JP2799474B2 - Code division multiplex signal reception method - Google Patents

Description

Description: TECHNICAL FIELD The present invention is applied to, for example, mobile communication, in which a base station receives L (L is 2 or more) communicators whose spectrum is spread by a short-period spreading sequence and a long-period spreading sequence, respectively. In particular, the present invention relates to a receiving method for receiving at least one signal and separating at least one of the signals, and more particularly to a receiving method of obtaining a despread output from which interference has been removed by performing de-correlation filtering on a despread sequence of the received signal.

2. Description of the Related Art Due to the excellent interference resistance and confidentiality of spread spectrum communication, studies for practical use of a code division multiplexing communication system (CDMA) using spread spectrum communication in various communication systems are becoming increasingly popular. The problem with the CDMA system is that there is a distance problem in which the power of the received signal received by the central station varies greatly depending on the location of the communication party. CDMA
In the method, since the same frequency band is shared by a plurality of communication parties, a transmission signal from another communication party becomes an interference wave that deteriorates communication quality with respect to a transmission signal from an arbitrary communication party.

For example, when a communicator near the base station and a distant communicator simultaneously communicate, the base station receives a large signal power from a near communicator, while a signal power from a distant communicator is received. Will be received small. This means that communication between a distant communicator and the base station is greatly degraded in reception characteristics due to interference from communication from a nearby communicator. Transmission power control has been conventionally studied as a technique for solving the near-far problem. In the transmission power control, the signal power received by the receiving station, or the signal power to interference power ratio determined from the received power is controlled so as to be constant regardless of the location of the communicator. Thus, uniform communication quality can be obtained.

A representative communication system in which the near-far problem is a main factor of characteristic deterioration is a mobile communication system. When the above-described transmission power control is performed in this mobile communication system, the improvement effect of the transmission power control on the proportion (area ratio) of the area where communication can be performed with a predetermined communication quality in a zone is described in WCYLe.
e, “Overview of Cellular CDMA”, IEEE Trans.VT, Vol.V
Analyzed by T-40, pp.291-302,1991. It has also been reported that if high-speed transmission power control that can follow the fading fluctuations that occur in the mobile communication radio wave propagation environment can be realized, the effective frequency utilization efficiency can be increased to about 20 times that of the North American AMPS mobile communication system. (See K.
S. Gilhousen, IMJacobs, R. Padovani, AJViterbi, LA
Weaver, Jr. and CEWheatley III, “On the Capacity o
fa Cellular CDMA System ", IEEE Trans.VT, Vol.VT-4
0, pp. 303-312, 1991).

However, the location ratio after transmission power control is greatly affected by control errors generated by various factors. For example, reference: E. Kudoh and T. Matsumoto, “Effect of Transmitter
Power Control Imperfections on Capacity in DS / CDM
A Cellular Mobile Radios ", Proc. Of IEEE ICC'92, Cica
go, pp. 310.1.1-6, 1992, discusses the effect of control errors on the relative frequency utilization efficiency of the mobile communication system described above. According to it, if there is a control error of 1dB,
It is shown that the relative frequency use efficiency drops to 29% (up) and 31% (down).

Meanwhile, recently with Ruxandra Lupas of Princeton University, USA
Sergio Verdu unveils a class of linear filters for binary asynchronous CDMA systems that suffer from additive Gaussian noise, which can estimate the transmitted signal from the received signal from each communicator, even if the received signal power varies. . This class is called an inverse correlation filter. The processing amount of this inverse correlation filter increases only in proportion to the number N of simultaneous communication parties,
It does not increase exponentially significantly. This means
Reference: R. Lupas and S. Verdu. “Near-Far Resistance of
Multiuser Detectors in Asynchronous Channels ", IEE
E Trans.COM, Vol.COM-38, pp.496-508, 1990 (hereinafter referred to as Reference 1).

Another effect other than the improvement of the effective frequency use efficiency when the CDMA is applied to the mobile communication system is that code management free can be realized. That is, in conventional FDMA and TDMA, interference power from other communication parties using the same frequency and the same time slot (these are referred to as the same channel) is reduced to a predetermined level or less. Reuse in multiple zones that are far apart. For this reason, conventional FDMA and TDMA require channel management for controlling co-channel interference. Channel management includes optimization of how the service area is divided into zones, how many channels are assigned to each zone, and at which time each channel is reused in which zone. Inevitably, it becomes extremely difficult for a plurality of operators to operate different systems using a certain frequency band.

