JP2007074211A - Multi-user receiving apparatus and method of receiving multi-users - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To reduce deterioration in the weight estimation accuracy of a succeeding stage due to the propagation of a weight estimation error of a preceding stage in a multi-user receiving apparatus for separately receiving signals from a plurality of users by blind processing. <P>SOLUTION: Receiving signals of a plurality of antennas 1_1 to 1_M are sent to a whitening circuit 3, which whitens these receiving signals and sends whitened signals to information series extraction circuits 4_1 to 4_N connected as a plurality of stages. Each of the information series extraction circuits 4_1 to 4_M estimates weight for extracting information series included in an input signal from the input signal, separates the signal by using the estimated weight, reproduces the information series from the separated signal, outputs the reproduced information series from a first output terminal, calculates the likelihood of weight to the weight estimation result of the preceding stage, outputs a coordinate transformation result performed so that a receiving signal space parallel with the weight estimated in accordance with the likelihood is reduced from a second output terminal to the information series extraction means of the succeeding stage. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a multiuser receiving apparatus and a multiuser receiving method capable of separately receiving radio signals from a plurality of users existing on the same frequency channel using a plurality of antennas in a digital radio communication system.

  In order to improve frequency utilization efficiency, multi-user detection technology that uses a plurality of antennas to separately receive radio signals from a plurality of users existing on the same frequency channel has attracted attention. In multi-user detection using multiple antennas, each received signal from multiple antennas is multiplied by a predetermined weighting factor, and the results are combined to enhance the signal from a specific direction, and a different specification. A technique called an adaptive array that suppresses signals from the direction is used. Therefore, in multi-user detection using a plurality of antennas, a method for estimating a plurality of weight coefficients (hereinafter, weights) for controlling the adaptive array is required in order to separate radio signals of a plurality of users.

  The weight estimation method can be broadly divided into a method in which the transmission side assigns a known preamble independently to each other by transmission and reception on the transmission side, and the reception side performs estimation using that, and the reception side does not use the preamble. There is a technique for performing estimation using only an unknown information sequence. However, particularly in packet communication, there is a problem that the throughput is greatly reduced due to the assigned preamble, and when communication is performed to assign an independent preamble for each user, the throughput is further reduced due to the communication. Therefore, the latter method is promising. The latter is called blind processing because an unknown information sequence is used on the receiving side.

As a method of estimating a plurality of weights by blind processing, a method of sequentially estimating weights by CMA (Constant Modulus Algorithm) while imposing a certain orthogonal condition is known (for example, Non-Patent Document 1). This technique uses the feature that transmission signals from different users are uncorrelated, and the output signals when the adaptive array is controlled by the ideal weight of each user are uncorrelated with each other, that is, the received signals are converted to X ₀ (t), the weight of the _mth user is w _m

By imposing the constraint condition, and performing weight estimation sequentially by CMA, weight estimation of a plurality of users is performed.

  FIG. 6 is a diagram showing a configuration of a conventional multi-user receiving apparatus 201 having a function of separating and receiving radio signals from a plurality of users existing in the same frequency channel using a plurality of antennas by blind processing.

  In FIG. 6, the multi-user receiving device 201 includes antennas 51_1 to 51_M, radio units 52_1 to 52_M, and information series extraction circuits 54_1 to 54_N.

  The information series extraction circuits 54_1 to 54_N all have the same configuration. For example, the n-stage information series extraction circuit 54_n includes a weight estimation circuit 61_n, a signal separation circuit 62_n, and a demodulation circuit 63_n, as shown in FIG.

  The multi-user receiving apparatus 201 in the prior art receives radio signals with the antennas 51_1 to 51_M, and performs frequency conversion and conversion to digital signals on the signals received by the radio units 52_1 to 52_M. Output signals of the radio units 52_1 to 52_M are supplied to information series extraction circuits 54_1 to 54_N connected in multiple stages. Since all the information series extraction circuits 54_1 to 54_N perform the same operation, the operations of the weight estimation circuit 61_n, the signal separation circuit 62_n, and the demodulation circuit 63_n constituting the n-th information series extraction circuit 54_n will be described below with reference to FIG. Will be explained.

