JPH0530338B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION TECHNICAL FIELD OF THE INVENTION This invention relates to spread spectrum multiplexing communication systems.

[Technical background of the invention and its problems]

2. Description of the Related Art Spread spectrum multiplexing has been known as one of the transmission technologies for wireless and wired communications. Spread spectrum multiplexing allows multiple parties to communicate using the same frequency at the same time.

Therefore, in spread spectrum multiplexing communication, since many people use the same frequency at the same time, a technique is required to separate only desired waves from the radio waves transmitted simultaneously.

Therefore, in spread spectrum multiplexing communication, in order to separate those communicating simultaneously (channels), a different code sequence is assigned to each channel, and matching (correlation) is performed using the same code as the code used by the other party. ) is done.

However, depending on how the code sequences are assigned, interference with the desired wave, so-called unnecessary crosstalk, may occur.
For example, since there is almost no difference between the code sequence (1, 1, 1, 1, 1, 1, 1) and the code sequence (1, 1, 1, 1, 1, 1, -1), Very large interference occurs. Therefore, code orthogonality (how to reduce sidelobes) is important to maintain communication quality.

For this reason, codes such as the M sequence and the Gold sequence are used as code sequences with good orthogonality, but these codes have a small side lobe of the cyclic autocorrelation function, but the side lobe of the autocorrelation function as a finite length sequence is small. Since the lobes are not necessarily small, there is a drawback that side lobes cause deterioration when the signal is inverted.

Furthermore, since the cross-correlation function between different sequences is not small, there is also the drawback that crosstalk is unavoidable.

[Object of the Invention] In order to eliminate these drawbacks, the present invention aims to provide a synchronous multiplex communication system that is free from deterioration and crosstalk and can be used under conditions where synchronization is achieved over the entire communication channel. purpose.

[Summary of the invention]

The present invention is a multiplex communication system that transmits synchronized information signals on a communication path based on a predetermined synchronization signal, and which maintains multiple difference orthogonal sequences that are in a mutually different mate relationship. a plurality of transmitting devices that generate transmission signals by modulating signals based on; a synthesizer that combines and multiplexes each of the transmission signals generated by the plurality of transmitting devices;
an adaptive filter that holds each multiple difference orthogonal sequence corresponding to the multiple difference orthogonal sequence held in each of the plurality of transmitting apparatuses, the multiple difference orthogonal sequence held in the adaptive filter corresponding to a desired transmitting apparatus; and a receiving device that selectively receives desired transmission signals based on orthogonal sequences.

〔Effect of the invention〕

According to the present invention, the above-mentioned deterioration and crosstalk during signal inversion can be completely eliminated.

[Embodiments of the invention]

Before explaining the embodiments of the present invention, the principle of the present invention will be explained.

First, we define a multiple difference orthogonal sequence, a mate, and a complementary sequence system.

Definition of multiple difference orthogonal series The autocorrelation of an infinite length series consisting of a finite length series consisting of complex numbers with an absolute value of 1 and a 0 series connected before and after is
The value normalized to 1 for 0 shift is defined as the autocorrelation function of this finite length series. Further, the cross-correlation functions of two finite-length sequences are similarly normalized and defined at the same ratio.

A finite-length sequence whose components are complex numbers with an absolute value of 1,
A sequence whose autocorrelation function becomes 0 with a multiple shift of N excluding a 0 shift is defined as a multiple difference orthogonal sequence of order N. This is an extension of the concept of even-difference orthogonal sequences.

For example, the autocorrelation function of the sequence (+++-), (+ indicates 1, - indicates -1) with a sequence length of 4 is (-1/4, 0, 1/4, 1, 1/4, 0, -1/4
) Therefore, this series is a multiple difference orthogonal series of order 2.

An example of a multiple difference orthogonal sequence of order 4 (+++-++
-++-+++--) is shown in FIG. 1.

It is clear that the multiple difference orthogonal sequence of order N includes all sequences whose sequence length is N or less.

Definition of Mate When the cross-correlation function of two orthogonal sequences with multiple differences of order N becomes 0 with a multiple shift of N including a 0 shift, these two orthogonal sequences with multiple differences are mutually mates of order N. Define that there is. Also, if there are M multiple difference orthogonal sequences of order N, and any two sequences among these M sequences are mates in order, then these M sequences are mutually Define it as a mate of order N.

