JP5013771B2 - MIMO-OFDM noise power measurement method and receiver - Google Patents

Description

  The present invention relates to a method and a receiving apparatus for measuring noise power included in a received signal in a multiple input multiple output-orthogonal frequency division multiplexing (MIMO-OFDM) system.

  One of the wireless transmission systems that has attracted attention in recent years is the MIMO technology. This is because a transmission device is provided with a plurality of antennas to transmit a plurality of transmission signals at the same frequency, and a reception device is also provided with a plurality of antennas to receive a propagation channel response between transmission / reception antennas. This is a technique for estimating a plurality of transmission signals based on this propagation channel response. By using this technology, it is expected that high frequency utilization efficiency can be realized. Hereinafter, a propagation channel between a plurality of transmission antennas and a plurality of reception antennas is referred to as a MIMO channel.

A transmission signal of each transmission antenna is a vector X = [X ₁ X ₂ ... X _m ] ^T , and a reception signal of each reception antenna is a vector Y = [Y ₁ Y ₂ ... Y _n ] ^T , If the included noise component is a vector N = [N ₁ N ₂ ... N _n ] ^T , the MIMO channel response H is an n-by-m matrix expressed by the propagation equation Y = HX + N. However, here, the number of transmitting antennas is m, the number of receiving antennas is n, and T represents transposition. The i row j column element h _ij of the channel response H is a complex number representing the amplitude gain and phase rotation of the transmission signal component output from the j th transmission antenna included in the reception signal received by the i th reception antenna. is there. The MIMO receiving apparatus estimates the channel response matrix H and demodulates the transmission signal vector X from the received signal vector Y using the estimated channel response matrix H (solves Y = HX + N).

  Such a MIMO technique can be easily applied to a narrow-band single carrier system. The MIMO technique can also be applied to wideband transmission by combining with an OFDM transmission scheme in which a signal is divided into a number of orthogonal subcarriers and transmitted. Hereinafter, a transmission system combining the MIMO technology and the OFDM transmission scheme is referred to as a MIMO-OFDM system.

  In general, in OFDM, a known signal for channel estimation is included in a specific subcarrier (hereinafter referred to as pilot carrier) of a transmission signal, and channel estimation is performed from the ratio of the received signal to the known signal. In MIMO-OFDM, the channel response of each transmission system in each received signal can be estimated by using a unique known signal for each transmission system.

By the way, the noise component N in the MIMO propagation equation is a vector representing the noise component of each received signal. Noise figure of the receiving system (Noise Figure: NF) and the average power of each element N _i in the noise component N by matching the gain is the same, as it is a channel estimation result of each receiving system as an element of the channel response matrix H Can be used.

  However, in OFDM, generally, since the received signal is AD converted and demodulated by digital signal processing, the gain of each received signal is automatically adjusted so that the dynamic range can be increased within the input voltage range of the AD converter. Since the power of the signal reaching each receiving antenna is generally not the same and fluctuates with time, non-uniform gain adjustment is performed in each receiving system. For this reason, when channel estimation results obtained with respect to received signals having different noise powers in the respective receiving systems are used as they are as elements of H, accurate MIMO demodulation cannot be realized. In order to solve this problem, after detecting the noise power of each receiving system, in order to remove the difference in noise power, the received signal or channel after FFT is set so that the noise power becomes the same level in each receiving system. The estimation result needs to be adjusted.

  Even in normal OFDM, when maximum ratio combining and directivity control are performed using a plurality of receiving antennas, or when soft decision Viterbi decoding is performed, noise power is measured. As an example of a technique for measuring the noise power, a method is known in which the power at a frequency slightly away from the OFDM signal band is used as the noise power. However, in general, in order to improve channel selectivity in the receiving apparatus, the received signal is passed through a steep band pass filter such as a SAW filter. For this reason, the signal attenuates in a frequency band far from the OFDM signal band, and it is often difficult to measure accurate noise power.

