KR20030068782A - Method for daptive beamforming system of smart antenna - Google Patents

Abstract

The present invention relates to a multi-input multi-output adaptive beamforming system and method for detecting multiple users at the same time, unlike conventional single-user detection methods among adaptive beamforming methods of smart antennas. After multiplying the signal received by the array antenna by the search vector to find the correlation, the search step of finding the incidence angle with a high correlation is used. It has a convergence step of converging each signal into a real signal by applying the single user detection method.

Description

Adaptive beamforming method of smart antenna {METHOD FOR DAPTIVE BEAMFORMING SYSTEM OF SMART ANTENNA}

The present invention relates to a method of using a smart antenna, and more particularly, to an adaptive beamforming method for detecting multiple users of a smart antenna.

An antenna that improves wireless communication efficiency by forming an optimal beam pattern by adjusting a beam pattern by multiplying a received signal by a weight using an array antenna is called a smart antenna. By using a plurality of array antenna elements to adjust the gain (Gain) and phase (Phasor) of the signals received from each antenna element, the base station receives only the signal propagated from the direction of the desired user and propagates in other directions By greatly reducing the noise signal level due to the multiple access interference, the system performance is improved and the channel capacity of the base station is increased. The signal processor automatically changes the system arrangement to reduce interference and noise, and to increase signal quality and capacity.

Such smart antennas are emerging as an alternative to solve the shortage of available frequency resources. In particular, the use of smart antennas can achieve the same or better performance than existing systems with less power, increasing system capacity. The advantages of smart antennas can be explained in three major ways.

First, the gain of the signal can be increased since the signal gathers where it is desired without being dispersed. Therefore, the area that can be covered (Cover) per base station is increased, and the power consumption of the terminal can be reduced by increasing the gain, thereby increasing the battery usage time.

Secondly, since unwanted signals are effectively removed, interference signals can be removed. In particular, a system that is weak to interference signals such as a code division multiple access (CDMA) system can be expected to have a great effect. Therefore, in the CDMA system, more subscribers can be accommodated in the case of voice communication and high-speed data communication is possible in the case of data communication.

Third, since the smart antenna performs spatial filter effects as well, the effects of multipath can be greatly reduced. The signal processing method that controls the beam pattern, which is a key part of the smart antenna, is called an adaptive beamforming method.

The basic principle of the smart antenna is to extract only the desired signal from the interference signal, which gives a large gain in the direction of the desired signal and a small gain in the other direction so that the transmitting and receiving end can obtain more power for the same transmission power. will be. In other words, where there is a desired subscriber, constructive interference occurs, and when an unwanted subscriber acts as an interference signal, a destructive interference occurs. This method is called beamforming.

The beam forming method is divided into the fixed beam forming method and the adaptive beam forming method. The fixed beam shaping method is a method in which the pattern of the antenna is fixed, and the adaptive beam shaping method is a method in which the pattern of the antenna changes according to time or the surrounding environment. Compared to fixed beam forming in which a beam is fixed, adaptive beam forming has an advantage of forming a beam directly to a user. That is, the adaptive beamforming antenna has an advantage of being able to adapt to the environment more intelligently than the fixed beamforming antenna.

The conventional adaptive beamforming method is a method for detecting only one signal among several incident user signals. These single user detection methods include the following.

First, in order to obtain a weight that maximizes the SNR / SIR of a received signal, there is a method of repeatedly finding an eigenvector corresponding to the maximum eigenvalue of the autocorrelation matrix of the received signal. A method for obtaining a weight that maximizes the SNR / SIR of the received signal as described above is described in the paper "Ayman F. Naguib," Adaptive Antennas for CDMA Wireless Networks "[Ph. D. Dissertation, Dept of Electrical Engineering, Stanford University, Aug. 1996, for example.

An implementation of this is the MCGM (LCGM) method. Such a method related to MCGM is disclosed in Korean Application No. 1998-0057416 (Signal Processing Method and Apparatus for Optimal Weight Vector Calculation of Adaptive Array Antenna System based on Conjugation Slope Method, Applicant: Seungwon Choi) For example.

Secondly, there is an SGM which is a method of analyzing a received signal using the position of an array of receive antennas. The method of analyzing the received signal using the position of the array of the receiving antenna as described above is published in Korean Application No. 2001-0026629 (name of the invention: weight vector detection method of smart antenna, Applicant: Jaedon Park, Kiwan Wan, Kim Jewoo) For example.

