JPH0951294A - Data receiving device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To attain equalization processing minimizing a residual error in the case of presence of a noise which properly traces even a fast fluctuation in a channel. SOLUTION: The data receiver having a demodulator 4 with an adaptive control device and a tap coefficient correction device 5 correcting a tap coefficient of the demodulator by an adaptive algorithm is provided with a parameter setting means 8 controlling independently a parameter to specify the correction value of each tap coefficient corrected by the tap coefficient correction device for each tap coefficient. In the case of correcting the tap coefficient of the adaptive control device by the algorithm such as the LMS or the RLS, when the direction of the correction of the tap coefficient is unchanged between the preceding correction and the correction this time, it is decided to be a state before convergence, then a large correction coefficient or a large obliteration coefficient is set. When the direction of the correction of the tap coefficient is changed between the preceding correction and the correction this time, it is decided to be a state after convergence. Then a small correction coefficient or a small obliteration coefficient is set.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data receiving apparatus for performing adaptive equalization processing for compensation of line distortion in wireless communication such as mobile communication, and more particularly, to follow line fluctuation. And is capable of adaptive processing with a small residual error.

[0002]

2. Description of the Related Art In mobile communication or the like, a line changes due to fading, and a receiving side receives a signal whose waveform is distorted. In a data receiving device, an equalizer is used to remove the waveform distortion of the received signal, and an interference canceller applying the equalizer has been developed to remove an interference signal included in the received signal.

As a data receiving apparatus having such an adaptive equalization processing function, there is known one using a decision feedback equalizer (DFE) or a maximum likelihood sequence estimator (MLSE) as a demodulator.

These equalizers are equipped with a digital filter having a large number of taps and a coefficient controller for controlling the tap coefficient of this filter by an adaptive algorithm, and the impulse response of the line fluctuates with time. Also,
By correcting the tap coefficient according to the algorithm, the fluctuation is tracked and the distortion and interference of the received signal are removed to maintain the reception quality.

As the adaptive algorithm, LMS is used.
(Least Mean Square) and RLS (Recursive Least Squ)
ares) are often used. The LMS is an algorithm in which the sum of squares of the prediction error between the output of the equalizer and the expected signal waveform is used as the evaluation amount, and the tap coefficient is determined so that this evaluation amount is minimized. Further, the RLS has a concept of increasing the tracking characteristic by increasing the weight related to the prediction error with respect to the latest data rather than the past observation data when the estimation target changes with time. A time weight is introduced into the evaluation amount to be performed. In this RLS, the forgetting factor (ρ) is defined, and if ρ = 1, the data up to the present is uniformly weighted, whereas when ρ <1, the estimated value that attaches importance to the weight of the latest data. Becomes

FIG. 4 is an explanatory diagram of a data receiving apparatus having an equalizer for compensating for line distortion. The tap coefficient of this equalizer is controlled using LMS. The transmission side is provided with a modulator 52 that modulates transmission data 51, and the reception device is a line
Received data by equalizing the signal received through 53
The equalizer 54 for demodulating 56 and the tap coefficient of the equalizer 54 are set to LM.
The LMS corrector 55 is controlled according to the S algorithm, and the LMS corrector 55 is provided with a constant correction coefficient α.

The operation of the conventional data receiving apparatus will be described with reference to FIG. The transmission data 51 is modulated by a modulator 52 in accordance with a predetermined law and transmitted as a radio wave. The transmitted radio wave reaches the receiving device through the line 53, but since the line 53 is affected by fading, the waveform is distorted. If demodulation is performed with this distortion added, the error rate characteristic will deteriorate. Therefore, the equalizer 54 compensates for this.

The equalizer 54 compensates for the distortion on the line 53 to obtain high-quality received data 56, but if the characteristics of the line 53 fluctuate over time, it is necessary to follow it. Therefore, the LMS corrector 55 sequentially corrects the tap coefficient of the digital filter of the equalizer 54 using the LMS.

The LMS corrector 55 updates the tap coefficient so that the sum of squares of the difference (error) between the output of the equalizer 54 and the expected signal waveform is minimized. Now, assuming that the number of taps is N, the values of N tap coefficients at time t are collectively N
The input signal input to each tap at time t is expressed as a (t) by a dimensional vector, and X (t) is expressed by an N-dimensional vector.
And the error at that time is represented by ε (t), LM
In S, the tap coefficient will be updated by Equation 1.