  In order to solve this problem, a method of measuring noise power within the OFDM signal band has also been proposed. For example, an OFDM subcarrier includes a non-signaled NULL carrier that does not transmit information due to a large distortion due to a frequency relationship, and a technique that pays attention to this point, that is, a reception power of a NULL carrier is reduced. There is a known technique for noise power. However, in general, the number of NULL carriers is very small, and since it is strongly influenced by distortion, it is difficult to measure accurate noise power.

  As another method, at the time of demapping the signal after waveform equalization of the received signal using the channel estimation result, the distance from the normal symbol point according to the modulation method such as QPSK or 16QAM is used as the noise component. Methods for measuring noise power are known. In this method, when the noise component is large or when channel estimation is not performed correctly, the received symbol jumps out to a different symbol area and is not the distance from the original symbol point. Therefore, it is difficult to measure accurate noise power with this method.

  Further, as other methods, those described in Patent Documents 1 to 5 have been proposed. In Patent Document 1, channel estimation is performed with the first OFDM symbol using two OFDM symbols and a known signal, and the next OFDM symbol is waveform-equalized using the channel estimation result, and the received symbol point obtained is known. The noise power is measured using the difference from the symbol point of the signal as a noise component. In Patent Document 2, after the pilot carrier signal is waveform-equalized, the noise power is measured using the distance from a known symbol point as a noise component. Further, in Patent Document 3, a complex correlation between a received signal and a known signal is obtained for each subcarrier, subtracted by adjacent subcarriers to detect a noise component, and noise power is measured.

  Such a conventional noise measurement method in the OFDM signal band can be applied when a single OFDM signal is transmitted, but a plurality of OFDM signals different at the same frequency as in a MIMO-OFDM system. Cannot be applied if they are transmitted and received in a complex mix. Specifically, the channel response matrix H cannot be obtained correctly unless the noise power of each received signal is the same for the method of measuring at the time of demapping and the methods of Patent Documents 1 and 2. . For this reason, interference removal and waveform equalization cannot be performed, and noise power cannot be measured.

  For the technique of Patent Document 3, in the case of burst OFDM in which all subcarriers of a specific OFDM symbol are modulated with known signals for channel estimation, the ratio of known signals mixed by channel estimation is calculated. It can be applied by reproducing correctly. However, the present invention is not applicable to a continuous pilot MIMO-OFDM system in which there is no such specific OFDM symbol and a pilot signal is inserted at a predetermined subcarrier interval in each symbol.

  Further, in Patent Document 4, the integral value of the guard correlation performed for detecting OFDM symbol synchronization is used as signal power, and the difference between this and the target power of AGC (Automatic Gain Control) is used as noise power. . Further, in Patent Document 5, a signal in an OFDM guard period and a signal delayed by an effective symbol therefrom (the signal in the guard period at the time of generating an OFDM signal is a replica of a signal delayed by an effective symbol). The absolute value of the difference is obtained, and an appropriate window is provided and averaged, and this is used as noise power.

  In the method of the above-mentioned Patent Document 4, when the CN ratio of the guard correlation value becomes small, the noise component cannot be ignored, and the integrated value does not represent accurate signal power. In an actual analog circuit, it becomes difficult to make the target power of AGC uniform in each receiving system due to device variations and temperature characteristics. For this reason, the calculated noise power tends to include an error.

  Further, in Patent Document 5 described above, it is necessary to optimally select a window so that the calculated noise power does not include a multipath component. In other words, if there is a long-delay multipath, the averaging period must be considerably shortened, resulting in a decrease in calculation accuracy. Also, since the OFDM signal is passed through a steep filter and transmitted so that the spectrum outside the band is below the specified value, the signal is distorted in the vicinity of the symbol switching (the beginning of the guard period and the end of the symbol). Yes. Therefore, in Patent Document 5, since noise power is measured using this period, an error is likely to occur.

