation device that can improve transmission characteristics and is easy to downsize and simplify manufacturing and maintenance.

Another object of the present invention is to provide a single sideband frequency division multiplex signal demodulator that can convert between FDM and TDM signals with high quality and economically.

The single sideband frequency division multiplexing signal demodulation device of the present invention comprises SS
The frequency fs (unit: Hz) whose sampling phase is sequentially shifted from the actual sample value series of frequency N-fs of the B-FDM signal
(will be omitted hereafter), create a series of 2N real sample values, and
The present invention is characterized in that these 2N real sample value sequences are subjected to arithmetic operations using a digital filter and offset discrete IJ processing to obtain N baseband signal sample value sequences of frequency fs.

Next, the principle and configuration of the single sideband frequency division multiplex signal modulation/demodulation apparatus of the present invention will be explained in detail with reference to the drawings.

In the following description, it is assumed that a sample value sequence generally means a sequence of digitized code words.

This digitized codeword sequence is converted into an analog sample value sequence, that is, a PAM (Pulse Amplitude Module) by a digital-to-analog converter.
ation) signal and can be further filtered to convert it into a continuous analog waveform.

In addition, conversely, a continuous analog waveform is sampled and PAM
It can be converted into a signal and further converted into a digitized sequence of sample values by means of an analog-to-digital converter.

Since the conversion between the analog waveform and the digitized sample value series can be easily performed in this way, the following explanation will be limited to the conversion process between the digitized sample value series. does not lose its generality.

FIG. 1 is a diagram showing an example of a single sideband frequency division multiplex signal modulation device that supplies a modulation signal to a demodulation device of the present invention;
The figure is a diagram showing the frequency spectrum of the signals of each part in FIG. 1.

In FIG. 1, input terminals 201, 202 . ...,
2ON includes N-channel baseband signal samples having a frequency spectrum with a periodic structure that repeats at frequency fs as shown in Figure 2A, which is obtained by sampling a baseband signal with a bandwidth of fs/2 or less at frequency fs. A value series is given, and an SSB-FDM signal sample value series having a frequency spectrum as shown in FIG. 2 is output from the output terminal 261.

For predetermined N/2 sequences among the N-channel baseband signal sample value sequences, the preprocessing circuit 200 performs a multiplication operation of (-1)n, as shown in FIG. 2B. Inverts the spectrum (however,
n is the time index).

The output of the preprocessing circuit 200 is connected to terminals 210, 21□...2
1N to the offset discrete Fourier processing circuit 210.

The offset discrete Fourier processing circuit 210 also includes c
c terminal 21N't 21. j-1-, 214, 21; f
tsramo N inputs are supplied, and the contents of these N inputs may all be O inputs, but the terminal 21A
.

21N-1,..., 21≦, 21Z to terminal 21Nt2
1N-1+...

In some cases, N inputs are obtained by commonly connecting terminals 21□ and 211 with corresponding numbers.

Kokote, Z = exp (j2yrf/CN・f s)
), and terminal 2
1.

212, -...21N, 21〈,...,21≦,21
! The Z transformation of the sample value series given by Xk(ZN)
, the Z transformation of the SSB-FDM signal sample value series appearing at the terminal 261 is denoted as Y (Z).

The arithmetic operations performed in FIG. 1 are based on the following equation.

In equation (1), Gi (-Z2N) is the Z transfer function of the real digital bandpass filter, and Ai (ZN) is the output of the offset discrete tool processing circuit 210.

The offset discrete Fourier processing circuit 210 performs the arithmetic operation of equation (2) on 2N inputs Xk(ZN) (k=0.1..., 2N-1) to produce 2N outputs Ai(ZN).
) (i=0.1..., 2N1).

If the phase offset exp(j2πi/4N) is extracted from equation (2), the remainder becomes a 2N-point discrete Fourier inverse transform equation.

Also, in equation (2), instead of (+1 to 2), -(2
+1), a phase offset of exp(-j2πi/4N) and a discrete Fourier transform equation of 2N points are obtained.

In this case, the same results can be expected. In short, there is no essential difference between discrete Fourier transform and inverse transform, and therefore, in the present invention, these are generally referred to as offset discrete Fourier processing.

