JP3501767B2 - Frequency error estimating apparatus and frequency error estimating method - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a frequency error estimating device for estimating a frequency error between a local oscillation frequency and a carrier frequency of a received signal in a receiver used in satellite communication / mobile satellite communication / mobile communication. Is related to
The present invention relates to a frequency error estimation device capable of ensuring high estimation accuracy without impairing the followability of a Doppler frequency with respect to time variation in frequency error estimation.

[0002]

2. Description of the Related Art A conventional frequency error estimating device will be described below. For example, some receivers used for mobile communication use a synchronous detection method that can obtain good detection characteristics even in a low C / N propagation path. In this synchronous detection system, a carrier wave synchronized with the carrier frequency of the received signal is reproduced, and a detection output is obtained based on this carrier wave. However, in a receiver adopting the synchronous detection method, the oscillation frequency fluctuates due to fluctuations in the accuracy of the oscillator, temperature fluctuations, and the like, and as a result, a frequency shift between transmission and reception, that is, a frequency error occurs. When such a frequency error exists, the phase of the signal point on the IQ plane (the complex plane represented by the real axis and the imaginary axis) rotates.

Therefore, in the receiver, in order to reduce the frequency error and realize the improvement of the synchronization characteristic, a frequency error estimator for measuring the phase rotation amount from the received signal and estimating the frequency error from the measurement result is required. Becomes

FIG. 8 shows a configuration of a conventional frequency error estimating device, for example, "Doppler-Correc".
ted Differential Detection of MPSK ", IEEE Trans. C
ommun., Vol. COM-37, 2, pp99-109, Feb., 1989. In FIG. 8, 1 is a received signal, 21 is an M multiplier that removes the modulation component of the received signal 1, and 22 is a D symbol delay detector that delay-detects the output of the M multiplier 21 between D symbols. Yes, 101 is an averaging filter for averaging the output of the D-symbol delay detector 22 to suppress noise components, and 24 is an averaging filter 10.
A coordinate converter that calculates a phase component from the output of 1.
Reference numeral 5 is a divider that calculates a frequency error from the phase component output from the coordinate converter 24, and reference numeral 2 is an estimated frequency error value output from the divider 25.

Further, FIG. 9 shows the averaging filter 101.
Is an example of the internal configuration of the first-order IIR (Infinite Impulse
Response) filter. In FIG. 9, 111 and 112 are multipliers that multiply an input signal by a specific coefficient, 32 is an adder that adds two inputs, and 33 is a delay device that delays the signal by one symbol.

The conventional frequency error estimation circuit configured as described above is a D symbol delay detection waveform frequency error estimation device for estimating the phase change amount due to the frequency error by performing delay detection of the received signal between D symbols. is there.

Next, the operating principle will be described with reference to FIGS. 8 and 9. The reception signal 1: r (nT) is expressed by Expression (1), for example, when the modulation method is M-phase PSK (Phase Shift Keying). r (nT) = A (nT) exp [j {θ (nT) + ΔωnT}] (1) However, in Expression (1), the received signal 1: r (nT)
Is a complex baseband signal sampled at the symbol period T. A (nT) represents an amplitude component,
Δω represents the angular frequency error. Further, θ (nT) represents a modulation component, for example, 2πk / M (k = 0, 1, ..., M
-1) M values are taken. Note that, here, for simplicity of description, it is assumed that there is no noise component.

The M multiplier 21 multiplies the received signal 1 by a modulation multilevel number M in order to remove the modulation component. The signal after multiplication: r1 (nT) is expressed by the equation (2). r1 (nT) = A (nT) exp [jM {θ (nT) + ΔωnT}] (2) However, in Expression (2), Mθ (nT) is a multiple of 2π and can be ignored. That is, the formula (2) is the formula (3)
Can be replaced with r1 (nT) = A (nT) exp (jMΔωnT) (3)

The D-symbol delay detector 22 delay-detects the output r1 (nT) from the M multiplier 21 between D symbols. The signal after differential detection: d1 (nT) is expressed by the equation (4).
It is represented by. d1 (nT) = r1 (nT) r1 * (nT-DT) = A (nT) A (nT-DT) exp (jMDΔωT) (4) However, in Expression (4), r1 * (nT- DT) is r
It is a conjugate complex number of 1 (nT-DT).