Japanese Patent No. 3445773 Patent No. 3662579 JP-A-2005-204307 JP 2004-112155 A JP 2005-175878 A

  Thus, in the MIMO-OFDM system, when gain adjustment is performed independently for each reception system, in order to obtain an accurate channel response matrix H, the noise power level of each reception system becomes the same. In addition, it is necessary to adjust the received signal after the FFT. However, the conventional method has a problem that it is difficult to accurately measure the noise power of each receiving system.

  Therefore, the present invention has been made in view of such a situation, and an object thereof is to provide a method and a receiving apparatus capable of accurately measuring noise power included in a received signal in a MIMO-OFDM system. There is to do.

  In order to achieve the above object, the present invention obtains accurate noise power by measuring noise power in a band using a transmission pattern of a pilot carrier signal before performing channel estimation. Then, by normalizing the received signal using the measured noise power, the noise power of all reception systems is set to the same level. Further, the channel response matrix is accurately estimated by using the normalized received signal.

  Regarding the principle of the noise power measurement method according to the present invention, a MIMO-OFDM system in which pilot carriers of each transmission signal are code-division multiplexed over a plurality of symbols by orthogonal codes, where transmission is composed of two systems, A MIMO-OFDM system in which orthogonal codes “1, 1, 1, −1” are assigned to “1, 1, 1, 1” and transmission system 2 will be described as an example.

The transmission signal x _1i (t) of the i-th pilot carrier of the transmission system 1 is the following repetition in a 4-symbol period, where CP _i is a known signal assigned to this pilot carrier. Here, t represents a symbol number.

Similarly, the output x _2i (t) of the i-th pilot carrier of the transmission system 2 is a repetition of the following four symbols.

The received signal y _ji (t) of the i-th pilot carrier of the receiving system j is such that h _j1i (t) is a channel response to the output of the transmission system 1, h _j2i (t) is a channel response to the output of the transmission system 2, n _ji (t) can be expressed by the following equation as N (0, σ _ji ² ) complex Gaussian noise.

(1) is substituted into equation and (2) a (3), the received signal sequence _{h J1i} between successive 4 symbols (t) and _{h j2i} (t) _{h j1i,} a constant value of _{h J2i} If it represents, it will become the following formulas.

パラメータｓ_ｊｉをＬ個のパイロットキャリアについて求め、ｓ_ｊｉの平均値Ｓ_ｊを（６）式により求める。

Here, since y _ji (1) and y _ji (3), and y _ji (2) and y _ji (4) are received signal sequences that differ only in noise components and can be regarded as the same, the following parameter s _ji is set: Introduce only noise components. However, | Z | ² = Z · Z ^* , ^* represents a complex conjugate.

The parameter s _ji is obtained for L pilot carriers, and the average value S _j of s _ji is obtained from the equation (6).

When L is sufficiently large, the average value S _j becomes an ensemble average, and each term becomes an expected value. That is, the square (power) of the magnitude of the noise component becomes the variance σ _ji ² , and the product of the different noise components becomes zero by the central limit theorem. Since the noise is flat white noise on the frequency axis, σ _ji ² can be a constant σ _j ² for all i. Therefore, when the average value S _j is calculated, the equation (7) is obtained, and the noise power σ _j ² of the receiving system j can be obtained.

Based on the principle of the noise power measurement method according to the present invention described above, a receiver that executes this noise power measurement method uses a received signal sequence of symbols that can be regarded as the same except for the noise component of the received signal of the pilot carrier. By averaging the square of the magnitude of the difference, the noise power σ _j ² of the received signal is obtained. Then, the received signals of all subcarriers are normalized by dividing by the square root σ _j of the noise power σ _j ² . As a result, the result of channel estimation performed thereafter reflects the SN ratio of the received signal, and a correct channel response matrix can be obtained.

