JPH08242196A - Diversity equalizer - Google Patents

Abstract

PURPOSE: To eliminate propagation estimation errors by an easy method without increasing operation quantity, and to improve an error rate by making the delay time of a matched filter for estimating a propagation state different between a first receiving system and a second receiving system. CONSTITUTION: The configuration of a demodulator 100 is space diversity configuration, and the demodulator is constituted so as to be provided with the first receiving system 21, the second receiving system 22, a frame detecting part 6, an adaptive equalizer 7, and an 8-phase code judging part 8. Then, in the first matched filter part of the first branch of the diversity branch of the adaptive equalizer 7, the matched filter of one symbol interval is used, but in the second matched filter part of the second branch of the diversity branch, the matched filter to deal with 1/2-symbol delay is used, and an equalizing characteristic is averaged for the dispersion of the error rate owing to the state of a space propagation path. Accordingly, in regard to the portion of a code error caused by inter-code interference by a delayed wave, the receiving system in which the delay time of the delayed wave is shorter operates chiefly.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to digital time division multiple access (TDMA) in land mobile communications.
: Time-division multiple access) signal transmission, in particular, a diversity equalization device using an adaptive equalization method in high-speed digital signal transmission that is greatly affected by spatial propagation path distortion such as Rayleigh fading including multipath fading.

[0002]

2. Description of the Related Art The currently standardized Japanese digital automobile / mobile phone system (PDC system: Public Digital Cel)
1) and the digital cordless telephone system (PHS system: Personal Handy phone System), the transmission speed is suppressed to the extent that the effect of spatial propagation distortion can be ignored and no adaptive equalizer is used. However, when the transmission capacity is increased or high-speed data is transmitted in the near future, it is necessary to increase the transmission speed, and the influence of spatial propagation distortion cannot be ignored, so an equalizer to compensate for it is essential. become. Various equalization schemes have been devised, but all of them estimate the distortion of the propagation path and judge the received data. Therefore, how to match this propagation estimation for various multipath fading is the key to determine the performance of the equalizer. Among the equalization methods, the method of performing the maximum likelihood sequence estimation from the replicas of the received signals of the two systems has a large amount of calculation but is superior in the equalization ability to the other methods, and is usually a transversal type having taps with symbol intervals. The adaptive filter of is used as a matched filter. If performance is important, taps of the matched filter may be arranged at fractional intervals to improve the propagation state estimation accuracy.
However, in the case of multi-phase such as 8-phase PSK (Phase Shift Keying) modulation, the number of signal points is large, so that the calculation amount becomes enormous and realization becomes difficult.

[0003]

In high-speed digital mobile communication, the transmission / reception point constantly changes, and the position of the delayed wave also changes accordingly. Under such a propagation condition, an estimation error occurs due to the structural limitation of the matched filter used for the propagation estimation, and the error varies depending on the position of the delayed wave, resulting in a difference in the compensation amount.

The present invention has been made to solve the above problems, and an object of the present invention is to prevent a propagation estimation error and improve an error rate by a simple method without increasing the amount of calculation. It is to provide a diversity equalizer.

[0005]

In order to solve the above-mentioned problems, the diversity equalizer according to claim 1 has a first structure.
A diversity equalization device that receives received signals of a receiving system and a second receiving system, estimates respective propagation states, performs maximum likelihood sequence estimation based on the results, and performs equalization demodulation. The delay time of the matched filter for estimating the situation is configured to be different between the first reception system and the second reception system. The diversity equalizer according to claim 2 is the diversity equalizer according to claim 1, wherein
The first reception system has a first matched filter in which equalization reference data is prepared at 1-symbol intervals, and the second reception system is optimized with a ½ symbol delay wave,
A second matched filter that multiplies the estimated value of the channel distortion by a fixed value during replica generation is configured. The diversity equalization apparatus according to claim 3 is the diversity equalization apparatus according to claim 1, wherein the first reception system prepares reference data for equalization at 1-symbol intervals.
It has a matched filter, and the second receiving system is configured to have a second matched filter which is optimized with a 1/2 symbol delay wave and stores equalization reference data in a read-only memory.

