JPH0779207A - Frequency multiplexing separation circuit with digital filter - Google Patents

Abstract

PURPOSE: To obtain a multiplex/separation circuit allowed to be applied to the variation of traffic derived from changes in both of the number of channels and band width and capable of utilizing a multi-phase network. CONSTITUTION: After sampling a multi-carrier signal SM, a group of modulated and frequency-multiplexed carriers having the same band width is separated by filters 1021 , 1022 and their burst frequency bands are declined by decliners 1021 ', 1022 ' depending upon their own band width. The multi-phase networks 105a to 105d annexed to a delay line stage 104 execute the processing of the indivisual group of frequency-multiplexed carriers having the same band width by programming one Fourier transforming composite multiplex circuit 1055 . Respective multiplex/separation circuits 102 to 105 can be programmed so as to deal with traffic variation.

Description

Detailed Description of the Invention

[0001]

FIELD OF THE INVENTION The present invention relates generally to frequency multiplexing / demultiplexing by digital processing.

[0002]

BACKGROUND OF THE INVENTION The first digital frequency multiplexing / demultiplexing circuit was used at the heart of a transmultiplexer to transfer data between multiplexed lines of different nature inside the telephone network. It was to solve the problem. For example, such a transmultiplexer is an analog telephone line 60 having a frequency-multiplexed bandwidth of 4 kHz.
Second-order group (GS) multiplexing consisting of books and 2048 Mbi
Secure a connection between two t / s primary MIC (TN1) multiplex transmission lines. Thus, very diagrammatically, a transmultiplexer is a frequency division multiplex transmission line (MR).
F) and the time division multiplexing transmission line (MRT).

In the field of satellite communications, this transmultiplexer technology has proved to be of particular interest for satisfying two conditions for optimal utilization of satellite communications networks. For one, frequency division multiple access (AMRF) for uplink connections set up from ground stations to satellites saves the transmit power and transmit antenna size required by the ground station, It has proved to be a promising technology in terms of providing greater flexibility regarding the location of ground stations. On the other hand, for downlink connections set up from satellites to ground stations, time division multiplexing (in order to avoid intermodulation noise when multiple carriers modulated by the transmitted signal are sent by the satellite, AMRT) technology is recommended. This intermodulation noise increases as the number of carriers increases. This intermodulation noise problem does not exist if only one carrier is modulated by one digital frame of a time division multiplex access.

The satellites used to perform this kind of conversion between frequency division multiplex access and time division multiplex access are commonly referred to as regenerative satellites, because they are , Because it "regenerates" the received signal in the form of a multiplexed line as a function of each individual frequency band of signal components.

A satellite microstation communication network in which connections between the microstation and the central office and between the central office and the microstation are ensured by one time division multiple access carrier and one time division multiplex transmission line carrier, respectively (English and American) The usage of
Very Small Aperture Terminal (VSAT))
Compared to, the technology of the transmux onboard the regenerative satellite reduces the transmission time between the two stations by half, because the data transmission between the stations is done directly without going through the central station. is there.

Despite all of the above advantages derived from the use of regenerated satellites, satellite developers are somewhat hesitant about their use.

The reason must be satisfied for optimal utilization of the total bandwidth assigned to the communication network,
In the prior art, there are two requirements that are completely contradictory to each other: the two are

The transmultiplexer must be programmable from the ground in a "simple" manner so that it can be reallocated in response to changes in traffic in the satellite network, and

The transmultiplexer has to receive frequency-multiplexed signals with different bandwidths.

The core component of the transmultiplexer is the frequency multiplexing / demultiplexing circuit that is the subject of programming from the ground. In the prior art, various structures of demultiplexing circuits based on digital filters have been proposed. The notable ones are:

Implementation method by individual processing of channels. This is very complicated and expensive to implement as the number of frequency-multiplexed carriers increases, and
Has the drawback that programming becomes complicated;

Implementation of the "Fast Fourier Transform". This has the drawback that it is only really efficient when the number of carriers is large, otherwise it takes a long processing time compared to the number of carriers;

Implementation of a polyphase network consisting of delay lines and digital filters. This cannot be dealt with in the prior art if the frequency-multiplexed signals have different bandwidths.

[0014]

The present invention can be used for reconstructing initially frequency division multiplexed channels having different frequency bandwidths, and thus, in particular,
The purpose of the present invention is to provide a demultiplexing circuit using a polyphase network which can adapt to fluctuations in traffic resulting from changes in both the number of channels and the bandwidth.

[0015]

To this end, the circuit according to the invention demultiplexes a frequency-multiplexed channel in a multi-carrier sampled signal having a predetermined effective frequency bandwidth. A channel is defined by the modulated carrier contained in the effective frequency band and the frequency bandwidth juxtaposed in the effective frequency band. The frequency bandwidth of the channel is an integral fraction of the maximum frequency bandwidth of the channel, and the latter itself is also an integral fraction of the effective bandwidth and an integral multiple of the minimum bandwidth of the channel.