For example, the cross-correlation function of two order 2 multiple difference orthogonal sequences (+++-) (++-+) is (1/4, 0, 1/4, 0, 3/4, 0, -1/4 ) Therefore, these two series are mutually mates of order 2.

When the inner product of two sequences with sequence length N or less is 0,
These two series are mutually mates of order N.

Definition of Self, Mutual and Completely Complementary Sequence Systems When the sum of the autocorrelation functions of N sequences is 0 for all shifts other than the 0 shift, these N sequences are defined as a self-complementary sequence system.

Also, there are two sets of N series, and the N series of both sets are numbered from 1 to N.
When the sum of the cross-correlation functions (there are N) of sequences with the same number becomes 0 at every shift, these two sets of sequence systems are defined as mutually complementary sequence systems.

Furthermore, when there are N sets of self-complementary series consisting of N series, and the series in each set are numbered from 1 to N, if any two of the N sets are mutually complementary series, then , this sequence system of N sets is defined as a completely complementary sequence system of order N.

For example, the sum of the autocorrelation functions of two sequences (Figure 2) A ₀ = (+++-) A ₁ = (+-++) is (0, 0, 0, 2, 0, 0, 0), so {A ₀ , A ₁ } is a self-complementary series system.

Also, if B ₀ = (++-+) B ₁ = (+---), then {B ₀ , B ₁ } is also a self-complementary series system,
Moreover, the sum of the cross-correlation function of A ₀ and B ₀ and the cross-correlation function of A ₁ and B ₁ is (0, 0, 0, 0, 0, 0, 0), so {A ₀ , A ₁ } and {B ₀ , B ₁ } are mutually complementary sequence systems, and at the same time are completely complementary sequence systems of order 2.

Figure 2 shows the structure of a complete complementary sequence system consisting of {A ₀ , A ₁ } and {B ₀ , B ₁ }.

Next, the order N whose components are complex numbers with absolute value 1
We will show how to derive multiple difference orthogonal sequences with sequence length N ² and mate and perfect complementary sequence systems.

Multiple difference orthogonal sequence with order N and sequence length N ² A regular matrix whose components are complex numbers and whose inner product of two rows at different positions is 0 is called a "complex orthogonal matrix." Now, the N-th order complex orthogonal matrix whose components are complex numbers with an absolute value of 1 is A=a ₁₁ a ₂₁・ ・ ・ a _N1 a ₁₂ a ₂₂・ ・ ・ a _N2 ………… ………… …………a _1N a _2N・・・a _NN . |a _iJ |=1, and the inner product of any two different rows and the inner product of any two different columns are 0.

Now, by arranging each row of the matrix A in the order of rows, the sequence A with sequence length N ² = (a ₁₁ , a ₁₂ , ......, a _1N , a ₂₂ , a ₂₂ , ......, a _2N , ………… a _N1 , a _N2 …………, a _NN ), then the inner product of any two different rows of matrix A is 0.
Therefore, it is clear that the autocorrelation function of the series A〓 becomes 0 when shifted by a multiple of N.

The first to Nth rows of N mate matrices A with order N and sequence length N ² are respectively A ₁ . . .
......A _N , then A = (A ₁ , A ₂ , ......, A _N ), but a new N-order complex orthogonal matrix whose components are complex numbers with an absolute value of 1 B = [b _iJ ] Using, C ₁ = (b ₁₁ A ₁ , b ₁₂ A ₂ , ......, b _1N A _N ) C ₂ = (b ₂₁ A ₁ , b ₂₂ A ₂ , ......, b _2N A _N )・ ・ ・ If C _N = (b _N1 A ₁ , b _NN A ₂ , ......, b _NN A _N ), then N sequences C ₁ ...... C _N of sequence length N ² are mutually mate. It becomes a multiple difference orthogonal series in the relationship.
Because C ₁ ...... any different 2 of C _N
The cross-correlation function of the series becomes 0 because the inner product of any two different rows of the matrix A is 0 for a multiple of N shift other than a 0 shift, and for a 0 shift, A ₁ ~A This is because the respective norms of _N are equal to N and the inner product of any two different rows of matrix B is 0, so it becomes 0.

Completely complementary sequence system of order N and sequence length N ² derived using the procedure in the previous section. A complementary sequence system consisting of N sequences is derived from each of N multiple difference orthogonal sequences of order N and sequence length N ² that are mates, and a complete complementary sequence consisting of N ² sequences is derived. Show how to derive the system.