[0006]

According to the first aspect of the present invention having the above-mentioned structure, in regard to the amount of code error caused by the inter-code interference due to the delayed wave, the receiving system whose delay time of the delayed wave is closer operates mainly. . According to the second aspect of the present invention having the above configuration, the number of reference data in the second matched filter of the second reception system is reduced. According to the invention of the third aspect having the above configuration, since the equalization reference data is stored in the read-only memory, the operation speed is improved.

[0007]

An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 shows a basic configuration of a demodulation device including a diversity equalization device according to an embodiment of the present invention. As shown in the figure, the demodulation device 100 has a space diversity structure, and includes a first reception system 21, a second reception system 22, a frame detection unit 6, an adaptive equalization unit 7, and 8. The phase code determination unit 8 is provided.

The first receiving system 21 is composed of the frequency conversion unit 1
It has a quasi-synchronous detection unit 2, an A / D conversion unit 3, a clock recovery unit 4, and an AFC unit 5. The same applies to the second reception system 22.

Next, the operation of the above demodulator 100 will be described. Eight-phase P received through different propagation paths
The SK modulated waves A and B are input to the first reception system 21 and the second reception system 22, respectively. Since the internal configurations of the receiving systems 21 and 22 are the same, the first receiving system 21 will be described below.

First, the received harmonic signal A is converted into an IF signal C by the frequency conversion unit 1 and input to the quasi-coherent detection unit 2. The quasi-synchronous detection unit 2 quadrature-detects the IF signal C in the internal asynchronous local, extracts a baseband signal D consisting of two sequences of I-CH and Q-CH, and AD this.
Input to the conversion unit 3. The output of the A / D conversion unit 3 is input to the clock recovery unit 4 and is punched out at the symbol timing F created by the clock recovery unit 4 to obtain a sampling baseband quantized signal E.

Since this sampling baseband quantized signal E is detected in asynchronous local, a frequency error with the transmission carrier wave is superimposed. If the error is small, the equalizer can absorb this frequency error in principle, but if the error becomes large, the characteristics are significantly deteriorated. Therefore, the AFC unit 5 detects the frequency error in the signal E and corrects this error to the extent that it does not affect the operation of the equalizer. The signal G whose frequency error has been corrected is input to the adaptive equalization unit 7.

FIG. 2 shows the detailed structure of the adaptive equalizer 7. As shown in the figure, the adaptive equalization unit 7 includes a first matched filter unit 23 and a second matched filter unit 24.
, Sequence generator 11, and branch metric synthesizer 15
And an ACS (Add Compare Select) calculator 16 and an equalized data memory 17.

The first matched filter section 23 includes a received data memory section 9, an error extraction section 10, a replica generation section 12, a branch metric calculation section 14, and a CIR.
The update unit 13 is included. The second matched filter unit includes a received data memory unit 9 and an error extraction unit 10A.
, A replica generation unit 12A, a branch metric calculation unit 14, and a CIR update unit 13.

The signal G input to the first matched filter unit 23 is once written in the reception data memory 9, then read at the frame timing detected by the frame detection unit 6, and input to the error extraction unit 10. It Here, the reception data candidate sequence M generated by the sequence generator 11 and the channel impulse response (CIR) N, which is the estimation result of the channel distortion, are used to perform subtraction with the reception replica signal O replica-generated 12. .

The square of the absolute value of the actual received signal obtained as a result of the subtraction and the replica error P generated by the equalizer is called a branch metric (BM). The calculation for obtaining this is the branch metric calculator. Perform at 14. The branch metric calculation result Q is set so that the CI is minimized.
It is used to control the R update unit 13.

