JP4488868B2 - Waveform equalization method, waveform equalizer, radio apparatus and radio communication system - Google Patents

Description

  The present invention relates to a waveform equalization method for a waveform equalizer that compensates for multipath fading of a radio propagation path that occurs in a radio communication system such as a digital cellular phone. The present invention also relates to a waveform equalizer, a radio apparatus, and a radio communication system to which this waveform equalization method is applied.

  In a digital wireless communication system, when a signal transmission rate is high or when a communication service area covered by one base station is wide, multipath fading that makes it difficult to receive a specific frequency component due to intersymbol interference. (Also called frequency selective fading) occurs. In such a case, it is common to mount a waveform equalizer in the wireless device to compensate for multipath fading.

  Typical examples of waveform equalizers include transversal type, decision feedback type (DFE), and maximum likelihood estimation type (MLSE). Among these, the case where a transversal type and decision feedback type waveform equalizer is used is assumed here.

  In a receiver, a transversal type or decision feedback type waveform equalizer generally receives a predetermined known signal, performs filter training so that the known signal approaches a desired signal, and determines the radio propagation path. The filter operates so that a filter having a frequency inverse characteristic is adaptively configured. In general, the more known signals that are received, the more accurately the multipath state of the radio propagation path can be estimated, and the training performance and equalization performance of the waveform equalizer improves. However, since a known signal is not a data signal that is originally desired to be transmitted, increasing the number of known signals leads to a decrease in transmission efficiency of the data signal. For this reason, in wireless formats, known signals are often shorter than data signals, and the technical issue is how to properly equalize data signals with such short known signals. It has become.

  Here, as a conventional technique, a decision feedback type waveform equalizer and its operation will be briefly described, and an example of an equalization method when a known signal is short in a wireless format will be shown. FIG. 14 is a block diagram showing a basic configuration of a decision feedback type waveform equalizer. The decision feedback type waveform equalizer includes an equalization filter unit 1401, a discriminator 1404, a training signal generator 1405, a switch 1406, an equalization error calculator 1407, a data decoder 1408, and a tap coefficient updater 1409. .

  The equalization filter unit 1401 includes a feedforward filter (FF filter) 1402 and a feedback filter (FB filter) 1403. The FF filter 1402 and the FB filter 1403 are filters each composed of a tapped delay element. The delay time of the delay element is 1/2 symbol time for the FF filter 1402 and 1 symbol time for the FB filter 1403.

  FIG. 15 is a diagram showing a waveform equalization method in a decision feedback type waveform equalizer. The burst signal received by the decision feedback type waveform equalizer includes a known signal and a data signal. The known signal has a pattern that is known in advance on both the transmission side and the reception side, and is used for training the tap coefficient of the decision feedback type waveform equalizer. The data signal is a signal including data to be demodulated. In the decision feedback type waveform equalizer, training is first performed using a known signal. That is, training is performed on the received signals at times T01 to T02 in FIG. First, the operation of this training mode is shown.

  The IQ sample signal is input to the FF filter 1402 of the equalization filter unit 1401 at a sample rate that is ½ of the symbol rate, and feedback data obtained via the switch 1406 is input to the FB filter 1403. In the equalizing filter unit 1401, the data accumulated in the delay element is subjected to a product-sum operation including tap coefficients, and an equalized output signal is obtained. The equalization output signal is input to the equalization error calculator 1407.

  In the training mode, the switch 1406 is connected to the B side, and the equalization output signal is output from the training signal generator 1405 in a desired training pattern to be converged and input to the equalization error calculator 1407. In the equalization error calculator 1407, the equalization output signal from the equalization filter unit 1401 is compared with the training pattern from the training signal generator 1405, and this difference is input to the tap coefficient updater 1409 as an equalization error signal. The

  The tap coefficient updater 1409 updates the tap coefficients of the FF filter 1402 and the FB filter 1403 so that the equalization error of the next equalization output signal is as small as possible. LMS and RLS are well known as update algorithms at this time. On the other hand, the training pattern from the training signal generator 1405 is fed back to the FB filter 1403 via the switch 1406 and input.

  As described above, in the training mode, equalization is performed from the time T01 to the time T02 shown in FIG. 15, and when the tap coefficients of the equalization filter converge at the time T02, the subsequent data Signal equalization can be performed satisfactorily.

  Next, the operation in the tracking mode is shown. In the decision feedback type waveform equalizer, tracking is performed on the next data signal. That is, tracking is performed on the received signals from time T02 to time T03 in FIG. In the tracking mode, the switch 1406 is connected to the A side. The equalization output signal from the equalization filter unit 1401 is input to the discriminator 1404, and symbol determination data is output from the discriminator 1404. The symbol determination data is input to the equalization error calculator 1407 via the switch 1406. In the equalization error calculator 1407, the equalization output signal from the equalization filter unit 1401 is compared with the symbol determination data from the discriminator 1404, and this difference is input to the tap coefficient updater 1409 as an equalization error signal. The The tap coefficient updater 1409 updates the tap coefficients of the FF filter 1402 and the FB filter 1403 so that the equalization error of the next equalization output signal is as small as possible. On the other hand, the symbol determination data from the discriminator 1404 is fed back to the FB filter 1403 via the switch 1406 and input. Further, the symbol determination data is input to the data decoder 1408, and a decoded bit string is output.

  As described above, in the decision feedback type waveform equalizer, the known signal is often shorter than the data signal in the wireless format. In the training mode, when the tap coefficient of the equalization filter unit 1401 does not sufficiently converge, the equalization performance of the data signal in the tracking mode is deteriorated, so that the technique is to converge the tap coefficient with a short known signal. It is a typical problem.