  In FIG. 7, the received digital signal is input to the input terminal IN101 of the information series extraction circuit 54_n, and this digital signal is supplied to the weight estimation circuit 61_n and the signal separation circuit 62_n and output from the output terminal OUT102. The weight estimated up to the previous stage is input to the input terminal IN102 of the information series extraction circuit 54_n.

The weight estimation circuit 61_n uses the weights estimated up to the previous stage from the input terminal IN102, that is, the weights estimated by the weight estimation circuits 61_1 to 61_n-1, while imposing the constraint condition expressed by the equation (1). The weight is estimated by CMA (Constant Modulus Algorithm). The estimated weight is output from the output terminal OUT103 and sent to the next stage. The signal separation circuit 62_n extracts a signal based on the weight estimated by the weight estimation circuit 61_n. In the demodulation circuit 63_n, each information series is reproduced from the extracted signal and output from the output terminal OUT101.
Tuan Nguyen and Zhi Ding, "CMA beamforming for multipath correlated sources" ICASSP-97, vo.3, pp.2521-2524, 2002

  However, in general, the above-described weight estimation by CMA deteriorates the estimation accuracy when the signal sequence length to be used is short or when the SNR (Signal to Noise Ratio) is low. In the multi-user receiving apparatus described above, weights are estimated by CMA under the constraint condition shown in equation (1). However, if the weight estimation accuracy in the previous stage deteriorates, each weight does not satisfy equation (1). The estimation error is propagated to the subsequent stage, and the weight estimation accuracy in the subsequent stage is greatly degraded. As a result, the operation of the adaptive array becomes inaccurate, the SINR (Signal to Interference plus Noise Ratio) of the adaptive array output signal decreases, and the probability that the information sequence is correctly reproduced decreases.

  The present invention has been made in view of the above, and an object of the present invention is to improve the weight estimation accuracy of the subsequent stage by propagation of the weight estimation error of the previous stage in a multi-user receiving apparatus that separately receives signals from a plurality of users by blind processing. An object of the present invention is to provide a multi-user receiving apparatus and a multi-user receiving method capable of avoiding a decrease in SINR of an adaptive array output signal and extracting more information sequences by reducing degradation.

  In order to solve the above-described problem, the present invention is a multi-user receiving apparatus that receives signals transmitted from a plurality of users, and includes M (M ≧ 2) receiving antennas and M-sequence input signals. A whitening means for whitening and outputting, a first output unit for inputting the whitened M-sequence signal, extracting and outputting an information sequence included in the input signal, and a subsequent-stage information sequence extracting means And N (N ≧ 2) information sequence extraction means provided with a second output unit for outputting an M-sequence signal. The information sequence extraction means converts the input signal into the input signal. Weight estimation means for estimating weights for extracting the included information series, signal separation means for separating signals from the input signal using the estimated weights, and information series from the signals separated by the signal separation means Play and Demodulating means for outputting to one output unit, likelihood calculating means for calculating and outputting the likelihood of the estimated weight from the estimated weight and the signal separated by the signal separating means, and the input signal And a coordinate conversion means for performing coordinate conversion so as to reduce the reception signal space parallel to the estimated weight in accordance with the calculated likelihood, and then outputting to the second output unit. A multi-user receiving device.

  The present invention is the multiuser receiving apparatus according to the invention described above, wherein the likelihood calculating means calculates the likelihood based on an average value of amplitude distortion of the signal separated by the signal separating means. Features.