The demultiplexing circuit is similar to the known circuit with polyphase networks in that it has a predetermined number equal to the ratio of the effective bandwidth to the minimum bandwidth of the channel, one delay line and one delay line, respectively. Each of the above filters includes a circuit of parallel placed digital filters in series and a Fourier transform digital processing mechanism for individually reconfiguring the above channels connected to the outputs of the filters. Is derived from one of the digital low-pass filters which filters within the minimum band of the above channel.

However, known circuits with polyphase networks can only process channels with the same bandwidth. In other words, the minimum and maximum bandwidth are equal to the bandwidth of all channels and each set of delay line and digital filter is
It is assigned to process one channel.

According to the invention, the demultiplexing circuit comprises:

A plurality of filtering and frequency-decreasing mechanisms for separating channels in a multi-carrier sampled signal into a predetermined number of channel groups, each channel group of the other channels Includes juxtaposed bands of all channels with the same bandwidth that are different from the bandwidth,

A plurality of parallel delay lines, the number of which is a predetermined number equal to the ratio of the effective bandwidth to the maximum bandwidth of the channel, and each channel group has a bandwidth of the group and a maximum bandwidth of the channel. Is delayed by a number of parallel delay lines each of which is greater than the ratio and is equal to the nearest integer, and

A plurality of identical parallel polyphase networks, each connected to the output of a delay line, each polyphase network comprising a plurality of said parallel delay lines and digital filter pairs, the maximum bandwidth of the channel; Contains a predetermined number equal to the ratio of

By arranging the sets of delay lines and digital filters independently in the same polyphase network, the polyphase network is distributed between maximum and minimum bandwidths with different bandwidths. It becomes possible to process channels. The orientation of channels to these polyphase networks is realized as a function of the characteristics of the channels within the groups, ie the number and bandwidth of the channels within each group, but especially the filtering and frequency reduction prior to the polyphase networks. The characteristics of the filtering and decrementing and delaying of the mechanism and the delay line, as well as the characteristics of the matrix in the processing mechanism of the Fourier transform, are adapted to the characteristics of the channels in the group.

These adaptations in the demultiplexing circuit according to the invention manifest themselves in the following properties:

The filtering and frequency-decreasing mechanism for separating one channel group from the multi-carrier signal is a digital filter having the bandwidth of the group, and then the bandwidth and effective bandwidth of the group. Including a reducer having a reduction ratio equal to the ratio;

Parallel delay lines for delaying a group of channels, each delayed in proportion to an integer multiple of zero or more equal to the ratio of the effective bandwidth to the bandwidth of the group;

The delay lines included in the polyphase network are each delayed by a multiple of an integer greater than or equal to zero equal to the ratio of the maximum bandwidth of the channel to that of the minimum;

The digital processing mechanism includes, for each channel group, a plurality of square matrix mechanisms of the coefficients of the Fourier transform, the number of which is equal to the ratio of the bandwidth of the group to the minimum bandwidth, and which is attached to the group. The order of each of the schemes is equal to the number of channels in the group.

The various characteristics described above allow the demultiplexing circuit to adapt to variations in traffic. In other words, the demultiplexing circuit is rather programmable.
In this case, a filtering / decreasing digital mechanism, parallel delay lines,
And the digital processing mechanism is programmable by a control mechanism that receives a predetermined number of channels in the effective band and characteristics regarding division and the number of channels in the group, and the multiplexing circuit includes a plurality of switching mechanisms. However, they are each connected to the output of a filtering / reducing digital mechanism, and these digital mechanisms are selectively connected to parallel delay lines depending on the division of the group and the number of channels in the group. Controlled by the control mechanism.

[0029]

DESCRIPTION OF THE PREFERRED EMBODIMENT Referring to FIG. 1, a network of regenerative satellites typically includes a satellite 1, N ground stations S ₀ to S _N-1 , and an assigning station SC for programming satellite 1 from the ground. .

The assigning station SC serves to program the satellite 1 when there is a fluctuation or change in the traffic between the ground stations S ₀ to S _N-1 . Typically, the assigning station SC
The connection set up between the satellite and the satellite is a dedicated connection LS, which transmits the entire programming data created by the allocating station SC.

The ground station S _i (where i is an integer from 0 to N-1) is connected to satellite 1 through the upstream connection.
Transmit each carrier with a given total bandwidth in the order of gigahertz, frequency modulated and frequency multiplexed. The bandwidth assigned to the various carriers is a few megahertz (eg 6 MHz). This frequency multiplexing is shown diagrammatically in FIG. The satellite 1, which receives these modulated and multiplexed carriers, utilizes the function of the transmultiplexer and the satellite 1 and the ground station S _0.
To S _N-1 AMRF (Frequency Division Multiple Access) format characterizing the upstream connection and M for the downstream connection.
Convert between RT (time division multiple access) formats. The digital data transmitted on each upstream connection by the modulated and frequency multiplexed carrier is
In this way, downlink connections are multiplexed in time. Upon reception, the stations S _i are respectively assigned to the time division channels of the downlink connection and reproduce the data transmitted on the carrier of the uplink connection.