When 2N (substantially N) inputs Xk (ZN) are given every sampling period 1/fs, the offset discrete Fourier processing circuit 210 performs the calculation of equation (2) and calculates 2N every '/fs. outputs Ai(ZN) are produced.

The output Ai (
The real part Re(Ai(ZN)) of zN) is then passed through a real digital bandpass filter (hereinafter abbreviated as RDBPF) 221+
1 (i20, 1..., 2N-1).

This RDBPF22 has an amplitude i for all i
-+-i This is a filter circuit in which the frequency characteristics are the same and only the phase frequency characteristics differ by an amount proportional to i, and the amplitude remains unchanged and only a phase shift occurs.

Filter circuits having such characteristics will be collectively referred to as polyphase circuits hereinafter.

According to equation (1), the signal passing through the RDB PF 221-+-t having the Z transfer function Gi(-Z2N) has Z-1
It is shown that the SSB-FDM signal Y(Z) can be obtained by adding these delayed signals for all i after giving a delay of .

To perform this operation, delay circuits 23□, 232 .・・・
., 23N, adders 240, 242...

24N and a parallel converter 250 are provided.

Delay circuits 230, 232. . . , 23N gives Z-ゞ, that is, a delay of one sample in the sample value series of the sampling frequency fs.

The parallel-to-serial converter 250 time-division multiplexes N sample value series of sampling frequency fs to obtain a sampling frequency of N−.
This means that one sample value sequence of fs is obtained, equivalently giving a delay of Z-t to the output of the adder 24i+t (i=0.1...,N-1).

RDBP F 22N+1.22N +2. ..., 2
Since 22N is further delayed by the delay circuits 231, 23□, ..., 23N before it is input to the parallel-to-serial converter 250, all of the RDBPFs 221
+1 (i=0.1...

2N-1), a delay of Z-1 is given.

RDBP among the 2N sample value series after giving a delay of Z-1 to the output of RDB P
Since the sampling time slots of the outputs of F22i+1 and 22N+i+1 overlap, >a of both outputs is taken in adder 241+1.

The fact that the SSB-FDM signal is obtained by the calculations of equations (1) and (2) performed in FIG. 1 will now be explained with reference to FIG.

In FIG. 3, reference numeral 420 is the Z transfer function G of RDBPF22□.

R DB P F221+1 based on (-Z2N)
of the Z transfer function Gi (-Z2N) (i.e., Gi (
-Z2N)/Go (-z2N)), reference numeral 420 indicates the phase characteristic (with respect to zO) of a 2-1 delay circuit equivalently entered after RDBPFGi (-Z2N).

Combining the phase characteristics of reference numerals 410 and 420 yields
The overall phase characteristic is a step-like characteristic indicated by reference numeral 430.

This is the output terminal of the polyphase circuit (relative)
This is the overall phase characteristic.

The overall (relative) phase characteristic from the input of the offset discrete Fourier processing circuit 210 is obtained by adding the phase shift according to equation (2) to the characteristic indicated by reference numeral 430.

Reference numeral 440 is the overall (relative) phase characteristic when 20 is set in equation (2), reference numeral 450 is -1, and reference numeral 460 is the overall (relative) phase characteristic when 2 is set as 2. shows.

When considering such phase characteristics, it is assumed for convenience that all operations after the offset discrete Fourier processing circuit 210 are complex operations, and only the real part of the operation result is finally taken.

Of course, in reality, the offset discrete Fourier processing circuit 2
Since the operations after 10 are performed independently on the real and imaginary parts, it is clear that the operations on the imaginary part, which is discarded at the end, need not be performed from the beginning.

According to the overall phase characteristic of FIG. 3, k-0, that is, input signal X.

Band (0 to fs/2) for (ZN). k-1, that is, the band (fS/
2 to fs), i.e., for input signal X2 (ZN) only in the band (fs to 3fs/2) (relative)
The phase becomes O regardless of i, and in other bands the (relative) phase takes a value that is an integral multiple of 2πi/2N.

When the phase is an integer multiple of 2πi/2N, the vector sum for all i=0, 1° . . . 2N cancels out and becomes O.