  A waveform equalization method for solving this problem has already been proposed (see Patent Document 1). FIG. 16 is a diagram showing another conventional waveform equalization method. In this waveform equalization method, training is repeatedly performed using the same known signal. That is, the first training is performed from time T01 to time T02, and at time T02, the second training is performed again from time T01 to time T02 with the tap coefficient being left unchanged. This is repeated, and after N times of training, the data signal is tracked from time T02 to time T03.

  Thus, in the past, it seems that the training period has become longer by repeating training using the same known signal, and the tap coefficient converges compared to the case where training is performed only once, and equalization performance Had improved.

JP-A-3-235511

  However, in the above conventional waveform equalization method, the same known signal is used for repeated training, so the convergence improves up to a certain number of repetitions. There was a limit to the degree of convergence of the coefficients. In general, since the waveform equalizer has a large amount of calculation, if the number of repetitions is increased in this way, the amount of calculation per unit time increases, current consumption increases, and real-time processing cannot be performed. .

  The present invention has been made in view of the above circumstances, and in a waveform equalizer that compensates for multipath fading in a radio propagation path of a digital radio communication system, uses only a short known signal in the radio format and increases the amount of calculation. It is an object of the present invention to provide a waveform equalization method, a waveform equalizer, a wireless device, and a wireless communication system that can improve the equalization performance by accurately converging tap coefficients.

The waveform equalization method of the present invention includes a delay circuit with a tap, performs an operation including a tap coefficient on an input signal and outputs an equalized output signal, and taps of the equalization filter unit A waveform equalizer method for a waveform equalizer , comprising: a tap coefficient control means for setting a coefficient; a tap coefficient storage means for storing the tap coefficient; and an equalization performance monitoring means for monitoring the equalization performance , The tap coefficient stored in the tap coefficient storage means is set to an initial value, the known signal received at the first time is used to train the tap coefficient of the equalization filter unit, and the training is performed at the end of the training. A first step of storing a tap coefficient of the result in the tap coefficient storage means, and using the tap coefficient of the training result stored in the tap coefficient storage means, the first time Second time following a second step of performing a tracking of the data signals received, the equalization performance is monitored by the equalization performance monitoring means third of determining whether reaches a predetermined level In the first step, when the equalization performance has reached a predetermined level based on the determination of the equalization performance in the third step, it is stored in the tap coefficient storage means. The previous tap coefficient is set to an initial value, and when the equalization performance has not reached a predetermined level, the previous tap coefficient stored in the tap coefficient storage means is replaced with the previous tap coefficient. stored in the tap coefficient storage unit, a recent tap coefficients earlier that the equalization performance has reached a predetermined level set to an initial value, repeat the first step and the second step One in which to perform.

Thereby, only a short known signal on the radio format is used, and the tap coefficient can be accurately converged and the equalization performance can be improved without increasing the calculation amount. In other words, since training is performed continuously over a plurality of known signals, this is equivalent to training with known signals of an infinite length, and the tap coefficients of the equalization filter section are sufficiently converged. As a result, the equalization performance is improved. In particular, when the propagation path has a static characteristic, the improvement effect is remarkable. Further, since the training is not performed using the same known signal as in the prior art, the convergence value of the tap coefficient becomes more accurate, and good equalization performance can be obtained without increasing the amount of calculation at all. In addition, if the equalization performance does not reach a predetermined level, it is possible to avoid setting the tap coefficient of the training result as an initial value, and even if there is burst-like degradation in the radio propagation path, the equalization performance is degraded. Disappear.

Further, the present invention is the above waveform equalization method, wherein in the third step, the equalization performance is obtained by using a comparison result obtained by comparing the reception level of the received signal with a threshold value by the equalization performance monitoring unit. Shall be judged.
Also, the present invention is the waveform equalization method described above, wherein, in the third step, a comparison result obtained by comparing an equalization error during tracking of the data signal with a threshold by the equalization performance monitoring unit is used. Based on this, the equalization performance is judged.
The present invention is the above waveform equalization method, wherein in the third step, equalization is performed by using a comparison result obtained by comparing the bit error rate of the decoding result with a threshold value by the equalization performance monitoring unit. The performance shall be judged.

The present invention is also the waveform equalization method of any one of the above waveform equalizers, wherein reception level monitoring means for monitoring the reception level of the reception signal is used, and in the first step, the current training is started. As an initial value of the tap coefficient at the time, a value obtained by multiplying the tap coefficient of the previous training result according to the reception level is set.
Thereby, the magnitude of the tap coefficient at the start of training can be adjusted in accordance with the change in the reception level, and the tap coefficient is stabilized even when the reception signal level is not constant, so that equalization performance is improved.

Further, the present invention provides a waveform equalization method for any of the above waveform equalizers, the reception level monitoring means for monitoring the reception level of the reception signal, and the level of the input signal of the waveform equalizer is multiplied. In the first step, the level of the input signal of the waveform equalizer is multiplied according to the reception level.
Thus, by adjusting the level of the input signal of the waveform equalizer according to the change in the reception level, the tap coefficient is stabilized even when the level of the reception signal is not constant, so that the equalization performance is improved.

The present invention is also the waveform equalization method of any of the above waveform equalizers, wherein the reception level monitoring means for monitoring the reception level of the reception signal, and the discriminator for determining the symbol of the equalization output signal In the first step, the reference level of the discriminator and the output level of the training signal are changed according to the reception level.
This makes it possible to change the reference level of the discriminator and the output level of the training signal while leaving the level of the equalized output signal as it is, and even if the received signal level is not constant, the tap coefficient is stable, so equalization Performance is improved.