In the case of CDMA, a "channel" corresponds to a spreading code. Therefore, the magnitude of the interference from different channels corresponds to the magnitude of the cross-correlation between the spreading codes. Since the cross-correlation between the spreading codes is not completely zero for a plurality of spreading codes used in a certain zone and a zone adjacent thereto, each channel in those zones receives interference from other channels. In CDMA using spread spectrum communication, interference from any other channel inside or outside such a zone with respect to an arbitrary channel is regarded as equivalent noise, and a desired signal is selected from these synthesized signals in the process of despreading a received signal. Is extracted. In other words, there is no distinction between interference from inside the zone and interference from outside the zone as long as interference from other communication parties is regarded as equivalent noise. That is, in the mobile communication system "designed to regard interference from other communication parties as equivalent noise", the problem of spreading code assignment does not occur. Therefore,
Code management free can be realized. Multiple operators can operate different systems using the same frequency band if these systems are "systems designed to regard interference from other communication parties as equivalent noise". .

Therefore, it is necessary to thoroughly randomize the spreading sequence in order to “consider interference from other communicators as equivalent noise”. This is achieved by performing spectrum spreading using both a short-period spreading sequence whose period is one information symbol time length of the information symbol to be transmitted and a long-period spreading sequence whose period is multiple information symbol times. it can.
In this case, the chip rates of the short-period spreading sequence and the long-period spreading sequence are the same, and after normal spreading by the short-period spreading sequence, the long-period spreading sequence is multiplied for each chip,
Or, conversely, spread by a short-period spreading sequence after spreading by a long-period spreading sequence achieves spectrum spreading by both.

By the way, it is theoretically possible to configure the above-described inverse correlation filter even in a system that performs spectrum spreading using both the short-period spreading sequence and the long-period spreading sequence. However, no concrete method has been shown before.

An object of the present invention is to receive a plurality of asynchronous CDMA signals spectrum-spread by both a short-period spreading sequence and a long-period spreading sequence at a base station, and detect a signal of each communication party by inverse correlation filtering. It is an object of the present invention to provide a code division multiplex signal receiving method which enables the above.

DISCLOSURE OF THE INVENTION The receiving method according to the present invention provides an L person and an L person spread spectrum by a short period spreading sequence and a long period spreading sequence, respectively.
Is a code division multiplexed signal receiving method for receiving a signal from a communication party of an integer of 2 or more and separating at least one of the signals, including the following steps: (a) converting the received signal to the L Each of them is despread with a spreading sequence for a human communicator to obtain L despread output sequences. (B) For symbol timing k, each of the ranges from the (k−g) th symbol timing to the (k + g) th symbol timing L × L-dimensional partial correlation matrix R ^{k + h} representing the cross-correlation of the spread sequences of the L communication parties at symbol timing
(1), R ^{k + h} (0), R ^{k + h} (−1) are ^defined as h = −g,..., 0,.
Where k is an arbitrary integer, g is a constant integer of 1 or more, and a correlation matrix R _k in the range of the symbol timing defined by these partial correlation matrices is generated, and its inverse correlation matrix R _k ^-1 is calculated. (C) multiplying the inverse correlation matrix R _k ⁻¹ by the L despread output sequence vectors from the (k−g) th symbol timing to the (k + g) th symbol timing obtained in the step (a) (D) The symbol judgment identification is performed on the multiplication result corresponding to at least one of the desired L communication parties in the multiplication result of the step (c).

In the above receiving method, each symbol timing k + after the symbol timing at which the inverse correlation matrix R _k ^-1 is obtained.
In contrast, the process for obtaining the inverse correlation matrix in the step (b) is as follows: (b-1) the partial correlation matrix R ^{k + g} (−1) and R
^{Using k + g + 1} (0) and the inverse correlation matrix R _k ^-1 obtained at the symbol timing k, an extended inverse correlation matrix R _{k, k + 1} ^-1 extended from the above inverse correlation matrix by one symbol timing is obtained. (B-2) calculating the inverse correlation matrix R _k ^-1 at symbol timing k + 1 from the extended inverse correlation matrix R _{k, k + 1} ^-1 .

