JP2010283419A - Multi-mode switching adaptive equalizer - Google Patents

Abstract

In a decision feedback type adaptive equalizer, error propagation in an equalization circuit, a failure of a feedback filter and a Ford forward filter, a training signal cannot be obtained due to a poor environment, and tracking cannot be performed due to an unforeseen external factor Assuming conditions, etc., it should be possible to continue transmission while maintaining a certain level of quality, and be robust and resistant to environmental fluctuations.
A blind feedback equalizer has a blind mode in addition to a training mode and a tracking mode in a decision feedback type adaptive equalizer, and a switching control unit selects an optimum mode according to environmental information and equalizer internal information. The present invention proposes a multi-mode switching type equalizer that switches to 1).
[Selection] Figure 6

Description

  The present invention relates to an adaptive equalization technique and an adaptive equalizer that compensate for a fading phenomenon in a wireless communication environment.

  In recent years, digital communication broadcasting systems such as mobile phones, high-speed wireless LANs, and terrestrial digital TVs are rapidly spreading. However, in a wireless communication environment, in addition to the effects of thermal noise, fading that causes a sudden change in the envelope (amplitude) and phase of the received signal due to reflection, diffraction, scattering, and traveling of a moving object. The effect becomes significant. Such effects of fading are classified into two types: frequency non-selective fading in which the frequency components constituting the signal are uniformly attenuated, and frequency selective fading that is subject to different attenuation depending on the frequency components of the signal. Although depending on the characteristics of the transmission path, the latter is generally fading to be considered when transmitting a wideband signal, and the multi-wave propagation delay cannot be ignored. Therefore, in a fading channel such as wireless communication, it is essential to consider some fading compensation measures. Examples of main fading compensation techniques include adaptive equalization techniques, diversity techniques, and adaptive array antenna techniques. Of these, the diversity and adaptive array antennas require a plurality of receiving antennas, which is not suitable for downsizing the apparatus. On the other hand, with adaptive equalization, distortion generated by fading can be reduced by digital signal processing of a signal generated at one receiving antenna. The adaptive equalization technique is one technique for compensating for frequency selective fading. The present invention relates to this adaptive equalization technique and adaptive equalizer (Patent Document 1).

  The adaptive equalizer adjusts the filter coefficient using the training signal and then switches to the tracking mode.In other words, the coefficient is estimated from the input / output signal of the unknown system and output, and the training signal is not used. There are two methods: a method that estimates and outputs the coefficient only from the input signal of the unknown system. The former can be broadly classified as a training type adaptive equalizer and the latter can be broadly classified as a blind type adaptive equalizer.

  However, as shown in FIG. 1, if a training equalizer and a blind equalizer are used within the same constrained processing time, the actual data transmission time is longer for the blind equalizer. Even if a transmission error occurs at an early stage of the convergence, there is a possibility that improvement in communication efficiency can be expected.

A decision feedback adaptive equalizer is widely used as a training equalizer. The basic operation is described in Non-Patent Document 1 and the like. It is basically composed of the following four parts.
(1) Equalizing section It consists of a feedforward filter and a feedback filter. It operates to reduce intersymbol interference by feeding back the output from the FB filter via the data decision unit.
(2) Determination part The transmission symbol is estimated by determining the output of the equalization unit.
(3) Error estimation unit The error e (t) between the ideal value r (t) of the output of the equalization unit and the actual equalization output z (t) is the actual equalization output z from the ideal value r (t). It can be obtained by subtracting (t). Here, the ideal value r (t) is a known transmission symbol sequence y (nD) (D: delay amount) in the training mode, and the equalized determination output y _d ( n)
(4) Coefficient update unit
The tap coefficients of the feedforward filter and feedback filter are updated using an adaptive algorithm such as the RLS algorithm or LMS algorithm.

FIG. 2 shows a training mode state in the decision feedback type adaptive equalizer.

The tracking mode state in the decision feedback adaptive equalizer is shown in FIG.

  In the training equalizer, before transmitting information to be transmitted, parameters of the equalizer are adjusted according to a certain procedure using the training signal. Thereafter, transmission of information to be transmitted is started, and adaptive equalization using the equalizer output (reproduced signal) is performed. However, when the transmission path characteristic changes greatly due to fading or the like, it is necessary to readjust the equalizer parameters using the training signal. This causes inconveniences in a polling system and a mobile communication system. Therefore, a method for adjusting the parameters of the equalizer without using the training signal has been desired. One answer to this is blind equalization (Patent Document 2).