  The present invention is a multi-user receiving method for receiving signals transmitted from a plurality of users, the step of receiving signals by M (M ≧ 2) receiving antennas, and a white signal for M-sequence input signals. A whitening step for generating and outputting, a first output unit for inputting the whitened M-sequence signal, extracting and outputting an information sequence included in the input signal, and a subsequent information sequence extracting step And an N number (N ≧ 2) information sequence extraction step including a second output unit for outputting a signal, and the information sequence extraction step is included in the input signal from the input signal. A weight estimation step for estimating a weight for extracting an information sequence, a signal separation step for separating the input signal using the estimated weight, and a signal separation step. The likelihood of the estimated weight is calculated and output from the demodulation step of reproducing the information sequence from the received signal and outputting it to the first output unit, and the estimated weight and the signal separated by the signal separation step. A likelihood calculating step, and coordinates to be output to the second output unit after the input signal is subjected to coordinate conversion so as to reduce the reception signal space parallel to the estimated weight according to the calculated likelihood A multi-user reception method comprising: a conversion step.

  The present invention is the multiuser reception method according to the invention described above, wherein the likelihood calculating means calculates the likelihood based on an average value of amplitude distortion of the signal separated by the signal separating means. Features.

According to the present invention, the whitening means performs whitening processing on the received signal, and the information series extraction means estimates the weight for extracting the information series included in the input signal from the input signal, and estimates Signals were separated using weights, an information sequence was reproduced from the separated signals and output from the first output terminal, and coordinate conversion was performed so as to reduce the reception signal space parallel to the previous weights. The result is output from the second output terminal and output to the information sequence extraction means at the next stage.
As a result, it is possible to sequentially perform weight estimation for all users while relaxing the constraint condition shown in Equation (1) according to the likelihood of the weight estimated up to the previous stage, so that there is an estimation error in the previous stage weight. Even if it occurs, it is possible to reduce the influence on the subsequent stage and realize highly accurate signal separation reception.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a functional configuration of a multiuser receiving apparatus 101 according to an embodiment of the present invention.

  As shown in FIG. 1, the multi-user receiving apparatus 101 according to the embodiment of the present invention includes antennas 1_1 to 1_M, radio units 2_1 to 2_M, a whitening circuit 3, and N (N ≧ 2) multistage connections. Information series extraction circuits 4_1 to 4_N.

  The antennas 1_1 to 1_M are M (M ≧ 2) receiving antennas that receive radio signals from a plurality of users existing in the same frequency channel.

  The radio units 2_1 to 2_M perform frequency conversion and conversion to digital signals on the signals received by the antennas 1_1 to 1_M.

  The whitening circuit 3 inputs digital signals from the radio units 2_1 to 2_M and whitens the input signals.

  The information series extraction circuits 4_1 to 4_N all have the same configuration. For example, as shown in FIG. 2, the first information series extraction circuit 4_1 includes a weight estimation circuit 21_1, a signal separation circuit 22_1, and a demodulation circuit 23_1. The likelihood calculation circuit 24_1 and the coordinate conversion circuit 25_1 are configured. The number of stages N of the information series extraction circuit is preferably equal to or greater than the number of user stations that transmit data simultaneously.

  As shown in FIG. 2, each of the information series extraction circuits 4_1 to 4_N includes an input terminal IN1, a first output terminal OUT1, and a second output terminal OUT2. From the input signal from the input terminal IN1, A weight for extracting an information sequence included in the signal is estimated, a signal is separated using the estimated weight, an information sequence is reproduced from the separated signal, and is output from the first output terminal OUT1. The likelihood of the estimated weight is calculated from the estimated weight and the separated signal, and coordinate conversion is performed to reduce the reception signal space parallel to the estimated weight according to the calculated likelihood. Output from the second output terminal OUT2.

  Of the information series extraction circuits 4_1 to 4_N, the last information series extraction circuit 4_N only needs to have the first output terminal OUT1 for outputting the reproduced information series, and the second output terminal OUT2 Is unnecessary.

  Next, the operation of the multiuser receiving apparatus 101 according to the embodiment of the present invention will be described. In FIG. 1, M antennas 1_1 to 1_M receive radio signals from a plurality of users. Received signals of the antennas 1_1 to 1_M are supplied to the radio units 2_1 to 2_M, respectively. Radio units 2_1 to 2_M perform frequency conversion and conversion to digital signals on the received signals.

  The radio units 2_1 to 2_M output M series of digital signals. The M series of digital signals are supplied to the whitening circuit 3. This signal is whitened by the whitening circuit 3.