FIG. 2 shows a block diagram of a known embodiment of the transmultiplexer onboard the satellite 1.
Trans multiplexer, in series, a frequency demultiplexing circuit 10, N pieces of the demultiplexer 11 ₀ 11 _N-1, exchange matrix 12, the time division multiplexer, and a modulator 1
Including 4. The N carriers that have been modulated, transmitted by the ground stations S ₀ to S _N-1 , and frequency-multiplexed are received by the antenna of the satellite 1 via the upstream connection, and are input to the demultiplexing circuit 10. Applied to. Demultiplexing circuit 10
Demultiplexes the N carriers, which are transposed to the baseband frequency and reproduced at each of the N outputs of the demultiplexing circuit. The N carriers modulated to these basebands are N demultiplexers 11 ₀ to 11 to recover the carrier modulated digital data.
It is applied to each input of _N-1 and demodulated. The exchange matrix 12 assigns the data thus retrieved to the time division channels. These various channels are multiplexed by the time division multiplexer 13 and, after modulation of only one carrier P _{AMRT in} the modulator 14, the N stations in the downlink connection. Sent again to.

As shown in FIG. 3, a network of regenerative satellites of this kind has, for example, a base access of 144 Kbit / sec (2B + D) and 1544 Kbit / sec (23B + D) and / or 2
It can belong to a service integrated digital network that provides primary access of 048 Kbit / sec (30B + D). Communication volume
144 Kbit / s, 1544 Kbit / s (23B + D), or
Carriers modulated with different digital signals of 2048 Kbit / sec will be frequency-multiplexed within the frequency band provided by the upstream connection and will be characterized by different bandwidths depending on the traffic.

Two known digital demultiplexing circuits 10A and 10B according to the prior art will now be introduced with reference to FIGS.

The first demultiplexing circuit 10 depicted in FIG.
A relates to a frequency multiplexer / demultiplexer including one polyphase network and one Fourier transformer.
L OF SATELLITE COMMUNICATIONS, Vol. 6, March 1988, p. 267-281, Enric
It is based on the articles by o del Re and Romano Fantacci. A polyphase network consists of N parallel branches. Row n (n
Is an integer contained between 0 and N-1), and each branch of the series has a delay line with an impedance that can be expressed in the form z ^-1 of the transformation in the z plane, and E _n (z ) Includes one impedance filter. The modulated and multiplexed carrier is sampled and then applied to each branch,
The output of each branch is applied to the input of a Fourier transformer with N inputs and N outputs.

The general function of the first demultiplexing circuit 10A is to shift to the frequency domain of the element low pass filter with a gauge of bandwidth equal to the same bandwidth of the modulated carrier. Therefore, a functional drawback of this type of technology is that the modulated and frequency multiplexed carriers must have the same bandwidth. The functions of the first demultiplexing circuit with a polyphase network are as follows:

[0037]

[Equation 1]

The meaning of this equation is the transfer function in the form of a transformation in the z-plane, that of an element low-pass filter with finite pulse response. H ₀ (z) can be written as:
H ₀ (z) = [h (0) z ^-0 + h (1) z ^-1 + h (2) z ^-2 + h (3) z
^-3 + ..... + h (M) z ^{- M + 1} ], that is, decomposed into N sequences:

[0039]

[Equation 2]

This is written as:

[Equation 3]

That is:

[0042]

[Equation 4]

Putting, H ₀ (z) = z ⁻⁰ [E ₀ (z ^−kN )] + z ⁻¹ [E ₁ (z ^−kN )]
+ ..... + z ^{-N + 1} [ ^EN _-1 (z ^-kN )].

This equation corresponds to a polymorphic representation with N components.

That is, there are N filters H _n (z),
n varies from 0 to (N-1), each filtering N modulated and frequency-multiplexed carriers with the same bandwidth, resulting in a total frequency bandwidth of Δf _T / 2. From the transfer function H ₀ (z) of the element low-pass filter, the frequency is (n
By shifting by Δf _T / 2 N) leads to an nth passband filter with gauges each corresponding to the frequency band assigned to the nth modulated carrier, which is the exponential complex exponential operator Z exp of variable z when pointing to
Corresponds to conversion to [j2πn / N]. The transfer function H _n (z) of the frequency “shifted” nth filter is thus
It is written as:

H _n (z) = H ₀ [z exp (j2πn / N)].