On the other hand, there is also an ILSP method that obtains a Least Square based on the Maxim Likelihood method. The method for obtaining a Least Square based on the Maxim Likelihood method is described in the article "Blind Separation of Synchronous Co-Channel Digital Signals Using an Antenna Array-Part 1: Algorithms" [T. Shilpa, V. Mats, and P. Arogyaswami, IEEE Trans. Signal Processing, 1996, vol. 44, no. 5, pp. 1184-1197, for example.

However, all the adaptive beamforming methods for the smart antenna as described above can detect only a single user.

Accordingly, an aspect of the present invention is to provide a method for adaptive beamforming of a smart antenna for detecting multiple users.

According to an aspect of the present invention, in order to achieve the above object, in the multi-user detection adaptive beamforming method of a smart antenna, a plurality of smart antennas for receiving user signals transmitted from a plurality of users, and received from the smart antenna An RF / demodulator for demodulating the received user signal, a searcher for searching for a location of a user by searching for a user signal received from the RF / demodulator, and a located user signal received from the searcher. And a plurality of converging units that converge to the user signal suitable for the user position.

According to another aspect of the present invention, in the adaptive beamforming method of the smart antenna, the processing of the search unit may include generating a preset search vector, receiving a user signal from an RF / demodulator, and performing the preset search vector. And calculating a correlation between and the user signal, identifying a location of the user signal with the calculated correlation, and transmitting the determined user signal to a converging unit. .

According to still another aspect of the present invention, in the adaptive beamforming method of the smart antenna, the processing of the convergence unit may include receiving a user signal located from the search unit, applying a PMMSE with the received user signal, and And converging the user signal to which the PMMSE is applied to a real user signal.

According to still another aspect of the present invention, in the adaptive beamforming method of the smart antenna, the method of applying the PMMSE during the processing of the converging unit may include estimating an initial value, calculating a prediction signal using the predicted initial value, Projecting the calculated prediction signal to a predetermined signal locus, updating a autocorrelation matrix and a cross autocorrelation matrix having a value projected with the initial value, and updating the updated autocorrelation matrix and cross autocorrelation matrix And calculating a gradient with and converging the weight vector to the user position with the calculated gradient.

1 is a schematic overall configuration diagram of a system built using an adaptive beamforming method of a smart antenna according to an embodiment of the present invention;

2 is a flowchart illustrating an operation of a search unit in a multi-user detection method of a smart antenna according to an embodiment of the present invention;

3 is a flowchart illustrating a processing operation of a converging unit in a multi-user detection method of a smart antenna according to an embodiment of the present invention;

4 is a flowchart illustrating an operation of Projection Minimum Mean Square Error (PMMSE) according to an embodiment of the present invention.

5 is a graph showing the performance of the adaptive beamforming method of the smart antenna according to an embodiment of the present invention in SNR versus BER,

6 is a graph showing the performance of the adaptive beamforming method of the smart antenna according to an embodiment of the present invention as a beam pattern according to the incident angle.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the following description, specific details such as specific components are shown, which are provided to help a more general understanding of the present invention, and it is understood that these specific details may be changed or changed within the scope of the present invention. It is self-evident to those of ordinary knowledge in Esau.

1 is a schematic overall system configuration diagram for providing an operation of a system constructed by using an adaptive beamforming method of a smart antenna according to an embodiment of the present invention for multi-user detection. Referring to FIG. 1, a plurality of signals sent by a plurality of users come through a plurality of smart antennas 110. The plurality of smart antennas 110 receives a signal by forming a beam pattern directly to a plurality of users in an adaptive beamforming method. A plurality of user signals received from the plurality of smart antennas 110 is transmitted to the RF / demodulator 120. The RF / demodulator 120 modulates the user signal sent by the transmitter and is a signal in a carrier frequency band. This signal can be multiplied by a cosine or sine function with the same frequency as the carrier frequency to shift the frequency band to the low band. This modulation is performed to move the signal in the high band to the low band. Thereafter, the user signal processed by the RF / demodulator 120 is transmitted to the searcher 130 according to an embodiment of the present invention. The search unit 130 determines the approximate location of the user using the user signal received from the RF / demodulator 120. The identified user signal is then transmitted to the plurality of converging units 140. The multiple converging unit 140 applies PMMSE (Projection Minimun Mean Square Error: PMMSE) according to an embodiment of the present invention to a plurality of user signals having a rough position, according to each user signal. Converge.