Let C ₁ = (C ₁₁ , C ₁₂ , ......, C _1N ² ) be the N multiple difference orthogonal sequences of order N and sequence length N ² that are in a mate relationship, derived by the procedure in the previous section. C ₂ = (C ₂₁ , C ₂₂ , ………, C _2N ² ) ・ ・ ・ C _i = (C _i1 , C _i2 , ………, C _iN ² ) ・ ・ ・ C _N = (C _N1 , C _N2 , ......, C _NN ² ). Here, from the new complex orthogonal matrix D = [djk] whose components are complex numbers with an absolute value of 1 and the i-th multiple difference orthogonal sequence C _i = (C _i1 , C _i2 , ......C _iN ² ), the following In this way, N sequences of sequence length N ² are derived. The j-th sequence among the N derived sequences is Eij=(C _i1 d _j1 , C _i2 d _j2 , ......, C _iN d _jN , (C _i(N+1) d _j1 , C _i(N+2) d _j2 ,...CC _i(2N) d _jN , ・ ・ ・ C _i(N2-N+1) d _j1 , C _i(N2-N+2) d _j2 ,..., C _iN ² d _jN ).That is, the m-th component of Eij is the m-th component of Ci multiplied by the component of the m-th (mod N) column of the j-th row of D. The N sequences of E _i1 ...E _iN obtained in this way become a self-complementary sequence system of order N.In the next chapter, we will discuss the order N and sequence length N ⁿ
We will prove that the sequence derived from the multiple difference orthogonal sequence using the same procedure is a self-complementary sequence, so we will omit the proof here.

N sets of self-complementary sequence systems each consisting of N sequences obtained in this way {E ₁₁ ,...,
E _1N }, ..., {E _i1 , ..., E _iN }, ..., {E _N1 , ...
..., E _NN } are mutually complementary sequence systems. This proof is also omitted for the same reason as the proof of the self-complementary series system.

Example Here, an example of N=4 is shown. The series is a four-phase series, and 1, j, -1, -j are respectively 0, 1,
It is written as 2, 3,.

If A=0000 0123 0202 0321=A ₁ A ₂ A ₃ A ₄ , then A = (0000012302020321). A〓 is a multiple difference orthogonal sequence of order 4.

If B = 0000 0123 0202 0321, it will be C ₁ = (000001230202020321) C ₂ = (000012302020203210) C ₃ = (000023010202022103) c ₄ = (0000301220201032). C ₁ , C ₂ , C ₃ , and C ₄ are mutually mates of order 4.

Furthermore, if D=0000 0123 0202 0321, then E ₁₁ = (0000012302020321) E ₁₂ = (0123020203210000) E ₁₃ = (0202032100000123) E ₁₄ = (0321000001230202) E ₂₁ = (0000123020203210) E ₂₂ = (0123131321033333) E ₂₃ = ( E ₃₄ = ( ₀₃₂ 1222201232020 ₎ E ₄₁ = (0000301220201032) E _{42 =} (0123313121031111) E ₄₃ = ( _{0202321022221230} ) E ₄₄ = (0321333323011313). {E ₁₁ , E ₁₂ , E ₁₃ , E ₁₄ }, {E ₂₁ , E ₂₂ ,
E ₂₃ , E ₂₄ }, {E ₃₁ , E ₃₂ , E ₃₃ , E ₃₄ }, {E ₄₁ , E ₇₂ ,
E ₄₃ , E ₄₄ } are each a self-complementary sequence system, and are mutually complementary sequence systems. i.e. E ₁₁
The 16 sequences ~ _E44 are a completely complementary sequence system.

Next, a method of deriving a complete complementary sequence system of order N and sequence length N n from a complete complementary sequence system of order N and sequence length ^{N n -1} will be described. In the previous chapter, we showed how to derive a complete complementary sequence system of order N and sequence length N ² , so we will show how to derive a complete complementary sequence system of arbitrary order N and sequence length N ⁿ using mathematical induction. It turns out.