  The output signal of the whitening circuit 3 is supplied to the information series extraction circuits 4_1 to 4_N connected in multiple stages. Each of the information series extraction circuits 4_1 to 4_N extracts each information series included in the whitened signal, and sequentially outputs the demodulated information series signals from the first output terminal OUT1. Also, each of the information series extraction circuits 4_1 to 4_N calculates the likelihood of the weight with respect to the previous weight estimation result, and coordinates so as to reduce the reception signal space parallel to the weight estimated according to the likelihood. The data is converted and output from the second output terminal OUT2, and sent to the information sequence extraction circuit at the next stage.

A multi-user receiving apparatus 101 according to an embodiment of the present invention includes a whitening circuit 3 and N information series extraction circuits 41_ to 4_N connected in multiple stages. In the whitening circuit 3, the received signal is whitened. Processing is performed, and in each of the information series extraction circuits 4_1 to 4_N, the likelihood of the weight is calculated for the previous weight estimation result, and the reception signal space parallel to the weight estimated according to the likelihood is reduced. The result of the coordinate conversion is output from the second output terminal OUT2 and output to the next stage.
As a result, it is possible to sequentially perform weight estimation for all users while relaxing the constraint condition shown in Equation (1) according to the likelihood of the weight estimated up to the previous stage, so that there is an estimation error in the previous stage weight. Even if it occurs, it is possible to reduce the influence on the subsequent stage and realize highly accurate signal separation reception.

Next, the operation of each circuit constituting the whitening circuit 3 according to the embodiment of the present invention will be described. The whitening circuit 3 calculates a reception correlation matrix R _xx from the input signal X ₀ (t) by the following equation.

  Here, the subscript H represents a conjugate transposition, and E [] represents an average operation.

  Then, a whitening matrix Q satisfying the following expression is calculated using, for example, Cholesky decomposition.

Further, an output signal X ₁ (t) is calculated from the input signal X ₀ (t) and the whitening matrix Q by the following equation, and is output.

  Next, each circuit constituting the information series extraction circuits 4_1 to 4_N according to the embodiment of the present invention will be described.

Since all the information series extraction circuits 4_1 to 4_N perform the same operation, only the operation of the information series extraction circuit 4_1 will be described below. In FIG. 2, the weight estimation circuit 21_1 estimates a weight w ₁ that separates an information sequence included in the input signal X ₁ (t) from the input signal X ₁ (t). The signal separation circuit 22_1 separates the signal y ₁ (t) from the input signal X ₁ (t) and the weight w ₁ estimated by the weight estimation circuit 21_1 by the following equation.

The demodulation circuit 23_1 reproduces the information series from the extracted signal y ₁ (t) and outputs it to the first output terminal OUT1.

The likelihood calculating circuits 24_1 calculates a likelihood alpha ₁ of the estimated weights w1. By the way, the whitening circuit 3 makes the above-described constraint condition a simple constraint condition in which the weights are orthogonal to each other.

Therefore, when the weights w ₁ which is estimated there is no error, which the may be performed subsequent weight estimation in space perpendicular.

However, generally the weight w ₁ that is estimated to have errors, because no existing in a space orthogonal subsequent weights weights w ₁ and necessarily estimated, also taking into account the space parallel to the weights w ₁ estimated, subsequent Needs to be estimated.

Such order to realize the operation, the coordinate transformation circuit 25_1, based on the likelihood alpha ₁ calculated weights w ₁ and estimated to calculate the coordinate transformation matrix F ₁ that satisfies the following expression.

  Here, E represents a unit matrix, and ‖‖ represents a vector norm calculation.

Then, from the input signal X ₁ (t) and the coordinate transformation matrix F ₁ , an output signal X ₂ (t) obtained by reducing the space parallel to the estimated weight according to the likelihood is calculated by the following equation.

  Then, the calculated signal is output to the second output terminal OUT2, and the next stage process is performed using the signal in which the space parallel to the estimated weight is reduced.

