JP4403877B2 - Method and apparatus for compensating time variation of amplitude in diversity reception, and method and apparatus for compensating intra-symbol time variation in multicarrier diversity reception - Google Patents

Description

  The present invention relates to a method and apparatus for compensating for temporal fluctuations in the amplitude direction caused by fluctuations in propagation path characteristics and the like when a modulated signal is received. The present invention is suitable for diversity reception, and can be used in combination with phase fluctuation compensation in multicarrier diversity reception.

For example, in an orthogonal frequency division multiplexing (OFDM) receiver, propagation path characteristics are estimated and compensated by extracting pilot symbols after fast Fourier transform (FFT). This execution method is described in the following Patent Documents 1 and 2. Further, the following Patent Document 3 and Non-Patent Document 1 describe an execution method for compensating for fluctuations in the symbols of amplitude and phase before fast Fourier transform (FFT) using a guard interval.
JP 2001-44963 A JP 2002-261729 A JP 2000-286817 A Atsumori Hashizume, Minoru Okada, Shozo Komaki “High-speed fading compensation method using quadrature multicarrier modulation guard section”, IEICE Technical Report CS97-20, RCS97-8 (1997-05), pages 9-14

  On the other hand, the present inventors have found and applied for a simple configuration that eliminates the need for a symbol synchronization signal when compensating for intra-symbol variations in amplitude and phase before fast Fourier transform (FFT) (Japanese Patent Application 2003). -404366). This method has the following configuration when combined with diversity reception.

  FIG. 6 is a block diagram showing a configuration of a multicarrier diversity receiver 9000 described in Japanese Patent Application No. 2003-404366, which has a configuration for compensating for variations in amplitude and phase in symbols in diversity reception. The multicarrier diversity receiver 9000 includes orthogonal demodulation units 1a and 1b and complex multipliers 22a and 22b corresponding to two antennas, a weight coefficient calculation unit 21 and an adder 23, an intra-symbol time variation estimator 31 and a complex number. An intra-symbol time variation compensator 30 comprising a multiplier 32, a synchronization circuit 4, a guard interval (GI) remover 5, a serial-parallel converter (S / P) 6, an N-point fast Fourier transformer (FFT) 7, a pilot It comprises a symbol extractor 81, a channel characteristic estimator 82, a channel characteristic frequency interpolator 83, a frequency domain equalizer 8, and a parallel / serial converter (P / S) 9. The weighting factor calculation unit 21, the complex multipliers 22a and 22b, and the adder 23 are portions that perform diversity combining. This diversity synthesis is, for example, by maximum ratio synthesis.

  The characteristic part of Japanese Patent Application No. 2003-404366 is the intra-symbol time variation compensator 30 of the multicarrier diversity receiver 9000 shown in FIG. 6, where the amplitude compensation coefficient (real number) and the phase compensation coefficient (complex number) are calculated. The output signal is multiplied by the combined signal of the adder 23 and output.

  In diversity reception, the weighting factor is not determined so that the average power of the combined signal is always constant. That is, the multicarrier diversity receiver 9000 in FIG. 6 does not determine the weighting factor by the weighting factor calculation unit 21 so that the average power of the combined signal output from the adder 23 is not necessarily constant. In other words, the intra-symbol time variation estimator 31 also calculates the amplitude compensation coefficient so that the amplitude variation in time is eliminated. However, as is well known, it is not easy to completely eliminate fluctuations in the amplitude of the diversity combining output.

  Therefore, the inventors have intensively studied to develop a diversity method capable of outputting a combined signal (output of the adder 23) in which the amplitude variation in the symbol is compensated by combining these, and compensating for the novel amplitude variation in time. Completed the method.

The invention according to claim 1 combines the received waves received by a plurality of antennas, compensates for fluctuations in the amplitude direction of the combined signal, and makes amplitude time in diversity reception to keep the average power of the combined signal constant. In the variation compensation method, a complex cross-correlation between a received signal of each antenna and a combined signal obtained by multiplying each received signal by a complex weighting factor is added corresponding to each antenna, and Smoothing each complex cross-correlation to obtain a smoothed correlation value, and dividing each smoothed correlation value by the sum of squares of all smoothed correlation values is used as a new updated value of each weighting factor. This is a characteristic time variation compensation method for diversity reception.