The waveform equalizer of the present invention has a tapped delay circuit, performs an operation including tap coefficients on an input signal and outputs an equalized output signal, and a tap of the equalizing filter unit A tap coefficient control means for setting a coefficient; a tap coefficient storage means for storing the tap coefficient; and an equalization performance monitoring means for monitoring the equalization performance. It is something to execute.
Thereby, only a short known signal on the radio format is used, and the tap coefficient can be accurately converged and the equalization performance can be improved without increasing the calculation amount.

  The present invention also provides a radio apparatus equipped with the waveform equalizer. The present invention also provides a wireless communication system including the wireless device.

  According to the present invention, in a waveform equalizer that compensates for multipath fading in a radio propagation path of a digital radio communication system, only a short known signal on the radio format is used, and the tap coefficient is accurately determined without increasing the amount of calculation. It is possible to improve the equalization performance by converging.

(First embodiment)
FIG. 1 is a block diagram showing a basic configuration of a waveform equalizer according to the first embodiment of the present invention. In the present embodiment, a case where the present invention is applied to a decision feedback type waveform equalizer is shown. Is not limited.

  The decision feedback type waveform equalizer of this embodiment includes an equalization filter unit 101, a discriminator 104, a training signal generator 105, a switch 106, an equalization error calculator 107, a data decoder 108, and a tap coefficient control unit 109. The tap coefficient storage unit 110 and the frequency offset amount estimation unit 111 are configured.

  The equalization filter unit 101 includes a feed forward filter (FF filter) 102, a feedback filter (FB filter) 103, and the like. The FF filter 102 and the FB filter 103 are filters each composed of a tapped delay element. The delay time of the delay element is 1/2 symbol time for the FF filter 102 and 1 symbol time for the FB filter 103.

  As shown in FIG. 15, the burst signal received by the decision feedback type waveform equalizer includes a known signal and a data signal. The known signal has a pattern that is known in advance on both the transmission side and the reception side, and is used for training the tap coefficient of the decision feedback type waveform equalizer. The data signal is a signal including data to be demodulated. In the decision feedback type waveform equalizer, training is first performed using a known signal. That is, training is performed on the received signals at times T01 to T02 in FIG. First, the operation of this training mode is shown.

  The IQ sample signal is input to the FF filter 102 of the equalization filter unit 101 at a sample rate that is ½ the symbol rate, and the feedback data obtained through the switch 106 is input to the FB filter 103. In the equalizing filter unit 101, the data accumulated in the delay element is subjected to a product-sum operation including tap coefficients, and an equalized output signal is obtained. The equalization output signal is input to the equalization error calculator 107.

  In the training mode, the switch 106 is connected to the B side, and the equalization output signal is output from the training signal generator 105 in a desired training pattern to be converged and input to the equalization error calculator 107. In the equalization error calculator 107, the equalization output signal from the equalization filter unit 101 and the training pattern from the training signal generator 105 are compared, and this difference is input to the tap coefficient control unit 109 as an equalization error signal. The The training pattern from the training signal generator 105 is fed back to the FB filter 103 via the switch 106 and input.

  As described above, in the training mode, equalization is performed from the time T01 to the time T02 shown in FIG. 15, and when the tap coefficients of the equalization filter converge at the time T02, the subsequent data Signal equalization can be performed satisfactorily.

  Next, the operation in the tracking mode is shown. In the decision feedback type waveform equalizer, tracking is performed on the next data signal. That is, tracking is performed on the received signals from time T02 to time T03 in FIG. In the tracking mode, the switch 106 is connected to the A side. The equalization output signal from the equalization filter unit 101 is input to the discriminator 104, and symbol determination data is output from the discriminator 104. The symbol determination data is input to the equalization error calculator 107 via the switch 106. In the equalization error calculator 107, the equalization output signal from the equalization filter unit 101 is compared with the symbol determination data from the discriminator 104, and this difference is input to the tap coefficient control unit 109 as an equalization error signal. The The symbol determination data from the discriminator 104 is fed back to the FB filter 103 via the switch 106 and input. Further, the symbol determination data is input to the data decoder 108, and a decoded bit string is output.

Below, the structure of the characteristic part of this embodiment is demonstrated in detail.
The tap coefficient control unit 109 has a function of a tap coefficient control unit, and performs control related to tap coefficient update of the FF filter 102 and the FB filter 103 of the equalization filter unit 101. Specifically, the tap coefficient control unit 109 performs the FF filter 102 so that the next equalization error becomes as small as possible based on the equalization error information input from the equalization error calculator 107 during the training operation and the tracking operation. And each tap coefficient of the FB filter 103 is updated. Further, at the training start time based on the known signal, the tap coefficient at the end of the previous training is read from the tap coefficient storage unit 110, the tap coefficient is corrected as necessary, and set in the FF filter 102 and the FB filter 103.

  The tap coefficient storage unit 110 has a function of a tap coefficient storage unit. When one training is completed, the tap coefficient at that time is read from the FF filter 102 and the FB filter 103 and stored. The frequency offset amount estimation unit 111 has a function of a frequency offset amount estimation unit. The frequency offset amount estimation unit 111 estimates the frequency offset amount between the transmission side and the reception side radio devices as needed, and the estimated frequency offset amount is used as the tap coefficient control unit 109. To enter. The estimation of the frequency offset amount can be performed using the received IQ sample signal, the equalized output signal from the equalizing filter unit 101, and the like.

  FIG. 2 is a diagram showing a radio format of a received signal and a waveform equalization method in a decision feedback type waveform equalizer. A waveform equalization method in the case of continuously receiving slots composed of known signals and data signals will be described. In FIG. 2, the radio format is continuous, but there is no problem even if it is a burst or a data signal portion is an empty line as long as a known signal is periodically transmitted.

  First, a waveform equalization method when the frequency offset amount estimation unit 111 is not used will be described. In step S201, the waveform equalizer uses the known signal received at time Ta to perform training from time T01 to time T02. At time T02 at the end of training, the tap coefficient storage unit 110 stores the tap coefficients of the FF filter 102 and the FB filter 103.