Introducing the above equation into a decomposition into N polyphase components yields the following equation:

[0048]

[Equation 5]

In this way, for N filters, the relation can be written as:

[0050]

[Equation 6]

Here, V is equal to exp [-j2π / N].

The central arithmetic matrix represents the Fourier transform, while the filters E ₀ (z ^-nN ) and z ^-1 E ₁ (z
^-kN ), .... z ^{-1 (N-1)} E _N-1 (z ^-kN ) transfers delay lines z ^-0 , z ^-1 , ... z ^{-1 + N} of the transfer function, respectively. Function filters E ₀ (z ^-nN ), E ₁ (z ^{-n N} ), ... E _N-1 (z ^-nN ).
It is characterized by putting in series with. These considerations give the conditions for implementation of the demultiplexing circuit 10A shown in FIG.

A second embodiment 10B of the demultiplexing circuit according to the prior art will now be described with reference to FIG.
"INTERNATIONAL JOURNAL OF SATELLITE COMMUNICATION
S (International Satellite Communication Journal) "Volume 6, 1988 243-25
See “Multi-carrier demodulators for on-board: roces” on page 1
WH YIM, entitled "sing satellites), is introduced in line with other papers. For this approach, called" by channel ", there are typically two methods proposed. , Each of which is frequency-shifted in order to transpose each of the modulated and multiplexed carriers first to baseband and then to low-pass filtering, a second equivalent method of which is to separate each other by element low-pass filters. It consists of using a pass-band filter with a frequency-shifted gauge.In the block diagram of the demultiplexer there are several branches, each containing a channel filter.

According to the first method above, two methods are typically proposed for this approach, called "channel-by-channel", which filters carriers modulated with a given bandwidth. The first method consists of frequency shifting each modulated and multiplexed carrier to transpose first to baseband and then to lowpass filtering. A second equivalent method consists of using passband filters with gauges that are frequency shifted with respect to each other by an elemental lowpass filter. In the block diagram of the demultiplexer there can be seen multiple branches, each containing one channel filter.

According to the first method above, a low-pass filter for filtering the carrier modulated by a given bandwidth
The transfer function of the filter h _i (z) can be written in the form of a transformation in the z plane as:

[0056]

[Equation 7]

Given i-th branch filter F _i (i
Is an integer contained between 0 and (N-1)), the first digital processing is to shift the carrier associated with the i-th channel to baseband. The sampled digital input sequence of the modulated and multiplexed carrier is called x (k). The first step is to create the following sequence from the sequence x (k):

X _i (k) = x (k) exp (-j.ω _i kM)

Where exp represents the complex exponential operator, ω _i is the angular center frequency of the i th channel,
And M represents the diminishing rate. The second digital processing is then low-pass filtering this sequence x _i (k) for the i th channel with the given bandwidth. Low-pass filter h _i (z) performs this process.

Thus, for the i-th channel, the digital sequence Y _i (k) of the carrier demultiplexed to the baseband is written as:

[0061]

[Equation 8]

With respect to the second method of the demultiplexer of FIG. 5, for the i-th branch, the z-plane of the z-plane, obtained from the transfer function of the low-pass filter in the z-plane, is written as The use of passband filters with transfer functions is proposed:

[0063]

[Equation 9]

Then, a result similar to the result obtained by the first method is obtained.

FIG. 6 shows a demultiplexing circuit 10 according to the present invention.
Figure 3 shows a schematic block diagram of the functionality of the.

Modulated, frequency-multiplexed carriers relating to channels having different bandwidths are applied to the inputs of a first filtering stage containing two filters of the above "channel-by-channel filter" type, Each output is applied to a polyphase demultiplexing circuit 105, which has a role similar to that originally introduced in the description of the prior art.

It should be noted that in the embodiment shown in FIG. 6, the number of groups each containing channels with modulated carriers of the same bandwidth is two. It is within the scope of the invention to expand the number of fractions each containing channels of the same bandwidth to any number, which is for two embodiments of demultiplexing circuits with programmable filters according to the invention. This is done by adding a circuit similar to the one that belongs to one of the channels defined below. The two embodiments shown with reference to Figures 7A and 7B also relate to the case where two groups of modulated carriers, each of the same bandwidth, are frequency-multiplexed.

Also, for the sake of simplicity of explanation, the channels of the modulated carriers of _one group, called group ₁ , have a frequency band f ₁ , which is 4 of the frequency band f ₂ of the other group, called group 2. It is assumed to be one-third. Bandwidths f ₁ and f ₂
Is considered here to be equal to the minimum bandwidth and the maximum bandwidth of the frequency-multiplexed channels that the frequency-multiplexing / separation circuit can handle. Bandwidth f ₁ is effective frequency band f _u integral submultiple one, as typically is f ₁ = f _u / 16, also, there is also an integer fraction of the bandwidth _{f 2, f 1 = f 2} /
Is 4. However, the demultiplexing circuit can be used in other applications. For example, in the RNIS network for baseband access (144 Kbit / sec) and primary access (1544 Kbit / sec and / or 2048 Kbit / sec), the channel bandwidth ratio is not 4 and the bandwidth is f
It can also be used when given by ₁ , f ₂ = 11f ₁ and f ₃ = 15f ₁ .