In the present invention, by considering only the influence of the information symbol vector at the k-th symbol timing on the despread output vector before and after it, that is, only the time range in which the intersymbol interference sufficiently converges, by ignoring other symbol timings, Information symbols can be detected by inverse correlation filter processing.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a time chart showing transmission symbol sequences by a plurality of communicators, and FIG. 2 is a diagram showing a transmission symbol sequence in consideration of a cross-correlation of spread sequences by a plurality of communicators and a despread output of a reception signal thereof. The figure which shows the determinant which shows a relationship.

FIG. 3 is a diagram showing a determinant showing a relationship between a transmission symbol sequence and a despread output of a received signal, which are the basis of the present invention.

FIG. 4 is a block diagram showing a configuration of a transmitting side device of spread spectrum communication using a short-period spread sequence and a long-period spread sequence, FIG. 5 is a block diagram showing a configuration example of a receiving device to which the present invention is applied, and FIG. FIG. 7 is a diagram illustrating an example of simulation results of reception according to the method of the present invention and the conventional method. FIG. 7 is a diagram illustrating another example of simulation results of reception according to the method of the present invention and the conventional method.

BEST MODE FOR CARRYING OUT THE INVENTION The biggest reason that it is difficult to construct an inverse correlation filter for an asynchronous CDMA system using both short-period spreading sequences and long-period spreading sequences is that the cross-correlation between sequences is temporally ( (For each symbol). That is, in a CDMA system using only a short-period spreading sequence, the cross-correlation with another received signal during one symbol period is the same for each symbol. The cross-correlation in each symbol period within the period of the long-period spreading sequence is different and fluctuates. Now, consider a case where the base station simultaneously communicates with L (L is an integer of 2 or more) communicators in an asynchronous CDMA environment. A vector in which the despread output for the received signal from each communication party in the base station receiver is arranged in time order Is given by equation (1) shown in FIG.
Can be considered from. However, Is a despread output y obtained by despreading the received signal from each communicator at the k-th symbol timing using both long-period and short-period spreading codes for each channel.
_i is a vector in which (k) is arranged for i = 1 to L,
Next formula And t is the transpose. Also, Is an array of symbol vectors, Is an information symbol vector at the k-th symbol timing. Lines # 1 to #L show transmission information symbol sequences from correspondents # 1 to #L in FIG. In this case, without loss of generality, the received power of the signal from each communication party is 1
Has been normalized to

If the received power from each communicator is different, the information symbol vector Weighted vector instead of Obviously, it is sufficient to use. Here, the weight coefficient W is L × L in which the received power from each communication party is arranged in a diagonal component.
It is a diagonal matrix.

Is a noise vector. R ^k (0), R ^k (1), R ^k (-1)
^Are the communication parties ^#i and ^#j (1 ≦ i, ^#j ) forming the L × L-dimensional complex number space C ^LxL at the k-th symbol timing, respectively.
j ≦ L) is the corresponding spread sequence mutual partial correlation matrix between, those elements the formula ^{_{R ij k (m) = ∫S}} i k (t-τ i) S * j k (t + mT-τ j ) Dt, m = -1,0,1
(2) given by Here, * represents a complex conjugate, T represents a symbol length, and ∫ represents an integration of the time t from −∞ to ∞. Also, τ ₁ is the relative delay time of the i-th communicator, and 0 = τ ₁ ≦ τ ₂ ... ≦ τ _L <T without loss of generality. These partial correlation matrices satisfy the following equation: R ^k (−1) = R ^{k + 1} (1) ^H H represents a complex conjugate transpose. S _i ^k (t) is a spreading sequence at the k-th symbol timing of the i-th communicator (the product of the long-period spreading sequence and the short-period spreading sequence in the symbol section), and is a time section [(k−1) T, kT]. Is set to 0 outside the symbol section defined by. Therefore, the integration of equation (2) may be performed in the range of the section [(k-1) T, kT]. As described above, since the long period spreading sequence is a period of several symbol times in length, the spreading sequence S _i ^k for each symbol in the plurality of symbols times different.