According to the second aspect of the present invention, the time length T _{S of} one symbol obtained by copying the waveform of the portion of the time length T _GI at the end of the signal and adding it to the head of the signal of the effective symbol length T _f is T _S = T _The intra-symbol time variation compensation method in multi-carrier diversity reception for receiving a multi-carrier modulation signal of _GI + T _f compensates for intra-symbol time variation in the amplitude direction using the amplitude time variation compensation method according to claim 1. The phase variation at time T _f is obtained by a complex correlation between the combined signal whose amplitude is compensated for the time length T _S and the delayed signal obtained by delaying the combined signal whose amplitude is compensated by T _f as the delay time. to estimate, by calculating the phase variation of each short time T _a than the effective symbol length T _f, the compensation child symbol within the time variation of the phase seeking correction coefficient of the phase at that time based on its T _a And an intra-symbol time variation compensation method in multicarrier diversity reception.

The invention described in claims 3 and 4 is an apparatus invention corresponding to the method invention described in claims 1 and 2 .

The invention according to claim 3 synthesizes the received waves received by the plurality of antennas, compensates for fluctuations in the amplitude direction of the synthesized signal, and maintains the average power of the synthesized signal in the diversity receiver. In a time variation compensation device, a plurality of complex multipliers for multiplying a reception signal of each antenna by a weighting factor that is a complex number calculated for each reception signal, and outputs of the plurality of complex multipliers are added. A complex adder, a plurality of complex cross-correlation calculators each for calculating a complex cross-correlation between a reception signal of each antenna and a composite signal that is an output of the complex adder, and the plurality of cross-correlation calculators A plurality of smoothers for smoothing the output of the plurality of smoothing correlation values obtained by dividing the smoothed correlation values, which are the outputs of the plurality of smoothers, by the sum of squares of all smoothed correlation values; Multiple weights to Is a time-varying compensation device of amplitude in diversity reception, characterized in that it comprises a coefficient updater.

According to the fourth aspect of the present invention, the time length T _{S of} one symbol obtained by copying the waveform of the portion of the time length T _GI at the end of the signal and adding it to the head of the signal of the effective symbol length T _f is T _S = T 4. An intra-symbol time variation compensator for multi-carrier diversity reception for receiving a multi-carrier modulation signal of _GI + T _f , the time variation compensator for amplitude according to claim 3 , and a time length T _S as an integration interval. Phase variation at time _Tf is estimated by a complex correlation between the compensated synthesized signal and a delayed signal obtained by delaying the compensated synthesized signal of amplitude by _Tf as a delay time, and is shorter than the effective symbol length _Tf. And a phase time variation compensator for calculating a phase variation coefficient at each time T _a , obtaining _a phase correction coefficient at the time T _a based on the phase variation, and compensating for the phase intra-symbol time variation. This is an intra-symbol time variation compensator for multi-carrier diversity reception.

と置く。今、補償係数を実数としてｗと置くと、補償後の信号ｗｒの平均電力は

となる。入力信号ｒと補償後の信号ｗｒとの相互相関は

である。相互相関を求める九端では補償係数を更新しないため一定値とする。本発明ではこの逆数

を新たな補償係数の値として更新値とするものである。 The present invention will be described using mathematical expressions. The main feature of the present invention is smoothing. First, a case where there is no smoothing is assumed. In the case of one input signal, the average power of the input signal over a certain period of time, for example, a fraction of the symbol length, is calculated with respect to the complex input signal r.

Put it. Now, assuming that the compensation coefficient is w as a real number, the average power of the compensated signal wr is

It becomes. The cross-correlation between the input signal r and the compensated signal wr is

It is. At the nine ends where cross-correlation is obtained, the compensation coefficient is not updated, so a constant value is set. In the present invention, this reciprocal is

Are updated values as new compensation coefficient values.

が、補償係数の更新後も同じ場合に、新たな補償後の信号

の平均電力は、

となる。入力信号ｒと新たな補償後の信号

の相互相関は１／ｗとなり、この逆数ｗが補償係数の次の更新値となる。 Then the average power of the input signal r

Is the same after the update of the compensation factor, the new compensated signal

The average power of

It becomes. Input signal r and new compensated signal

Is 1 / w, and this reciprocal w becomes the next updated value of the compensation coefficient.

の場合に補償後の信号の平均電力が１となる。一方、補償係数ｗが他の値の場合は補償係数を更新する度ごとに補償後の信号の平均電力が

とを交互に繰り返すことがわかる。そこで平滑化のステップを設けることで、繰り返しの振動を抑え、補償係数ｗについて、

の場合の補償係数ｗの値に収束させるとするものである。 In such cases,

In this case, the average power of the compensated signal is 1. On the other hand, when the compensation coefficient w is another value, the average power of the signal after compensation is updated every time the compensation coefficient is updated.