As shown in FIG. 4, blind equalization estimates the reverse characteristics of the transmission path from only the a priori knowledge of the transmission signal and the received signal, and reproduces the transmission signal using the reverse system of the transmission path and the received signal. It is a method. Conventionally proposed blind equalization methods are roughly classified into the following four types.
(1) Method based on probability distribution joint method (2) Minimum entropy method (3) Method using higher-order statistics (mainly third-order and fourth-order moment) of received signal (4) Obtained by oversampling received signal Using the periodic steadiness of generated signals

  Sato's algorithm is known as the most representative algorithm of (1). This method adaptively matches the probability distribution of the transmitted signal and the reproduced signal (equalizer output) on the assumption that the transmitted signal is independent and strongly stationary (hereinafter referred to as IID for short). This is an adaptive equalization method that realizes equalization. However, the above method generally has the advantage that adaptive equalization is possible and the amount of calculation is small, but has the disadvantage that the convergence speed of the algorithm is slow and the reliability of the reproduced signal is low (Non-patent Document 2) .

  The method (2) is a method that enables blind equalization by setting the equalizer parameters so that the entropy of the equalizer output decreases when the transmission signal is IID (Non-Patent Document). 3, 4). Based on this, Non-Patent Document 5 gives a specific evaluation amount and an equalization algorithm. Although this method has the same amount of computation as the Sato algorithm method and has made a slight improvement with respect to the convergence speed and the reliability of the reproduced signal, there is room for further improvement in the algorithm in this document. There is something to say.

  With the method (3), if accurate high-order statistics can be calculated, the transmission path characteristics can be estimated with high accuracy, and as a result, accurate equalization can be performed. However, in order to realize this method, an enormous amount of calculation is required, and online equalization is difficult.

  The method (4) uses the periodic steadiness of the signal obtained by oversampling the received signal, and performs equalization only from the secondary statistics of the signal (Non-Patent Document 6). However, this method is a deterministic method, and research on adaptive methods is rarely seen.

Japanese Patent Application No. 2008-184383 Japanese Patent Laid-Open No. 2002-33682 JP 2001-345743 A JP 2006-109501 A JP 2006-33663 A Japanese Patent Laid-Open No. 2002-135180

Shuichi Sasaoka, “Mobile Communications”, 3rd edition, Ohmsha, Ltd., 1998, p.201-222 Yoichi Sato, “Linear Equalization Theory”, 1st Edition, Maruzen Co., Ltd., 1991, p.154-176 R.A. Wiggins, “Minimum Entropy Deconvolution, 1st Edition, Geoexploration, Volume 16 (Vol.16), 1978 D. Donoho, “On Minimum Entropy Deconvolution”, 1st edition, Applied Time Series Analysis II, Academic Press, 1981, Hiroyoshi Oda and Yoichi Sato, “About speeding up blind equalization”, System Control Information, vol6 No. 7, July 1993, By L.Tong, G.Xu, and T.Kailath, “Nessesary and Sufficient Conditions of Channeled Fidelity Beast on Second-Order Cyclostainary Staties Sticks (Necessary and Sufficient condtionsof Channelidentifiability Based on Scond-order CyclostationaryStatistics) ”, (Proc.ISIT), 1993

  For example, as seen in Patent Documents 3 to 6, various inventions have been made so far with respect to decision feedback adaptive equalizers for the purpose of reducing the amount of calculation in the equalizer and realizing interference cancellation.

  In Patent Document 3, the correctness of the data is determined according to the degree of the influence of noise according to whether the equalized signal one symbol before exists in the determination region or in the erasure region in the equalization circuit. The decision data is fed back. In the case of an error, the decision data is not fed back, and the decision data is determined so that the equalized signal of the current symbol belongs to the decision region from the decision data one symbol before, The configuration is such that error propagation is reduced by feeding back the judgment data.

  Further, in Patent Document 4, an output value is determined based on a result of comparing a received threshold value equalized in an equalizer with a positive threshold (determination value) and a negative threshold (determination value). ing. Further, when it is determined that the output value is 0, the equalizer adjusts the stored correction value depending on whether the equalized reception value is positive or negative, thereby causing a defect in the equalization circuit line. Even if an error occurs, the error propagation is reduced by learning it.

  However, the methods disclosed in Patent Document 3 and Patent Document 4 are methods focused on reducing propagation when error propagation occurs in the equalization circuit. Therefore, depending on the state, when an error that is difficult to reduce occurs, it may be difficult to expect the reduction effect.