The averaging filter 101 suppresses the noise component by averaging the output: d1 (nT) from the D-symbol delay detector 22. For example, when the first-order IIR filter of FIG. 9 is used as the averaging filter,
The output of the averaging filter 101: d2 (nT) is expressed by equation (5). d2 (nT) = [alpha] d1 (nT) + (1- [alpha]) d2 (nT-T) (5) However, in the equation (5), the first term in the multiplier 111 is the input signal: d1 ( As a result of multiplying nT) by the coefficient α, the second term is the output of the averaging filter 101 delayed by one symbol in the multiplier 112: d2 (nT−
This is the result of multiplying T) by the coefficient 1-α.

If the sample timing is assumed to be the Nyquist point, that is, the amplitude component is assumed to be 1, the output of the averaging filter 101: d2 (nT) is
It is expressed by equation (6). d2 (nT) = exp (jMDΔωT) (6)

In the coordinate converter 24, the averaging filter 10 is used.
The output of 1: d2 (nT) is converted from rectangular coordinates to polar coordinates, and the phase component MDΔωT is calculated. Finally, the divider 2
5, the phase component output from the coordinate converter 24: MDΔ
ωT is divided by MD to obtain the angular frequency error Δ for one symbol
Calculate and output ωT.

[0013]

However, the above
In the conventional frequency error estimation device, in order to accurately estimate the frequency error, it is necessary to set the coefficient α of the multiplier in the averaging filter as small as possible to enhance the averaging effect. On the other hand, when the received signal undergoes a large Doppler shift and the Doppler frequency fluctuates with time, as in mobile satellite communication, the frequency error estimator follows the fluctuation and detects the frequency error. Need to estimate. That is, in order to improve this tracking ability, it is necessary to set the coefficient α of the multiplier in the averaging filter as large as possible.

As described above, in the conventional frequency error estimating apparatus, since the coefficient α of the multiplier is fixed, it is difficult to make the estimation accuracy of the frequency error and the followability with respect to the time fluctuation of the Doppler frequency compatible. There was a problem that there was.

The present invention has been made in view of the above, and in the frequency error estimation processing in the receiver,
An object of the present invention is to obtain a frequency error estimation device and a frequency error estimation method capable of ensuring high estimation accuracy without impairing the followability to time variation due to the Doppler frequency.

[0016]

[Means for Solving the Problems]
In order to achieve the object, the present invention provides a station in a receiver.
Frequency of local oscillation frequency and carrier frequency of received signal
In the frequency error estimation device that estimates the error,
Included in the received signal according to the filter coefficient input to
Suppresses noise components and is based on multiple phase variations of the received signal
Frequency error estimation means for estimating the frequency error by
Calculates the difference between the estimated values of frequency error between symbols
Difference calculation means and the frequency calculated by the difference calculation means.
Calculate the filter coefficient based on the difference in the estimated wave number error
Determining means for inputting the filter coefficient to the filter
And are provided.

According to the present invention, the characteristics of the filter for suppressing the noise component of the received signal are changed according to the magnitude of the time variation of the Doppler frequency, so that the followability and the frequency error with respect to the time variation of the Doppler frequency are obtained. It is possible to achieve both the estimation accuracy of

In the frequency error estimation method according to the next invention, the local oscillation frequency in the receiver and the received signal
Frequency error that estimates the frequency error between the carrier frequency
Difference estimation method, which is included in the received signal using a filter
Based on the phase fluctuation of the received signal
Frequency error estimation process for estimating frequency error and symbol
Difference operation to calculate the difference between the estimated values of the frequency error between
Frequency error calculated by calculation process and difference calculation process
Based on the difference between the estimated values of
Error frequency changes depending on the magnitude of the time variation.
Is included .