It turns out that and are repeated alternately. Therefore, by providing a smoothing step, repetitive vibration is suppressed, and the compensation coefficient w is

In this case, the value is converged to the value of the compensation coefficient w.

と置く。簡単のため、アンテナを２個とし、各アンテナにおける伝搬路応答を複素係数で表されるとしてｃ₁とｃ₂とすると、複素数である受信信号ｒ₁とｒ₂は次の式で表される。

Next, the case of maximum ratio combining diversity reception will be described. First, s is the original transmission signal, and the average signal power is

Put it. For simplicity, assuming that there are two antennas and the channel response at each antenna is represented by complex coefficients and c ₁ and c ₂ , the complex received signals r ₁ and r ₂ are represented by the following equations. .

ここで、＊は複素共役を表す。この関係が成り立つ場合、各アンテナにおける受信信号は位相を揃えて合成される。振幅変動がまだ補償されていない場合、合成信号ｙの平均電力は１となっていない。そこで、合成信号ｙは、元の送信信号に対して振幅方向の大きさが変動していることを表すある実数係数ｋを用いることにより、次式で表すことができる。

Here, in order to explain the process of compensating for the amplitude fluctuation, the maximum ratio synthesis can be realized, but the case where the amplitude fluctuation is not yet compensated is considered. The weighting factors w ₁ and w ₂ , which are complex numbers when the maximum ratio combining is realized, have the following relationship with the propagation path response.

Here, * represents a complex conjugate. When this relationship holds, the received signals at the respective antennas are combined with their phases aligned. If the amplitude variation is not yet compensated, the average power of the combined signal y is not 1. Therefore, the synthesized signal y can be expressed by the following equation by using a certain real number coefficient k indicating that the magnitude in the amplitude direction varies with respect to the original transmission signal.

ここで、元の送信信号電力を１とおいているため、次のようになる。

The complex cross-correlation x _i (i = 1, 2) between the synthesized signal y and the received signal r _i (i = 1, 2) is expressed by the following equation.

Here, since the original transmission signal power is set to 1, it is as follows.

Next, when a new update value w _i ′ of the weighting coefficient is obtained by dividing this complex cross-correlation value by the sum of squares of all cross-correlation values, the following is obtained.

The composite signal y ′ after being updated to this new weighting coefficient is expressed by the following equation.

Here, when the synthesized signal y before updating the weighting coefficient is compared with the synthesized signal y ′ after updating, it can be seen that the value of the coefficient k representing the amplitude ratio with respect to the original transmission signal changes to an inverse number. That is, when smoothing is not performed, the composite signal alternately repeats ks and s / k every time the weighting factor is updated. This repetition occurs due to the update of the weighting coefficient, and the complex cross-correlation value when obtaining the updated value of the weighting coefficient is oscillating in conjunction. Therefore, by providing a smoothing step so as to suppress the vibration of the complex cross-correlation value, this vibration is suppressed and converged when k = 1. As a result, the weighting factor converges to a value that satisfies the following equation.

  In the following, description will be made centering on the configuration where the simulation was performed. In addition, this invention is not limited to a following example.

  FIG. 1 is a block diagram showing a configuration of a multicarrier diversity receiver 1000 according to a specific embodiment of the present invention. The multicarrier diversity receiver 1000 includes quadrature demodulation units 1a and 1b and complex multipliers 202a and 202b corresponding to two antennas, a weight coefficient calculation unit 201, an adder 203, an intra-symbol phase fluctuation estimator 301, and a complex number. In-symbol phase fluctuation compensation device 300 including a multiplier 302, a synchronization circuit 4, a guard interval (GI) remover 5, a serial-parallel converter (S / P) 6, an N-point fast Fourier transformer (FFT) 7, a pilot It comprises a symbol extractor 81, a channel characteristic estimator 82, a channel characteristic frequency interpolator 83, a frequency domain equalizer 8, and a parallel / serial converter (P / S) 9. The weighting factor calculation unit 201, complex multipliers 202a and 202b, and adder 203 constitute an amplitude time variation compensating apparatus 200 that performs diversity combining and amplitude time variation compensation. In the amplitude time variation compensator 200, two weights are calculated from the received signals that are the outputs of the orthogonal demodulation units 1a and 1b and the combined signal that is the output of the adder 203, and are output to the complex multipliers 202a and 202b.