The despread output Y of the signals from the L communicators can be expressed as in the above equation (1). Therefore, a vector Y in which the despread outputs are arranged in time order is obtained. Can be obtained in the order of time. However, since equation (1) means a linear equation of infinite dimension, it cannot be solved directly.

Therefore, when ignoring the other symbol timing and cutting out only the portion including the influence of the k-th symbol from each term of Expression (1), Expression (3) shown in FIG. 3 is obtained. Here, 2g + 1 is a time range in which the intersymbol interference sufficiently converges, g may be a constant value of, for example, about 2 to 4, and the truncation length (Truncati
on length). The despread output vector on the left side of equation (3) is
Y ^k , the matrix of the partial correlation matrix on the right side (simply called the correlation matrix)
Is represented by R _k , the symbol vector is represented by B ^k , and the noise vector is represented by N ^k , Equation (3) can be represented as Y ^k = R _k B ^k + N ^k . Therefore, when the inverse matrix of the correlation matrix R _k is represented as R _k ⁻¹ , the transmission symbol vector
B ^k can be expressed by the following equation.

_{^{B k = R k -1 Y k}} -R k -1 N k (4) in the first embodiment of the method according to the invention, the correlation matrix R _k
Is a vector in which the inverse matrix R _k ^-1 (referred to as the inverse correlation sequence) is calculated for each symbol timing, and the despread outputs are arranged. Multiplied by the information symbol vector The estimated vector of Get. As is apparent from equation (4), the noise vector N ^k
If each element is sufficiently smaller than the despread output and the truncation length is sufficiently large, B ′ (k) can be regarded as matching B (k).

Equation (3) is an information symbol vector at the k-th symbol timing. , K ± g, which are simultaneously obtained by multiplying the inverse correlation matrix R _k ^-1 by Y ^k. Symbol vector about Are not guaranteed. Therefore, In order to estimate the above, it is necessary to find the inverse correlation matrix of Expression (3) at other symbol timings for each symbol timing. However, the matrix R _k is (2g + 1) L × (2g + 1)
It is a matrix of L, and it is not practical to perform such a large size inverse matrix operation at each symbol timing because the amount of calculation is extremely large.

In a second embodiment of the method according to the invention, (2g + 1)
Perform the inverse matrix operation of L × (2g + 1) L matrix only once,
For subsequent timings, the amount of calculation is significantly reduced by sequentially updating the inverse correlation matrix by the following method. This technique is called a sliding escalator algorithm.

Sliding escalator algorithm Now, it is assumed that the inverse correlation matrix R _k ^-1 is known. This inverse correlation matrix
Consider finding an inverse correlation matrix R _{k + 1} ^-1 at a symbol timing advanced by one symbol timing from R _k ^-1 . Referring to FIG. 3, the upper left 2 gL × 2 gL submatrix of the correlation matrix R _{k + 1} (within the broken line block 3 _{k, k + 1} ), that is, one column of the rightmost partial correlation matrix and one column of the lowermost partial correlation matrix matrix excluding the row are consistent with submatrix lower right 2GL × 2GL in the correlation matrix R _k. Therefore, (2g + 1) L × (2g + 1) L created by overlapping the common part of the correlation matrices R _k and R _{k + 1.}
A matrix, that is, an extended correlation matrix R _{k, k + 1 obtained} by extending the size of the correlation matrix R _k by one symbol timing is represented by the following equation.

Here, if the inverse correlation matrix R _k ^-1 is used, the extended inverse correlation matrix R
It is mathematically easy to show that _{k, k + 1} ^-1 is expressed by the following equation from the right side of the first equality sign of the equation (5).

However, Also, Similarly, R _{k, k + 1} ^-1 from the right side of the second equality sign of equation (5)
Can be expressed as the following equation.

However, Also, This equation (9) is redefined as the following equation.