  In step S202, the waveform equalizer performs tracking from time T02 to time T03. As an initial value of the tap coefficient at the start of tracking (time T02), the training result using the known signal at time Ta is used as it is.

  In step S <b> 203, the tap coefficient control unit 109 reads the tap coefficient at time T <b> 02 stored in the tap coefficient storage unit 110 and sets it in the FF filter 102 and the FB filter 103. After that, the waveform equalizer performs training from time T03 to time T04 using the known signal received at time Tb. At time T04 at the end of the training, the tap coefficient storage unit 110 stores the tap coefficients of the FF filter 102 and the FB filter 103.

  In step S204, the waveform equalizer performs tracking from time T04 to time T05. As an initial value of the tap coefficient at the start of tracking (time T04), the training result based on the known signal at time Tb is used as it is.

  In step S <b> 205, the tap coefficient control unit 109 reads the tap coefficient at time T <b> 04 stored in the tap coefficient storage unit 110 and sets it in the FF filter 102 and the FB filter 103. Then, the waveform equalizer performs training from time T05 to time T06 using the known signal received at time Tc. At time T06 at the end of training, the tap coefficient storage unit 110 stores the tap coefficients of the FF filter 102 and the Fb filter 103.

  In step S206, the waveform equalizer performs tracking from time T06 to time T07. As an initial value of the tap coefficient at the start of tracking (time T06), the training result based on the known signal at time Tc is used as it is.

  As described above, the training result based on the known signal is stored in the tap coefficient storage unit 110, and the operation of setting the initial value of the tap coefficient at the start of the next training based on the known signal is repeated. Equivalent to training.

  Therefore, according to the waveform equalization method of the present embodiment, since the tap coefficients of the equalization filter unit 101 are sufficiently converged, the data signal equalization performance is improved. In particular, when the propagation path has a static characteristic, the improvement effect is remarkable. Further, since the training is not performed using the same known signal as in the prior art, the convergence value of the tap coefficient becomes more accurate. Also, good equalization performance can be obtained without increasing the amount of calculation.

  Next, a waveform equalization method when the frequency offset amount estimation unit 111 is used will be described. FIG. 3 is a flowchart showing an operation processing procedure of the tap coefficient control unit 109. In general, there is a frequency offset between radio devices on the transmission side and the reception side. Therefore, for example, at time T03 in FIG. 2, the tap coefficient control unit 109 does not set the tap coefficients read from the tap coefficient storage unit 110 as tap coefficients of the FF filter 102 and the FB filter 103 as they are. It is better to set after performing processing considering the frequency offset amount. By doing so, the phase relationship between the received signal and the tap coefficient can be maintained. The other operations are exactly the same as when the frequency offset amount estimation unit 111 is not used.

  In the process at time T03 in FIG. 2, the frequency offset amount estimated by the frequency offset amount estimation unit 111 is foff [Hz], the symbol time of the signal is Ts [sec], and the data signal from time T02 to time T03 is processed. Assume that the number of symbols is M [symbol]. FIG. 3 shows processing at time T03.

  When the tracking of the data signal from time T02 to time T03 in FIG. 2 ends, first, the tap coefficient control unit 109 reads the tap coefficient from the tap coefficient storage unit 110 (step S301). Then, the tap coefficient control unit 109 rotates the phase by an angle represented by Expression (1) for each read tap coefficient (step S302).

2π × foff × Ts × M [rad] (1)
The tap coefficient control unit 109 sets each phase-tapped tap coefficient to the tap coefficient of the FF filter 102 and the FB filter 103, and sets it as an initial value at the start of the next training (step S303). Thereafter, the waveform equalizer starts training with the received known signal between time T03 and time T04 (time Tb). Although only the process at time T03 is shown here, the same process is repeated at time T05 and time T07.

  Thus, according to the waveform equalization method of the first embodiment, the tap coefficient control unit 109 is stored in the tap coefficient storage unit 110 based on the frequency offset amount estimated by the frequency offset amount estimation unit 111. The tap coefficient is rotated in phase to be the initial value of the tap coefficient at the start of the next training. Thereby, the phase relationship between the received signal and the tap coefficient can be maintained, and compared with the case where the frequency offset estimation unit 111 is not used, the tap coefficient is further converged, and the equalization performance can be further improved.

(Second Embodiment)
FIG. 4 is a block diagram showing a basic configuration of a waveform equalizer according to the second embodiment of the present invention. The waveform equalizer of the second embodiment has a configuration in which an equalization performance monitoring unit having a function of an equalization performance monitoring unit is added to the waveform equalizer of the first embodiment. Specifically, an equalization error calculator 107, an RSSI calculator 415, an equalization performance determination unit 416, and a BER measurement unit 417 are provided as the equalization performance monitoring unit. The same components as those in the first embodiment are denoted by the same reference numerals, and the description thereof is omitted.

  The tap coefficient control unit 109 performs control related to the tap coefficient update of the FF filter 102 and the FB filter 103 of the equalization filter unit 101. Specifically, the tap coefficient control unit 109 performs the FF filter 102 so that the next equalization error becomes as small as possible based on the equalization error information input from the equalization error calculator 107 during the training operation and the tracking operation. And each tap coefficient of the FB filter 103 is updated. Further, at the training start time based on the known signal, a necessary tap coefficient is read from the tap coefficient storage unit 110 in accordance with the output information of the equalization performance determination unit 416, the tap coefficient is corrected as necessary, and the FF filter 102 is corrected. And FB filter 103.

  The input IF signal in the intermediate frequency band is quadrature demodulated by the quadrature demodulator 412, digitally sampled by the A / D converter 413, baseband band-limited by the root Nyquist filter 414, and this output IQ sample The signal is input to the equalization filter unit 101.