  In addition to the error propagation in the equalization circuit, abnormal situations that should be considered can be mentioned. For example, failure of feedback filter (hereinafter abbreviated as FB filter) and Ford forward filter (hereinafter abbreviated as FF filter), inability to acquire training signal due to poor environment, untrackable state due to unforeseen external factors, etc. It is. If the training signal cannot be acquired, it means that the training is not performed for the training time in FIG. 1, and if the tracking is impossible, the tracking is not performed for the tracking time in FIG. It can be said that the function of the information transmission system using the equalizer is greatly impaired.

  Regarding the blind equalizer, various inventions have been made as described in the background art. However, up to now, training equalizers and blind equalizers have been improved and improved separately, and various techniques have been invented. Therefore, the invention proposal of the system which complements each other's advantage and fault is not seen.

  In order to achieve the above object, the multimode switching type adaptive equalizer described in claim 1 of the claims includes a feedforward filter, an error calculation unit, and a determination part for estimating a transmission symbol in this order. And the output of the determination part is fed back to the error calculation unit via a feedback filter, and the error calculation unit uses the training signal in a training mode, and the error calculation unit outputs the output of the feedforward filter and the feedback filter. A decision feedback type adaptive equalizer that switches to the tracking mode after setting the coefficients of the two filters so that the calculated value is minimized, and a blind adaptive equalizer, It is characterized by having a switching control unit that switches to either of these adaptive equalizers to operate.

  The multimode switching type adaptive equalizer described in claim 2 is the one described in claim 1, wherein the switching control unit has an abnormality in the operation of the decision feedback type adaptive equalizer. And / or by referring to external environment information, the decision feedback adaptive equalizer is switched to the blind adaptive equalizer.

  The multi-mode switching type adaptive equalizer described in claim 3 is the one described in claim 1, wherein the switching control unit determines that the determination feedback is not necessary when the training signal is not required to be acquired. It is characterized by switching from the shape adaptive equalizer to the blind equalizer.

  The multi-mode switching type adaptive equalizer described in claim 4 is the one described in claim 2, wherein the switching control unit detects that at least one of the feedforward filter and the feedback filter is abnormal. If it is determined, the decision feedback adaptive equalizer is switched to the blind equalizer.

  The multimode switching type adaptive equalizer described in claim 5 is the one described in claim 4, wherein the switching control unit is configured to determine whether at least one of the feedforward filter and the feedback filter is abnormal. It is characterized in that it is determined whether the training mode is performed and / or after the training mode is performed.

  The multi-mode switching type adaptive equalizer described in claim 6 is the one described in claim 2, wherein the switching control unit determines that the blind mode or the like when the operation in the training mode is abnormal. It is characterized by switching to a generator.

  The multi-mode switching type adaptive equalizer described in claim 7 is the one described in claim 6, wherein the switching control unit is provided for a predetermined training time set in advance in the training mode. When the training is not performed normally, it is determined that the operation in the training mode is abnormal.

  The multi-mode switching type adaptive equalizer described in claim 8 is the one described in claim 2, wherein the switching control unit has a continuous determination error in the determination part. Then, it is judged as abnormal, and the decision feedback type adaptive equalizer is switched to the blind equalizer.

  According to the invention proposal, one information transmission system has three modes, is robust to the surrounding environment, and even when an unexpected situation occurs, transmission is interrupted all at once. However, even if the quality deteriorates due to the mode switching using the environmental information and the internal information of the equalizer, it can be expected that the transmission can be continued while maintaining a certain quality.

  That is, the invention proposal does not necessarily use the average bit error rate characteristic and throughput characteristic, which have been emphasized in the conventional information transmission system, as an evaluation index, but focuses on robustness and environmental fluctuations.

In a training equalizer and a blind equalizer, it is a figure which shows the ratio which the real transmission time occupies with respect to the total transmission time. FIG. 10 is a diagram illustrating a training mode state in the decision feedback adaptive equalizer. FIG. 6 is a diagram illustrating a tracking mode state in the decision feedback adaptive equalizer. It is a figure which shows the state of a blind equalizer. It is a figure which shows the system configuration | structure of the multi-mode switching equalizer proposed by the invention. It is a figure which shows the processing details of the multi-mode switching equalizer proposed by the invention. It is a table | surface which shows the simulation specification in computer simulation. It is a figure which shows the result of a computer simulation.

  Embodiments of the present invention will be described below.