At this time, when Expression (9) and Expression (12) are compared, It is. From this, Is obtained, so Becomes

R _k ^-1 is known as described above. First, using the equation (2), the partial correlation R ^{k + g} (−
1) Calculate R ^{k + g + 1} (0) respectively. Based on the result, Equation (6) is calculated from Equations (7) and (8) (2g
+2) L × (2g + 2) L spread inverse correlation matrix R _{k, k + 1} ⁻¹
Is obtained, and the lower right (2g + 1) L × (2g + 1) L submatrix is _obtained as Q _{k + 1} in equation (12). L × on it
From the (2g + 1) L submatrix, the left (2g + 1) L × L submatrix, and the upper left L × L submatrix, Are obtained respectively. From these equations (1
5) is calculated to obtain an inverse correlation matrix R _{k + 1} ^-1 . Use this to R
_{k + 1} ^-1 Y ^{k + 1} is the symbol vector at symbol timing k + 1 Is calculated as an estimate of In the above description, the case where the inverse correlation matrix R _{k + 1} ^-1 for the next symbol timing k + 1 is obtained assuming that the inverse correlation matrix R _k ^-1 for the symbol timing k has already been obtained.
As an inverse correlation matrix R _k ^-1 is obtained for symbol timing k-1 immediately before by replacing the k-1, completely and to obtain the inverse correlation matrix R _k-1 for the current symbol timing k Are equivalent.

Accordingly, once the correlation matrix _k-1 in the equation (3) shown in FIG. 3 is calculated once at the first symbol timing, the calculation of the equation (6) and the calculation of the equation (15) are performed thereafter. Without calculating directly from _k , the inverse correlation matrix R _k ^-1 is
Each immediately preceding inverse correlation matrix R _k−1 ⁻¹ and partial correlation matrix R ^{k + g} (0)
And R ^{k + g−1} (−1), and can be updated at each symbol timing. In that case Equation (8) for calculating the inverse matrix of the L × L matrix and the equation (15) And an inverse matrix operation of an L × L matrix for calculating the matrix is required. However, since the processing amount required for the inverse matrix operation increases by the cube of the size of the matrix, the matrix is (2g + 1) L × (2g + 1) L matrix. Compared with the inverse matrix operation of the correlation matrix _Rk, the amount of operation can be significantly reduced.

In the above description, the received power from each communicator is normalized to 1, but when the received power from each communicator is different, a diagonal matrix W in which the received power from each communicator is arranged in a diagonal component is described. Use the information symbol vector May be used.

FIG. 4 shows a transmitting apparatus in a code division multiplex communication system in which a receiving method according to the present invention is implemented, and FIG.
Are shown below. As shown in FIG. 4, in the transmitting device of each communication party #i, the transmission symbol information from the input terminal 11 is a multiplier.
At 12 the spectrum is spread by being multiplied by the short-period spread sequence SSC _i from the terminal 13 and its spread output is further multiplied by the long-period spread sequence LSC _i from the terminal 15 by the multiplier 14 and further spread by the spectrum to obtain the spread output. Are transmitted from the transmitter 16 as radio waves. The period of the short-period spreading sequence SSC _i is equal to the symbol period T of the transmission information, the period of the long-period spreading sequence LSC _i is a plurality of symbol periods, and the chips of both spreading code sequences are synchronized. After spreading by the long period spreading sequence LSC _i above, it may be spread with a short period spreading sequence SSC _i. The configuration on the transmitting side is the same as the conventional one.

As shown in FIG. 5, the configuration of a receiving apparatus to which the present invention is applied is such that a spread spectrum signal from L communicators is
The received output is correlated with the spreading sequences SSC ₁ , LSC ₁ -SSC _L , and LSC _L corresponding to each of the communication parties # 1-# L generated from the spreading code generator 23 in the despreader 22. There are despread respectively by a matched filter or sliding correlator 22 ₁ through 22 _L at timing t ₁ ~t _L with the maximum.
A despread output vector composed of these L-sequence despread outputs is output at each symbol timing. The despread output vector at the k-th symbol timing is It is. This despread output vector Is input to a (2g + 1) first-in-first-out register, that is, a shift register 24, and each of the shift stages 23 _-g
Despread output vector to ~ 23 _g Are respectively supplied to the multiplier 25.
The multiplier 25 includes a partial correlation matrix calculator 26 and an inverse correlation matrix calculator.
Together with 27, an inverse correlator 30 is formed.