  On the other hand, the IF signal is input to the RSSI calculator 415, and the received signal level (RSSI: Received Signal Strength Indicator) is sequentially calculated in the RSSI calculator, and the calculated received level is input to the equalization performance determiner 416. . The equalization error signal that is the output of the equalization error calculator 107 is input to the equalization performance determination unit 416.

  The decoded bit string that is the output of the data decoder 108 is input to the BER measuring device 417. The BER measuring device 417 sequentially calculates the bit error rate (BER) of the received signal, and the BER information is equal. Is input to the conversion performance determination unit 416. The equalization performance determination unit 416 determines whether the equalization performance is good or not at a certain time. When the equalization performance determination unit 416 determines whether the equalization performance is good or bad based on the reception level of the received signal, the equalization performance is good when the reception level input from the RSSI calculator 415 is larger than a certain threshold, and when the reception level is small. It is determined that the equalization performance is poor, and the determination result is output to the tap coefficient control unit 109 as equalization performance information.

  In addition, when the equalization performance is judged based on the equalization error during tracking, for example, a total value of the equalization error signal level is calculated using the equalization error signal input from the equalization error calculator 107. When the total value (error level) is smaller than the threshold, it is determined that the equalization performance is good, and when it is large, the equalization performance is bad, and this determination result is output to the tap coefficient control unit 109 as equalization performance information. . Also, when determining whether the equalization performance is good or not based on the bit error rate, the BER information input from the BER measuring device 417 is used. When this value is smaller than the threshold value, the equalization performance is good, and when the value is large, etc. It is determined that the equalization performance is poor, and the determination result is output to the tap coefficient control unit 109 as equalization performance information. Then, the tap coefficient control unit 109 controls the tap coefficients of the FF filter 102 and the FB filter 103 based on the equalization performance information input from the equalization performance determination unit 416.

  FIG. 5 is a diagram showing a radio format of a received signal and a waveform equalization method. FIG. 5 also shows equalization performance information obtained from the equalization performance determiner 416. First, in step S501, the waveform equalizer performs training from time T01 to time T02 using a known signal received at time Ta. During training, the equalization performance determination unit 416 outputs equalization performance information to the tap coefficient control unit 109 as necessary. At time T02 at the end of training, the tap coefficient storage unit 110 stores the tap coefficients of the FF filter 102 and the FB filter 103.

  In step S502, the waveform equalizer performs tracking from time T02 to time T03. As the initial value of the tap coefficient at the start of tracking (time T02), the training result of the known signal at time Ta is used as it is. During tracking, the equalization performance determiner 416 outputs equalization performance information to the tap coefficient control unit 109 as necessary.

  In step S503, the tap coefficient control unit 109 refers to equalization performance information. Since the equalization performance information is “good” from time T01 to time T03, the tap coefficient control unit 109 reads the tap coefficient at time T02 stored from the tap coefficient storage unit 110, and the FF filter 102 and FB Set in the filter 103. After that, the waveform equalizer performs training from time T03 to time T04 using the known signal received at time Tb. During training, the equalization performance determiner 416 outputs equalization performance information to the tap coefficient control unit 109 as necessary. At time T04 at the end of the training, the tap coefficient storage unit 110 stores the tap coefficients of the FF filter 102 and the FB filter 103.

  In step S504, the waveform equalizer performs tracking from time T04 to time T05. As the initial value of the tap coefficient at the start of tracking (time T04), the training result based on the known signal at time Tb is used as it is. During tracking, the equalization performance determiner 416 outputs equalization performance information to the tap coefficient control unit 109 as necessary.

  In step S505, the tap coefficient control unit 109 refers to equalization performance information. Since the equalization performance information is “bad” from time T03 to time T05, the tap coefficient control unit 109 does not read the tap coefficient stored at time T04 from the tap coefficient storage unit 110. Instead, the tap coefficient control unit 109 reads the tap coefficient at the end of training of the slot whose equalization performance information was “good” recently before that. That is, the tap coefficient control unit 109 reads the tap coefficient at time T02 stored in the tap coefficient storage unit 110, and sets it in the FF filter 102 and the FB filter 103. Then, the waveform equalizer performs training from time T05 to time T06 using the known signal received at time Tc. During training, the equalization performance determination unit 416 outputs equalization performance information to the tap coefficient control unit 109 as necessary. At time T06 at the end of training, the tap coefficient storage unit 110 stores the tap coefficients of the FF filter 102 and the FB filter 103.

  In step S506, the waveform equalizer performs tracking from time T06 to time T07. As the initial value of the tap coefficient at the start of tracking (time T06), the training result based on the known signal at time Tc is used as it is. During tracking, the equalization performance determination unit 416 outputs equalization performance information to the tap coefficient control unit 109 as necessary.

  Here, the training length is equivalently increased by continuously using known signals that are transmitted periodically, but if there is bursty reception degradation, the known signal received at the time of degradation is used. Training should not be performed. By using the degraded known signal for training, the tap coefficients that have converged until now diverge, and this result also affects training from the next time on, leading to degradation of equalization performance.

  Therefore, in the second embodiment, the initial value of the tap coefficient at the start of training is controlled based on the equalization performance information. That is, at the start of a new training with a known signal, a tap coefficient which is a training result of a slot whose equalization performance information was “good” recently before that is set. As a result, it is possible to avoid a set of training results for slots where the equalization performance information is “bad”, and even when there is a burst-like degradation in the radio propagation path, the equalization performance is not degraded. Compared to the first embodiment, better equalization performance can be obtained. In the second embodiment, the case where the RSSI calculator 415, the equalization error calculator 107, and the BER measuring device 417 are exclusively used as the equalization performance monitoring unit has been described, but these may be used in combination. Good.