  In order to overcome the problems described in “Problems to be Solved by the Invention”, a transmission system in which a conventional decision feedback adaptive equalizer and a blind equalizer are fused is proposed. Specifically, in addition to the training mode and tracking mode in the conventional decision feedback type adaptive equalizer, it has a blind mode in the blind equalizer, and the optimum mode is selected according to the environmental information and the equalizer internal information. The invention proposes a multi-mode switching equalizer for switching. FIG. 5 is a system configuration diagram of the multi-mode switching equalizer proposed by the invention.

  As shown in the figure, the switching control unit switches between a decision feedback adaptive equalizer and a blind equalizer. The decision feedback type adaptive equalizer is connected in the order of the feedforward filter, the error calculation unit, and the determination part for determining the transmission symbol, and the output of the determination part is fed back to the error calculation part via the feedback filter, In the training mode, using the training signal, the error calculator calculates the difference between the output of the feedforward filter and the output of the feedback filter, sets the coefficients of both filters so that this calculated value is minimized, and then enters the tracking mode. It is to switch. The decision feedback adaptive equalizer and the blind adaptive equalizer are connected in parallel, and the switching control unit switches which of the two adaptive equalizers operates. In the figure, the switching control unit is shown inserted in three places, the latter stage of the feedforward filter, the latter stage of the determination part, and the latter stage of the feedback filter. It shows that the changeover switch is controlled based on the inserted position data to switch from the decision feedback adaptive equalizer to the blind adaptive equalizer. The change control unit can control the changeover switch based on environment information described later.

  FIG. 6 is a flowchart showing an example of processing details of the multimode switching equalizer proposed by the invention. First, in step 1, the transmission time is set to a variable T.

  Next, in step 2, it is determined whether a training signal is necessary. If it is determined that the training signal is unnecessary, the process proceeds to step 11 to switch to a blind equalizer. For example, when the switching control unit acquires external environmental information and refers to it, it is known that the equalizer is used as a blind equalizer or that it is difficult to acquire a training signal in advance. This is the case. The external environment information refers to, for example, a training signal, a fading element that prevents normal operation of the replacement control unit and the determination part.

  If it is determined in step 2 that a training signal is necessary, the process proceeds to step 3 to check whether both the FB filter and the FF filter are normal. If the coefficients of the FB filter and FF filter are updated to a predetermined value based on a predetermined adaptive algorithm, it is determined as normal.

  If it is determined in step 3 that either the FB filter or the FF filter is abnormal due to a failure or the like, the process proceeds to step 11 to switch to the blind equalizer. When the FB filter or FF filter coefficient is being updated based on a predetermined adaptive algorithm, if a lock-up (updating update) phenomenon or the like due to zero division or the like occurs, it is determined as abnormal.

In step 3, FB filter, if it is determined to be normal in both the FF filter, and proceeds to step 4 to set the training time to the variable t _0.

  After the process of step 4 is completed, the process proceeds to step 5 to set the training mode in the decision feedback adaptive equalizer shown in FIG.

After the process of step 5 is completed, the process proceeds to step 6 where it is checked whether training has been performed for the time set in step 4 (variable t ₀ ).

If it is determined in step 6 that training has not been performed for time t _0, it is determined that a sufficient training signal has not been acquired (determined as abnormal), and the process proceeds to step 11 to switch to the blind equalizer. .

If it is determined in step 6 that training has been performed for time t ₀ , the process proceeds to step 7 to set the tracking mode in the decision feedback adaptive equalizer shown in FIG.

  After the processing in step 7, the process transitions to step 8 to check whether both the FB filter and the FF filter are normal. At this time, the same determination as in step 2 is performed.

  If it is determined in step 8 that either the FB filter or the FF filter is abnormal due to a failure or the like, the process proceeds to step 11 to switch to the blind equalizer. At this time, the same determination as in step 3 is performed.

  If it is determined in step 8 that both the FB filter and the FF filter are normal, the process proceeds to step 9 to check whether consecutive errors have occurred in the determination part.

  If it is determined in step 9 that consecutive errors have occurred, it is determined that an abnormal situation such as error propagation in the equalization circuit has occurred (determined as abnormal), and the process proceeds to step 11. Switch to blind equalizer.

In step 9, if the continuous error is determined not to have occurred, the process proceeds to step 10 reduces the time t ₀ is set from the time T set in step 1 Step 4, the time period, the tracking Check if they have gone.

If it is determined in step 10 that tracking is not performed for the time obtained by subtracting the time t ₀ set in step 4 from the time T set in step 1, an information transmission system using a decision feedback equalizer It is determined that the function of the function is greatly impaired (determined as abnormal), and the process proceeds to step 11 to switch to the blind equalizer.