In order to explain one embodiment of the present invention, there are several equations that should be basically understood. First, the transmitted user signal vector is expressed as in Equation 1 below.

This is the case when there are M user signals. In the case of a CDMA (Code Division Multiple Access) environment, each of the above user signals is the value after demodulation and despreading.

In addition, a signal incident on each antenna is expressed by Equation 2 below.

Also in the CDMA environment, each signal is a signal after demodulation and despreading. Here, the number of antennas is N.

In addition, a steering vector is expressed as in Equation 3 below.

It is determined by the arrangement position of the antenna and the direction of the received signal. Each element is represented as Each element of Equation 3 is obtained as in Equation 4 below.

2 is a flowchart illustrating a process of a searching step of a multi-user detection method of a smart antenna according to an embodiment of the present invention. Referring to FIG. 2, first, a scan vector is generated in step 210.

The search vector is obtained as in Equation 5 below.

In step 220, the RF / demodulator 120 receives a user signal. Thereafter, in step 230, the correlation is calculated using Equation 2 and Equation 5. Correlation calculation method is the angle of Equation 5 To change the search vector from -90 to +90 degrees in 1 degree increments. And the correlation between the antenna incident signal x (t). The correlation calculation is calculated as in Equation 6 below.

Thereafter, in step 240, the correlation degree calculated in step 230 is observed to determine the position of the user by using the points having the maximum magnitude as the initial value of the incident position of the user signal. For example, a process of searching for a user's position with a correlation according to an aspect of the present invention will be described. The above correlation is obtained by varying theta by 1 degree from -90 to +90 degrees. The correlation will be high in the direction of the incident signal, and the value of the correlation will be small in the direction where the signal does not enter. That is, when a graph of correlation is drawn, a peak is formed in a direction in which a signal is incident. Therefore, the point representing the peak is regarded as the point where the signal is incident. Here we find all the peaks in the correlation graph. All of these peaks mean that the user signal is incident in that direction. If we consider the ideal case where there is no noise in the communication channel, each peak in the correlation graph is exactly a peak in the direction of the incident signal.

However, due to the influence of noise such as AWGN (Additive White Gaussian Noise) AWGN in the communication channel and Multi-Path Rayleigh Fading, the angle of the peak of the correlation found by the above method is used. Slight deviation from the angle of incidence of the signal.

AWGN means totally existing noise in all frequency bands. Here, white means that it includes all frequency bands, and is based on the composition principle of color that all the light adds up to white. That is, since light is a kind of electromagnetic wave having a high frequency, signals existing in all bands in frequency become white in color. In this case, the random variable has a Gaussian form, and the word Additive is added to mean the original signal. In other words, AWGN refers to the general concept used in those cases where electromagnetic waves propagating into the air must consider some noise that is always present in the air.

Multipath Rayleigh fading means a kind of interference noise generated by combining electromagnetic waves multi-reflected by surrounding objects with each other. These multi-reflected electromagnetic waves have a very short period of Rayleigh shape as random variables. In other words, multipath Rayleigh fading is a very short period of fading coming in on multiple paths.

After determining the approximate location of the user signal in step 240, the user signal is transmitted to the converging unit 140 in step 250.

3 is a flowchart illustrating a process of a convergence step in the multi-user detection adaptive beamforming method of a smart antenna according to an embodiment of the present invention. Referring to FIG. 3, the adaptive beamforming method is independently applied to each peak. As shown in FIG. 2, the conventional single user detection adaptive beamforming method is independently applied to each incident signal found to converge each user signal distorted due to channel noise into a signal of an actual user.

Referring to FIG. 3, in operation 310, a user signal transmitted from the search unit 130 is received. Thereafter, in step 320, the PMMSE method is applied to each user signal received to find an accurate user location. In step 330, the user signal that finds the user's location is converged into an actual user signal.

The cost function of the conventional minimum mean square error (MMSE) method is calculated by Equation 7 below.

Meanwhile, the weight vector update formula that minimizes the above cost function is calculated using the steepest-descent method. The weight vector update equation is expressed as Equation 8 below.

The gradient value is then obtained using Equation 9 below.

here, Is the autocorrelation matrix of the received signal, Is a cross-correlation matrix between the received signal and the reference signal.