Multiple difference orthogonal sequence of order N and sequence length N ⁿ and N mates Self-complementary sequence system of N sets of order N and sequence length N ^n-1 {F ₁₁ , ..., F _1N }, ..., {F _i1 ,…,F _iN },…, {F _N1
，
..., F _NN } and these are completely complementary sequence systems, let Gj be a sequence of sequence length N ⁿ obtained by interleaving F _i1 , ..., F _iN . That is,
The k-th component of Fij _j is expressed as F _ijk , and when Fij=(f _ij1 , f _ij2 , ..., f _ijN ^n-1 ), Gi=(f _i11 , f _i21 , ..., f _iN1 , f _i12 , f _i22 , ..., f _iN2 , ...... f _i1No-1 , f _i2N ^n-1 , ..., f _iNN ^n-1 ). Then, when l is an arbitrary integer, the lN shift component of the autocorrelation function of Gi is the lN shift component of each autocorrelation function of F _i1 , ..., F _iN
Since it is equal to the sum of the shift components, the lN shift components excluding the 0 shift of the autocorrelation function of Gi are 0.
Therefore, Gi is a multiple difference orthogonal sequence of order N and sequence length N ⁿ .

Also, the lN shift component of the cross-correlation function between Gi and Gk is 1 of the l shift component of the cross-correlation function between Fij and Fkj.
Since it is equal to the sum of ≦j≦N, Gi and Gk
The lN shift component (including 0 shift) of the cross-correlation function of is 0. Therefore, Gi and Gk each have order N
is the mate of

As a result of the above, N multiple difference orthogonal sequences G ₁ , . . . , G _N of order N and sequence length ^N n are mutually order N mates.

Completely complementary sequence system of order N and sequence length N ⁿ Consists of N sequences from each of the N multiple difference orthogonal sequences shown above that are mutually mates and have an order N and sequence length N ² Deriving a self-complementary series system,
The method of deriving a completely complementary sequence system consisting of ^two N sequences can be used even with sequence length N ⁿ , so the order N and sequence length N ⁿ derived in the previous section can be used in a mate relationship with each other. By applying it to N multiple difference orthogonal sequences, N ² of order N and sequence length N ⁿ is obtained.
A completely complementary sequence system consisting of multiple difference orthogonal sequences can be derived.

N multiple difference orthogonal sequences G ₁ ,..., which are in a mate relationship with each other and have order N and sequence length N ⁿ derived in the previous section.
G _N as G ₁ = (g ₁₁ , g ₁₂ , ..., g _1N ⁿ ) G ₂ = (g ₂₁ , g ₂₂ , ..., g _2N ⁿ ) ・ ・ G _i = (g _i1 , g _i2 , ... ..., g _iN ^N ) ・ ・ ・ G _N = (g _N1 , g _N2 , ..., g _NN ⁿ ). Here, from the new complex orthogonal matrix P = [pik] whose components are complex numbers with an absolute value of 1 and the i-th multiple difference orthogonal series G _i = (g _i1 g _i2 , ..., g _iN ⁿ ), the following formula is obtained. N sequences of sequence length N ⁿ are derived according to . The j-th sequence among the N derived sequences is Q _ij = (g _i1 p _j2 , g _i2 p _j2 , ..., g _iN p _jN , g _i(N+1) p _j1 , g _{i(N +2)} p _j2 ,..., g _i(2N) p _jN , 〓 〓 〓 g _i(No-N+1) p _j1 , g _i(No-N+2) p _j2 ,..., g _iN ⁿ p _jN ). That is, the m-th component of Qij is the product of the m-th component of Gi multiplied by the component in the j-th row and m-th (mod N) column of P. It will be shown below that the N sequences of Q _i1 , . . . , Q _iN obtained in this way form a self-complementary sequence system of order N. Q _i1 ，…，
The sum of l-shift components of the autocorrelation function of each series of Q _iN is expressed as _N 〓 ^j=1 P _Qij (l) = _N 〓 ^j=1 1/N ⁿ _N 〓 ^{m= 1} g _in p _j(n(npdN))・g ^* _i(n+l) p ^* _{j((n+l)(npdN))} =1/N ⁿ _N 〓 ^m=1 g _in g ^* _{i(n +l)}・_N 〓 ^j=1 p _j(n(npdN)) p ^* _{j((n+l)(npdN))} , but since the row matrix P is a complex orthogonal matrix, _N 〓 ^j=1 p _{j(n (npdN))} p ^* _{j((n+l)(npdN))} is N when l is a multiple of N, and is 0 otherwise. On the other hand, when l is a multiple of N other than 0, since Gi is a multiple difference orthogonal sequence of order N, _N 〓 ^m=1 g _in g ^* _i(n+l) = 0. Therefore, for any l other than 0, _N-1 〓 ^j=0 p _Qij (l)=0, and {Q _i1 , . . . , Q _iN } is a self-complementary sequence system.

