JPH09162847A - Adaptive spread spectrum receiver - Google Patents

Abstract

PROBLEM TO BE SOLVED: To allow the receiver to be in excellent operation with a desired wave multi-path component in which timing synchronization of a desired wave is incomplete or a delay time difference is small. SOLUTION: An input reception signal is inverse-spread at an inverse spread synthesis means 36 by using spread codes orthogonal to each other from a desired wave and other user waves respectively and its inverse spread output MPS is multiplied with a weight W and the results are synthesized. A coefficient control means 37 obtains an optimum weight Wo=αR<-1> T' satisfying W"T'=1. A rotary angle detection means 45 detects a phase angle P(i) making a phase of an inverse spread output x1 (i) of a desired wave zero, the P(i) is multiplied with each inverse spread output by multipliers 461 -464 and the result is timewise averaged to obtain a steering vector T'.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a receiver used in a direct sequence code division multiple access system in spread spectrum communication and adapted to adaptively remove interference waves.

[0002]

2. Description of the Related Art In recent years, various spread spectrum systems have been studied in order to effectively use frequencies in digital mobile communication (MKSimon, JKOmura, RAScholtz.
and BKLevitt, "Spread Spectrum Communicatio
n ", Computer Science Press, 1985). Especially CDMA using the direct sequence (DS) method.
Since the (Code Division Multiple Access) method has a relatively simple configuration, a practical method is being studied.

In the conventional DS-CDMA receiver, the received wave from the antenna is amplified by the amplifier and input to the quasi-coherent detection circuit. The quasi-coherent detection circuit performs quadrature detection using a local carrier signal whose phase is not synchronized but whose frequency is synchronized as a reference signal, and the amplitude I (t) of the in-phase component of the received wave.
And the amplitude Q (t) of the quadrature component. Below I
(T) and Q (t) are collectively used as a received signal. Also, this received signal is usually displayed as a complex number with I (t) as the real part and jQ (t) as the imaginary part, where j is the imaginary unit.
Therefore, various arithmetic processes are complex operations. In the DS system, a despreading process is performed on a received signal to extract a despread signal of a desired wave. There are two methods for this despreading process.
One is a method using a matched filter matched to a spreading code, and the output signal of this filter is a despread signal. The other is a method in which the spread code is multiplied in synchronism with the timing of the received signal and then the DC component is extracted by a low-pass filter, where the DC component becomes a despread signal. In the following, a method using a matched filter will be described, but the same result can be obtained by other methods. The despread signal is demodulated in baseband to extract a transmission code sequence.

In such a DS-CDMA system, a plurality of users simultaneously use the same carrier frequency. Each user uses different spreading codes, but since these spreading codes are correlated with each other, the components of other users are mixed in the despreading signal even when despreading with the spreading code of the desired wave. . Therefore, when the number of other users is large, the level of the interference wave component mixed into the despread signal increases, and the transmission characteristics deteriorate significantly. This deterioration becomes even more problematic when the reception level of other users becomes higher than the reception level of the desired wave. Therefore, it is conceivable to control the transmission power of each user and make the level at the reception point of each user constant, but it is extremely difficult to completely control this transmission power. The deterioration of the transmission characteristics due to the cross-correlation between spreading codes can be solved by adding an interference canceling function to the receiver. The conventional solution and its disadvantages are described below.

A conventional interference canceller (K. Fukawa, and H. Suzuki, “Orthogonalizing M”) is particularly close to the present invention.
atched Filter (OMF), detection for DS-CD
MAmobile radio systems ”, IEEE Globecom 199
4, pp. 385-389, Nov. 1994 Japanese Patent Application No. 6-16836) shows the basic structure thereof. This receiver operates as follows. The received signal SIG having the in-phase component amplitude and the quadrature component amplitude of the received wave is input from the input terminal 31 to the sampling means 32, is sampled at regular time intervals, and is sampled signal SPS
And output to the terminal 11. The signal extraction means 33 receives the sampled signal SPS, performs despreading and linear combining operations, and outputs a combined signal DCS to the terminal 26. The demodulation means 34 demodulates the combined signal DCS and outputs the determination signal to the terminal 17. The timing control unit 35 controls the operation timing of each unit.