On the other hand, if the phase is independent of i, the vectors are added in phase.

As a result, the input signal X. (ZN) is the band (O~, fS/
2), XI (ZN) is the band (fs/2 to fs), X2
(ZN) is the band (fs~3 fs/2),...
It occurs only in these bands and disappears in other bands.

Also, in equation (2), (+1 to 2) is changed to -(+1 to 2)
Even so, it can be easily determined from FIG. 3 that the signal remains only in the band uniquely determined by the value of k.

As a result, in FIG. 1, the input signal X.

For (ZN), Figure 2 C1 human power signal XI (ZN)l
..., the second diagram for X2N (ZN), ...
, G, respectively, are obtained which are equal to the transfer characteristics of a complex digital bandpass filter (abbreviated as CDBPF).

Therefore, the input signal X.

(zN) , X2(ZN) ,...,XN-2(Z
N) is Fig. 2 A1 human power signal X1 (ZN), X3 (ZN)
.

. . ,
N+1(ZN),...t O for X2N-1(ZN)
is input, the output obtained from the output terminal 261 in FIG. 1 will have a frequency spectrum as shown in FIG. 2 H.

However, since the output that is actually required is only the real part of the complex numbers as described above, it has a frequency spectrum as shown in FIG. 2 (2).

After decoding, the sample value series having the spectrum shown in Figure 2 (■) is passed through an analog bandpass filter with a passband of N'fs/2 to N4s.
A signal spectrum is obtained.

Furthermore, if necessary, a signal in the band O"-N-f s/2 can be extracted.

Furthermore, when the direction of the spectrum is required to be reversed, the above-mentioned preprocessing can be performed on even channels in the opposite direction. 'fs) is assumed to have a fixed relationship, but by assuming an appropriate number of dummy channels in the number N of multiplexed channels, the relationship between the actual multiplicity of the SSB-FDM signal and the sampling frequency can be changed freely. can be changed to

As a concrete example, considering a 60-channel FDM supergroup,
This SSB-FDM signal exists in the 312-552 KHz band.

Further, the bandwidth fs/2 of one channel is 4 KHz.

In this case, 12 dummy channels are assumed to be N near 2, and the band of the 72 channels of 5SB-F'DM signal is 288 to 288 (including the above 312 to 552 KHz).
If it is 576KH2, the sampling frequency is N-f
s=576 KHz, and no spectrum overlap occurs.

As will be described later, when the FFT operation is used, the number of multiplications can be particularly reduced when N is a power of two.

In the above example, to make N a power of 2, four dummy channels can be included to make N264, and the sampling frequency can be N'fs=512 KHz.

In this case, it is necessary to shift the frequency of the supergroup band 312-552KH2 to, for example, the 8-248KH2 band in advance.

Next, an embodiment of the 5SB-immediate M signal demodulation device according to the present invention will be described with reference to FIG.

When a sample value series Y (Z) of frequency N' f s of the SSB-FDM signal is given to the input terminal 561, first, in the serial/parallel converter 550, the given temporally serial sample value series is is converted into a series of N sample values in parallel.

This operation involves a substantial delay, but here, terminal 5
The delay from 61 to terminal 54□ is Zo, the delay from terminal 561 to terminal 542 is Z-1, and from terminal 561 to terminal 54N
Assume that the delay until then is Z - (N-1).

That is, it is assumed that the sample value inputted earliest in terms of time appears at the terminal 541, and the sample value inputted earliest in terms of time appears at the terminal 54N at the same time.

Terminals 541, 542. . . , 54N are RDBPFs 520, 522 . . . , 54N, respectively.・・・
, 52N, and delay circuits 531, 532., which provide a delay of Z-N. ..., RDBPF52 N+1 via 53N.

52N+2. ..., is input to 522N.

From the input terminal 561 to RDBPF521-1-t (i=0
．． 1, 2° .

RDBPF 521-4-t (i=0.1, 2...
・, 2N-1) sample value series at the input point is YN-t
−i(Z).

The Z transfer function of RDBPF 52□+1 is the above-mentioned Gi(z
2N).