N sets of self-complementary sequence systems {Q ₁₁ ,..., each consisting of N sequences obtained in this way
Q _1N },…, {Q _i1 ,…, Q _iN },…, {Q _N1 ,…, Q _NN }
It is shown below that are mutually complementary sequences.

As i≠h, l of the cross-correlation function of Qij and Qhj
The sum of the shift components for 1≦j≦N is _N 〓 ^j=1 1/N ⁿ _N 〓 ^m=1 g _in p _j(n(npdN)) g ^* _h(n+l) p ^* _{j((n +l)(npdN))} =1/N ⁿ _N 〓 ^m=1 g _in g ^* _h(n+l)N 〓 ^j=1 p _j(n(npdN)) p ^* _{j((n+l)( npdN))} , but since matrix P is a complex orthogonal matrix, _N 〓 ^j=1 p _j(n(npdN)) p ^* _{j((n+l)(npdN))} is N when l is a multiple of N. , and 0 otherwise. On the other hand, when l is a multiple of N (including 0), since Gi and Gh are mutually mates, _N 〓 ^m=1 g _in g ^* _h(n+l) = 0. Therefore, two sets of self-complementary series systems (Q _i1 ,
..., Q _iN ) and (Q _h1 , ..., Q _hN ) are mutually complementary sequence systems, and since i and h can be any combination such that i≠h, N sets of self-complementary sequence systems (Q ₁₁ , ... ，
Q _1N ),…, (Q _i1 ,…, Q _iN ),…, (Q _N1 ,…, Q _NN )
is a completely complementary series system.

Also, the l shift component of the cross-correlation function of Qij and Qhj is 1/N _oN 〓 ^m=1 g _in p _j(n(npdN)) g ^* _h(n+l) p ^* _{h((n+l)( npdN))} , but when l is a multiple of N, p _j(n(npdN)) = p ^* _{((n+l)(npdN))} , so p _j(n(npdN)) p ^* _{j ((n+l)(npdN))} = 1, and since Gi and Gh are mutually mates, _N 〓 ^m=1 g _in g ^* _(n+l) = 0. Therefore, the N multiple shift component of the interaction function of Qij and Qhj is 0, and Qij and Qhj are mutually mates.

The structure of the completely complementary sequence system derived in this way is shown in FIG.

The complex orthogonal matrix used to expand the sequence length is
Since any complex orthogonal matrix whose components are complex numbers with an absolute value of 1 can be used, there is a large degree of freedom in designing the complete complementary sequence system. For example, if only the Hadamard matrix is used, a completely complementary system having only ±1 as a component can be obtained.

Now, let's consider an "adaptive filter" that outputs a single pulse with no side lobes when inputted with a multiple difference orthogonal sequence of order N and sequence length N ⁿ in a synchronized state with this sequence mod N. I can do it.

This adaptive filter has a built-in N-order complex orthogonal matrix whose components are complex numbers with an absolute value of 1, and the adaptive multiple difference orthogonal sequence is synchronized with mod N (iN
There is no need to know the difference between +k time and jN+k time,
When input in the state (knowing only k), a self-complementary sequence system (consisting of N sequences) is generated from the input multiple difference orthogonal sequence and the built-in complex orthogonal matrix.
generate. Furthermore, matching filters for each of these N series are built in, and the N series are input to each matching filter. When the sum of the autocorrelation functions of each of the N sequences input and output from each matched filter is taken, a single pulse with no side lobes is obtained.

FIG. 4 shows the configuration of the adaptive filter for the multiple difference orthogonal series Gi. Details of this filter can be found in the patent application filed in 1983.
As shown in No. 97515.

Mod the matching series mate to this filter
The output response when input is synchronized with N will be explained.

A multiple _{difference orthogonal} sequence signal _C = (C ₁ , C ₂ , . . . C _No ), the output signal becomes 0. because,
For the multiple difference orthogonal sequence signal C, N multiple _difference orthogonal sequence signals (C ₁ , ..., C _j , ..., C _No ) is the N multiple difference orthogonal sequence system (A ₁ , ..., A _j , ..., A _No ) generated from the multiple difference orthogonal sequence signal A and the orthogonal matrix B. Since they are mutually complementary series, A _j
N compatible filters that meet (1≦j≦N),
If we input the corresponding sequences C _j (1≦j≦N) among the sequences generated from the input sequence C and take the sum of the output signals, that is, the cross-correlation function signals of C _j and A _j , we get It becomes 0 (no output).