(Third embodiment)
FIG. 6 is a block diagram showing a basic configuration of a waveform equalizer according to the third embodiment of the present invention. Compared with the waveform equalizer of the second embodiment, the waveform equalizer of the third embodiment deletes the equalization performance determination unit 416 and the BER measurement unit 417, and instead of the multiplication having the function of a multiplication unit. A device 616 is added. The same components as those in the first and second embodiments are denoted by the same reference numerals, and the description thereof is omitted. In the third embodiment, the RSSI calculator 415 has a function of reception level monitoring means. Further, the multiplication factor of the multiplier 616 is a value of 1 unless otherwise specified.

  FIG. 7 is a diagram showing a radio format of received signals and a waveform equalization method. FIG. 7 also shows the received signal level (RSSI) calculated by the RSSI calculator 415. The operation of FIG. 7 is almost the same as that of FIG. The only different operations are the operations performed by the tap coefficient control unit 109 before starting each training at time T03 and time T05. Here, three processing methods at time T03 are shown.

  First, a first waveform equalization method is shown. FIG. 8 is a flowchart showing an operation processing procedure of the tap coefficient control unit 109 in the first waveform equalization method. At this time, the multiplication factor of the multiplier 616 is fixed to the value 1. When the tracking of the data signal from time T02 to time T03 is completed, first, the tap coefficient control unit 109 reads the tap coefficient from the tap coefficient storage unit 110 (step S801). Assume that the RSSI value of the training known signal (received signal at time Ta in FIG. 7) used for calculating the tap coefficient is RSSI = Pa (true number).

  The tap coefficient control unit 109 acquires the RSSI value of the known signal (received signal at time Tb in FIG. 7) used for the current training from the RSSI calculator 415 (step S802). It is assumed that the acquired RSSI value is RSSI = Pb (true number).

  The tap coefficient control unit 109 determines the multiplication factor of the tap coefficient to be set based on the obtained RSSI value. In order to make the level of the equalized output signal of the equalization filter unit 101 constant, the filter tap coefficient is decreased as the RSSI value of the received signal is increased, and the filter tap coefficient is increased as the RSSI value of the received signal is decreased. That's fine. Considering that the tap coefficient has an I component and a Q component, the multiplication factor is (Pa / Pb) 1/2 power. Therefore, the tap coefficient control unit 109 multiplies each tap coefficient (I component and Q component) read out from the tap coefficient storage unit 110 by the 1/2 power of (Pa / Pb) (step S803).

  The tap coefficient control unit 109 sets the multiplied tap coefficients to the tap coefficients of the FF filter 102 and the FB filter 103, and sets them as initial values at the start of the next training (step S804). Thereafter, the waveform equalizer starts training with the known signal received from time T03 to time T04 (time Tb). Thereby, the level of the equalized output signal can be kept constant, and when the received signal level is not constant, the tap coefficient is stabilized, so that the equalization performance is improved.

  Next, a second waveform equalization method is shown. FIG. 9 is a flowchart showing an operation processing procedure of the tap coefficient control unit 109 in the second waveform equalization method. When the tracking of the data signal from time T02 to time T03 is completed, the tap coefficient control unit 109 reads the tap coefficient from the tap coefficient storage unit 110 (step S901). Assume that the RSSI value of the training known signal (received signal at time Ta in FIG. 7) used for calculating the tap coefficient is RSSI = Pa (true number).

  The tap coefficient control unit 109 acquires an RSSI value of a known signal (received signal at time Tb in FIG. 7) used for the current training from the RSSI calculator 415 (step S902). Assume that RSSI = Pb (true number).

  The tap coefficient control unit 109 adjusts the reception input level of the IQ sample sequence to the FF filter 102 based on the obtained RSSI value. That is, the multiplication factor of the multiplier 616 is determined. In order to make the level of the equalization output signal of the equalization filter unit 101 constant, the reception input level of the FF filter decreases as the RSSI value of the reception signal increases, and the reception input of the FF filter decreases as the RSSI value of the reception signal decreases. Just increase the level. Considering that the received signal has an I component and a Q component, the multiplication rate is (Pa / Pb) 1/2 power. Therefore, the tap coefficient control unit 109 sets the multiplication factor of the multiplier 616 to (1/2) times (Pa / Pb) (step S903).

  The tap coefficient control unit 109 sets the tap coefficient read from the tap coefficient storage unit 110 to the tap coefficient of the FF filter 102 and the FB filter 103, and sets it as an initial value at the start of the next training (step S904). Thereafter, the waveform equalizer starts training with the known signal received from time T03 to time T04 (time Tb). Thereby, the level of the equalized output signal can be kept constant, and even when the received signal level is not constant, the tap coefficient is stabilized, so that the equalization performance is improved.

  Next, a third waveform equalization method will be described. FIG. 10 is a flowchart showing an operation processing procedure of the tap coefficient control unit 109 in the third waveform equalization method. At this time, the multiplication factor of the multiplier 616 is fixed to the value 1. When the tracking of the data signal from time T02 to time T03 ends, first, the tap coefficient control unit 109 reads the tap coefficient from the tap coefficient storage unit 110 (step S1001). The RSSI value of the training known signal (received signal at time Ta in FIG. 7) used for calculating the tap coefficient is acquired from the RSSI calculator 415 (step S1002). Assume that RSSI = Pb (true number).