Referring to FIG. 7A, the frequency demultiplexing / separating circuit 10 according to the present invention includes two filters 102 ₁ and 102 for two groups of channels having the same bandwidth.
₂ and _two subsequent reducers 102 ₁ ′ and 102
Stage 102 having ₂ ', stage 103 of _two connection switching circuits 103 ₁ and 103 ₂ each having 4 outputs, delay line 104
a, 104b, 104c and 104d stages 104, 4
Individual polyphase networks 105a, 105b, 105c and 105
stage 105 including d, and one complex multiplexing circuit 105
Includes ₅ (Fourier Transform). These stages are placed in series in the order listed above, and the group filters 102 ₁ and 102 ₂ receive the modulated and frequency multiplexed carriers in sampled form and the modulated and demultiplexed carriers to baseband. Are produced at the output of the composite multiplexing circuit 105 ₅ .

The first embodiment of the demultiplexing circuit shown in FIG. 7A derives from the traffic of the upstream connection whose frequency is shown in FIG. 7B. As shown in this figure, the total effective frequency band f _u allocated to the upstream connection is shown in the form of standardized f _{no r} for the sake of simplicity, and is modulated and frequency-multiplexed. The channel carriers are reorganized into two groups, depending on their bandwidth respectively. There are 8 modulated carriers P ₁ ¹ to P ₈ ¹ seen in the _first group and 2 modulated carriers P ₁ ² to P ₂ ² in the second group. In this embodiment, the channels of the second group each have a bandwidth f ₂ equal to 4 times the bandwidth f ₁ of each of the channels of the first group, ie a total of 8f.
₁ + 2f ₂ = and the it remembered is f _u.

A sampling circuit 101 is placed in front of the demultiplexing circuit 10 according to the invention, the circuit 101 receiving the multicarrier signal SM of the upstream connection, (8 + 2) = 10.
The modulated and frequency multiplexed carriers are sampled at an initial frequency of 1 / T to create a sample of the multi-carrier signal. The modulated and frequency multiplexed carrier thus sampled is applied to the inputs of the _two filters 102 ₁ and 102 ₂ of the stage 102. These filters, which are typically as already introduced along with the prior art, additionally include passband filters or elemental baseband filters. The function of each of them is to filter two groups of carriers of the same bandwidth. Therefore, filter 102 ₁ is assigned to group 1 and filters all the carriers in this group, while filter 102 ₂ filters the modulated carriers in the second group, group 2. Filter 1
The reducers 102 ₁ ′ and 102 ₂ ′ connected in series to 02 ₁ and 102 ₂ are filters 102 ₁ and 10 ₂ , respectively.
Decrease the frequency of the signal leaving 2 ₂ . In fact, for each group of carriers, namely group 1 and group 2, it is possible to reduce the initial sampling frequency without affecting the respective modulated and multiplexed carriers (Sha.
nnon's theorem). Decreasing this signal frequency by a factor of _{2 at} the outputs of filters 102 ₁ and 102 ₂ corresponds to reducing the number of samples at the output of the filter by a factor of 2, which results in a sample at the output of stage 102. It maintains a number, ie the total frequency band f _u of the multi-carrier input signal SM. The gradual reduction particularly contributes to reducing the amount of calculation required in the complex multiplexing circuit 105 ₅ . The reduction rate of the reduction filters 102 ₁ ′ and 102 ₂ ′ is therefore ½.

Before introducing stages 103 and 104, stage 105, consisting of polyphase networks 105a to 105d and complex multiplexing circuit 105 _5, will now be described.

The four multiphase networks 105a to 105d are each constituted by four parallel branches. Each branch contains one delay line and one element digital filter in series, and the transformation Hp0 to H in the z plane of this digital filter
p15 is expressed as a 16th degree polynomial in z. The output of each of the 16 digital filters is applied to the input of a complex multiplexing circuit 105 ₅ for fast Fourier transform (TFR). A modulated and baseband demultiplexed carrier for each group is produced at each output of the circuit 105 ₅ . The 16th degree polynomial of z for the digital filter in the polyphase network fits the description of the prior art, NP = NP for group 1.
16 modulated carriers, ie minimum bandwidth f ₁
It _follows from the fact that the NP = 16 carriers that make up the set of carriers contained in the given total frequency band f _u of the input signal can be demultiplexed. Filter Hp0
Element low-pass filter for guiding each of Hp15 from, therefore, the multi-carrier input signal SM has a bandwidth equal to the channel minimum bandwidth f ₁ can be supported.