The output of RDBPF52 is then offset discrete i
+1 is given to the Fourier processing circuit 510.

If the output of RDBPF521+1 is Xl(z, then A'i (Z)-G i (-Z2N) ・'YN-1
-i (Z) (3) and the offset discrete Fourier processing circuit 510 is! It is defined as finding the output Xk (71) from Vi (Z) using the following equation.

The 2N outputs of the offset discrete Fourier processing circuit 510 are connected to terminals 511, 512 . ..., 51N, 51'N, 5
1'N-1...51'1.

The fact that the SSB-FDM signal can be demodulated in the embodiment shown in FIG. 4 can be explained with reference to FIG. 3 again.

As in the case of Fig. 1, the Z transfer function Gi(-Z2N)/G
The phase characteristics of o(-Z2N) are as indicated by reference numeral 410 in FIG.
Since the delay until input to F52i-H is Z-', its phase characteristic is represented by reference numeral 420 in FIG.

Therefore, the overall phase characteristic of the delay of Z-' and the RDBP F521-h is as shown by reference numeral 430 in FIG. 3 (however, the phase characteristic of the RDBPF 52 is used as a reference).

Furthermore, considering the phase shift caused by the offset discrete Fourier processing circuit 510, from the input terminal 561 to the terminal 51
0, terminal 51□, terminal 513. The overall phase characteristics up to... are shown in Figure 3 with reference numbers 440, 450 and 460, respectively.
,... is shown.

As is clear from equation (4), since Xk(η is given as a vector sum for all i, the phase is 2πi
/(2N), the vector sum is O regardless of the amplitude value.

Therefore, even if an SSB-FDM signal covering the entire band is given, only the signal component in the band where the phase is 0 is extracted.

That is, the output X. (η is the phase characteristic 44 shown in Figure 3.
Only signals within the band (0, fS/2) from 0, output X1 (
Zl has a band (fs/
2. fs) will be extracted.

This means that the signal in Figure 2 (■) can be converted to Figure 2 C2D...
This is equivalent to passing the signal through a G filter and taking the real part of the output.

As in the case of the modulator described above, (2 in equation (7) is +1
It goes without saying that it is also possible to set ) to -(2 +1).

The output rk(Z) is then sampled at the frequency fs, and □
If a sample value series of frequency fs (ZN) is obtained, its spectrum will be as shown in FIG. 2 A or B.

Since sampling at the frequency fs means that only one out of N samples of the output Xkη is effectively used, the offset discrete Fourier processing circuit 5
There is no need to perform arithmetic operations on unnecessary samples in RDBPF 10 and RDBPF 521+1.

As a result, the output of the serial-to-parallel converter 550 can be a sample value series of frequency fs.

Terminals 541, 542. ..., 54N, even if it is a sample value series of frequency fs,
Since all the sample values of the SSB-FDM signal applied to the terminal 561 are included without excess or deficiency, there is no loss of information due to this.

Therefore, YN-1-1(Z), Ai in the above explanation
(Z)X'k(7> is hereafter referred to as YN-1-1(ZN)
, A/1(ZN), X4c(ZN) and delay time, bribe 530, 53□,..., 53N.

It can be considered that the RDBPF 52□+1 and the offset discrete Fourier processing circuit 510 all operate at the sampling frequency fs.

Outputs X'k (ZN) with ≧N are unnecessary, and the endogenous number of X'k (ZN) where k<N causes frequency inversion, so in the post-processing circuit 500, For the sample value series causing these frequency inversions (-1
) The original baseband signal can be restored by multiplying by n.

Terminals 501, 502. .... In this way, N-channel baseband signal sample value sequences are obtained from the 5ON.

Next, offset discrete Fourier processing circuits 210 and 5
10 will be explained in more detail.

Offset discrete Fourier processing circuit 210 in FIG.
As shown in the calculation formula (2), when applying 0 manual force to the terminals 21, ..., 21'2, 211 as described above, for k≧N, the input Xk (ZN)20 The output Ai (ZN) of the circuit 210 is the sum of N terms.

Also, since all inputs Xk (ZN) are real signals,
There is a relationship of 4N-i(ZN)-Ai''(ZN) (. is a conjugate complex number).