A (t + 1) = a (t) −α · (ε (t) *) · X (t) (Equation 1) where (ε (t) *) is a complex conjugate of ε (t) Is represented. Further, α is a correction coefficient 57, which is a fixed value of a real number from 0 to 2. If a large value in this range is taken as α, the tap coefficient update width increases and the followability to line fluctuations increases, but it is easily affected by the noise included in the input data, and the tap coefficient It lacks stability (in other words, the convergence speed is fast, but the residual error in the steady state is large). On the other hand, if α takes a small value,
It is not easily affected by noise and is stable in the convergent state, but the followability to the line fluctuation is poor (the residual error in the steady state is small, but the convergent speed is slow). In consideration of these points, the value of α is selected so that the performance of the equalizer 54 is optimized.

The LMS corrector 55 sequentially updates the tap coefficient of the equalizer 54 according to the equation 1, so that the equalizer 54 follows the fluctuation of the line 53 and receives the received data 56 of good quality. Can be demodulated.

[0012]

However, in the conventional data receiving apparatus, no matter how properly the fixed value of the correction coefficient α is determined, the equalization processing can follow the line fluctuations to the maximum extent, and The residual error cannot be kept to a minimum. This is because the followability to the line and the residual error have a trade-off relationship. Therefore, if the correction coefficient α is small, the error rate is deteriorated because it cannot follow the case where the line fluctuation is fast, and if the correction coefficient α is large, the tap coefficient is changed due to noise and the error rate is deteriorated. .

The same applies to the forgetting factor of the RLS algorithm.

The present invention solves the above-mentioned conventional problems, can appropriately follow the case where the line fluctuates rapidly, and minimizes the residual error even when there is noise. It is an object of the present invention to provide a data receiving device that enables stable demodulation of high quality received data by performing equalization processing that can perform the above.

[0015]

Therefore, according to the present invention, in a data receiving apparatus including a demodulator having an adaptive control mechanism and a tap coefficient modifier for modifying the tap coefficient of the demodulator by an adaptive algorithm, the tap coefficient modification is performed. The parameter setting means is provided for independently controlling the value of the parameter that defines the correction amount of each tap coefficient corrected by the device.

The parameter setting means controls the parameter value to increase the correction amount of the tap coefficient when there is no change with time in the correction direction of the tap coefficient. When a significant change is observed, the parameter value is controlled so as to reduce the correction amount of the tap coefficient.

Further, the parameter setting means includes a multiplier setting means for setting a multiplier based on a change over time in the correction direction of each tap coefficient, a latch for storing the parameter value one time before, and a latch for storing. The value of the parameter is multiplied by the multiplier set by the multiplier setting means, and the value calculated by the multiplying means is used as the parameter at the next time.

Further, when the multiplier setting means shows no change with time in the correction direction of the tap coefficient, it sets a value of 1 or more as the multiplier, and the change with time in the correction direction of the tap coefficient was observed. At this time, the multiplier is set to a value of 1 or less.

Further, the multiplier setting means is configured to set the reciprocal (1 / X) of the multiplier (X) of 1 or more as the multiplier of 1 or less.

The multiplier setting means sets 1.1 as X.

Further, the parameter setting means includes an addend setting means for setting an addend based on a temporal change in the correction direction of each tap coefficient, a latch for storing a parameter value one time before, and a latch. And the adder that adds the addend set by the addend setting means, and the value calculated by the adder is used as the parameter at the next time.

Further, when the addend setting means shows no change with time in the correction direction of the tap coefficient, a positive value is set as the addend, and the change with time in the correction direction of the tap coefficient is seen. When this is done, a negative value is set as the addend.

Further, the addend setting means sets a value (-Y) whose absolute value is equal to that of the positive addend (Y) and whose sign is different as a negative addend.

As the demodulator, DFE or MLS is used.
E is used.

Further, the adaptive algorithm is LMS, and the parameter setting means is configured to control the value of the correction coefficient.

Further, the adaptive algorithm is RLS, and the parameter setting means is configured to control the value of the forgetting factor.

[0027]

Therefore, when the tap coefficient of the adaptive control mechanism is modified by an algorithm such as LMS or RLS,
If the direction of tap coefficient modification does not change between the previous modification and the current modification, it is determined that the tap coefficient is in the state before convergence (state still far from the convergence value), and the tap coefficient is modified. , Set a large correction coefficient or forgetting coefficient.
On the other hand, if the direction of tap coefficient modification differs between the previous modification and the current modification, it is determined that the tap coefficient has been converged, and a small modification coefficient or forgetting coefficient is set when modifying the tap coefficient. To do.

Further, the device having the multiplier setting means and the addend setting means in the parameter setting means updates the correction coefficient to a value larger than the previous correction coefficient for the tap in the state before the convergence, and after the convergence, For the tap in the state, the correction coefficient is updated to a value smaller than the previous correction coefficient.