According to the present invention, the characteristics of the filter for suppressing the noise component of the received signal are changed according to the magnitude of the time variation of the Doppler frequency, so that the followability and the frequency error with respect to the time variation of the Doppler frequency are obtained. It is possible to achieve both the estimation accuracy of

[0020]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of a frequency error estimating apparatus according to the present invention will be described below in detail with reference to the drawings. The present invention is not limited to this embodiment.

Embodiment 1. FIG. 1 is a block diagram showing a configuration example of a frequency error estimation device according to the first exemplary embodiment. In FIG. 1, 1 is a received signal, 11 is a frequency error estimator that estimates a frequency error between the local oscillation frequency and the carrier frequency of the received signal 1, 2 is an estimated frequency error value, and 12 is an estimated frequency. The filter coefficient determiner calculates the filter coefficient information of the averaging filter included in the frequency error estimator 11 based on the error value 2, and 3 is the filter coefficient information.

Further, FIG. 2 shows the frequency error estimator 11 described above.
Is an example of the internal configuration of the D symbol delay detection frequency error estimator 11a. In addition, in the present embodiment, the same configurations as those of the conventional technique described above are denoted by the same reference numerals, and description thereof will be omitted. D shown in FIG.
The symbol delay detection waveform frequency error estimator 11a is based on the same principle as the D symbol delay detection frequency error estimation circuit described in the prior art, and the filter coefficient information 3 is input to the averaging filter 23. And that the filter characteristic in the averaging filter 23 changes depending on the filter coefficient information 3.

FIG. 3 shows an internal configuration example of the averaging filter 23, which is a first-order IIR filter like the averaging filter 101 described above in the prior art.
The averaging filter 23 shown in FIG.
Is input, and multipliers 31 and 34 are provided instead of the multipliers 111 and 112.
The difference is that the multiplier coefficients of 1 and 34 change according to the filter coefficient information 3.

Further, FIG. 4 shows the filter coefficient determiner 1 described above.
2 is an example of the internal configuration of FIG. In FIG. 4, 41 is a difference calculator for calculating the difference between the estimated frequency error values 2, 42 is an averaging filter for averaging the output of the difference calculator 41 to suppress noise components, and 43 is an averaging filter. 42 is a coefficient determiner that determines the filter coefficient information 3 based on the output of 42.

The operation of the frequency error estimating apparatus of the present embodiment configured as described above will be described below. First, the received signal 1 is the frequency error estimator 11, that is, the D symbol delay detection waveform frequency error estimator 11a shown in FIG.
Entered in. At this time, the D-symbol delay detection waveform frequency error estimator 11a that has received the reception signal 1 operates in the same manner as the D-symbol delay detection waveform frequency error estimation circuit shown in FIG. 8 (corresponding to equations (1) to (5)). ), Estimated frequency error value 2
Is output. The output of the averaging filter 23: d2
Formula (5) is used as a calculation formula for obtaining (nT), and here, the multiplier coefficient α is the filter coefficient information 3
It will change depending on.

Next, the operation of calculating the filter coefficient information 3 in the filter coefficient determiner 12 will be described. In addition,
Here, it is assumed that the frequency error changes in proportion to the time due to the time variation of the Doppler frequency. Further, it is assumed that the estimated frequency error value 2 output by the frequency error estimator 11 is an optimum (correct) value. Therefore, the estimated frequency error value 2 is output as the angular frequency error between one symbol: ΔωT and is represented by the equation (7). ΔωT = AnT + B (7) As described above, the estimated frequency error value 2 changes in proportion to time: nT as in the above equation. That is, as the time variation of the Doppler frequency becomes faster, the absolute value of the variable A becomes larger accordingly.