However, in the present invention, the digital delay line arranged upstream of each digital filter is
It has no transformation in the z plane that changes from z ⁻⁰ to z ⁻¹⁵ . This difference compared to prior art polyphase networks is made clear below.

The stage 105 of the polyphase network must simultaneously demultiplex the group 1 modulated carriers and the group 2 modulated carriers, but they are not carriers of channels having equal bandwidth. This demultiplexing cannot be done by using only one polyphase network. For this,
The delay line stage 104 includes four delay lines 104a, 104b, 104c and 104d, each output of which, in accordance with the preferred embodiment shown, is a polyphase network 10.
5a to 105d, respectively. In the preferred embodiment, the four delay lines in each polyphase network have transformations in the z-plane that can be written as z- ⁰ , z- ⁴ , z- ⁸ , and z- ¹² , respectively. Under these conditions, the same bandwidth
If the demultiplexer derived from the prior art for demultiplexing the 16 modulated carriers is stage 104,
Would include delay lines denoted as z- ⁰ , z- ¹ , z- ² , and z- ³ . In each polyphase network, placing four delay lines in parallel, each connected to a digital filter derived from an elemental low-pass filter (associated with the complex multiplexing mechanism of the Fourier transform), gives a bandwidth f ₂ / f ₁ To obtain a frequency filter equal to 4 times the bandwidth of the elemental filter, and a modulated carrier of group 2 having a bandwidth of 4 times the bandwidth of the carrier of group 1 filtered by this elemental filter. Can be reconfigured.

Delay lines 104a through 104d, represented in modeled form of the transformation in the z plane in FIG. 7A, are actually shift register based digital circuits and other base digital components, which are initially It is bursting at a frequency equal to the sampling frequency 1 / T. As seen above, the initial frequency of the sequence of modulated multiplexed carrier samples at the output of the sampling circuit 101 is the output of the reducers 102 ₁ ′ and 102 ₂ ′, reduced by the respective respective rates given. To be done. According to the frequency division of the carrier shown in FIG. 7B, the switching circuits 103 ₁ and 1
03 ₂ connects the outputs of reducers 102 ₁ ′ and 102 ₂ ′ with delay lines 104a, 104b and 104c, 104d, respectively, which is exactly the two reducers 102 ₁ ′ and 1
Sequence of samples coming from 02 ₂ ', of being carried out in accordance with a frequency which is diminishing to 1/2 as compared with the initial sampling frequency. Therefore, in reality, the exchange mechanism 10
3 ₁ and 103 ₂ have the outputs of the downconverters 102 ₁ ′ and 102 ₂ ′ at the input of the delay line of stage 104, depending on the frequency of the sequence coming from said downconverter, ie the modulated and multiplexed carrier. Depending on the total bandwidth devoted to each group of
Connect selectively.

The delay provided by the delay line of stage 104 is a frequency that is reduced relative to the burst frequency of the delay line (equal to the initial sampling frequency 1 / T) of the sequence that the delay line receives through stage 103 of the exchange. Each is programmed depending on. Thus, for the two groups, the frequency of the downscaled sequence is one-half the sampling frequency, so the sample delay of the downscaled sequence is half the frequency of the sampled multi-carrier signal. Equals one sample period.
In the embodiment of FIG. 7A, two carrier groups have equal bandwidths, so two delay lines 104a and 104b and 104c and 104d having transfer functions z- ⁰ and z- ² are used.
Applies to each carrier group.

The outputs of the delay lines 104a to 104d are four multiphase networks 10 irrespective of the traffic of the upstream connection.
5a to 105d, respectively.

The outputs of all the filters Hp0 to Hp15 in the polyphase networks 105a to 105d are composite multiplexing circuits 105.
It is applied to the input of the _5, 16 output circuits 105 ₅ 2
It is divided into two groups, of which eight outputs are applied to the eight baseband carriers of group 1, and the other eight outputs are separated into two sets of four, the two bases of group 2. It is applied to band carriers, because the group 2 carriers have a bandwidth corresponding to four times the bandwidth of the group 1 carriers. In the circuit 105 ₅ , the digital composite multiplexing process performed between the sequences exiting the branches of the polyphase network 105a to 105d consists of one 8 × 8 matrix mechanism and 4
Implemented by a 2 × 2 matrix mechanism.

The 8 × 8 matrix mechanism receives the sequences exiting the filters Hp0 to Hp7 contained in the polyphase networks 105a and 105b, and at the first eight outputs of the circuit 105 ₅ the carriers P ₁ ¹ to P ₈ ¹ Reconstruct the eight digital channels of group 1 corresponding to. This matrix mechanism can be expressed by using a transfer function, for example, V = exp with N = 8 above.
Perform a digital calculation derived from a coefficient matrix, such as that defined as a function of [−j2π / N].