In this way, the correction coefficient and the forgetting coefficient are independently and momentarily controlled for each tap, so that an algorithm having a convergence suitable for the situation of the line can be obtained even when the line fluctuation is fast and the noise is large. You can
The performance of the equalizer can be improved.

[0030]

(First Embodiment) As shown in FIG. 1, a data receiving apparatus according to a first embodiment is an equalizer 4 which demodulates received data 6 by performing equalization processing on a signal received through a line 3. And an LM that controls the tap coefficient of the equalizer 4 according to the LMS algorithm.
The S correction device 5 and the correction coefficient setting device 8 for controlling the value of the correction coefficient in the LMS are provided.

Each time the tap coefficient of each tap is controlled by the LMS corrector 5, the correction coefficient setter 8 determines the correction direction data 9, that is, whether the value of each tap coefficient is increased or decreased by the control. The data that represents is output,
The correction coefficient setting unit 8 outputs a correction coefficient vector 7 in which a correction coefficient to be used for the next control of each tap coefficient is set based on the correction direction data.

The transmission data 1 modulated by the modulator 2 according to a predetermined law is sent to the data receiving device by radio waves. The transmission data has a waveform that is distorted due to fading on the line 3. The equalizer 4 is
The LMS corrector 5 performs equalization processing on the received signal using tap coefficients that are sequentially corrected to compensate for the distortion received on the line 3.

The LMS corrector 5 successively updates the tap coefficient of each tap in the equalizer 4 by the equation 1 (however, for the correction coefficient α, the value given from the correction coefficient setter 8 is used. The correction coefficient setting unit 8 gives the initial value of the correction coefficient α to the LMS correction unit 5) and outputs the correction direction data 9 to the correction coefficient setting unit 8.

The correction coefficient setter 8 uses the correction direction data 9 to change the correction direction vector (the correction direction of the tap coefficient is different for each tap, and in the case of the same tap or complex number, the real part and the imaginary part are different. Treat it as a vector with twice the number of dimensions). Then, the previous correction direction vector is compared with the current correction direction vector, and components having the same direction are judged to be in a state before convergence, and the correction coefficient for the component of the correction coefficient vector 7 is increased. A value that is set to a value and a component facing in the opposite direction is determined to be in a state after convergence, and the correction coefficient for that component of the correction coefficient vector 7 is set to a small value.

As a result, in the next control of the tap coefficient by the LMS corrector 55, the tap coefficient update range is widened and the small correction coefficient value is set for the tap to which the large correction coefficient value is set. With taps, the width of updating the tap coefficient becomes smaller.

In this way, by controlling the size of the correction coefficient vector 7 independently for each tap and changing it momentarily, it is possible to combine the followability with which the condition of the line 3 is quickly responded and the stable convergence. The equalizer 4 provided can be obtained.

Even when another algorithm such as RLS other than LMS is used to control the tap coefficient, the size of the parameter corresponding to the correction coefficient vector 7 (the forgetting coefficient in RLS) is similarly set. It is possible to realize equalization characteristics having both followability to fluctuations and stable convergence by controlling the taps independently for each tap and changing them momentarily.

The correction direction data 9 may be data including not only the correction direction but also the correction amount.

As described above, in the data receiving apparatus according to the first embodiment, the correction coefficient at the time of tap coefficient control is independently controlled by the correction coefficient vector 7 so as to be optimized at that time. Since the demodulator has, the stable and high quality received data can be demodulated even if the line condition changes.

(Second Embodiment) As shown in FIG. 2, the data receiving apparatus of the second embodiment uses the correction direction data 19 as the means for setting the correction coefficient, and sets the correction coefficient vector 17 of the previous time based on the correction direction data 19. A multiplier setter 20 for setting a multiplier to be multiplied, a latch 18 for storing the previous correction coefficient vector 17, and a latch
A multiplier 21 for multiplying the previous correction coefficient vector 17 stored by 18 by the multiplier set by the multiplier setter 20. Other configurations are the same as those of the first embodiment (FIG. 1).

The equalizer 14 of this data receiving device is an LMS.
The modifier 15 performs equalization processing on the received signal using the tap coefficient that is sequentially modified, and compensates for the distortion received on the line 13. Further, the LMS corrector 15 calculates the tap coefficient of each tap in the equalizer 14 by using the correction coefficient vector output from the multiplier 21.
The correction direction data 19 is output sequentially to the multiplier setter 20 by using Equation 17 to sequentially update it by the equation 1.