  Next, a detailed operation of the weight estimation circuit 21_1 will be described. In the following description, LS-CMA is used as the weight estimation algorithm, but the weight estimation algorithm is not limited to this. In LS (Least Squares) -CMA, a weight that minimizes the distortion of the envelope of the output signal after signal separation, that is, minimizes the following equation is estimated.

  Here, p and q are positive integers and usually take 1 or 2.

  Specifically, when P = 1 and q = 2, the processing shown in the flowchart in FIG. 3 is performed to estimate the weight.

3, at step S1, setting an initial weight _{w 1.} Next, in step S2, the output signal y ₁ (t) is calculated from the weight w ₁ and the input signal as follows.

Next, in step S3, a reference signal δ ₁ (t) is generated from the output signal y ₁ (t) as follows.

Next, in step S4, the estimated weights _{w 1} according to the following equation.

  In step S5, it is determined whether or not the operations in steps S2 to S4 have been repeated a predetermined number of times. If the predetermined number of times has not been reached, the process returns to step S2. When the operations in steps S2 to S4 are repeated a predetermined number of times, the weight estimation process is terminated.

In LS-CMA, it is common to multiply the inverse matrix of the correlation matrix _Rxx from the left in step S4 for estimating the weight w1, but in the multi-user receiving apparatus 101 according to the embodiment of the present invention, white is used. Since the correlation matrix R _xx becomes a unit matrix by the conversion circuit 3, this operation is not necessary.

  Next, a detailed operation of the likelihood calculating circuit 24_1 will be described. In the following example, the likelihood is calculated based on the average value of the amplitude distortion of the signal separated by the signal separation circuit, but the configuration of the likelihood calculation circuit is not limited to this. .

  As shown in FIG. 4, the likelihood calculation circuit 24_1 includes an amplitude distortion detection circuit 31_1, an averaging circuit 32_1, and a likelihood conversion circuit 33_1.

The amplitude distortion detection circuit 31_1 detects the amplitude distortion σ ₁ from the input signal, that is, the signal separated by the signal separation circuit 22_1 by the following equation.

Averaging circuit 32_1 is averages the amplitude distortion sigma _1, to obtain an average amplitude distortion. The likelihood conversion circuit 33_1 converts the average amplitude distortion into likelihood. An example of the likelihood conversion method is shown in the following equation, but the likelihood conversion method is not limited to this.

Here, σ ₁ bar is the average amplitude distortion, and β is a positive constant.

FIG. 5 is a simulation result of the conventional multiuser receiving apparatus and the multiuser receiving apparatus according to the embodiment of the present invention, and shows the output signal y _n (t) of the signal separation circuit, that is, the SINR of the input signal of the demodulation circuit. Complementary cumulative probability. Since the error rate of the demodulated information sequence is determined by the SINR of the input signal of the demodulation circuit, the probability of extracting a correct information sequence (information sequence extraction probability) can be analogized by this simulation. Note that the SINR of each user is defined as the highest SINR among the SINRs for all weights estimated by the weight estimation circuit. The simulation conditions were set as shown in the table below.

  As can be seen from FIG. 5, the multi-user receiving apparatus according to the embodiment of the present invention has a higher SINR than the conventional multi-user receiving apparatus. If QPSK is used for the modulation scheme and convolutional codes (r = 1/2, k = 7) are used for error correction, the demodulated information sequence is almost error-free if the SINR is 3 dB or more.

  From FIG. 5, the probability of satisfying this condition is about 84% in the conventional multiuser receiving apparatus and about 90% in the multiuser receiving apparatus according to the embodiment of the present invention, and the information sequence extraction probability of about 6% according to the present invention. Improvement can be expected.

  As described above in detail, the multi-user receiving apparatus according to the embodiment of the present invention performs incomplete orthogonalization according to the likelihood of the weight estimated in the previous stage when separately receiving signals from a plurality of users. After that, the subsequent weight estimation is performed. By operating in this way, it is possible to prevent the weight estimation error in the previous stage from propagating to the subsequent stage, and to reduce the decrease in the weight estimation error in the subsequent stage.