For the two carriers P ₁ ² to P ₂ ² of group 2, the division of the digital processing in the four square matrix arrangement in circuit 105 ₅ is such that each of the channels of group 2 has a basic bandwidth f. ₁ corresponds to four juxtapositions, so that each set containing two identical fundamental bandwidths in the channels of carriers P ₁ ² and P ₂ ² is correlated in stage 102.

In this way, each of the 2 × 2 matrix schemes is connected to the output of two filters in the networks 105c and 105d, respectively, which are the carriers P ₁ ² and P ₂ ² Of the channels corresponding to the same frequency sub-baseband with bandwidth f ₁ , ie network 1
It receives the same delayed digital sequence in 05c and 105d.

The first 2 × 2 matrix mechanism is the network 105.
Process the sequence exiting the filters Hp8 and Hp12 placed after the delay line z- ⁰ in c and 105d.
The same is true for the two 2x2 matrix schemes. The fourth and final 2x2 matrix mechanism is therefore the net 1
Process the sequence leaving filters Hp11 and Hp15 placed after delay line z- ¹² in 05c and 105d. Two
× Each 2 network reconstructs a quarter of the bandwidth of one square coefficient matrix by calling the carrier P ₁ ² and P ₂ ² channels.

Referring to FIG. 8A, a second embodiment of the demultiplexer 10 is introduced for the upstream traffic shown in FIG. 8B. In this traffic, group 1
Includes four successive channels with modulated carriers P ₁ ¹ to P ₄ ¹ and a frequency bandwidth of f ₁ , while group 2 has carriers P ₁ ² to P ₃ ² and a frequency bandwidth of f ₂ = 4f ₁
Including three consecutive channels of.

This second embodiment is, in practice, identical in structure to the first embodiment shown in FIG. 7A, but in terms of function some of the demultiplexing mechanisms 10 are shown. In order to be able to program parts according to the division of traffic,
It differs from that of FIG. 7A. Programmable components are filters 102 ₁ and 102 ₂ , reducers 102 ₁ ′ and 10
2 ₂ ′, the delay lines 104 a to 104 d, and the matrix mechanism in the complex multiplexing circuit 105 ₅ . In the preferred embodiment, programming of these various components is performed using the control unit UC shown in FIG. 2, which receives all features of each of the new divisions of traffic through the upstream dedicated connection LS. To do.

In this second allocation of traffic:
The filters 102 ₁ and 102 ₂ have bandwidths f _u / 4 and 3 f _u / 4.
The two groups 1 and 2 are programmed to separate the two groups so that they no longer have the same bandwidth as the allocation of FIG. 7B.
Decreasing rate of downconverter 102 ₁ 'and 102 _2' is adjusted in accordance with the new bandwidth f _u / 4 and 3f _u / 4, becomes 1/4 and 3/4.

With respect to the delay line of stage 104, the delay line is programmed according to the new diminishing rate. Multi-phase network 105a
Is again applied to the processing of carriers P ₁ ¹ to P ₄ ¹ , so that the transfer function of the first delay line 104a is still z ⁻⁰ . Since 3 out of 4 samples of the multi-carrier signal must be processed at the output of the reducer 102 ₂ ′, 3
The number of delay lines 104b, 104c and 104d is 3
Two outputs are connected to the output of the reducer 102 ₂ ′ through the switching circuit 103 ₂ used. Delay line 104b, 1
04c and 104d are (4 × 0) / 3, (4 ×
1) / 3, and (4 × 2) / 3 ratio, sampling frequency 1
Apply a delay using / T.

In this case, the composite multiplexing circuit 105 ₅
Programmed in four 4x4 matrix schemes and four 3x3 matrix schemes. The 4 × 4 matrix mechanism is connected to the outputs of the filters Hp0 to Hp3 in the polyphase network 105a and reconstructs the carriers P ₁ ¹ to P ₄ ¹ of group 1. Four
The 3 × 3 matrix mechanism has three carriers P ₁ ² , P 3.
_{Of the 2} ² and P ₃ ² channels, a circuit Hp4-Hp8-H with three filters for the four sub-bands of bandwidth f ₁ so as to reconstruct these channels.
p12, Hp5-Hp9-Hp13, Hp6-Hp10-Hp14, and Hp7-Hp11-Hp15, respectively.

[Brief description of drawings]

FIG. 1 diagrammatically shows a known network of regenerative satellites.

FIG. 2 shows a block diagram of a known transmultiplexer included in a regenerative satellite.

FIG. 3 shows the frequency-multiplexed base access signal and primary signal RNIS in the frequency domain.

FIG. 4 shows a block diagram of a conventional digital frequency multiplexing / demultiplexing circuit.

FIG. 5 shows a block diagram of a digital frequency multiplexing and demultiplexing circuit according to the prior art.

FIG. 6 is a schematic block diagram of a programmable frequency multiplexing and demultiplexing circuit according to the present invention.