Further, the signal extraction means 33 comprises a despreading / synthesis means 36 and a coefficient control means 37, which operate as follows. First, the despreading / combining means 36 performs despreading and linear combination processing of the input sampled signal using the weighting coefficient W to obtain and output a combined signal DCS, and the weighting coefficient W is a product. It outputs the multiplied signal MPS to be performed. Next, the coefficient control means 37 receives the multiplied signal MPS and the combined signal DCS as input, and under the constraint condition of the weighting coefficient W, sets the weighting coefficient W calculated by the algorithm that minimizes the average power of the combined signal DCS. Output to the despreading / combining means 36.

Various algorithms are conceivable for minimizing the average power of the combined signal under the constraint condition of the weighting coefficient. A simple method is LMS with constraint condition.
rost method (Frost, OL, “An algorithm for lin
early constrained adaptive array processing ”, Pro
c. IEEE, vol. 60, No.8, PP.926-935, August197
2) is known.

As the despreading / combining means 36, a configuration example in which a despreading means 38 and a linear combining means 39 are connected in cascade is shown in FIG.
Shown in The despreading means 38 outputs a plurality of despreading signals obtained by despreading the input sampled signal SPS with a plurality of despreading codes. Also, a plurality of these despread signals are output as the multiplied signals. The linear combining means 39 is the despreading means 38.
After multiplying the plurality of despread signals obtained in step 1 by weighting coefficients, they are combined and output as a combined signal DCS. In the figure, the spreading factor is set to 4 and the number of users using the same frequency is set to 4 to simplify the explanation.

First, the input sampling signal from the input terminal 11
Are the four matched filters 12 forming the despreading means 38.
₁ ~ 12_Four Supplied to Matched filter 12₁ ~ 12_Four 
Then, the correlation calculation between the sampled signal and the spread code is performed.
The despread signal x of the calculation result₁(i) ~ x_Four(i) is multiplied
It is output as the signal MPS. One matched filter 12
₁ Is the correlation calculation with the spread code of the desired wave, and other matching
Filter 12_Two ~ 12_FourIs orthogonal to the spreading code of the desired wave,
In addition, the correlation calculation is performed with the spread codes that are orthogonal to each other. linear
The synthesizing unit 39 includes the multiplier 13₁ ~ 13_Four And adder 15
And a plurality of despread signals x from the despreading means 38.
₁(i) ~ x_FourWeighting coefficient W for (i)₁ ~ W_Four Multiply by
Then, the multiplication result is added by the adder 15 to obtain the combined signal DC.
S is generated and output to the output terminal 26. Coefficient control means 3
7 is a plurality of despread signals x₁(i) ~ x _Four(i) and synthesized signal DC
With S as input, combine under the constraint of weighting coefficient
Weight with an algorithm that minimizes the average power of the signal DCS
Coefficient W₁ ~ W_Four And output. Matched filter 1
2₁ ~ 12_Four Can be replaced by a correlator,
The same applies to the matched filter described below.

Four-dimensional weighting coefficient vector at this time
The optimum value of Wo = [Wo₁Wo_TwoWo_ThreeWo_Four]^TAnd then
The optimum value Wo under the constraint condition of the fitting coefficient is as follows.
Is clear from the Frost literature and the like. Wo = αR^-1T (1) where α is a scalar value and R is a 4 × 4 despread signal
Correlation matrix, T is a four-dimensional steering vector
You. R is the despread signal vector X (i) = [x₁(i) x _Two(i) x
_Three(i) x_Four(i))^TUsing R = <X (i) X^H(i)> It becomes like (2). Where i is the symbol period T as a unit
Time, Wo_jIs W_jThe optimal value of x_j(i) is the j-th (j
= 1, 2, 3, 4) matched filter 12_jTime in
the despread signal of i,^TIs the transposed matrix,^HIs the complex conjugate transposed row
The column, <>, represents the set average. This R is approximated as
can do.

R = [X (1) X ^H (1) + X (2) X ^H (2) + ... + X (N _t ) X ^H (N _t )] / N _t (3) where N _t is large It is a natural number. The larger N _t is, the higher the degree of approximation is. However, it is preferable that the N _t has a value that does not change the communication status such that another user newly starts communication during this N _t, and it depends on the system used.