  The tap coefficient control unit 109 changes the reference level of the discriminator 104 and the output level of the training signal generator 105 based on the obtained RSSI value. The level of the equalized output signal of the equalization filter unit 101 increases as the RSSI value of the received signal increases, and decreases as the RSSI value of the received signal decreases. Therefore, the tap coefficient control unit 109 considers that the equalized output signal includes an I component and a Q component, and sets the reference level of the discriminator 104 and the output level of the training signal generator 105 to (Pa / Pb) is set to 1/2 power (step S1003). As an example, the case of 16QAM modulation is shown. FIG. 11 is a diagram showing changes in the reference level of the discriminator 104 and the output level of the training signal generator 105 in 16QAM modulation.

  The tap coefficient control unit 109 sets the tap coefficient read from the tap coefficient storage unit 110 to the tap coefficient of the FF filter 102 and the FB filter 103, and sets it as an initial value at the start of the next training (step S1004). Thereafter, the waveform equalizer starts training with the known signal received from time T03 to time T04 (time Tb). Thereby, even when the received signal level is not constant, the tap coefficient is stabilized, so that equalization performance is improved. The above is the three processing methods at time T03.

  Here, known signals that are separated in time are used for training of the waveform equalizer. However, if the received signal level varies, the tap coefficient level to be converged differs from slot to slot, so the tap coefficient is sufficient. May not converge. In the first embodiment and the second embodiment, the equalization performance is improved and the effect is great depending on the condition of static characteristics, but when the reception level varies, the improvement of the equalization performance is small. Become.

  Therefore, in the third embodiment, in order to solve such a problem, the RSSI calculator 415 is provided as reception level monitoring means for monitoring the reception level. Thereby, the tap coefficient control unit 109 can adjust the magnitude of the tap coefficient at the start of training or the reception input level to the FF filter 102 according to the change in the reception level. Even in training with known signals that are separated in time, the level of the equalized output signal can be kept constant, and even when the received signal level is not constant, the tap coefficient is stabilized, so that equalization performance is improved. Alternatively, by changing the reference level of the discriminator 104 and the output level of the training signal generator 105 while leaving the level of the equalized output signal as it is, the tap coefficient is stabilized even when the received signal level is not constant. Performance is improved.

  Therefore, the waveform equalization method of the third embodiment is not limited to the case where the radio propagation path has a static characteristic, but also an excellent waveform equalization that provides good equalization characteristics even for dynamic characteristics in which fading exists. Is the method. In the third embodiment, the case where reception level monitoring means is provided for the first embodiment is shown. However, the same thing can be said by providing reception level monitoring means for the second embodiment. If there is an effect and there is bursty reception degradation, good equalization performance can be obtained.

(Fourth embodiment)
FIG. 12 is a block diagram showing a configuration of a wireless device according to the fourth embodiment of the present invention. This radio apparatus includes a waveform shaper 1201, a root Nyquist filter 1202, a D / A converter 1203, a quadrature modulator 1204, a mixer 1205, an amplifier 1206, a duplexer 1207, an antenna 1208, a mixer 1209, a quadrature demodulator 1210, an A / A A D converter 1211, a root Nyquist filter 1212, a synchronization unit 1213, and a waveform equalizer 1214 are included.

  Here, the transmission system corresponds to the waveform generator 1201 to the antenna 1208. In the transmission system, when the transmission bit string U is input to the waveform generator 1201, control data is added to generate modulation data. After the baseband band is limited by the root Nyquist filter 1202, the modulated data is input to the D / A converter 1203 and converted from a digital signal to an analog signal. Further, in the quadrature modulator 1204, the converted analog signal is quadrature modulated. The signal quadrature modulated by the quadrature modulator 1204 is up-converted by a mixer 1205 that converts it to a predetermined transmission frequency, and the up-converted signal is amplified by an amplifier 1206. When the amplified signal is input to the duplexer 1207 that branches the transmission / reception signal, the amplified signal is transmitted from the transmission / reception shared antenna 1208 as a radio signal.

  On the other hand, the reception system corresponds to the duplexer 1207 to the waveform equalizer 1214. In the reception system, the received signal received through the transmission / reception shared antenna 1208 and the duplexer 1207 is down-converted by the mixer 1209 and further orthogonally demodulated by the orthogonal demodulator 1210. The quadrature demodulated analog signal is converted to a digital signal by the A / D converter 1211, band-limited by the root Nyquist filter 1212, and then input to the synchronization unit 1213 and the waveform equalizer 1214. The waveform equalizer 1214 has the same configuration as the waveform equalizer in any of the first to third embodiments, and the waveform equalizer 1214 includes any of the first to third embodiments. The waveform equalization method is applied. The synchronization unit 1213 acquires synchronization using a digital signal, and a symbol timing signal is input to the waveform equalizer 1214. The waveform equalizer 1214 performs waveform equalization using the digital signal and the symbol timing signal described above, and outputs a received bit string D.

  According to the radio apparatus of this embodiment, since the waveform equalization method of any of the first to third embodiments is applied to the waveform equalizer 1214, multipath fading exists in the radio propagation path. In this case, the equalization performance described above is improved, and a device having good reception performance is obtained. Further, the waveform equalization method applied to the waveform equalizer 1214 does not repeatedly perform calculations as in the prior art, so that a device with a small calculation amount and low current consumption can be obtained.

  FIG. 13 is a diagram illustrating a configuration of a wireless communication system including the wireless device according to the present embodiment. This radio communication system is composed of mobile station or fixed station radio apparatuses RS1, RS2,... RSm and a base station radio apparatus BS. Among the radio devices RS1, RS2,... RSm of mobile stations or fixed stations, at least one is composed of the radio devices shown in FIG. Therefore, a system having good reception performance can be constructed as the entire system. In addition, a system with reduced current consumption can be constructed.

  The configuration of the present embodiment described above can be applied to an apparatus equipped with a waveform equalizer that equalizes multipath fading in a digital radio communication system that periodically transmits known signals (synchronization symbols and pilot symbols). .