Furthermore, only the real part needs to be found, and Re(AN(
ZN))-〇 is clear and Lucara, so in the end, instead of equation (2), i = 0.1..., the following equation (5
) or equation (6), and for i>N, Re(A
2N-, (ZN))-Re(A1(ZN)E relationship).

In addition, when N can be further decomposed into products of prime numbers, the fast Fourier transform (FF"T) method, which is well known as time decimation or frequency decimation, is applied to equation (5) to reduce the number of multiplications. can do.

Apart from the above method, the terminals 21A, . . . , 216, 2
11' to terminals 21N, . . . , 212, 21, respectively.
It is also possible to connect X2N-t-k(ZN)=Xk(ZN) and perform the calculation of equation (2) using the circuit shown in FIG.

In FIG. 5, reference numerals 211.212. ..., 2
1N.

21Z, 21 child, ..., 21 i, 210 and reference alphabetic character A.

, A1°...j A2N-1 have the same meaning and content as the corresponding ones in FIG.

Reference numeral 610 is an N-point real input discrete Fourier converter circuit that performs the following calculation, and for its N complex outputs, reference numeral 621-4-1 (i = 0.1..., N-1
), exp(j2
πi/4N).

Since only the real number output should be obtained by the offset discrete Fourier processing circuit 510, the phase offset circuit 62i+
For t, calculations related only to real number outputs may be performed.

The output of the phase offset circuit 62i+t is Ai(
1==0. i, . . . , N-1) and simultaneously input to the output circuit 630.

The output circuit 630 generates AN+1t AN+21 . . . , -1 from Ai by setting A2N-1=A1 and outputs it.

Note that the adders 241, 242 . ..., 24N
If subtraction is performed in the output circuit 630, it is only necessary to connect the output circuit 630 so that A2N-1=Ai.

The N-point real-input discrete Fourier (inverse) transform circuit 610 in FIG. 5 only needs to perform half the number of multiplications performed by a normal N-point complex manual discrete Fourier (inverse) transform circuit.

The number of multiplications in real force FF'T is expressed as general complex human force F'
There are already some known examples of methods that reduce the number of FT multiplications by half, so detailed explanations will be omitted here.

One of these known examples is detailed in the following document, for example.

Publication "Commun i -" published in October 1968
cations of the ACM, Voll,
410J, pages 703-710”A
Fast Fouier Transform Algo
rithm for Real-ValuedSeri
es” (Reference 3).

Note that the fact that the circuit in FIG. 5 performs the offset discrete Fourier processing shown by equation (2) means that
)Xk(ZN), it is confirmed that equation (2) is equivalent to the following equation.

Next, the offset discrete Fourier processing circuit 51 in the embodiment of the single sideband frequency division multiplex signal demodulation device shown in FIG.
It will be explained that 0 can also be realized by the configuration shown in FIG.

In FIG. 6, reference numbers 510, 511, 51□.

..., 51N and reference alphabetic characters A≦, AI, ..., A
! M-1, wind.

X'1. ...,X! 2N-1 represents the same meaning and content as the corresponding one in FIG.

The input A′i is i=0.1. ..., A1 for N-1
N complex numbers (A'1(
ZN)+jA/N+i (zN)) is added to the phase offset circuit 71,+1.

In the phase offset circuit 711+1, exp(j2πi/
4N) is multiplied by B10(ZN) - (A'1(ZN)
-1-JAli(ZN))ehojthme4N) (9) is calculated.

Complex output B/ of phase offset circuit 71i-h, □□□
N) is then provided to an N-point discrete Fourier (inverse) transform circuit 720 to calculate .

The calculation in the offset discrete Fourier processing circuit 510 is to obtain the following, but since A'1(ZN) is a real signal, the relationship of X'2N-0-k(ZN)-X'k(ZN) is Yes, in the end, X2 becomes zN) or +1(Z
It is sufficient to find the inclination of N), and it can be seen that the component 510 in FIG. 6 can be replaced with the component 510 in FIG. 4.