  The present invention is not limited to the above-described embodiments, and various modifications and applications can be made without departing from the gist of the present invention.

  The present invention can be used to separate and detect radio signals from a plurality of users existing in the same frequency channel using a plurality of antennas.

It is a block diagram which shows roughly the structure of the multiuser receiver which concerns on embodiment of this invention. It is a block diagram which shows roughly the structure of the information series extraction circuit in the multiuser receiver which concerns on embodiment of this invention. It is a flowchart used for description of the estimation of the weight in the multiuser receiver which concerns on embodiment of this invention. It is a block diagram which shows roughly the structure of the likelihood calculation circuit in the multiuser receiver which concerns on embodiment of this invention. It is a graph which shows the simulation result of the conventional multiuser receiver and the multiuser receiver which concerns on this invention, Comprising: The complementary cumulative probability of SINR of the output signal of a signal separation circuit. It is a block diagram which shows schematically the structure of the conventional multiuser receiver. It is a block diagram which shows roughly the structure of the information series extraction circuit in the conventional multiuser receiver.

Explanation of symbols

1_1 to 1_M: antenna,
2_1 to 2_M: wireless unit,
3: Whitening circuit,
4_1 to 4_N: information series extraction circuit,
21_1 to 21_N: weight estimation circuit,
22_1 to 22_N: signal separation circuit,
23_1 to 23_N: demodulation circuit,
24_1 to 24_N: likelihood calculation circuit,
25_1 to 24_N: coordinate conversion circuit,
31_1 to 31_N: amplitude distortion detection circuit,
32_1 to 32_N: averaging circuit,
33_1 to 33_N: likelihood conversion circuit

Claims (4)

A multi-user receiving device that receives signals transmitted from a plurality of users,
M (M ≧ 2) receiving antennas;
Whitening means for whitening and outputting an M-sequence input signal;
A first output unit that inputs the whitened M-sequence signal, extracts and outputs an information sequence included in the input signal, and outputs an M-sequence signal to a subsequent information sequence extraction unit. And N (N ≧ 2) information series extracting means having two output units,
The information series extracting means includes
Weight estimation means for estimating a weight for extracting an information sequence included in the input signal from the input signal;
Signal separation means for separating the input signal using the estimated weight;
Demodulating means for reproducing an information sequence from the signal separated by the signal separating means and outputting it to a first output unit;
Likelihood calculating means for calculating and outputting the likelihood of the estimated weight from the estimated weight and the signal separated by the signal separating means;
Coordinate conversion means for converting the input signal into a second output unit after performing coordinate conversion so as to reduce the reception signal space parallel to the estimated weight according to the calculated likelihood. A multi-user receiving device. The multi-user receiving device according to claim 1,
The likelihood calculating means includes
A likelihood is calculated based on an average value of amplitude distortion of the signal separated by the signal separation means. A multi-user reception method for receiving signals transmitted from a plurality of users,
Receiving signals by M (M ≧ 2) receiving antennas;
A whitening step for whitening and outputting an M-sequence input signal;
A first output unit that inputs the whitened M-sequence signal, extracts and outputs an information sequence included in the input signal, and outputs an M-sequence signal for the subsequent information sequence extraction step. N (N ≧ 2) information sequence extraction steps each having two output units,
The information series extraction step includes:
A weight estimation step for estimating a weight for extracting an information sequence included in the input signal from the input signal;
A signal separation step of separating the input signal using the estimated weight;
A demodulation step of reproducing an information sequence from the signal separated by the signal separation step and outputting the information sequence to a first output unit;
A likelihood calculating step of calculating and outputting the likelihood of the estimated weight from the estimated weight and the signal separated by the signal separation step;
A coordinate conversion step of performing coordinate conversion on the input signal so as to reduce a reception signal space parallel to the estimated weight according to the calculated likelihood, and then outputting the converted signal to a second output unit. A multi-user receiving method characterized by the above. The multi-user receiving method according to claim 3,
The likelihood calculating means includes
A likelihood of calculating a likelihood based on an average value of amplitude distortion of a signal separated by the signal separation means.

Images