The multiplier setter 20 compares the current correction direction vector obtained from the correction direction data 19 with the previous correction direction vector, and regarding components that are in the same direction,
It is determined before convergence and outputs a value of 1 or more as a multiplier. For components facing in the opposite direction, it is determined as after convergence and a value of 1 or less is output as a multiplier.

The multiplier 21 adds the multiplier setter 20 to each component of the correction coefficient vector 17 currently stored in the latch 18.
Multiply each multiplier set in step (3) and output the next correction coefficient vector 17 thus obtained to the LMS corrector 15. The correction coefficient vector 17 is also stored in the latch 18 and then multiplied by the multiplier set by the multiplier setter 20. Then, such an operation is sequentially repeated.

As a result, for the component determined by the multiplier setter 20 before convergence, the multiplier 21 multiplies the value by 1 or more, and a value larger than the previous correction coefficient is output as the correction coefficient. The component determined by the setter 20 to be after convergence is multiplied by a value of 1 or less in the multiplier 21, and a value smaller than the previous correction coefficient is output as the correction coefficient.

In this way, by controlling the convergence speed of the correction coefficient vector 17 by changing the tap speed independently for each tap, it is possible to combine the followability with which the condition of the line 13 is quickly responded with and the stable convergence. The equalizer 14 provided can be obtained.

The value of the multiplier set by the multiplier setter 20 is X when the value of the multiplier of 1 or more is Y and the value of the multiplier of 1 or less is Y:
Set Y = 1 / X. Further, it has been confirmed that when X is set to a value of about 1.1 at this time, an extremely good result is obtained.

As described above, the data receiving apparatus of the second embodiment makes a simple determination and calculation of the correction coefficient at the time of tap coefficient control independently for each component of the tap coefficient based on the previous correction direction data 19. Since the demodulator has the equalizer set by and, it is possible to demodulate the stable received data of high quality even if the condition of the line changes.

(Third Embodiment) As shown in FIG. 3, the data receiving apparatus according to the third embodiment uses the correction direction data 39 as the means for setting the correction coefficient, and sets the correction coefficient vector 37 to the previous correction coefficient vector 37. An adder setter 40 that sets the addend to be added, a latch 38 that stores the previous correction coefficient vector 37, and a latch
An adder 41 for adding the addend set by the addend setter 40 to the previous correction coefficient vector 37 stored by 38 is provided. Other configurations are the same as those of the first embodiment (FIG. 1).

The equalizer 34 of this data receiving device is an LMS.
The modifier 35 performs equalization processing on the received signal using the tap coefficient that is sequentially modified to compensate for the distortion received on the line 33. Further, the LMS corrector 35 calculates the tap coefficient of each tap in the equalizer 34 from the correction coefficient vector output from the adder 41.
Using 37, the correction direction data 39 is sequentially updated by the equation 1, and the correction direction data 39 is output to the addend setter 40.

The addend setter 40 compares the correction direction vector of this time obtained from the correction direction data 49 with the correction direction vector of the previous time, and regarding the components facing in the same direction,
It is determined to be before convergence, and a positive value is output. For components facing in the opposite direction, it is determined to be after convergence, and a negative value is output.

The adder 41 adds the adder setter 40 to each component of the correction coefficient vector 37 stored in the latch 38 at that time.
The values output from the above are added, and the next correction coefficient vector 37 thus obtained is output to the LMS corrector 35. This correction coefficient vector 37 is also stored in the latch 38 and then added to the value output from the addend setter 40. Then, such an operation is sequentially repeated.

As a result, with respect to the component determined by the addend setter 40 before convergence, a positive value is added by the adder 41 and a value larger than the previous correction coefficient is output as a correction coefficient,
Further, with respect to the component determined by the addend setter 40 after convergence, a negative value is added by the adder 41, and a value smaller than the previous correction coefficient is output as the correction coefficient.

In this way, by controlling the convergence speed of the correction coefficient vector 37 independently for each tap and changing it momentarily, it is possible to combine the followability that responds quickly to the conditions of the line 33 and the stable convergence. The equalizer 34 provided can be obtained.

The value set by the addend setter 40 can be set to Y = -X, where X is a positive value and Y is a negative value.

As described above, the data receiving apparatus of the third embodiment makes a simple determination and calculation of the correction coefficient for tap coefficient control independently for each component of the tap coefficient based on the previous correction direction data 19. Since the demodulator has the equalizer set by and, it is possible to demodulate the stable received data of high quality even if the condition of the line changes.

In the present invention, the tap coefficient is controlled by L
It can also be applied to the case of using other algorithms such as RLS other than MS, and for each tap shown in each embodiment with respect to the parameter (forgetting coefficient in RLS) corresponding to the correction coefficient vector of LMS. By applying a means for independently and momentarily controlling to, it is possible to realize equalization characteristics having both followability to fluctuations and stable convergence.