FIG. 7A shows a first embodiment of a programmable demultiplexing circuit.

FIG. 7B shows a frequency representation of the traffic for the corresponding upstream connection in FIG. 7A.

FIG. 8A shows a second embodiment of a programmable demultiplexing circuit.

FIG. 8B shows the frequency of the corresponding upstream connection traffic in FIG. 8A.

Claims (6)

[Claims] 1. A - in predetermined effective frequency bandwidth of (f _u) multi-carrier of the sampled signals with (SM), a device for demultiplexing the channels that are frequency multiplexed, the The channels of the carrier are modulated carriers (P ₁ ¹ to P ₄ ¹ , P ₁ ² to P ₃ ² ) included in the effective frequency band and a frequency bandwidth (f ₁ , F
₂ ), the frequency bandwidth of the channel is a fraction of the maximum frequency bandwidth of the channel (f ₂ ), but the latter itself is also a fraction of the effective bandwidth ( _fu ),
And becomes an integral multiple of the minimum bandwidth of the channel (f ₁ ), and the above circuit (10) has a predetermined number ( _fu / f ₁ = 16) equal to the ratio of the effective bandwidth to the minimum bandwidth of the channel. ,
A parallel placed circuit, each including a delay line and a digital filter in series, and a Fourier transform for individually reconstructing the above channels connected to the outputs of the above filters (Hp0 to Hp15) of includes a digital processing mechanism (105 _5), each of said filters is directed from one of the digital low-pass filter for filtering in the minimum bandwidth of the channel (f _1), - a plurality of This is a mechanism for separating channels in a sampled signal (SM) of multi-carrier into a predetermined number of channel groups (group 1, group 2) by a filtering / frequency-reducing mechanism (102). The group comprises juxtaposed bands of all channels having the same bandwidth, different from the bandwidths of the other channels, -with a plurality of parallel delay lines (104), the number of which is Effective bandwidth and maximum bandwidth a predetermined number equal to the ratio of the width of the channel (f
_u / f ₂ = 4) and each channel group (group 1, group 2)
The delay lines in parallel each number equal to greater than the nearest integer ratio of bandwidth _{(f u / 4,3f u / 4} ) and channels of maximum bandwidth _{(f 2 = f u / 4} ) of the group Delayed by (104a, 104b, 104c, 104d), and-in a plurality of identical parallel polyphase networks (105a, 105b, 105c, 105d) the output of the delay lines (104a, 104b, 104c, 104d). Each of the polyphase networks is connected to a plurality of parallel delay lines and digital filter pairs, each of which has a predetermined number (f ₂ , f ₂₎ equal to the ratio of the maximum bandwidth to the minimum of the channel.
Circuits characterized by including only ₁ ). 2. A filtering and frequency-decreasing mechanism for separating one channel group from a multi-carrier signal (SM),
Digital filters (102 ₁ , 1 having the bandwidth of the group)
02 ₂ ), followed by the bandwidth ( _fu /
4,3f _u / 4) and the downconverter of equal decreasing ratio to the ratio of the effective bandwidth _{(f u) (102 1 '} , 102 2') include, characterized by circuit according to claim 1 . 3. Parallel delay lines (104b, 104c, 104d) for delaying one channel group (group 2),
Effective bandwidth (f _u) and applying a respective proportional to the delay to zero or a multiple of an integer equal to the ratio (4/3) of the bandwidth of the group (3f _u / 4), characterized by, wherein The circuit according to Item 1 or 2. 4. Each polyphase network (105a, 105b, 10)
5c and 105d) include delay lines (z- ⁰ , z- ⁴ ,
z ^-8 , z ^-12 ) are each delayed in proportion to an integer multiple of zero or greater equal to the ratio (f ₂ / f ₁ ) of the maximum bandwidth to the minimum of the channel, The circuit according to claim 1. 5. A digital processing mechanism (105 ₅ ) comprises, for each channel group (group 2), a plurality of square matrix mechanisms of Fourier transform coefficients, the bandwidth (f ₂ ) of said group.
And a minimum bandwidth (f ₉ ) ratio equal to the number of channels in the group, the order of each matrix mechanism belonging to the group is equal to the number of channels in the group. 5. The circuit according to any one of 4 to 4. 6. A digital mechanism for filtering / decreasing (10)
2), parallel delay lines (104), and a digital processing mechanism (105 _5), control of receiving the number of channels of the characteristics and Gun'nai for a given number and division of the group of channels in the effective band (f _u) Is programmable by the mechanism (UC), and the multiplexing circuit has multiple switching mechanisms (1
03 ₁ , 103 ₂ ), each of which is connected to the output of a filtering / decreasing digital mechanism (102), and these digital mechanisms (102) are divided into groups of Circuit according to any of claims 1 to 5, characterized in that it is controlled by a control mechanism (UC) for selectively connecting parallel delay lines depending on the number of channels.