The high-frequency signals received from the two antennas are converted into intermediate frequency bands by the quadrature demodulation units 1a and 1b, respectively, quadrature demodulated and analog-to-digital converted, and output as the in-phase component I and the quadrature component Q. Input to the fluctuation compensation device 200. Here, diversity combining and amplitude fluctuation variation are compensated and input to the intra-symbol phase fluctuation compensator 300. The in-phase component I and the quadrature component Q whose phases have been compensated by the intra-symbol phase fluctuation compensation device 300 are input to the synchronization circuit 4 and the GI remover 5. The timing of the fast Fourier transform (FFT) window period (time length T _f ) is calculated by the synchronization circuit 4, and the timing signal is input to the intra-symbol phase fluctuation compensator 300 and the GI remover 5. The GI remover 5 removes the guard interval, outputs a valid symbol, and inputs it to the S / P 6.

  The serial signal input to S / P 6 is output as a parallel signal and input to an N-point fast Fourier transformer (FFT) 7. The parallel signal input to the N-point FFT 7 is fast Fourier transformed and output as a signal corresponding to each carrier, and is input to the frequency domain equalizer 8 and the pilot symbol extractor 81. Of the signals input to the pilot symbol extractor 81, only the pilot symbols are output to the channel characteristic estimator 82. Channel characteristic estimator 82 estimates the channel characteristic of the carrier having the pilot symbol and outputs it to channel characteristic frequency interpolator 83. Channel characteristic frequency interpolator 83 interpolates the channel characteristics of carriers that do not have pilot symbols, and outputs the channel characteristics of all carriers to frequency domain equalizer 8. The frequency domain equalizer 8 compensates the channel characteristics of all the carriers and outputs them to the P / S 9. In P / S9, a parallel signal is output as a serial signal, and subsequent signal processing is performed.

  The calculation contents of the weighting factor calculation unit 201 are shown in FIG. Here, for example, a configuration in which calculation is performed by one CPU or the like is shown, but each calculation process may be performed by an individual calculation device.

  As shown in FIG. 2, there are calculation process sequences of calculation processes SD1a, SD2a and SD4a and calculation process sequences of calculation processes SD1b, SD2b and SD4b corresponding to the two antennas. Is placed.

  First, in the calculation step SD1a, a complex cross-correlation between the received signal that is the output of the orthogonal demodulator 1a and the combined signal that is the output of the adder 203 is obtained. Next, in the calculation step SD2a, the forgetting factor is set to 0.9 and smoothing processing is performed. The smoothed correlation value is used in the calculation step SD3 and the calculation step SD4a. Exactly in the same way, in the calculation step SD1b, the complex cross-correlation between the received signal that is the output of the quadrature demodulator 1b and the combined signal that is the output of the adder 203 is obtained. Is processed. The forgetting factor was used in the following meaning. That is, when the forgetting factor is γ (where 0 ≦ γ ≦ 1), the input signal to the smoothing circuit is p (n), and the output q (n), q (n) = γq (n−1) + ( 1-γ) p (n). Accordingly, the greater the forgetting factor γ, the greater the smoothing. When the forgetting factor γ is 0, the input signal is directly used as the output signal without any smoothing.

  In the calculation step SD3, the sum of squares (real number) is calculated from the two smoothed complex cross-correlation values. In the calculation step SD4a, the smoothed complex cross-correlation value that is the output of the calculation step SD2a is divided by the square sum obtained in the calculation step SD3. In this way, the weight coefficient compensated for amplitude fluctuation is updated and output to the complex multiplier 202a. Exactly in the same manner, in the calculation step SD4b, the smoothed complex cross-correlation value that is the output of the calculation step SD2b is divided by the sum of squares obtained in the calculation step SD3, and the weight coefficient corrected for amplitude fluctuation is updated. It is output to the multiplier 202b.

FIG. 3 shows a comparative example (FIG. 3.A, FIG. 3) of a simulation (FIG. 3.C) of diversity combined signal power (average power of the output of the adder 203) using the amplitude time variation compensator 200 shown in FIG. .B, FIG. 3.D). The simulation conditions are flat fading, Doppler frequency 10 Hz, Eb / No 30 dB, effective symbol length T _f of 1.008 msec, and 8192 terrestrial digital TV (mode 3) OFDM signal. The sample was 126 μsec and 1024 samples. FIG. 3 is a comparative example. A is a simulation in the case where the weighting coefficient is obtained by dividing the value of the complex cross-correlation by the square root of the square sum of the value of the complex cross-correlation instead of the configuration of FIG. In this case, the sum of squares of all weighting factors is 1. FIG. 3 is a comparative example. B is a simulation in the case where the weighting coefficient is obtained by dividing the value of the complex cross-correlation by the sum of the absolute values of the values of the complex cross-correlation instead of the configuration of FIG. In this case, the sum of absolute values of all weighting factors is 1. FIG. 2 omits the calculation steps SD2a and SD2b in the configuration of FIG. 2, and uses the values of the complex cross-correlation as they are in the calculation steps SD3, SD4a, and SD4b without being smoothed.