  In accordance with the principle described above, the noise power measurement method of the MIMO-OFDM system according to the present invention is based on the MIMO-OFDM system under the MIMO-OFDM system in which the pilot carrier of each transmission signal is code division multiplexed over a plurality of symbols. In a method of measuring noise power contained in a received signal by a receiving device that constitutes, a received signal of a symbol that can be regarded as the same except for a noise component according to the transmission pattern of a pilot carrier modulated with a known signal The difference between the sequences is squared, the squared result is averaged for a plurality of pilot carriers, and the averaged value is used as the noise power.

  In addition, the MIMO-OFDM system receiver according to the present invention measures the noise power included in the received signal in the MIMO-OFDM system in which the pilot carrier of each transmission signal is code division multiplexed over a plurality of symbols. Noise power measurement means, and the noise power measurement means extracts a pilot carrier from a received signal and a received signal sequence of symbols that can be regarded as the same except for noise components with respect to the extracted pilot carrier. A difference stage that squares the difference between the two, and an averaging stage that averages the squared result for a plurality of pilot carriers and outputs the averaged value as noise power.

  The receiving apparatus of the MIMO-OFDM system according to the present invention further includes normalizing means for normalizing the received signal based on the noise power measured by the noise power measuring means.

  As described above, according to the present invention, the noise power is measured using the received signal sequence of the symbols that can be regarded as the same except for the noise component of the received signal of the pilot carrier. Thereby, it is possible to accurately measure the noise power of the MIMO-OFDM system, which is difficult to measure with the conventional method. The received signal is normalized based on the measured noise power. As a result, even if gain control is performed independently in each receiving system, the received signal is normalized by the noise power measured in that receiving system. Can be at the same level. Accordingly, it is possible to accurately estimate the channel response matrix reflecting the reception S / N ratio of each reception system.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. The MIMO-OFDM system shown below is a system in which pilot carriers of each transmission signal are code-division multiplexed over a plurality of symbols by orthogonal codes. Here, the transmission apparatus includes two systems, and the transmission system 1 includes A system in which orthogonal codes of “1, 1, 1, 1” and “1, -1, 1, −1” are assigned to the transmission system 2 will be described as an example.

FIG. 1 is an overall configuration diagram of a MIMO-OFDM system including a receiving apparatus according to an embodiment of the present invention. The MIMO-OFDM system 100 includes two transmission apparatuses 101 and n reception apparatuses 102. Note that interleaving and error correction are omitted. The MIMO-OFDM system 100 is a system that allocates transmission signals x ₁ and x ₂ to each subcarrier and transmits them, and estimates a transmission signal from the received signal. The transmission apparatus 101 on the transmission side has a configuration example as shown in FIG. 3, and for two transmission systems, IFFT means 14-1 and 14-2 perform IFFT on the input signals x ₁ and x ₂ respectively, and guard is added. Guard signals are added by means 15-1 and 15-2, D / A conversion is carried out by D / A conversion means 16-1 and 16-2, and U / C (up-conversion) means 17-1 and 17-2. The frequency is converted to a radio frequency and output from the transmitting antennas 18-1 and 18-2.

  As shown in FIG. 2, the receiving device 102 on the receiving side includes n receiving systems 1 to n each having the same function, and includes receiving antennas 1-1 to 1-n and D / C (down converter). Means 2-1 to 2-n, AGC means 3-1 to 3-n, A / D conversion means 4-1 to 4-n, synchronization means 5-1 to 5-n, FFT means 6-1 to 6- n, normalizing means 7-1 to 7-n, noise power measuring means 8-1 to 8-n, channel estimating means 9-1 to 9-n, and MIMO demodulating means 10.