The signal Q is also input to the branch metric synthesizing unit 16 and is synthesized through the second matched filter unit 24 of the second receiving system 22 with the data R similarly calculated as the branch metric. The combined branch metric value S is input to the ACS calculator 16 and treated as likelihood data in sequence estimation. As a result, the most probable sequence (maximum likelihood sequence) is estimated from the candidate sequences and the equalization result is obtained. The equalized data is written in the equalized data memory 17, input to the 8-phase code determination section 8 at an appropriate timing, converted into binary data, and becomes demodulated output data.

Here, the configuration of the matched filter of the adaptive equalizer will be described in more detail. FIG.
Shows the configuration of the first matched filter 23, which is a 1-symbol-interval type matched filter. In the figure, CIR (n) D and CIR (n-1) U are the CIR of the direct wave and the delayed wave of the multipath fading, respectively.
Is a complex number that is updated by the received data at time (n-1) and is used when estimating the received signal point at time n. Further, T represents a delay line having a symbol interval.

With such a configuration, the replica generator 12 operates assuming a delayed wave with a delay of one symbol, so that when the delay time of the delayed wave is one symbol or less, the equalization performance is reduced. Will be reduced. Therefore, in the demodulation device 100 of the present invention, the first matched filter unit 23 of the first branch of the diversity branch uses the matched filter of one symbol interval, but the second matched filter of the second branch of the diversity branch is used. The unit 24 uses a matched filter corresponding to 1/2 symbol delay to average the equalization characteristics against the error rate variation due to the situation of the spatial propagation path.

FIG. 6 shows an impulse response of a roll-off filter which is a matched filter corresponding to 1/2 symbol delay. The figure shows the following formula

[Equation 1] Is a discrete time representation of the result calculated with α = 0.5 in T / 2 interval. Strictly speaking, when equalizing the 1/2 symbol delay, side lobes occurring in 3T / 2, 5T / 2, ... Must be taken into consideration in replica generation, but in order to reduce the calculation amount of the equalizer, The method does not handle these, and targets only T / 2. Therefore, although a residual error occurs in the channel estimation by that amount, this method can tolerate this error because it adopts the method of sequence estimation.

FIG. 4 shows an example of the configuration of a matched filter for 1/2 symbol delay. This configuration example is a configuration method when the reference signal points for replica generation are prepared at 1-symbol intervals. As shown in FIG. 6, the response of the roll-off filter at the T / 2 timing is about 0.822 times the T timing (the value when the roll-off rate α = 0.5, and if the value of α changes, The ratio changes.)
Therefore, CI is set to a fixed value when the replica is generated.
Multiplies the value of R. In the case of 1/2 symbol delay, intersymbol interference from the (n-1) time point and intersymbol interference that the signal at the n time point propagates to the n time point at the (n + 1) time point during equalization at the n time point. Both must be considered. From this, as the CIR at the time of equalization, (n-1)
CIR (n-1) U for intersymbol interference from time
However, CIR (n) U is necessary for the intersymbol interference in which the signal at the time point n propagates to the time point n at the time point (n + 1), but since the fluctuation period of fading is sufficiently long with respect to the symbol period, CIR (n-1) ≠ CIR (n) U
Can be The configuration of FIG. 4 has an advantage that the number of reference data is reduced and the circuit scale can be reduced.

In FIG. 5, the reference signal point for replica generation is 1
7 is another configuration example of a matched filter for 1/2 symbol delay when a table of ROM (Read Only Memory) is prepared at an interval of / 2 symbols. The circuit scale is larger than that of Fig. 4, but 1
There is an advantage that the calculation of scaling for the / 2 symbol reference becomes unnecessary.

The present invention is not limited to the above embodiment. The above-described embodiment is an exemplification, and has substantially the same configuration as the technical idea described in the claims of the present invention, and any device having the same function and effect can be realized by the present invention. It is included in the technical scope of the invention.