  In a configuration in which the sum of squares of all weighting factors is 1, FIG. As shown in A, the fluctuation of the value of the synthesized signal (the amplitude of the output of the adder 203) is extremely large, and it cannot be said that the fluctuation of the amplitude is compensated. In a configuration in which the sum of absolute values of all weighting factors is 1, FIG. As shown in FIG. 3B, the fluctuation of the value of the composite signal (the amplitude of the output of the adder 203) is changed as shown in FIG. If it is not as large as A, it cannot be said that fluctuations in amplitude are compensated. In a configuration that does not use smoothing, FIG. Although the fluctuations are within a certain range as shown in D, the fluctuations are extremely severe, and it seems that a new fluctuation in amplitude is added. On the other hand, in the present invention having the configuration of FIG. As shown in C, the amplitude is stable around 1, which is a very satisfactory result.

  FIG. As is apparent from a comparison between C and other simulation results, the configuration of the present invention can sufficiently suppress amplitude fluctuations while performing diversity combining, in which it is difficult to compensate for amplitude fluctuations.

Next, the phase fluctuation compensation performed by the intra-symbol phase fluctuation compensating apparatus 300 will be described with reference to FIG. The relationship between the received signal r (t) and the delayed signal r (t−T _f ) delayed by the effective symbol length T _f is as shown in the upper half of FIG. The last interval of the effective symbol of the received signal r (t) coincides with the guard interval with the delayed signal r (t−T _f ) delayed by the effective symbol length T _f . If there is no phase fluctuation of the propagation path characteristic during the time T _f , the guard interval (GI) in the delayed signal r (t−T _f ) and the effective symbol of the received signal r (t) as the copy source The waveform at the end is exactly the same.

Therefore, the complex correlation between the received signal r (t) including the end of the effective symbol corresponding to the guard interval and the delayed signal r (t−T _f ) including the guard interval with an integral interval length T _S at the start position. Calculate. Of the integration interval length T _s , in the integration interval other than the complex correlation between the end of the effective symbol corresponding to the guard interval of the received signal r (t) and the guard interval of the delayed signal r (t−T _f ), the carrier Since the number may be as large as several thousand, it can be considered that the complex correlation is extremely small statistically. That is, the complex correlation between the end of the effective symbol corresponding to the guard interval of the received signal r (t) and the guard interval of the delayed signal r (t−T _f ) occupies a large value in the integration interval length T _S. If the waveform of the guard interval (GI) in the delayed signal r (t−T _f ) and the end of the effective symbol that is the copy source coincide completely, the complex correlation becomes a real number, and the argument of the complex correlation becomes zero. . If the waveform of the guard interval (GI) in the delayed signal r (t−T _f ) and the end of the effective symbol that is the copy source do not completely match, the complex correlation has an imaginary part, and the complex correlation The declination is not zero. Therefore, this declination is defined as phase fluctuation. This phase fluctuation is a phase fluctuation during a time T _f from the start time of the guard interval (GI) to the start time of the copy source.

In the present invention, the integration interval for estimating the phase change is a time length T _S of one symbol, the received signal is always a guard interval (GI) in the delay signal r (t-T _f), the copy source The end of the effective symbol of r (t) is included. Also, the start timing of the integration interval for estimating the phase variation does not need to be synchronized with each symbol. In the present invention, each integration interval may include the guard interval (GI) in the delayed signal r (t−T _f ) and the end of the effective symbol of the received signal r (t) as the copy source, and one guard interval. (GI) It is not necessary to have the whole in one integration section. That is, the integration interval for the correlation calculation starts from the middle of a guard interval (GI) of a certain symbol in the delay signal r (t−T _f ) at the start, and the integration interval ends at the end of the delay signal r (t−T _f ). _{Even if} it ends in the middle of the guard interval (GI) of a symbol in _f ), the phase fluctuation can be calculated. This is because a part of the guard interval in one symbol and the last part of the effective symbol corresponding to that part, and a part of the guard interval in the next symbol and the last part of the effective symbol corresponding to that part are: This is because it always exists in the integration interval.