The present invention also relates to a decision feedback equalizer (DF
E) and the maximum likelihood sequence estimator (MLSE) can be applied to a data receiving apparatus that uses a device that performs adaptive equalization processing as a demodulator.

[0058]

As is apparent from the above description of the embodiments, in the data receiving apparatus of the present invention, the noise such as LMS and RLS for correcting the tap coefficient of the equalization processing mechanism is generated even if the line fluctuation is fast. Even if it is large, it can be changed to an algorithm having a convergence property suitable for the state of the line, and the performance of the equalization processing mechanism can be improved.
As a result, it is possible to obtain high-quality received data with no distortion and no residual error.

[Brief description of drawings]

FIG. 1 is an explanatory diagram of a data receiving device according to a first embodiment of the present invention,

FIG. 2 is an explanatory diagram of a data receiving device according to a second embodiment of the present invention,

FIG. 3 is an explanatory diagram of a data receiving device according to a third embodiment of the present invention,

FIG. 4 is a diagram illustrating a conventional data receiving device.

[Explanation of symbols]

 1, 11, 31, 51 Transmission data 2, 12, 32, 52 Modulator 3, 13, 33, 53 Line 4, 14, 34, 54 Equalizer 5, 15, 35, 55 LMS corrector 6, 16, 36,56 Received data 7,17,37 Correction coefficient vector 8 Correction coefficient setter 9,19,39 Correction direction data 18,38 Latch 20 Multiplier setter 21 Multiplier 40 Addend setter 41 Adder 57 Correction coefficient α

Claims (12)

[Claims] 1. A data receiving apparatus, comprising: a demodulator having an adaptive control mechanism; and a tap coefficient modifier for modifying a tap coefficient of the demodulator by an adaptive algorithm, wherein each tap coefficient modified by the tap coefficient modifier is A data receiving device comprising a parameter setting means for independently controlling a value of a parameter defining a correction amount for each tap coefficient. 2. The parameter setting means controls the value of the parameter so as to increase the correction amount of the tap coefficient when the tap coefficient correction direction shows no change with time, and the tap coefficient correction direction is controlled. The data receiving apparatus according to claim 1, wherein the value of the parameter is controlled so as to reduce the correction amount of the tap coefficient when a change with time is observed. 3. The parameter setting means sets a multiplier based on a temporal change in a correction direction of each tap coefficient, a multiplier setting means, a latch for storing a parameter value one time before, and the latch. And multiplying the value of the parameter stored in (4) by the multiplier set by the multiplier setting means, and outputting the value calculated by the multiplying means as the parameter at the next time point. 1. The data receiving device according to 1. 4. The multiplier setting means sets a value of 1 or more as the multiplier when there is no change with time in the correction direction of the tap coefficient, and the change with time in the correction direction of the tap coefficient. The data receiving apparatus according to claim 3, wherein a value equal to or less than 1 is set as the multiplier when is seen. 5. The data receiving apparatus according to claim 4, wherein the multiplier setting means sets an inverse (1 / X) of the multiplier (X) of 1 or more as the multiplier of 1 or less. 6. The multiplier setting means sets the X to 1.
The data receiving apparatus according to claim 5, wherein 1 is set. 7. The addend setting means for setting an addend based on a temporal change in the correction direction of each tap coefficient, and the latch for storing the value of the parameter one time before, And an adder that adds the value of the parameter stored in the latch and the addend set by the addend setting unit, and outputs the value calculated by the adder as a parameter at the next time point. The data receiving apparatus according to claim 1, wherein the data receiving apparatus is provided. 8. The addend setting unit sets a positive value as the addend when there is no change with time in the correction direction of the tap coefficient, and the addition coefficient setting unit sets a positive value in the correction direction of the tap coefficient with time. The data receiving apparatus according to claim 7, wherein when a change is observed, a negative value is set as the addend. 9. The addend setting means sets a value (−Y) having the same absolute value and a different sign as the value (Y) of the positive addend as a negative addend. 8. The data receiving device according to item 8. 10. The DFE (Decision) is used as the demodulator.
Feedback Equalizer) or MLSE (Maximum Likel)
10. The data receiving apparatus according to claim 1, wherein an ihood Sequence Estimator) is used. 11. The adaptive algorithm is LMS (Le
11. The data receiving apparatus according to claim 1, wherein the parameter setting means controls the value of the correction coefficient. 12. The adaptive algorithm is RLS (Re
11. The data receiving apparatus according to claim 1, wherein the parameter setting means controls the value of the forgetting factor.