In Equation 00), when N can be decomposed into products of prime numbers, the N-point discrete Fourier (inverse) transform circuit 720 in FIG.
This can be realized with a small number of multiplications using the FT calculation method.

Furthermore, since only the real part is required for the output of the N-point discrete Fourier (inverse) transform circuit 720, the number of multiplications can be reduced by omitting all operations related only to the imaginary part output.

Next, the polyphase circuit will be explained in detail.

Polyphase circuit consists of 2N RDBPF220.22□
.

..., 222N or 52□, 522. ..., 52
□.

The Z transfer function of RDB PF 221-1-t is Gi(-
G
(Z), it can be derived from G(Z).

That is, if standards for in-band transmission characteristics and out-of-band attenuation are given for an RDLPF having a Z transfer function G(power), and G(Zl is designed to satisfy those standards,
Next, Gi(Z2N) that satisfies the following equation is determined from G(Z).

Equation (12) shows that after giving a delay of Z-1 to the impulse response of G i (z2N), the sum of all i is G(Z)
This means that the impulse response of

In other words, the impulse response of Gi(Z2N) is (G(
This is equivalent to sampling the impulse response of Z).21) at every 2N sample points, that is, at a frequency of fs/2.

Therefore, if the amplitude frequency characteristic of G(Z) is as shown in FIG. 7A, the amplitude frequency characteristic of Gi(Z2N) is as shown in FIG. 7B.

In addition, the phase frequency characteristic of Gi(Z2N)/Go(Z2N) is exp (j2r, fi/N−f s ) within the band.
), and due to the periodicity due to sampling, the overall characteristic is as shown in FIG. 7C.

Z of RDB P F 22 i +1 in FIG.
The transfer function Gi (Z2N) is defined by the Z transfer function oi (z2N) defined by equation 02), where all z2N is -z2N
is equivalent to replaced by .

Conversion from Gi (Z2N) to Gi (-z2N) is fs7
4 frequency shift, Gi(z2N)
Figure 7 shows the amplitude frequency characteristics of D% G, (-Z2N)/G.
The phase frequency characteristic of o(-Z2N) is as shown in FIG. 7E.

In fact, if G(Z) is designed to satisfy desired standards such as in-band ripple and out-of-band attenuation, then G(
When Z) consists only of acyclic parts, G(
To obtain Gi(Z2N) from Z), the filter coefficients of G(Z) may be thinned out so as to satisfy the equation (1).

This operation is extremely simple. If RDLPF G(Z) also has a cyclic part and is generally given by the Z transfer function of the following formula consisting of M poles and M zeros, apply the identity % formula %) to the denominator. can be rearranged into the form of equation 02) as shown below.

The first term in the above equation is G.

(z2N), and the part excluding z−i from the second term onwards is Gi(Z
2N) (i=1, 2..., 2N-1).

Z transfer function G of RDBPF22i+1 or 52i+i
1(-Z2N)(i-0,1,...,2N-1) is Z2N in the formula of Gi(Z2N) obtained in this way.
It can be obtained by replacing.

A circuit implementation method of such an RDBPF can be implemented using a conventionally known technique.

By the way, RDBPF221-1-t or 52i+t
As can be easily understood with reference to equation (2), the Z transfer function Gi(-Z''N) has a common denominator for all i (that is, independent of %i).

In other words, if Qi (-z2N) is divided into a numerator and a denominator, then Gi (-Z2N) Pi (-z2N)/Q(-
Z”N).

It can be seen that Q(-Z2N) is common to all i, and it is the numerator part that determines the phase characteristics specific to the polyphase circuit.

From this fact, the calculation of 1/Q(-Z2N) can be calculated using RD in Figure 1.
B P F 221+t or RDBPF52H- in Figure 4
It also becomes possible to separate from the position t.

That is, in FIG. 1, if the 1/Q(-Z''N) calculation circuit is moved after the adder 24 i-h or before the offset discrete Fourier processing circuit, N calculations will be performed instead of 2N calculations. It can be done by calculating.

Further, referring to FIG. 5, even if the calculation of VQ(-Z2N) is performed after the complex multiplication circuit 62i+1, only N calculations are required.

Although it is equivalent to performing after the adder 241+1, it may also be performed after the parallel-to-serial converter 250.