This will be described using mathematical expressions. First, the phase fluctuation (phase difference at two times of the time difference T _f ) φ within the time interval T _{f of} the received signal r (t) is expressed by the following equation (1) by the complex correlation function of the integration interval length T _S. Is required. Note that r ^* (t) represents the complex conjugate of r (t), and arg represents the complex argument.

This will be briefly described. First, as shown in FIG. 4, the interval length T _s, delay signal r (t-T _f) from OFDM symbol a _m of the m carriers, the guard interval and the OFDM symbol of the OFDM symbol b _m is b _m to, signal r (t) is the OFDM symbol b _m of the m carriers, the guard interval and the OFDM symbol of the OFDM symbol c _m is the c _m, will shift. Therefore, the head where the OFDM symbol of the signal r (t) is b _m is set to n = 0, the symbol length T _f is equally divided into N, and the sampling interval is set to τ, and Expressions (2-1) and (2-2) ). G is a value obtained by dividing the guard interval length T _GI by the sampling interval τ. Further, ω is ω = 2π / _Tf in terms of the frequency interval of the subcarriers. That is, T _f = nτ, T _GI = Gτ, and Nτω = 2π. Further, j is an imaginary unit, and θ _{m, n and the} like are phases generated by propagation path characteristics in each m-th carrier n-th sampling. Further, a guard interval and overlap portions of the OFDM symbol b _m of OFDM symbols b _m and the delayed signal r (t-T _f) of the signal r (t) represents a Enutau∈T _g, of the interval length T _s The part before T _g is indicated as T _pre , and the part after T _g is indicated as T _post .

From Equations (2-1) and (2-2), first, a portion where the OFDM symbol of the signal r (t) overlaps the guard interval of b _m and the OFDM symbol of the delayed signal r (t−T _f ) is b _m When (nτ∈T _g ) is expanded, the complex integral part of Expression (1) is expanded as shown in Expression (3). Here, considering that the number of carriers (value range of m and l) is several thousand, the second term is uniformly distributed in the section 0 ≦ x <N, where x = (1−m) n (modN). It can be approximated that the complex numbers b _m ^* and b _l are independent and random and θ _{m, n} varies linearly with respect to n, that is, θ _{l, nN} −θ _{m, n} is approximately n with respect to n. If considered to be a constant, it can be considered to be statistically zero.

Next, for the part of nτ∈T _pre where the OFDM symbol of the signal r (t) is b _m and the OFDM symbol of the delayed signal r (t−T _f ) is a _m , the complex integral part of Equation (1) is When expanded, equation (4) is obtained. Here, since the complex numbers b _m ^* and a _m are independent and random, the first term can be considered to be statistically zero considering that the number of carriers m is thousands. Also, in the second term, b _m ^* and a _l are independent and random, and θ _{l, n−N + G} −θ _{m, n} is a constant with respect to n, and y = ( l−m) n + lG (modN) is uniformly distributed in the interval 0 ≦ x <N, and therefore, considering that the number of carriers (range of m and l) is several thousand, it is considered to be statistically zero. be able to.

Just as it can be considered that the formula (4) is statistically kinetically 0, OFDM symbol of the signal r delay signal OFDM symbol in c _m of (t) r (t-T _f) is a b _m The portion of nτεT _post can be considered to be statistically zero. Eventually, the complex integral part of equation (1) can be approximated by the first term of equation (3). Further, in equation (3), if θ _{m, nN} −θ _{m, n} is constant for each m and θ _{m, nN} −θ _{m, n} = Δθ _m , the declination of equation (3) can be obtained. Can be further expressed as the following (5). That is, Equation (1) represents a weighted average of the phase difference Δθ _m of each carrier by the square of the carrier amplitude. Note that _setting θ _{m, nN} −θ _{m, n} = Δθ _m indicates that the phase change is linear with respect to the time axis direction.

The phase fluctuation is smoothed so as not to cause a large fluctuation due to noise. This is expressed by equation (6).

Since the phase variation of the formula (6) is the phase variation between the time interval T _f, which is converted into a phase variation per time interval T _a determined compensation coefficient amplitude fluctuation described above. That is, Equation (7).

Compensation coefficient of the phase change at each time interval T _a is the cumulative value of the previous phase variation (shown with a symbol - on the phi), is obtained by adding phase change equation (7). Therefore, the compensation coefficient for the phase variation is expressed by Expression (8) as a complex number having an absolute value of 1.