On the other hand, the spread code generator 23 is long corresponding to each correspondent #. 1 to # L, product _{LSC 1 · SSC 1 · LSC 2} · spreading code pair short period
SSC ₂ ,..., LSC _L and SSC _L are generated, and spreading sequences s _{1 to} s _L are respectively generated.
To the partial correlation matrix calculator 26. The partial correlation matrix operation unit 26 uses the timing signals t _{1 to} t _L from the correlator 22 to calculate
Relative delay time τ for all correspondents # i = 1, ..., L
_{1 to} τ _L are obtained, and from the spreading sequences s _{1 to} s _L provided from the spreading code generator 23, partial correlation matrices are obtained for all combinations (i, j) of the communicating parties using Expression (2). In this case, in the above-described first receiving method of the present invention, all the partial correlation matrices R ^{g + h} (1), R ^{g + h} (0), R for the symbol timings k + h, h = −g,. ^{g + h} (−1) is obtained from the equation (2) and given to the inverse correlation matrix calculation unit 27. The inverse correlation matrix calculation unit 27 generates a correlation matrix R _k composed of all of these partial correlation matrices, obtains an inverse correlation matrix R _k ⁻¹ as an inverse matrix of the correlation matrix R _k ^−1, and supplies the inverse matrix to the multiplier 25. The multiplier 25 estimates the product of the inverse correlation matrix R _k ^-1 and the despread output vector Y ^{k as} symbol vector information. Vector of symbol timing k Each component b ₁ in the '(k), ..., b L' (k) of each determination unit 2
The level is determined in step 8, and the result of determination is output as decoded symbols of the signals received from the communication parties # 1 to #L.

When the above-described sliding escalator algorithm, which is the second receiving method, is applied, the partial correlation matrix operation unit 26 uses the spreading sequences s _{1 to} s _L provided from the spreading code generator 23.
From equation (2), the partial correlation R in equations (7) and (8) (where k in the equations referred to hereinafter is replaced by k-1) for all combinations of communicating parties
^{k + g + 1} (−1) and R ^{k + g} (0) are calculated and given to the inverse correlation matrix calculator 27. The inverse correlation matrix calculator 27 calculates equations (7) and (8) using the partial correlation matrix and the inverse correlation matrix R _k-1 ^-1 obtained for the previous symbol timing k-1. Further, the result of the calculation is used to calculate equation (6), and the extended inverse correlation matrix R of (2g + 2) L × (2g + 2) L is calculated.
_{k, k + 1} ^-1 is obtained and its lower right (2g + 1) L × (2g + 1)
Let the submatrix of L be Q _k and L × (2g + 1) L on it
Corresponding from the submatrix, the left (2g + 1) L × L submatrix, and the upper left L × L submatrix And the equation (15) is calculated from these to obtain an inverse correlation matrix R _k ^-1 . The inverse correlation matrix is supplied to the multiplier 24, multiplied by the (2g + 1) despread output vectors input in the same manner as in the first method, and the estimated vector in the multiplication result is obtained. Are determined and identified by the determination identifier 28, and an output at the k-th symbol timing is obtained from the L communicating parties.

As described above, according to the present invention, a signal that has been spread using both a short-period spread sequence and a long-period spread sequence can be received by the inverse correlation filter processing.

Next, results of a computer simulation performed to verify the effectiveness of the present invention will be described. In the simulation, the primary modulation was BPSK. A Gold sequence (process gain = 31) with a length of 31 chips is used for the short-period spreading sequence.
A Gold sequence of 511 chips was used for the long-period spreading sequence. g =
4 and the number L of synchronous communicators is 5, and it is assumed that all the members are received with equal amplitude. Communicate in an asynchronous CDMA environment.

FIG. 6 is a simulation result, in which the horizontal axis represents the signal power to noise power ratio (SNR) after despreading, and the vertical axis represents the error rate. ● indicates the reception characteristic of the conventional matched filter, and ○ indicates the reception characteristic of the present invention. Also,
The dotted line indicates the theoretical value for a single communicator. The error rate characteristics of reception by a conventional matched filter, which is affected by interference, are greatly degraded from the case of a single communication, whereas the characteristics of the reception method of the present invention are almost equal to the theoretical values of a single communication. I do.