In the difference calculator 41, the estimated frequency error value Δ
The difference between the R symbols of ωT is calculated. In the present embodiment, the output of the difference calculator 41: Δωd is expressed by equation (8). Δωd = ART (8)

In the averaging filter 42, the difference calculator 41
The output of is averaged and the noise component is suppressed. Note that, here, as the averaging filter, a first-order IIR filter shown in FIG. 9, a higher-order IIR filter and a FIR (Fi
nite Impulse Response) filter may be used. Further, as described above, since it is assumed that there is no noise component here, the output of the averaging filter 42 is held and is expressed by the equation (8) as it is.

In the coefficient determiner 43, the averaging filter 42
Output: The filter coefficient information 3 is determined according to the value of Δωd. Here, for example, the filter coefficient information 3 is set to the multiplier coefficient α of the multiplier 31, while the multiplier 34
Let 1 be the multiplier coefficient of.

Specifically, in the coefficient determiner 43, when the time variation of the Doppler frequency is slow, the variable A
Since the absolute value of Δωd is small, the absolute value of Δωd is also small.
Therefore, the absolute value of Δωd also increases. That is, Δωd is an index indicating the time variation amount of the Doppler frequency. In this way, the coefficient determiner 43 sets α as small as possible when the absolute value of Δωd is smaller than the predetermined value, and conversely sets α as large as possible when the absolute value of Δωd is larger than the predetermined value. Thereby, when the time variation of the Doppler frequency is slow, the averaging effect of the averaging filter 23 is enhanced, and the frequency error can be estimated with high accuracy. On the other hand, when the time variation of the Doppler frequency is fast, the frequency error can be estimated with high accuracy after following the variation.

In the present embodiment, as a specific example of the multiplier coefficient α to be set, for example, a correspondence table between the absolute value of Δωd and the multiplier coefficient α as shown in FIG. 5 is prepared. The optimum value of the multiplier coefficient α in this correspondence table is determined in advance by computer simulation or the like. An example of the correspondence table shown in FIG.
The multiplier coefficient α is changed in four steps according to the value of d, and 0 <W1 <W2 <W3. In addition, here
Although the correspondence table of four stages is used, the multiplier coefficient α may be changed not by four stages but by an arbitrary number of stages of two or more stages, for example.

Therefore, the coefficient determiner 43 refers to the above correspondence table and determines the multiplier coefficient α corresponding to the value of Δωd.
Can be selected, and thus, the optimum α can be selected according to the time variation of the Doppler frequency.
The multiplier coefficient α can be updated for each symbol, but considering the stability of the filter coefficient determiner 12, it is better to use a counter, for example, at a specific cycle. The period is preferably equal to or longer than the convergence time of the frequency error estimator 11, for example.

As described above, according to the present embodiment, the characteristics of the averaging filter for suppressing the noise component of the phase change information are changed according to the magnitude of the time fluctuation amount of the Doppler frequency, so that the Doppler frequency is changed. Regardless of the magnitude of the time variation of the frequency, it is possible to achieve both the followability with respect to the time variation of the Doppler frequency and the estimation accuracy of the frequency error.

Embodiment 2. The frequency error estimating apparatus according to the second embodiment is configured similarly to the frequency error estimating circuit in the embodiment shown in FIG.
1 has a different internal structure. FIG. 6 shows an example of the internal configuration of the frequency error estimator 11 in the present embodiment, which is, for example, a multiple open loop type frequency error estimator 11b.

In the present embodiment, as the frequency error estimator 11, for example, a multiple open loop type frequency error estimator 11b is used.
A wide estimation range and high estimation accuracy can be realized. For the multiple open loop type frequency error estimator, refer to "A
Multiple Open-Loop Frequency Estimation Based on D
ifferential Detection for MPSK ", IEICE Trans. Comm
un., Vol. E82-B, 1, pp136-144, Jan., 1999.

In FIG. 6, 51a, 51b, ..., 51
Reference numeral c denotes N D-symbol delay detection waveform frequency error estimators provided in parallel, each having the same configuration as the D-symbol delay detection waveform frequency error estimator 11a shown in FIG. .., 51c in the D-symbol delay detection waveform frequency error estimators 51a, 51b, ...
DN, which are different from each other. Also, 52 is D
Symbol delay detection waveform frequency error estimator 51a, 51b,
, 51c is a selector that outputs an optimum (presumably correct) estimated frequency error value based on the estimated frequency error value output from 51c.