In this case, the steering vector T is a vector whose element is the cross-correlation ρ _jk of the spreading codes of user j and user k, and T = [ρ ₁₁ ρ ₁₂ ρ ₁₃ ρ ₁₄ ] ^T (4) or Frost's In the above-mentioned document, for the adaptive array, the steering vector T is set to 1 for the element corresponding to the main antenna having the most directivity to the desired wave and zero for all other elements. If the signal spread by the spreading code of user 1 is the desired wave, the multiplier 13 ₁
The ratio of the desired wave included in the output is Wo ₁ ρ ₁₁ . Similarly, the ratios of the desired waves included in the outputs of the multipliers 13 _{2 to} 13 ₄ are Wo ₂ ρ ₁₂ , Wo ₃ ρ ₁₃ , and Wo ₄ ρ ₁₄ , respectively. Since the combined signal is the sum of the outputs of the multipliers 13 _{1 to} 13 ₄ , the ratio of the desired wave contained in the combined signal is Wo ₁ ρ ₁₁ + Wo ₂ ρ ₁₂ +
_{Wo 3 ρ 13 + Wo 4 ρ} 14 , and the corresponding W ^H T. Since the despreading codes of the matched filters 12 _{1 to} 12 ₄ are orthogonal to each other, the steering vector T is as follows.

T = [1000] ^T (5) The coefficient control means 39 controls the weighting coefficient so as to keep the signal level of the desired wave included in the combined signal constant. Since the spread codes are orthogonal to each other, the matched filters 12 ₂ to 1
The outputs x ₂ (i) to x ₄ (i) of 2 ₄ do not include the signal component of the desired wave. Considering this, the constraint condition of this weighting coefficient is expressed by ^WH T = 1 (6). Minimizing the average power of the combined signal under this constraint is equivalent to minimizing the average power of the combined signal while keeping the signal level of the desired wave constant.
By controlling the weighting coefficient in this manner, the level of the interference wave component included in the combined signal can be minimized. It should be noted that α in equation (1) is set so that Wo satisfies the constraint condition of equation (6).

[0014]

[Problems to be Solved by the Invention] Now, the timing of the desired wave
Hope that the ring synchronization is incomplete or the delay time difference is small
If there is a multipath component of the wave, the above matched filter 12
_Two~ 12_Four, Output signal x _Two(i) ~ x_FourSignal generation of desired wave in (i)
Minutes leak in. This leaked desired wave signal component is matched
Filter 12₁Output signal x₁Desired wave signal included in (i)
Correlates with ingredients. Therefore, the binding condition expressed by the formula (6)
The weighting factor to minimize the average power of the combined signal under
If you control the number, x₁The desired wave signal component included in (i) is x
_Two(i) ~ x_FourCanceled by the desired wave signal component leaking into (i)
This reduces the average power of the desired wave signal component included in the combined signal.
The error rate characteristic is significantly deteriorated.

In order to suppress this deterioration, it is necessary to newly estimate the steering vector T'in consideration of the leakage of the desired wave signal component, instead of T expressed by the equation (5). . An object of the present invention is to provide an adaptive spread spectrum receiver that operates well even when the timing synchronization of a desired wave is incomplete or there is a multipath component of the desired wave with a small delay time difference.

[0016]

An adaptive spread spectrum receiver according to the present invention comprises sampling means for sampling a received signal at regular time intervals and outputting a sampled signal, and despreading and linearization with the sampled signal as an input. Signal extraction means for performing a combining operation and outputting a combined signal; demodulation means for demodulating the combined signal and outputting a determination signal; and timing control means for controlling the operation timing of each means. The signal extraction means, which is the central part of these means, includes a composite signal generated by performing despreading and linear combination processing of the sampled signal using the weighting coefficient, and a multiplied signal to be multiplied by the weighting coefficient. Despreading / combining means for outputting, the multiplied signal and the combined signal, and the steering vector that determines the constraint condition of the weighting coefficient, are combined under the above constraint condition. A coefficient control means for outputting the average power weighting coefficients determined by the algorithm that minimizes the issue, and a steering vector controller for particular estimating a steering vector output in the present invention.