In the case of FIG. 4 as well, if the 1/Q(-Z2N) operation is performed before or after the serial-to-parallel converter 550 or after the offset discrete Fourier processing circuit, the amount of necessary calculations can be halved. .

As described above in detail, the single sideband frequency division multiplexing signal demodulation device according to the present invention basically includes an offset discrete Fourier processing circuit operating at the sampling frequency fs and two
It is characterized in that it can be realized by cascading a polyphase circuit composed of N real digital bandpass filters, and converts a baseband signal into a complex signal, as seen in the known examples shown in Documents 1 and 2 above. No operation required.

This eliminates the need for arithmetic operations for converting baseband signals and complex signals, and the input/output of the offset discrete Fourier processing circuit is only the actual signal, so overall the number of multiplications per unit time can be reduced. It turns out.

In addition, calculation operations are performed in two stages: offset discrete Fourier processing and polyphase, making it easier to downsize the device and simplify manufacturing and maintenance.Furthermore, the in-band shifting characteristics are determined only by the RDBPF characteristics of the polyphase circuit. Therefore, it has the advantage that it is easy to improve transmission characteristics.

The present invention also provides a baseband signal sample value series and an SS
Since it allows conversion between B-FDM signal sample value sequences, it can be applied very directly to the purpose of conversion between TDM and FDM signals.

In the above explanation, it was assumed that the input and output sample value series of the offset discrete Fourier processing circuit and the polyphase circuit exist in parallel, but a small number of FFT basic calculation elements, complex multiplication circuit, RDBPF', etc. It goes without saying that by using division multiplexing, the functions shown in FIG. 1 or 4 can be realized for any number of songs.

In fact, in Fig. 1, when N-channel baseband signals are input serially in time, the preprocessing circuit, offset discrete Fourier processing circuit, RDBPF, adder, etc. can be operated in series, and the Series converter 2
50 becomes unnecessary.

[BRIEF DESCRIPTION OF THE DRAWINGS] FIG. 1 is a diagram showing an example of a single sideband frequency division multiplexing signal modulation device that supplies a modulation signal to a demodulation device of the present invention, and FIGS. 2 and 3 are single sideband frequency division multiplexing signal modulation devices according to the present invention. FIG. 4 is a diagram showing an embodiment of the single sideband frequency division multiplexing signal demodulation device according to the present invention, and FIG. 5 is a diagram for explaining the operation of the single sideband frequency division multiplexing signal demodulation device, and FIG. FIG. 6 is a diagram showing an example of an offset discrete Fourier processing circuit used in the embodiment of FIG. 4, and FIG. 7 is a diagram for explaining a polyphase circuit used in the present invention. This is a diagram. 1 and 4, reference numeral 200 is a preprocessing circuit, reference numerals 210 and 510 are offset discrete Fourier processing circuits, and reference numeral 22□. 22□,..., 222N and 520,522.・・・
・, 522N is a real digital bandpass filter that constitutes a polyphase circuit, reference numbers 234, 23□, . . .
22N and 53□. 53□, ..., 53N are delay circuits, reference number 24□,
242°..., 24N is an adder, reference numeral 250 is a parallel-to-serial converter, reference numerals 201, 202, . ..., 2ON
is an input terminal for a base punt signal sample value series, reference numeral 261 is an output terminal for an SSB-FDM signal sample value series, reference numeral 561 is an input terminal for an SSB-FDM signal sample value series, reference numeral 550 is a serial-to-parallel converter, Reference numeral 500 represents a post-processing circuit and reference numerals 501, 502 .
. . , 5ON indicates the output terminal of the baseband signal sample value series.

Claims (1)

[Claims] 1 means for obtaining a sequence of 2N real sample values of frequency fs from a sequence of single sideband frequency division multiplexed signal sample values of sampling frequency N-fs; A polyphase circuit consisting of a real digital bandpass filter, and an offset discrete Fourier process that is supplied with a series of 2N real sample values at a frequency fs from the polyphase circuit as an input and outputs a series of N real sample values at a frequency fs. A single sideband frequency division multiplex signal demodulator comprising a circuit.