  The present invention has the effect of using only a short known signal on a wireless format and improving the equalization performance by accurately converging tap coefficients without increasing the amount of calculation. The present invention is useful for a waveform equalizer that compensates for multipath fading of a radio propagation path that occurs in a system, a waveform equalizer to which this waveform equalization method is applied, a wireless device, a wireless communication system, and the like.

The block diagram which shows the basic composition of the waveform equalizer in the 1st Embodiment of this invention. The figure which shows the radio | wireless format and waveform equalization method of the received signal in the waveform equalizer of 1st Embodiment The flowchart which shows the operation | movement process procedure of the tap coefficient control part in 1st Embodiment. The block diagram which shows the basic composition of the waveform equalizer in the 2nd Embodiment of this invention. The figure which shows the radio | wireless format and waveform equalization method of the received signal in the waveform equalizer of 2nd Embodiment. The block diagram which shows the basic composition of the waveform equalizer in the 3rd Embodiment of this invention. The figure which shows the radio | wireless format and waveform equalization method of the received signal in the waveform equalizer of 3rd Embodiment. The flowchart which shows the operation processing procedure of the tap coefficient control part in a 1st waveform equalization method. The flowchart which shows the operation processing procedure of the tap coefficient control part in the 2nd waveform equalization method. The flowchart which shows the operation processing procedure of the tap coefficient control part in the 3rd waveform equalization method. The figure which shows the change of the reference level of a discriminator and the output level of a training signal generator in 16QAM modulation. The block diagram which shows the structure of the radio | wireless apparatus in the 4th Embodiment of this invention. The figure which shows the structure of a radio | wireless communications system provided with the radio | wireless apparatus of this embodiment. Block diagram showing the basic configuration of a decision feedback type waveform equalizer The figure which shows the waveform equalization method in the decision feedback type waveform equalizer The figure which shows the other conventional waveform equalization method

Explanation of symbols

101 Equalization filter unit 102 Feed forward filter (FF filter)
103 Feedback filter (FB filter)
DESCRIPTION OF SYMBOLS 104 Discriminator 105 Training signal generator 106 Switch 107 Equalization error calculator 108 Data decoder 109 Tap coefficient control part 110 Tap coefficient memory | storage part 111 Frequency offset amount estimation part 415 RSSI calculator 416 Equalization performance determination part 417 BER calculator

Claims (10)

An equalizing filter unit having a tapped delay circuit, performing an operation including a tap coefficient on an input signal and outputting an equalized output signal, and a tap coefficient control means for setting a tap coefficient of the equalizing filter unit; A waveform equalizer method of a waveform equalizer comprising a tap coefficient storage means for storing the tap coefficient and an equalization performance monitoring means for monitoring the equalization performance ,
The tap coefficient stored in the tap coefficient storage means is set to an initial value, the known signal received at the first time is used to train the tap coefficient of the equalization filter unit, and the training is performed at the end of the training. A first step of storing the resulting tap coefficient in the tap coefficient storage means;
A second step of tracking a received data signal for a second time following the first time, using the tap coefficient of the training result stored in the tap coefficient storage means ;
A third step of determining whether the equalization performance monitored by the equalization performance monitoring means has reached a predetermined level ;
In the first step, when the equalization performance has reached a predetermined level based on the determination of the equalization performance in the third step, the previous tap coefficient stored in the tap coefficient storage means is determined. If the equalization performance does not reach a predetermined level, the initial value is stored in the tap coefficient storage means before that instead of the previous tap coefficient stored in the tap coefficient storage means. The previous tap coefficient before the equalization performance has reached a predetermined level is set to an initial value,
A waveform equalization method for repeatedly performing the first step and the second step. 2. The waveform equalization method according to claim 1 , wherein in the third step, the equalization performance is judged by using a comparison result obtained by comparing the reception level of the received signal with a threshold value by the equalization performance monitoring means. Waveform equalization method. The waveform equalization method according to claim 1 , wherein in the third step, by using a comparison result obtained by comparing an equalization error during tracking of the data signal with a threshold by the equalization performance monitoring unit, etc. Waveform equalization method to judge the equalization performance. 2. The waveform equalization method according to claim 1 , wherein in the third step, the equalization performance is determined by using a comparison result obtained by comparing the bit error rate of the decoding result with a threshold by the equalization performance monitoring unit. Waveform equalization method. A waveform equalizer method for a waveform equalizer according to claim 1 ,
Using the reception level monitoring means for monitoring the reception level of the reception signal, in the first step, the tap coefficient of the previous training result is multiplied according to the reception level as the initial value of the tap coefficient at the start of the current training. Waveform equalization method to set what is done. A waveform equalizer method for a waveform equalizer according to claim 1 ,
The reception level monitoring means for monitoring the reception level of the reception signal and the multiplication means for multiplying the level of the input signal of the waveform equalizer are used. In the first step, the waveform etc. Waveform equalization method that multiplies the level of the input signal of the equalizer. A waveform equalizer method for a waveform equalizer according to claim 1 ,
The reception level monitoring means for monitoring the reception level of the reception signal and the discriminator for determining the symbol of the equalized output signal are used. In the first step, the reference level of the discriminator is determined according to the reception level. And a waveform equalization method for changing the output level of the training signal. An equalizing filter unit having a tapped delay circuit, performing an operation including a tap coefficient on an input signal and outputting an equalized output signal, and a tap coefficient control means for setting a tap coefficient of the equalizing filter unit; A tap coefficient storage means for storing the tap coefficient, and an equalization performance monitoring means for monitoring the equalization performance ,
The waveform equalizer which performs the procedure of the waveform equalization method in any one of Claims 1 thru | or 7 . A radio apparatus equipped with the waveform equalizer according to claim 8 . A wireless communication system including the wireless device according to claim 9 .

Images