Hereinafter, description will be made using a reception system j representing the reception systems 1 to n. The reception antenna 1-j receives transmission signals transmitted through the MIMO channel from the transmission antennas 18-1 and 18-2 of the transmission apparatus 101, and the D / C (down-conversion) means 2-j wirelessly receives the reception signals. The AGC means 3-j automatically adjusts the gain of the amplifier so as to obtain a specified level for the received signal by converting the frequency to the IF frequency. A / D conversion means 4-j performs A / D conversion on the received signal subjected to gain adjustment, and synchronization means 5-j detects the start position of the OFDM symbol for the A / D converted reception signal. . FFT means 6-j, for A / D converted received signal is subjected to Fourier transform, except a guard period, and outputs the FFT output _{y j.}

The noise power measuring means 8-j receives the FFT output y _j and measures the noise power σ _j ² . Normalizing means 7-j receives the FFT output _{y j} from the FFT unit 6-j receives the noise power sigma _j ² from the noise power measurement unit 8-j, the FFT output _{y j} of the noise power sigma _j ² Normalize by dividing by the square root σ _j and output the normalized FFT output y ′ _j .

Channel estimation means 9-j is 'enter a _j, signal y of the pilot carrier' normalized FFT output y from the normalization unit 7-j extracts _ji (t), assigned to the transmission system 2 perpendicular Correlation processing is performed as follows according to the code. (I in y ′ _ji (t) indicates a pilot carrier number.)
(Y ′ _ji (1) + y ′ _ji (2) + y ′ _ji (3) + y ′ _ji (4)) / 4, and (y ′ _ji (1) −y ′ _ji (2) + y ′ _ji (3) -Y ' _ji (4)) / 4
Then, the channel estimation means 9-j calculates the ratio between the correlation output and the known signal CP _i , filters this on the frequency axis, and estimates the channel responses h ′ _j1 and h ′ _j2 of all subcarriers. To do.

The demodulating means 10 includes normalized FFT outputs y ′ ₁ ,..., Y ′ _j ,..., Y ′ _n and channel responses h ′ ₁₁ , h ′ ₂₂ ,. , H ′ _j1 , h ′ _j2 ,..., H ′ _n1 , h ′ _n2 are input, and estimated values x ₁ and x ₂ of the transmission signal are output by an algorithm such as zero forcing or maximum likelihood estimation.

  FIG. 4 is a block diagram showing the configuration of the noise power measuring means 8-j shown in FIG. The noise power measuring means 8-j has the same function as the noise power measuring means of other receiving systems. The noise power measuring means 8-j includes a pilot carrier extraction stage 11-j, a difference stage 12-j, and an averaging stage 13-j.

Pilot carrier extraction stage 11-j receives the FFT output _{y j} from the FFT unit 6-j, extracts a signal _y ji (t) of the pilot carriers of the FFT output _{y j.} The difference stage 12-j receives the pilot carrier signal y _ji (t) from the pilot carrier extraction stage 11-j, and the parameter s is obtained by extracting only the noise component from the L pilot carrier signals according to the above equation (5). _ji is calculated. The averaging stage 13-j receives the parameter s _ji from the difference stage 12-j, averages the parameters s _ji of the L pilot carrier signals according to the above equation (6), and uses the average value S _j as the noise power. Output to the normalizing means 7-j as σ _j ² .

As described above, according to the embodiment of the present invention, the noise power measuring unit 8-j extracts the pilot carrier signal y _ji (t) from the FFT output y _{j and} uses the four symbols according to the equation (5). The parameter s _ji is calculated using the pilot carrier signal of the minute, and the average value S _j of the parameters s _ji of the L pilot carrier signals is measured as the noise power σ _j ² . That is, since the noise power in the band is measured using the transmission pattern of the pilot carrier signal, accurate noise power can be obtained even in the MIMO-OFDM system.

Further, according to the embodiment of the present invention, the normalizing means 7-j normalizes the FFT output y _j by dividing by the square root σ _j of the noise power σ _j ² , and the channel estimation means 9-j The channel responses h ′ _j1 and h ′ _j2 are estimated based on the normalized FFT output y ′ _j . Thereby, even if the gain control is performed independently by the AGC means 3-1 to 3-n in each of the reception systems 1 to n, the FFT output of each of the reception systems 1 to n is the reception system 1 to 1. Since normalization is performed with the noise power of n, the noise power of all the receiving systems 1 to n can be set to the same level. Therefore, it is possible to accurately estimate the channel response matrix reflecting the S / N ratio of each receiving system 1 to n.