For example, in the above embodiment, the configuration example in which the target compensable delay time is 1 symbol in the first matched filter unit and the target compensable delay time is 1/2 symbol in the second matched filter unit has been described. The present invention is not limited to this, and the first matched filter unit is configured to target 1-symbol delay and the second matched filter unit is configured to target a-symbol delay (a: a real number satisfying 0 <a <1). You may. Alternatively, the first matched filter unit may target the a symbol delay (a: 0, a real number of 0 <a <1), and the second matched filter unit may target the b symbol delay (a real number of b: 0 <b <1). You may comprise.

[0024]

As described above, according to the first aspect of the present invention, with respect to the amount of code error caused by intersymbol interference due to delayed waves, the reception system whose delay time is closer to the delay time is mainly used. Works. Therefore, there is an advantage that an averaged compensation amount (improvement rate) can be obtained regardless of the difference between the presence and absence of the delayed wave and the time of the delayed wave within the equalization compensation range. According to the second aspect of the present invention having the above configuration, the number of reference data in the second matched filter of the second reception system is reduced. Therefore, there is an advantage that the circuit scale can be reduced. According to the third aspect of the present invention having the above configuration, since the equalization reference data is stored in the read-only memory, there is an advantage that the operation speed is improved.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a demodulation device including a diversity equalization device according to an embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration of an adaptive equalization unit in the demodulation device shown in FIG.

3 is a block diagram showing a configuration of a matched filter with a 1-symbol interval in the adaptive equalization unit shown in FIG.

4 is a block diagram showing a configuration example of a matched filter with a 1/2 symbol interval in the adaptive equalization unit shown in FIG.

5 is a block diagram showing another configuration example of a matched filter with a 1/2 symbol interval in the adaptive equalization unit shown in FIG.

6 is a diagram showing impulse response characteristics of a matched filter with a 1/2 symbol interval in the adaptive equalization unit shown in FIG.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 frequency conversion part 2 quasi-synchronous detection part 3 AD conversion part 4 clock recovery part 5 AFC part 6 frame detection part 7 adaptive equalization part 8 8 phase code determination part 9 received data memory 10, 10A, 10B error extraction part 11, 11B sequence generator 12, 12A, 12B replica generation unit 13 CIR update unit 14 branch metric calculation unit 15 branch metric synthesis unit 16 ACS calculation unit 17 equalized data memory 21 first reception system 22 second reception system 23 first matched filter Unit 24 Second matched filter unit 100 Demodulator A First receiving system receiving high-frequency signal B Second receiving system receiving high-frequency signal C IF signal D baseband signal E sampling baseband quantized signal F recovered clock G frequency-corrected sampling base Band quantized signal H Second receiving system frequency error correction Sampling baseband quantized signal I Frame timing signal J Equalized output signal K Demodulated digital signal L Sampling baseband quantized signal M Known training signal N Part of known training signal O Replica signal P Error between received signal and replica signal Q Branch metric data of the first receiving system R Branch metric data of the second receiving system S Branch metric combined value

Claims (3)

[Claims] 1. Diversity and the like in which received signals of a first receiving system and a second receiving system are received, respective propagation conditions are estimated, and maximum likelihood sequence estimation is performed based on the results to perform equalization demodulation. A diversity equalization device, wherein a delay time of a matched filter for estimating a propagation situation is made different between the first reception system and the second reception system. 2. The first reception system has a first matched filter in which reference data for equalization is prepared at 1-symbol intervals, and the second reception system is optimized with a 1/2 symbol delay wave. The diversity equalization apparatus according to claim 1, further comprising a second matched filter that multiplies an estimated value of propagation path distortion by a fixed value during replica generation. 3. The first reception system has a first matched filter in which equalization reference data is prepared at 1-symbol intervals, and the second reception system is optimized with a ½ symbol delay wave. The diversity equalization apparatus according to claim 1, further comprising a second matched filter in which equalization reference data is stored in a read-only memory.