If it is determined in step 10 that tracking has been performed for the time obtained by subtracting the time t ₀ set in step 4 from the time T set in step 1, the process proceeds to step 12, and transmission processing ends normally. Judge that

  This operation flow is an example for applying the present invention. As described above, the switching control unit switches between the decision feedback adaptive equalizer and the blind adaptive equalizer to operate. If so, it is included in the present invention. In particular, the switching control unit is abnormal in the operation of the decision feedback adaptive equalizer from the information in the equalizer (equalizer internal information) as determined in steps 2, 3, 6, 8, 9, and 10 above. It is preferable to switch from the decision feedback type adaptive equalizer to the blind adaptive equalizer when it is determined that the above has occurred. The present invention also includes that the switching control unit switches from the decision feedback adaptive equalizer operating in the training mode or the tracking mode to the blind adaptive equalizer by referring to external environment information. The switching control unit can also switch from the decision feedback adaptive equalizer to the blind equalizer by determining the internal information of the equalizer, the external environment information, or a combination thereof.

As an embodiment of the present invention, a computer simulation is performed to verify the effectiveness of the invention system. In this simulation, “S. Haykin, Book Title” Adaptive Filter Theory (Adaptive
Filter Theory ”, Publisher Prentice Hall
Hall), publication year 1991 ", etc.
A Cosine (Raised Cosine) channel was used. In this communication channel, the influence of intersymbol interference changes depending on the value of the parameter W, and it becomes difficult to equalize the communication channel as the value of W increases. The simulation specifications are shown in FIG. The RLS algorithm was used as the coefficient update algorithm for the decision feedback adaptive equalizer, and the SATO algorithm shown in Non-Patent Document 2 was used as the coefficient update algorithm for the blind equalizer. In order to show the effectiveness of the invention proposal, a burst length error occurred as a failure during communication.

The simulation result is shown in FIG. As comparisons, normal communication (通信 in the figure), abnormality (△ in the figure), and use of the invention proposal (◯ in the figure) are shown. When using the method proposed by the invention, SNR = 23 [dB]
An abnormal condition is detected at around and switched to a blind equalizer. Immediately after that, the characteristics deteriorate for a moment, but then SNR = 30 [dB]
The result toward the direction of improvement of characteristics was obtained. If the invention proposal method is not used, as shown by the time of occurrence of an abnormality (Δ in the figure), the result is that communication is performed with the characteristics deteriorated. From the above, the effectiveness of the invention proposal method was shown.

Claims (8)

  The feedforward filter, the error calculation unit, and the determination part for estimating the transmission symbol are connected in this order, and the output of the determination part is fed back to the error calculation part via the feedback filter, and the training signal is used in the training mode. Then, the error calculation unit calculates the difference between the output of the feedforward filter and the output of the feedback filter, sets the coefficients of the two filters so that the calculated value is minimized, and then switches to the tracking mode. A multi-mode switching type adaptive equalizer having a shape adaptive equalizer and a blind adaptive equalizer, and having a switching control unit that switches to and operates either of these adaptive equalizers.   When the switching control unit determines that an abnormality has occurred in the operation of the decision feedback adaptive equalizer, and / or by referring to external environmental information, the switching feedback control unit The multimode switching type adaptive equalizer according to claim 1, wherein switching is performed to the blind adaptive equalizer.   2. The multi-mode switching type adaptation according to claim 1, wherein the switching control unit switches from the decision feedback adaptive equalizer to the blind equalizer when it is determined that the acquisition of the training signal is unnecessary. Equalizer.   The switching control unit switches from the decision feedback type adaptive equalizer to the blind equalizer when it is determined that at least one of the feedforward filter and the feedback filter is abnormal. The multimode switching type adaptive equalizer described.   The switching control unit determines whether at least one of the feedforward filter and the feedback filter is abnormal before performing the training mode and / or after performing the training mode. The multimode switching type adaptive equalizer according to claim 4.   The multimode switching type adaptive equalizer according to claim 2, wherein the switching control unit switches to the blind equalizer when the operation in the training mode is determined to be abnormal.   The switching control unit determines that the operation of the training mode is abnormal when the training is not normally performed for a predetermined training time set in the training mode. Item 7. The multimode switching type adaptive equalizer according to item 6.   The switching control unit determines that it is abnormal when a continuous determination error occurs in the determination part, and switches from the determination feedback adaptive equalizer to the blind equalizer. The multimode switching type adaptive equalizer according to claim 2.

Images