Thus, in equation (8), the coefficients for compensating the time variation of the phase is obtained, which by multiplying the signal of the time interval T _a, it is possible to compensate for the time variation of the phase.

FIG. 5 shows the calculation inside the intra-symbol phase fluctuation estimator 301. FIG. 5 is a block diagram showing the procedure for calculating the phase fluctuation correction coefficient. In order to correspond to the above simulation, the effective symbol length T _f is 1.008 ms, the guard interval length T _GI = T _f / 8 = 126 μs, the time length of one symbol T _S = T _GI + T _f = 1.134 ms, T _a = T _f / 8 = 126 μs. Also, the FFT is 8192 point FFT and the effective symbol length _Tf is 8192 samples. Guard interval length T _GI and T _a is 1024 samples, the time length T _S of one symbol becomes 9216 samples.

As shown in FIG. 5, the procedure for calculating the phase fluctuation correction coefficient is as follows. The interval length is 1 symbol length T _s = 1.134 ms, and the digital complex signal and the signal are delayed by an effective symbol length T _f = 1.008 ms. A complex correlation with the signal is obtained (SD, SP1, SP2). In this complex correlation, the integration interval (length 1 symbol length T _s ) does not have to coincide from the start of the guard interval, which is the original 1 symbol, to the end of the effective symbol, but the delay of the effective symbol length T _f . Accordingly, the guard interval of the delayed signal inevitably overlaps with the end of the effective symbol that should correspond to the digital complex signal. Furthermore it is not necessary to the whole of one guard interval of the delayed signal is included in the integration interval (length 1 symbol length T _s), the integration interval beginning of one of the delay signals (length 1 symbol length T _s) It may start from the middle of the guard interval and the end of the integration interval may be up to the middle of the next guard interval of the delay signal.

In this complex correlation, there is no phase difference between the delayed signal and the received signal, that is, reception is started at a time t + T _f and a guard interval (section length T _GI ) at the beginning of one symbol at which reception started at a certain time t. If the end of the interval length T _GI of the corresponding one symbol matches completely, the complex correlation becomes real, declination i.e. phase difference is 0. On the other hand, when there is a phase difference between the delayed signal and the received signal, the complex correlation has an imaginary part, and its declination becomes the phase difference. When the phase difference of each carrier is sufficiently small (usually established because the effective symbol length _Tf is small), the declination is a power weighted average of the phase differences of each carrier. Using this, the phase fluctuation is estimated (SP3). This is a phase variation in the effective symbol length T _f = 1.008 ms, which is a delay time.

Next, the phase fluctuation is smoothed (SP4) and the amplitude fluctuation calculation time T _a is 1/8 of the effective symbol length T _f so as to match the update timing and update timing of the amplitude fluctuation. (SP5). This is accumulated nine times, and nine phase fluctuation accumulated values corresponding to one symbol length T _s are sequentially calculated (SP6). If necessary, the calculation result falls within the range of −π to π, or within the range of 0 to 2π. This is a phase variation correction coefficient (complex number, absolute value is 1) (SP7).

  Next, FIG. 7 shows a simulation for improving the error rate characteristics in diversity reception with time variation compensation. FIG. 7 shows average bit error rate characteristics when the configuration of FIG. 1 is 4-branch (antenna) diversity, Eb / No is 20 dB, and the Doppler frequency is changed from 2 Hz to 300 Hz. Other simulation conditions were the same as the simulation conditions of FIG. 3 of the first embodiment. In FIG. 7, when the Doppler frequency is small, the moving speed is slow, and when the Doppler frequency is large, the moving speed is fast.

  For comparison, FIG. 7 also shows a case where maximum ratio combining diversity is performed without adding time variation compensation (normal diversity) and a normal OFDM reception method without diversity and time variation compensation (without diversity). did. As shown in FIG. 7, the average bit error rate is reduced and the characteristics are improved when normal diversity is performed compared to the case where no diversity reception is performed. However, when the Doppler frequency exceeds 10 Hz, the average bit error rate deteriorates. This is because the time variation of the received signal becomes faster as the moving speed becomes faster. On the other hand, according to the present invention, when the Doppler frequency is 40 Hz or less, the average bit error rate hardly deteriorates. This indicates that by adding time variation compensation to diversity reception, the average bit error rate when the moving speed is high is reduced and the characteristics are improved. This is the improvement effect by suppressing the time fluctuation of the amplitude and phase according to the present invention.