There are two types of steering vector control means, (1) one that inputs only the multiplied signal, and (2) one that inputs the multiplied signal and a judgment signal or a known training signal. There is. The sampling means samples the received signal at regular time intervals and outputs a sampled signal.The signal extraction means receives the sampled signal as input, performs despreading and linear combining operations, and outputs a combined signal. , The composite signal is demodulated and a determination signal is output, and the timing control means controls the operation timing of each means.
Further, in the signal extraction means, the despreading / combining means is
The despreading and linear combination processing of the sampled signal using the weighting coefficient is performed to generate a synthesized signal, and the multiplied signal to be multiplied by the weighting coefficient is output, and the coefficient control means is
Estimate the weighting coefficient obtained by the algorithm that minimizes the average power of the combined signal under the above constraint conditions, with the steering vector that determines the constraint condition of the multiplied signal and the combined signal and the weighting factor as input. Output,
The steering vector control means includes one that inputs only the multiplied signal and one that inputs the multiplied signal and the determination signal or a known training signal, and estimates and outputs the steering vector.

The present invention differs from the prior art in that the steering vector that determines the constraint condition of the weighting coefficient is not fixed, but is estimated using the multiplied signal or the like.

[0019]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The basic structure of the present invention is shown in FIG. 1A, and the portions corresponding to those in FIG. This receiver operates as follows. The received signal SIG composed of the in-phase component amplitude and the quadrature component amplitude of the received wave is input terminal 31.
The sampling signal SPS is supplied to the signal extracting means 33. The signal extraction means 33 receives the sampled signal SPS as input, performs despreading and linear combining operations, and combines the combined signal DCS.
Is output to the terminal 26. In the demodulation means 34, the composite signal D
The CS is demodulated and the determination signal OPS is output to the terminal 17.
The timing control means 35 controls the operation timing of each means 33, 34.

Further, the signal extracting means 33 comprises not only the despreading / synthesizing means 36 and the coefficient control means 37 but also a steering vector control means 41 in the present invention, which operate as follows. First, despreading / combining means 3
Reference numeral 6 denotes terminals 26 and 42, respectively, for the synthesized signal DCS generated by performing despreading and linear combination processing of the sampled signal using the weighting coefficient and the multiplied signal MPS to be multiplied by the weighting coefficient W. Output to. Next, the coefficient control means 37 receives as input the steering vector T ′ that determines the constraint condition of the multiplied signal MPS, the combined signal DCS, and the weighting coefficient W, and the average power of the combined signal is minimized under the above-mentioned constraint condition. The weighting coefficient W obtained by the algorithm is output to the terminal 43. And steering
The vector control means 41 estimates the steering vector T ′ based on the multiplied signal MPS and outputs it to the terminal 44. The specific configuration of the despreading / combining means 36 is the same as that shown in FIG. 6, and a configuration in which the despreading means 38 and the linear synthesizing means 39 are replaced with a transversal filter in the figure is also possible. This is clear from the above-mentioned document published by the present inventor.

An example of the construction of the steering vector control means 41 shown in FIG. 1A is shown in FIG. 1B (claim 2). The multiplied signal MPS is input from the terminals 42 _{1 to} 42 ₄ ,
The multiplied signal input from the terminal 42 ₁ is the matched filter 12
The output signal of ₁ is x ₁ (i). The rotation angle extraction means 45 receives the multiplied signal from the terminal 42 ₁ and multiplies the phase angle signal P (i) which makes the phase of this signal zero to the multipliers 46 ₁ to 4 4.
Output to 6 ₄ . It becomes (7) | Expressing this phase angle signal P (i) in the formula ^{_{P (i) = x 1 *}} (i) / | x 1 (i). Where ^* is the complex conjugate. Terminals 42 ₁ to 4
The multiplied signal MPS from 2 ₄ is multiplied by the phase angle signal P (i) in the multipliers 46 _{1 to} 46 ₄ and then time-averaged by the averaging means 47 to obtain a steering vector T ′ at the terminal 44. It is output to the _{1-44 4.}

The steering vector T'is equal to the cross-correlation vector between the desired wave signal component and the multiplied signal MPS, and the above operation approximately corresponds to extracting this cross-correlation vector. This will be explained below. The output signal x ₁ (i) of the matched filter 12 ₁ is almost equal to the desired wave signal component as long as the interference wave power and noise power are low. Therefore, multiplying the phase angle signal P (i) corresponds to canceling the carrier phase and the modulation phase of the desired wave signal component. Then, multiplying the multiplied signal MPS by the phase angle signal P (i) and taking the time average corresponds to extracting the cross-correlation vector between the desired wave signal component and the multiplied signal MPS.