The operation of the frequency error estimating apparatus of the present embodiment configured as above will be described below. First, the received signal 1 is input to the frequency error estimator 11, that is, the multiple open loop type frequency error estimator 11b shown in FIG. In the multiple open loop type frequency error estimator 11b that has received the reception signal 1, the input reception signal 1 is branched into N systems, and D symbols with different delay detection symbol intervals (D1 to DN) are provided in parallel. Input to the delay detection waveform frequency error estimators 51a, 51b, ..., 51c.

Each D-symbol differential detection waveform frequency error estimator performs differential detection at different symbol intervals to perform frequency error estimation processing. The differential detection symbol interval is set so that D1 <D2 <... <DN. In the present embodiment, for example, when the differential detection symbol interval is small, the frequency error estimation range is wide, but the estimation accuracy is low. On the other hand, when the delay detection symbol interval is large, the frequency error estimation range is narrowed but the estimation accuracy is increased. As described above, in the multiple open loop type frequency error estimator 11b of the present embodiment, the D symbol delay detection frequency error estimator 51a having the smallest delay detection symbol interval causes the D symbol delay detection having the largest delay detection symbol interval. By providing a plurality of waveform frequency error estimators 51c in parallel and selecting each output by the selector 52, a wide estimation range and high estimation accuracy are realized.

Here, the operation of the multiple open loop type frequency error estimator composed of, for example, three D symbol delay detection frequency error estimators will be specifically described. In this multiple open loop type frequency error estimator, the differential detection symbol interval: Di (i = 1, 2, ..., N) is given by the equation (9).
Set like. Di = 2 ^ (i-1) (i = 1, 2, 3) (9) However, in Expression (9), ^ represents a power operation.

Therefore, assuming that the actual frequency error is π / 4, the estimated value candidates for the three D-symbol delay detection waveform frequency error estimators are delayed as shown in FIG. 7 due to frequency uncertainty. There are 1 (), 2 (Δ), and 4 (□) corresponding to the detection symbol interval. Then, the multiple open-loop type frequency error estimator 11b first uses the selector 52 to select the closest Δ (black triangle shown) from the two candidates (Δ). Next, from the four candidates (□), the □ (black square in the figure) closest to the previously selected Δ is selected. As a result, the multiple open loop frequency error estimator 11b outputs the selected square as an estimated value of the frequency error.

As described above, since the multiple open loop type frequency error estimator 11b operates while removing the frequency uncertainty, the estimation range of the frequency estimation error is D (D1) symbol delay detection waveform frequency error estimator 51a. Determined by
The estimation accuracy is determined by the D (DN) symbol delay detection frequency error estimator 51c. That is, the multiple open loop type frequency error estimator 11b of the present embodiment can simultaneously realize a wide estimation range and high estimation accuracy.

The multiple open loop type frequency error estimator 1
In the operation of 1b, the multiplier coefficient α of the averaging filter 23 in each D symbol delay detection waveform frequency error estimator,
That is, α of the calculation formula in the averaging filter 23 shown in the formula (5) changes depending on the filter coefficient information 3. At this time, the filter coefficient information 3 is determined by the filter coefficient determiner 12 based on the difference between the estimated frequency error values, as in the first embodiment. That is, when the time variation of the Doppler frequency is slow, the averaging effect can be enhanced by setting α small, and the frequency error can be estimated with high accuracy. On the other hand, when the time variation of the Doppler frequency is fast, α can be set to a large value and the variation can be tracked, and then the frequency error can be estimated with high accuracy.
Here, as in the first embodiment, a correspondence table as shown in FIG. 5 may be used.