FIG. 7 shows a simulation result. In this case, the number L of synchronous communicators is 2, and the reception power of the second communicator is set to be 10 dB higher than that of the first communicator. This situation is typical of a near-far environment. The horizontal axis is the signal power to noise power ratio (SN) after despreading for the first party.
R), the vertical axis represents the error rate of the first correspondent. The symbol ● represents the characteristics of reception by the conventional matched filter, and the symbol 特性 represents the characteristics of reception by the present invention. As can be seen from FIG.
The error rate characteristic of the matched filter is greatly degraded from the case of a single communication party due to the influence of the perspective problem, whereas the characteristic by the receiving method of the present invention is not affected by the perspective problem.

────────────────────────────────────────────────── ─── Continuation of the front page (56) References JP-A-6-177854 (JP, A) JP-A-6-268630 (JP, A) JP-A-7-170242 (JP, A) (58) Field (Int.Cl. ⁶ , DB name) H04J 13/00

Claims (6)

(57) [Claims] 1. A system according to claim 1, wherein each of said plurality of persons, each of which is spread over a short-period spreading sequence and a long-period spreading sequence, receives a signal from a communication person of an integer of 2 or more and separates at least one of the signals. The method of receiving a code division multiplexed signal includes the following steps: (a) despreading the received signal with spreading sequences for the L communicating parties to obtain L despread output sequences; ) An L × L-dimensional partial correlation matrix R representing the cross-correlation of the spread sequence of the L communication persons at each symbol timing in the range of the (k−g) th symbol timing to the (k + g) th symbol timing with respect to the symbol timing k. ^{g + h}
(1), R ^{g + h} (0), R ^{g + h} (−1) are defined as h = −g,..., 0,.
Where k is an arbitrary integer, g is a constant integer of 1 or more, and a correlation matrix R _k in the range of the symbol timing defined by these partial correlation matrices is generated, and its inverse correlation matrix R _k ^-1 is calculated. (C) multiplying the inverse correlation matrix R _k ⁻¹ by the L despread output sequence vectors from the (k−g) th symbol timing to the (k + g) th symbol timing obtained in the step (a) (D) The symbol judgment identification is performed on the multiplication result corresponding to at least one of the desired L communication parties in the multiplication result of the step (c). 2. The receiving method according to claim 1, wherein the correlation matrix R _k at the symbol timing k in the step (b) is _represented by the following equation: Ask by 3. The receiving method according to claim 1, wherein the step (b) is performed for each symbol timing k + 1 after the symbol timing at which the inverse correlation matrix R _k ^-1 is obtained.
The process for obtaining the inverse correlation matrix in (b) is performed by: (b-1) a partial correlation matrix R ^{k + g} (−1) and R
^{Using k + g + 1} (0) and the inverse correlation matrix R _k ^-1 obtained at the symbol timing k, an extended inverse correlation matrix R _{k, k + 1} ^-1 extended from the above inverse correlation matrix by one symbol timing is obtained. (B-2) calculating the inverse correlation matrix R _k ^-1 at symbol timing k + 1 from the extended inverse correlation matrix R _{k, k + 1} ^-1 . 4. The receiving method according to claim 3, wherein said extended inverse correlation matrix R in said step (b-1)
_The process to generate _{k, k + 1} ^-1 is Is a process of calculating It is. 5. The receiving method according to claim 4, wherein said extended inverse correlation matrix R in said step (b-2) is used.
_The process of calculating the above-described inverse correlation matrix R _k ^-1 at symbol timing k + 1 from _{k, k + 1} ^-1 is represented by the following expression representing the extended inverse correlation matrix: Then, the lower right (2g + 1) L × (2g + 1) L sub-matrix in the extended inverse correlation matrix calculated in the step (b-1) is _defined as Q _{k + 1,} and the L × (2g + 1) L Matrix, the left (2g + 1) L × L submatrix, and the upper left L × L submatrix Then Is a process of calculating by 6. The receiving method according to claim 1, wherein said partial correlation matrix at symbol timing k is given by the following equation: R _ij ^k (m) = ∫S _i ^k (t−τ _i ) S ^* _j ^k (T + mT−τ _j ) dt, m = −1,0,1 where ∫ is the integral of time t from (k−1) T to kT, and S _i ^k (t) is the i-th In the spreading sequence at the k-th symbol timing of the communication party, the time interval [(k−
1) T, kT] is 0, T is a symbol length, τ _i is a relative delay time of a signal received from the i-th communicator, and * represents a complex conjugate.