As a method of improving the conventional MMSE method, the PMMSE method according to an embodiment of the present invention is illustrated in FIG. 4. 4 is a flowchart illustrating a process of PMMSE according to an embodiment of the present invention. Referring to FIG. 4, the PMMSE method predicts the received signal y (k) using the received signal and then projects the signal to a predetermined signal constellation. It replaces the reference signal by performing. The preset signal locus means a signal locus closest to the predicted received signal.

First, the initial value in step 410 Predict. The initial value uses the incident position of the user signals in the search process in the search unit 130 as an initial value. Thereafter, the received signal y (k) is predicted using Equation 10 in step 420. The received signal prediction is predicted as shown in Equation 10 below.

Thereafter, in step 430, the signal predicted using Equation 10 is projected to a preset signal position. The preset signal locus is the signal locus closest to the signal predicted using Equation (10). The process of step 430 is performed as shown in Equation 11 below.

Thereafter, in step 440, the autocorrelation matrix and the cross autocorrelation matrix are updated and performed. The autocorrelation matrix and the cross autocorrelation matrix are calculated as in Equation 12 below.

After performing step 440, a gradient of 450 is calculated. The gradient calculation is calculated as in Equation 13.

After performing step 450, it is determined whether the weight vector converges to the user signal in step 460. The weight vector is calculated as in Equation 14 below.

As soon as the weight vector converges to the user signal in step 460, the PMMSE method terminates and performs the next step, step 330. However, if it is determined in step 460 and does not converge, step 420 is performed.

3 and 4 will be described in more detail with specific examples, starting from the peak of the actual number of users obtained in the first step. Assume that six peaks were obtained at -70 degrees, -30 degrees, 0 degrees, 20 degrees, 50 degrees, and 80 degrees. Then, the initial weight vector is obtained by first substituting theta = -70 into the direction vector of Equation 3. That is, the weight vector coincides with the direction vector. Then apply the PMMSE method just described. This process is applied to the predicted angles of the remaining users. That is, the same process is repeated for -30 degrees, 0 degrees, 50 degrees, 80 degrees, and so on to find all user signals.

5 is a graph showing the performance of the multi-user detection adaptive beamforming system according to an embodiment of the present invention as a value of BER compared to an SNR having a signal-to-noise ratio (SNR) and a bit error rate (BER) of the y-axis. to be. The dotted line through the white circle is a line graph showing the efficiency of ILSP in the absence of Rayleigh fading, and the solid line through the black circle is the efficiency of SMIMO according to the embodiment of the present invention in the absence of Rayleigh fading. This is a line graph showing. The double dashed line passing through the white inverted triangle is a line graph showing the efficiency of ILSP in the presence of Rayleigh fading, and the dashed line passing through the black inverted triangle is a line graph showing the efficiency of SMIMO in the presence of Rayleigh fading.

6 is a graph showing the performance of the multi-user detection adaptive beamforming system according to an embodiment of the present invention as the beam pattern of the smart antenna according to the incident angle of the incident angle of the x-axis and the beam pattern of the y-axis. The graph represented by the solid line passing through the black circle is a graph showing the beam pattern for the incident angle which varies depending on the user's position initially received by the smart antenna. The graph represented by the dotted line passing through the white circle is a graph representing the beam pattern for the incident angle depending on the user position after applying the PMMSE.

Meanwhile, in the above description of the present invention, specific embodiments have been described, but various modifications may be made without departing from the scope of the present invention. Therefore, the scope of the present invention should not be defined by the described embodiments, but by the claims and equivalents of the claims.

The present invention provides a multi-user detection adaptive beamforming system and a method of a smart antenna for detecting a multi-user. In addition, the PMMSE method according to an embodiment of the present invention is a method that does not require a reference signal because it is applied to each user signal used.

Claims (3)

In the adaptive beamforming method of the smart antenna, A search process of generating a search vector, calculating a correlation with the received user signal, and using the maximum points of the magnitude of the correlation as the initial values of the incident positions of the user signals; A convergence process of applying a PMMSE process to each user signal identified and converging to a position of a real user signal; In the convergence process, the prediction signal is calculated by predicting the initial value, and the projection signal is projected to a preset signal position to be used as a reference signal. And a PMMSE process of updating a weight vector having a gradient and converging to a user location. The method of claim 1, wherein the search vector is obtained using Equation (15). The method of claim 1, wherein the correlation is calculated by using Equation 16 below.