Another configuration of the steering vector control means 41 shown in FIG. 1A is shown in FIG. 2 (claim 3). The multiplied signal MPS is input from the terminal 42 to the autocorrelation matrix generation means 51. The autocorrelation matrix generation means 51 calculates and outputs the autocorrelation matrix R of the multiplied signal based on the multiplied signal MPS according to the equation (3). The autocorrelation matrix R is subjected to eigenexpansion by the eigenexpansion means 52, and eigenvalues and eigenvectors are obtained. There are two types of eigenvectors, one corresponding to a noise signal and one corresponding to a signal component including an interference wave. Regarding the magnitude of the eigenvalue, the one corresponding to the former eigenvector is substantially equal to the noise power, and the one corresponding to the latter eigenvector is substantially equal to the signal power. When the noise power is low, a plurality of eigenvectors {el} = corresponding to signal components based on the magnitude of the eigenvalue
Can easily be selected {e ₁ ···· e _M}.
Here, M is the number of eigenvectors corresponding to the signal components. The eigenexpansion means 52 uses the plurality of eigenvectors {e
l} is output.

Steering vector T'to be obtained
Has a small deviation from the vector T shown in the equation (5),
It exists in the space S spanned by the eigenvector {el}.
The vector obtained by projecting the vector T of the equation (5) onto the space S exists in the space S and has the smallest deviation from the vector T. Therefore, this vector is approximated to the steering vector T ′. Considering that the eigenvectors {el} are normalized and orthogonal to each other, the steering vector T ′ to be obtained is T ′ = (e ₁ ^H T) e ₁ + (e ₂ ^H T) e ₂ + ... + ( e _M ^H T) e _M (8). Here, (e ₁ ^H T) is the inner product of e ₁ and T. Each of {e ₁ } to {e _M } is a four-dimensional vector, and (e ₁ ^H T) is a vector and the inner product of the vectors is a constant, so T ′ is a four-dimensional vector.

The signal space projecting means 53 receives the above eigenvector {el} and the vector T as input, generates T'according to the equation (8), and outputs it from the terminal 44. Another basic configuration of the present invention is shown in FIG. 3A (claim 4). This configuration is different from FIG. 1A in that steering vector control means 41
, The steering vector is estimated by inputting the multiplied signal MPS and the training signal stored in the training signal memory 55 in the training signal section and the determination signal OPS from the demodulation means 34 in the subsequent data signal section. Is.

A concrete configuration example of the steering vector control means 41 shown in FIG. 3A is shown in FIG. 3B (claim 5). Training signal or judgment signal O from the terminal 56
PS is input to the complex conjugate means 57. In the training signal section in which the training signal is input, the terminals 42 ₁ ...
The multiplied signal MPS from 42 ₄ is directly applied to the multiplier 58 ₁
To 58 to ₄ is input. On the other hand, when the determination signal is inputted, the multiplication signal MPS from the terminal 42 ₁ to 42 ₄ are inputted to the multiplier 58 ₁ to 58 ₄ are respectively delayed by determination delay by the delay element 59 ₁ to 59 ₄ of the demodulation means 34 To be done.
The complex conjugate means 57 is a training signal or a decision signal OP.
The complex conjugate of S is output to the multipliers 58 _{1 to} 58 ₄ . In the multiplier 58 ₁ to 58 _4, and the multiplied signal MPS that the multiplied signal MPS or delay, a training signal or the determination signal O
After being multiplied by each complex conjugate of PS, the averaging means 61
Is time averaged and the steering vector T '
Are output from the terminals 44 _{1 to} 44 ₄ .