As described above, an adaptive filter for multiple difference orthogonal sequences outputs a single pulse without side lobes when a compatible sequence is synchronized with mod N, and outputs no signal when a mate is synchronized with mod N. It can be seen that Therefore, N multiple difference orthogonal sequences that are mutually mates are mod N
When the synchronized and linearly added signals are input to each of the N mates' adaptive filters, each adaptive filter will output a single pulse only for the compatible series.

FIG. 5 is a schematic configuration diagram of a spread spectrum communication multiplex communication system as an embodiment of the present invention. The transmitting devices 1a, 1b to 1n each modulate and transmit a signal using N multiple difference orthogonal sequences that are in a mutually different mate relationship. Synthesizer 2
linearly adds and multiplexes these transmission signals. Further, the receiving devices 3a, 3b to 3n are provided with matching filters (FIG. 4) that form pairs with the transmitting devices 1a, 1b to 1n, respectively, and selectively receive the transmitted signals. In addition, the synchronizer 4 is the transmitter 1
A synchronizing signal is supplied to each of the receiving devices 3a, 1b to 1n and the receiving devices 3a, 3b to 3n.

Next, multiple communication channels using the present invention will be explained.

Now, an N-fold communication channel is constructed using N mutually mate multiple difference orthogonal sequences of order N and sequence length N ⁿ . That is, N sets of communicators (N senders and
N multiple difference orthogonal sequences are allocated to each of the corresponding N receivers in such a way that they do not overlap.

A sender modulates its signal with its transmitter. Meanwhile, other (N-1) senders similarly modulate with their respective transmitters. The modulated signals of these N senders are synchronized by mod N according to the synchronization signal output from the synchronizer 4, and linearly added by the combiner 2.

When the linearly added received signals that have passed through the communication path are synchronized with mod N and received by the receiving device, the adaptive filter in the receiving device is
A single pulse is output for a matching multiple difference orthogonal sequence, but no output is produced for a non-matching (mate) multiple difference orthogonal sequence and a 0 sequence. The output signal of the adaptive filter is therefore a pulse train in which the pulses are present at times corresponding to one of the original signals of the corresponding transmitted signal, the original believer being restored and interference from other communicators being eliminated. I don't receive it. This also applies to the other (N-1) recipients, so an N-fold communication channel without crosstalk can be realized.

As described above, according to the present invention, when there is a communication channel that is synchronized over the entire communication channel, N multiple difference orthogonal sequences that are in a mate relationship with each other can be
The signals synchronized with mod N and linearly added are
When the signal is input to each matching filter of each mate, each filter outputs a single pulse only for matching sequences, making it possible to perform multiplex communication that is resistant to noise and free of crosstalk.

[Brief explanation of the drawing]

Figure 1 is a diagram showing the absolute value of the autocorrelation function of an example of a multiple difference orthogonal sequence, Figure 2 is a diagram showing the structure of a complete complementary sequence system of order 2, and Figure 3 is a diagram showing the structure of another complete complementary sequence system. FIG. 4 is a diagram showing the structure, FIG. 4 is a diagram showing an adaptive filter based on a multiple difference orthogonal system, and FIG. 5 is a diagram showing an embodiment of the present invention. 1a, 1b,..., 1n... transmitter, 2... combiner, 3a, 3b,..., 3n... receiver, 4... synchronizer.

Claims (1)

[Claims] 1. A multiplex communication system that transmits information signals synchronized based on a predetermined synchronization signal over a communication channel, which maintains multiple difference orthogonal sequences that are mutually different mates, and a plurality of transmitting devices that modulate signals based on differential orthogonal sequences to generate transmission signals; a combiner that combines and multiplexes each of the transmission signals generated by the plurality of transmitting devices; comprising an adaptive filter holding each multiple difference orthogonal sequence corresponding to the multiple difference orthogonal sequence held in each of the multiple difference orthogonal sequences held in each of the transmitting apparatuses; 1. A multiplex communication system comprising: a receiving device that selectively receives a desired transmission signal based on.