  As in the above embodiments, the present invention is very effective when applied to OFDM mobile reception, but is also effective when applied to other orthogonal communication systems.

The block diagram which shows the structure of the multicarrier diversity receiver 1000 which concerns on the specific 1st Example of this invention. The block diagram which shows the structure of the time fluctuation compensation apparatus 200 of 1st Example. The graph which shows the simulation result of the time fluctuation compensation apparatus 200 of 1st Example with a comparative example. The schematic diagram for demonstrating the calculation content of the phase variation compensation apparatus 300 of 1st Example. The block diagram which shows the structure of the phase variation compensation apparatus 300 of 1st Example. The block diagram which shows the structure of the multicarrier diversity receiver 9000 described in the application documents of the prior invention by the present inventors. The graph figure which shows the simulation result of the average bit error rate of the multicarrier diversity receiver which concerns on the specific 2nd Example of this invention.

Explanation of symbols

1000: Multi-carrier diversity receiver 200: Time variation compensator 201: Weight coefficient calculation units 202a, 202b, 302: Complex multiplier 203: Complex adder 300: Phase variation compensator 301: Intra-symbol phase variation estimator

Claims (4)

In the method of compensating for the time variation of amplitude in diversity reception to synthesize received waves received by a plurality of antennas, compensate for variations in the amplitude direction of the combined signal, and make the average power of the combined signal constant,
Obtaining the complex cross-correlation between the received signal of each antenna and the combined signal obtained by multiplying each received signal by a complex weighting factor and adding it, corresponding to each antenna,
Smoothing each complex cross-correlation corresponding to each antenna to obtain a smoothed correlation value,
A method for compensating for time fluctuation of amplitude in diversity reception, wherein a value obtained by dividing each smoothed correlation value by the sum of squares of all smoothed correlation values is newly set as an update value of each weighting coefficient. A multicarrier modulation signal in which the time length T _{S of} one symbol is T _S = T _GI + T _{f obtained} by copying the waveform of the portion of the time length T _GI at the end of the signal to the effective symbol length T _f and adding it to the head. In the intra-symbol time variation compensation method in multicarrier diversity reception for receiving
Compensating for intra-symbol time variation in the amplitude direction using the amplitude variation variation method according to claim 1 ,
Phase variation at time T _f is obtained by complex correlation between the combined signal with amplitude compensated for time length T _S and the delayed signal obtained by delaying the combined signal with compensated amplitude by T _f as delay time. It estimated, and calculates the phase variation of each short time T _a than the effective symbol length T _f, to compensate for the symbol in the time variation of phase seeking correction coefficient of the phase at the time T _a based on it Intra-symbol time variation compensation method in multi-carrier diversity reception characterized by the above. In a compensation apparatus for amplitude time fluctuations in a diversity receiver for synthesizing received waves received by a plurality of antennas to compensate for fluctuations in the amplitude direction of the synthesized signals and to keep the average power of the synthesized signals constant,
A plurality of complex multipliers for multiplying the received signal of each antenna by a weighting factor that is a complex number calculated for each received signal;
A complex adder for adding the outputs of the plurality of complex multipliers;
A plurality of complex cross-correlation calculators each for calculating a complex cross-correlation between a reception signal of each antenna and a composite signal that is an output of the complex adder;
A plurality of smoothers for smoothing the outputs of the plurality of cross-correlation calculators;
A plurality of weighting coefficient updaters, each of which is obtained by dividing each smoothed correlation value, which is an output of the plurality of smoothers, by the sum of squares of all smoothed correlation values. An apparatus for compensating for time fluctuation of amplitude in diversity reception as a feature. A multicarrier modulation signal in which the time length T _{S of} one symbol is T _S = T _GI + T _{f obtained} by copying the waveform of the portion of the time length T _GI at the end of the signal to the effective symbol length T _f and adding it to the head. In the intra-symbol time variation compensation apparatus in multicarrier diversity reception for receiving
An amplitude time variation compensating device according to claim 3 ,
Phase variation at time T _f is obtained by complex correlation between the combined signal with amplitude compensated for time length T _S and the delayed signal obtained by delaying the combined signal with compensated amplitude by T _f as delay time. The phase variation for each time T _a shorter than the effective symbol length T _f is calculated, and the phase correction coefficient of the phase at the time T _a is obtained based on the phase variation to compensate for the time variation in the symbol within the phase. An intra-symbol time variation compensating device in multicarrier diversity reception, comprising: a time variation compensating device.

Images