As described above, according to the present embodiment, the characteristics of the averaging filter for suppressing the noise component of the phase change information are set to the time variation amount of the Doppler frequency, as in the first embodiment described above. By changing the Doppler frequency with respect to time, it is possible to achieve both the followability with respect to the time variation of the Doppler frequency and the estimation accuracy of the frequency error. Further, by using the multiple open loop type frequency error estimator 11b in which a plurality of D symbol delay detection frequency error estimators are provided in parallel, higher estimation accuracy and wider estimation range can be realized at the same time.

Although the present invention has been described above with reference to the first and second embodiments, various modifications are possible within the scope of the gist of the present invention, and these modifications are not excluded from the scope of the invention.

[0045]

As described above, according to the present invention, the characteristics of the filter for suppressing the noise component of the received signal are
By changing the Doppler frequency according to the amount of time variation, it is possible to obtain a frequency error estimation device capable of achieving both the tracking of the Doppler frequency with respect to time variation and the estimation accuracy of the frequency error. Produce an effect.

According to the next invention, the characteristics of the filter for suppressing the noise component of the received signal are changed according to the magnitude of the time variation amount of the Doppler frequency, so that the followability with respect to the time variation of the Doppler frequency and the frequency are obtained. It is possible to obtain a frequency error estimation method that is compatible with the error estimation accuracy.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration example of a frequency error estimation device according to a first exemplary embodiment.

FIG. 2 is a diagram showing an internal configuration example of a frequency error estimator 11.

FIG. 3 is a diagram showing an internal configuration example of an averaging filter 23.

FIG. 4 is a diagram illustrating an internal configuration example of a filter coefficient determiner 12.

FIG. 5 is a diagram showing an example of a correspondence table.

FIG. 6 is a block diagram showing a configuration example of a frequency error estimation device according to a second exemplary embodiment.

FIG. 7 is a diagram showing a method of selecting a frequency error estimated by a plurality of D-symbol delay detection waveform frequency error estimators.

FIG. 8 is a diagram showing a configuration of a conventional frequency error estimation device.

FIG. 9 is a diagram showing an internal configuration example of a conventional averaging filter.

[Explanation of symbols]

1 received signal, 2 estimated frequency error value, 3 filter coefficient information, 11 frequency error estimator, 11a, 51a, 5
1b, 51c D symbol delay detection waveform frequency error estimator, 11b Multiple open loop type frequency error estimator, 12
Filter coefficient determiner, 21 M multiplier, 22 D symbol delay detector, 23 averaging filter, 24 coordinate converter, 25 divider, 31,34 multiplier, 32 adder, 33 delay device, 41 difference calculator, 42 averaging filter, 43 coefficient determiner, 52 selector.

Claims (4)

(57) [Claims] 1. A frequency error estimator for estimating a frequency error between a local oscillation frequency in a receiver and a carrier frequency of a received signal, wherein a noise component included in the received signal according to a filter coefficient input to a filter. To suppress the phase shift of the received signal.
Frequency error estimating means for estimating a frequency error based on motion , difference calculating means for calculating a difference between the estimated values of the frequency error between a plurality of symbols, and the estimated value of the frequency error calculated by the difference calculating means And a coefficient determining unit that inputs the filter coefficient to the filter based on the difference of the frequency error estimating apparatus. 2. The frequency error estimating means is configured to receive the received signal.
The frequency error is estimated by delay-detecting the signal, The coefficient determining means is a Doppler in the received signal.
Averaging effect of the filter according to the time variation of frequency
The frequency according to claim 1, characterized in that
Error estimation device. 3. A local oscillator frequency in the receiver, the carrier frequency of the received signal, a frequency error estimating method for estimating a frequency error, and suppresses noise components included in the received signal using a filter, A frequency error estimating step of estimating a frequency error based on a phase variation of the received signal, a difference calculating step of calculating a difference between the estimated values of the frequency error among a plurality of symbols, and the difference calculated by the difference calculating step. Based on the difference between the estimated values of frequency error, the characteristics of the filter are
And a step of changing the puller frequency according to the magnitude of the time variation amount of the puller frequency . 4. The frequency error estimating step comprises:
The frequency error by delay detection of the signal
The method for estimating frequency error according to claim 3, characterized in that
Law.