  The present invention has been described with reference to the embodiment. However, the present invention is not limited to the above embodiment, and various modifications can be made without departing from the technical idea thereof. For example, in the above-described embodiment, the transmission apparatus 101 includes two systems, and the transmission system 1 has “1,1,1,1” and the transmission system 2 has “1, -1,1, -1” orthogonal codes. However, the number of transmission systems and orthogonal codes are not limited to this, and the above equation (5) and the correlation processing are changed according to the number of transmission systems. Also, it can be applied by changing according to the assigned orthogonal code.

For example, when the transmission apparatus is configured by three systems and an orthogonal code “1, 1, −1, −1” is assigned to the added transmission system 3, reception signals y _ji (1) to 8 OFDM symbols are assigned. From y _ji (8), parameter s _ji is obtained by the following equation (8), and average value s _j is obtained by equation (9).

Note that the amount of calculation may be reduced by reducing the number of terms on the right side of equation (8). In this case, it is necessary to change the denominator of Expression (9) according to the number of y _ji (t) included in the right side of Expression (8). In this case, the correlation processing for estimating the channel of the transmission system 3 is as follows.
(Y ′ _ji (1) + y ′ _ji (2) −y ′ _ji (3) −y ′ _ji (4)) / 4

1 is an overall configuration diagram of a MIMO-OFDM system including a receiving apparatus according to an embodiment of the present invention. It is a figure which shows the structure of the receiver by embodiment of this invention. It is a figure which shows the structure of a transmitter. It is a block diagram which shows the structure of the noise power measurement means of FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Reception antenna 2 D / C (down-conversion) means 3 AGC means 4 A / D conversion means 5 Synchronization means 6 FFT means 7 Normalization means 8 Noise power measurement means 9 Channel estimation means 10 Demodulation means 11 Pilot carrier extraction stage 12 Difference Stage 13 Averaging stage 14 IFFT means 15 Guard addition means 16 D / A conversion means 17 U / C (up-conversion) means 18 Transmitting antenna 100 MIMO-OFDM system 101 Transmitting apparatus 102 Receiving apparatus

Claims (3)

Under a MIMO-OFDM system in which pilot carriers of transmission signals in each of a plurality of transmission systems are code-division multiplexed over a plurality of symbols, they are included in the reception signals of the respective reception systems by the receivers constituting the MIMO-OFDM system. In a method for measuring noise power,
Based on a transmission pattern over a plurality of symbols of pilot carriers that are modulated by a known signal and differ for each transmission system , square the difference of received signal sequences of symbols that can be regarded as the same except for noise components,
A method of measuring noise power, wherein the squared result is averaged for a plurality of pilot carriers, and the averaged value is used as the noise power of the receiving system . In a receiving apparatus constituting a MIMO-OFDM system in which pilot carriers of transmission signals in each of a plurality of transmission systems are code division multiplexed over a plurality of symbols,
Each of the plurality of reception systems includes noise power measurement means for measuring the noise power included in the received signal,
The noise power measuring means includes :
A pilot carrier extraction stage for extracting a pilot carrier from the received signal;
For the extracted pilot carrier , based on a transmission pattern over a plurality of symbols of different pilot carriers for each transmission system , a difference stage that squares the difference between received signal sequences of symbols that can be regarded as the same except for noise components;
And a averaging stage that averages the squared result of a plurality of pilot carriers and outputs the averaged value as noise power. The receiving device according to claim 2,
Each of the plurality of receiving systems is further based on the measured noise power by the noise power measurement unit of the receiving system, the receiving apparatus characterized by comprising a normalization means for normalizing the received signal.

Images