The steering vector T'is equal to the cross-correlation vector between the desired wave signal component and the multiplied signal MPS, and the above operation corresponds to the extraction of this cross-correlation vector. Finally, the computer simulation result of this invention is shown in FIG. The simulation condition is B
The PSK-modulated signal was spread 16 times, the number of users was 8, and the transmission path was a multipath propagation path of 2 paths under static conditions. The delay time difference is 0.5 chip cycle. In the prior art, the error rate characteristic is significantly deteriorated, but it can be seen that the present invention suppresses this deterioration and improves the error rate characteristic in any of the embodiments.

[0028]

As described above, the present invention operates well even when the timing synchronization of the desired wave is incomplete or there is a multipath component of the desired wave with a small delay time difference. It is effective to use the same carrier frequency in a wireless system shared by many users. In mobile communication, since the user's call changes in time series, such information is automatically extracted from the received wave, and a receiver adaptive to the change is particularly effective.

[Brief description of the drawings]

1 is a block diagram showing a functional configuration of an embodiment of the invention of claim 1, and B is a block diagram showing a functional configuration example of a steering vector control means 41 in FIG. 1A.

FIG. 2 is a steering vector control means 4 in FIG. 1A.
The figure which shows the other example of 1 functional structure.

FIG. 3 is a block diagram showing a functional configuration of an embodiment of the invention of claim 4, and B is a block diagram showing a functional configuration example of the steering vector control means 41.

FIG. 4 is a graph showing a computer simulation result.

FIG. 5 is a block diagram showing a functional configuration of an adaptive spread spectrum receiver that performs a conventional interference canceller.

FIG. 6 is a block diagram showing a functional configuration of signal extraction means in FIG.

Claims (5)

[Claims] 1. Sampling means for sampling a received signal at regular time intervals and outputting a sampled signal, and signal extracting means for outputting a combined signal by performing despreading and linear combining operations with the sampled signal as an input. , A demodulation means for demodulating the combined signal and outputting a decision signal, and a timing control means for controlling the operation timing of each means, wherein the signal extraction means uses a weighting coefficient for the sampling signal. Despreading / combining means for outputting the combined signal generated by performing despreading and linear combining processing, and the multiplied signal to be multiplied by the weighting coefficient, the multiplied signal, and the combined signal , The weighting coefficient is an algorithm that minimizes the average power of the combined signal under the constraint condition by inputting a steering vector that determines the constraint condition of the weighting factor. An adaptive spread spectrum receiver comprising a coefficient control means for obtaining and outputting the above-mentioned signal, and a steering vector control means for outputting the steering vector with the multiplied signal as an input. 2. The steering vector control means means for rotating the multiplied signal so that the phase of a specific element thereof becomes zero, and the time average of the phase-rotated signal as the steering vector. The adaptive spread spectrum receiver according to claim 1, further comprising output means. 3. The steering vector control means obtains an autocorrelation matrix of the multiplied signal, a means for uniquely developing the autocorrelation matrix, and an eigenvector corresponding to a signal component obtained by the unique expansion. 2. The adaptive spread spectrum receiver according to claim 1, further comprising means for estimating and outputting the steering vector. 4. Sampling means for sampling a received signal at fixed time intervals and outputting a sampled signal, and signal extraction means for outputting a combined signal by performing despreading and linear combining operations with the sampled signal as an input. , A demodulation means for demodulating the combined signal and outputting a decision signal, and a timing control means for controlling the operation timing of each means, wherein the signal extraction means uses a weighting coefficient for the sampling signal. Despreading / combining means for outputting the combined signal generated by performing despreading and linear combining processing, and the multiplied signal to be multiplied by the weighting coefficient, the multiplied signal, and the combined signal , The weighting coefficient is an algorithm that minimizes the average power of the combined signal under the constraint condition by inputting a steering vector that determines the constraint condition of the weighting factor. And a steering vector control means for outputting the steering vector with the judgment signal or a known training signal as an input. Type spread spectrum receiver. 5. The steering vector control means, means for multiplying the multiplied signal by a complex conjugate of the determination signal or the training signal, and means for outputting a time average of the multiplication signal as the steering vector. 5. The adaptive spread spectrum receiver according to claim 4, characterized by: