JP2019047180A - Transmission path estimator and equalizer - Google Patents

Abstract

To provide a transmission path estimator ensuring stable equalization performance, by estimating a transmission path adaptively for each tap.SOLUTION: A transmission path estimator outputting a transmission path estimate, calculated by a least square average algorithm, to a sequence estimate equalizer for equalizing the input signals includes a first transmission path estimation part for generating multiple first transmission path estimates corresponding to respective incoming waves forming the input signal, on the basis of known series signal included in the input signal, a weight generation part for generating multiple weight values corresponding to the respective incoming waves and imparting the weighting to step size in the least square average algorithm, on the basis of the multiple first transmission path estimates, and a second transmission path estimation part for calculating the multiple second transmission path estimates corresponding to the respective incoming waves on the basis of the multiple first transmission path estimates and the multiple weight values, and outputting to the sequence estimate equalizer.SELECTED DRAWING: Figure 3

Description

  The present invention relates to a channel estimator and equalizer having a function of channel estimation in adaptive equalization based on time domain sequence estimation, and more particularly to a technique for improving channel estimation accuracy.

  The conventional sequence estimation type equalizer calculates the likelihood value of the replica for the received signal based on the replica created for each sequence from the channel estimation value. The equalizer equalizes the received signal by estimating the sequence whose likelihood value is lowest. Thus, the accuracy of the channel estimate determines the equalizer performance.

  As one of the transmission path estimation methods, a least mean square (hereinafter referred to as LMS) method is well known. Since the LMS method in which the step size is fixed is fed back the calculated likelihood value at a constant rate to update the original channel estimation value, the LMS method has a feature that it can be easily realized with a small amount of calculation.

Japanese Patent Application Laid-Open No. 11-313013

  The channel estimation method of fixing the step size and updating the channel estimation value uniformly causes a constant estimation error with respect to taps having no delay wave in an environment where a long delay wave is generated. Therefore, Patent Document 1 proposes a transmission path estimation device and a transmission path characteristic correction device which fix the transmission path estimated value to 0 by thresholding the power of the delayed wave at each tap in advance and the power is below the threshold. It is done. However, these devices have a problem in terms of the performance stability of the equalizer because they cause an extreme estimation error in the channel estimation value when the threshold determination is wrong.

  The present invention has been made to solve the problems as described above, and an object thereof is to provide a transmission path estimator which can estimate a transmission path adaptively for each tap and obtain stable equalization performance.

  A channel estimator according to the present invention is a channel estimator that outputs a channel estimation value calculated by a least mean square algorithm to a sequence estimation equalizer that equalizes a reception signal, and is included in the reception signal. A first channel estimation unit for generating a plurality of first channel estimation values corresponding to each of a plurality of incoming waves forming the reception signal based on the known series signal, and corresponding to each of the plurality of incoming waves, A weight generation unit that generates, based on a plurality of first channel estimation values, a plurality of weight values for weighting step sizes in a least mean square algorithm; a plurality of first channel estimation values and a plurality of weight values And a second channel estimation unit configured to calculate a plurality of second channel estimation values corresponding to each of the plurality of incoming waves and output the calculated second channel estimation values to the sequence estimation equalizer.

  According to the present invention, it is possible to provide a channel estimator that can estimate a channel adaptively for each tap and obtain stable equalization performance.

FIG. 2 is a block diagram showing a configuration of an equalizer in Embodiment 1. FIG. 5 is a diagram showing a configuration of a reception signal in Embodiment 1. FIG. 2 is a block diagram showing a configuration of a transmission path estimator in Embodiment 1. FIG. 2 is a diagram showing an incoming wave in the first embodiment. FIG. 2 is a diagram showing a configuration of a weight generation unit in Embodiment 1. FIG. 2 is a diagram showing a configuration of a processing circuit in Embodiment 1. FIG. 2 is a diagram showing a configuration of a processing circuit in Embodiment 1. 5 is a flowchart showing a transmission channel estimation method in Embodiment 1. 7 is a flowchart showing a method of generating weight values in Embodiment 1. FIG. 10 is a diagram showing a configuration of a weight generation unit in Embodiment 2. 7 is a flowchart showing a method of generating weight values in Embodiment 2. FIG. 18 is a diagram showing a configuration of a weight generation unit in Embodiment 3. 15 is a flowchart showing a method of generating weight values in Embodiment 3.

  Embodiments of a channel estimator and equalizer according to the present invention will be described.

Embodiment 1
(Device configuration)
FIG. 1 is a diagram showing the configuration of the equalizer 10 according to the first embodiment. The equalizer 10 is an adaptive equalizer that sequentially estimates signals in the time domain.

  FIG. 2 is a diagram showing the configuration of the received signal 1 input to the equalizer 10. As shown in FIG. In the first embodiment, the received signal 1 is a radio frame, the known series signal 1a is disposed at the head of the radio frame, and the equalization target signal 1b is disposed behind the signal. The equalization direction is the direction from the known series signal 1a to the equalization target signal 1b. The received signal 1 is not limited to the configuration shown in FIG. 2, and may be configured such that the signal to be equalized 1b is disposed forward and the known sequence signal 1a is disposed posteriorly. In that case, the equalization direction is from back to front.

  As shown in FIG. 1, the equalizer 10 includes a channel estimator 20 and a sequence estimation equalizer 30. First, the reception signal 1 is input to the channel estimation unit 20 and the sequence estimation equalizer 30.

  The channel estimation unit 20 obtains a channel estimation value by cross correlation or normal equation based on the known series signal 1 a included in the received signal 1. Although a synchronization process for estimating the known sequence signal 1a is necessary as a precondition for generation of the channel estimation value, a detailed description of the synchronization process is omitted. The channel estimation unit 20 outputs the generated channel estimation value 21 to the sequence estimation equalizer 30.

  The sequence estimation equalizer 30 generates a replica based on the channel estimation value 21 input from the channel estimation unit 20 and the modulation scheme of the assumed system. Then, the sequence estimation equalizer 30 calculates an error between the equalization target signal 1 b included in the received signal 1 and the generated replica. The sequence estimation equalizer 30 outputs the calculated error information 12 and the sequence information 11 used for the error calculation to the channel estimation unit 20. At this time, the error information 12 and the sequence information 11 output to the channel estimation unit 20 correspond to any one of information corresponding to the smallest error among the plurality of errors calculated in the plurality of sequences. It may be each information or each information corresponding to each total error. Here, although the configuration of the equalizer 10 in which the error information 12 is generated by the sequence estimation equalizer 30 is shown as an example, it is an equalizer that generates the error information 12 in the transmission path estimator 20. May be

  The channel estimation unit 20 estimates and updates the channel estimation value 21 again based on the received signal 1 and the sequence information 11 and the error information 12 input from the sequence estimation equalizer 30. The update method will be described later. The channel estimator 20 outputs the updated channel estimation value 21 to the sequence estimation equalizer 30. The sequence estimation equalizer 30 equalizes the re-input equalization target signal 1b based on the updated channel estimation value 21, and outputs an equalization result 13.

  The equalizer 10 repeats, between the channel estimation unit 20 and the sequence estimation equalizer 30, the input of the received signal 1, equalization using the sequence estimation, and updating of the channel estimation value 21.

  Next, the configuration of the transmission path estimator 20 will be described. FIG. 3 is a diagram showing the configuration of the transmission path estimator 20 in the first embodiment. The channel estimator 20 includes an initial channel estimator 210, a weight generator 220, and an LMS channel estimator 230.

  The initial channel estimation unit 210 generates a first channel estimation value based on the known sequence signal 1 a included in the received signal 1. At this time, the initial channel estimation unit 210 generates a first channel estimation value by an initial channel estimation method represented by cross-correlation, normal equation, or the like.

  FIG. 4 is a diagram showing a channel impulse response (hereinafter referred to as “CIR”) corresponding to the received signal 1 inputted by the channel estimator 20. As shown in FIG. Here, it is assumed that the first channel estimation value obtained from the received signal 1 is expressed in the form of CIR, and the received signal 1 is composed of a plurality of incoming waves, among which the preceding wave and the plurality of delayed waves And shall be included. The LMS channel estimation unit 230 implements channel estimation using the LMS algorithm on each of these incoming waves.

  The initial channel estimation unit 210 outputs the generated first channel estimation value 211 for each incoming wave to the weight generation unit 220 and the LMS channel estimation unit 230.

  The weight generation unit 220 generates each weight value corresponding to the incoming wave based on the plurality of first channel estimation values 211. That is, the weight generation unit 220 generates each weight value for each tap. The weight value is a coefficient for weighting the step size in the LMS algorithm. The detailed generation method of each weight value will be described later. The weight generation unit 220 outputs the generated weight values 225 to the LMS channel estimation unit 230.

  The LMS channel estimation unit 230 determines each of the plurality of arrival waves based on the plurality of first channel estimation values 211 and the plurality of weight values 225, and the error information 12 and the sequence information 11 input from the sequence estimation equalizer 30. Calculating a plurality of second channel estimation values corresponding to. At the time of calculation of each second channel estimation value, the step size 231 of each tap is changed by each weight value 225. The LMS channel estimation unit 230 outputs each of the calculated second channel estimation values 212 as the channel estimation value 21 to the sequence estimation equalizer 30. That is, the LMS channel estimation unit 230 updates the input first channel estimation values 211 with the calculated second channel estimation values 212.

  FIG. 5 is a block diagram showing the configuration of the weight generation unit 220 in the first embodiment. Weight generation unit 220 includes transmission path estimated value storage unit 2210 and weight calculation unit 2211.

  The channel estimation value storage unit 2210 can distinguish the plurality of first channel estimation values 211 generated by the initial channel estimation unit 210 for each incoming wave and for each past reception timing of the reception signal 1 Store in For example, each first channel estimation value 211 is stored in the channel estimation value storage unit 2210 with a symbol identifying each incoming wave and a symbol identifying reception timing.

  The weight calculating unit 2211 reads a plurality of first channel estimation values 211 having different reception timings from the channel estimation value storage unit 2210, and calculates a time average value for each incoming wave. The weight calculating unit 2211 generates each weight value 225 for each incoming wave based on the time average value.

  FIG. 6 is a view showing an example of the processing circuit 90 provided in the transmission path estimator 20. As shown in FIG. Each function of initial channel estimation unit 210, weight generation unit 220, LMS channel estimation unit 230, channel estimation value storage unit 2210 and weight calculation unit 2211 is realized by processing circuit 90. That is, the processing circuit 90 includes an initial transmission path estimation unit 210, a weight generation unit 220, an LMS transmission path estimation unit 230, a transmission path estimated value storage unit 2210, and a weight calculation unit 2211.

  When the processing circuit 90 is dedicated hardware, the processing circuit 90 may be, for example, a single circuit, a complex circuit, a programmed processor, a parallel programmed processor, an application specific integrated circuit (ASIC), an FPGA (field-programmable) Gate Array) or a circuit combining these. Each function of initial channel estimation unit 210, weight generation unit 220, LMS channel estimation unit 230, channel estimation value storage unit 2210 and weight calculation unit 2211 may be realized individually by a plurality of processing circuits, It may be realized collectively by one processing circuit.

  FIG. 7 is a view showing another example of the processing circuit provided in the transmission path estimator 20. As shown in FIG. The processing circuit includes a processor 91 and a memory 92. Each function of initial channel estimation unit 210, weight generation unit 220, LMS channel estimation unit 230, channel estimation value storage unit 2210 and weight calculation unit 2211 by processor 91 executing a program stored in memory 92. Is realized. For example, software or firmware described as a program is executed by the processor 91 to realize each function. That is, the channel estimation unit 20 generates a plurality of first channel estimation values 211 corresponding to each of the plurality of incoming waves forming the reception signal 1 based on the known series signal 1a included in the reception signal 1; A plurality of weight values 225 corresponding to each of the plurality of incoming waves and weighting the step size 231 in the least mean square algorithm are generated based on the plurality of first channel estimation values 211, and the plurality of first channel A memory 92 for storing a program for calculating the plurality of second channel estimation values 212 corresponding to each of the plurality of incoming waves based on the estimation value 211 and the plurality of weight values 225 and outputting the second channel estimation values 212 to the sequence estimation equalizer 30. And a processor 91 that executes the program. In addition, the program stores the plurality of generated first transmission path estimated values 211 in the transmission path estimated value storage unit 2210 so as to be distinguishable for each arrival wave and for each reception timing of the reception signal 1 Each weight value 225 for each incoming wave is generated based on the plurality of first channel estimation values 211 at the past reception timing stored in the channel estimation value storage unit 2210. The transmission channel estimated value storage unit 2210 may be the memory 92. The program causes a computer to execute the procedure or method of the initial transmission channel estimation unit 210, weight generation unit 220, LMS transmission channel estimation unit 230, transmission channel estimated value storage unit 2210, and weight calculation unit 2211.

  The processor 91 is, for example, a central processing unit, a processing unit, an arithmetic unit, a microprocessor, a microcomputer, a DSP (Digital Signal Processor) or the like. The memory 92 is, for example, non-volatile or volatile such as random access memory (RAM), read only memory (ROM), flash memory, erasable programmable read only memory (EPROM), electrically erasable programmable read only memory (EEPROM) It is a semiconductor memory. Alternatively, the memory 92 may be any storage medium used in the future, such as a magnetic disk, a flexible disk, an optical disk, a compact disk, a mini disk, a DVD, and the like.

  The functions of the above-described initial transmission path estimation unit 210, weight generation unit 220, LMS transmission path estimation unit 230, transmission path estimated value storage unit 2210, and weight calculation unit 2211 are partially realized by dedicated hardware, A part of may be realized by software or firmware. Thus, the processing circuit implements each of the functions described above by hardware, software, firmware, or a combination thereof.

(Operation)
FIG. 8 is a flowchart showing a transmission channel estimation method in the first embodiment.

  In step S10, the initial channel estimation unit 210 of the channel estimation unit 20 receives the known sequence signal 1a included in the reception signal 1.

  In step S20, initial channel estimation unit 210 generates first channel estimation value 211 for each incoming wave, and outputs it to weight generation unit 220 and LMS channel estimation unit 230.

  At step S30, weight generation section 220 generates weight value 225 for each incoming wave based on each first channel estimation value 211. FIG. 9 is a flowchart showing a method of generating the weight values 225. Step S30 is performed according to the flowchart shown in FIG.

  In step S310, weight calculation section 2211 reads a plurality of first channel estimation values different in reception timing from transmission channel estimation value storage section 2210. The number of past channel estimation values to be read is arbitrary.

In step S320, the weight calculation unit 2211 calculates the time average value E _j of each of the first channel estimation values C _{(j, t)} for each incoming wave according to equation (1). In equation (1), j is the number of incoming waves, and in this case j = 0 to m. Also, t is the reception timing, and t = t _{0 to} t _{n -1} .

  In the calculation of the time average value, when the incoming wave is a complex number, the weight calculating unit 2211 may rotate the vector of the incoming wave in the reference direction and then perform averaging processing to convert it into a power value, or The averaging process may be performed after the power value is determined by the sum. Also, the number of radio frames used to calculate weight value 225 and the number of taps (the number of incoming waves) are arbitrarily set by the system provided with equalizer 10 and are described in the first embodiment. It is not limited to the configuration. Moreover, in Formula (1), although the time average value is calculated by the section average in the time-axis of each 1st transmission path estimated value, a time average value may be calculated by the accumulation average.

  In step S330, the weight calculation unit 2211 normalizes each time average value with the maximum value among the plurality of time average values. Each normalized time average value is a weight value for each incoming wave, that is, for each tap. Thus, the method of generating weight values is completed.

Next, in step S40 shown in FIG. 8, the LMS channel estimation unit 230 estimates the first channel estimation value 211, the weight value 225 generated by the weight calculation unit 2211, the step size, and the sequence estimation equalization. A second channel estimation value is calculated based on the error information and the complex conjugated sequence information input from the unit 30. More specifically, LMS channel estimation unit 230 determines weight value w _j , step size s, and error information R ₍ the first channel estimate C _(j, t ₁ ) at time t ₁ according to equation (2)). by adding the product of _t1) and the series information ^{u *,} to calculate a second channel estimation value _{C (j, t2)} at time _{t 2} is the next time. Thus, the LMS channel estimation unit 230 calculates the second channel estimation value for each tap.

  In step S <b> 50, LMS channel estimation unit 230 outputs each second channel estimation value 212 to sequence estimation equalizer 30. That is, the LMS channel estimation unit 230 updates each first channel estimation value 211 input from the initial channel estimation unit 210 with each second channel estimation value 212 and outputs it. Also, here, the second channel estimation value 212 is the updated channel estimation value 21 described above.

  In step S60, sequence estimation equalizer 30 equalizes received signal 1 with second channel estimation value 212, and outputs equalization result 13.

The method of sequence estimation in which weight values are used has been described above. Next, the method of sequence estimation in which weight values are not used will be described. An equation for calculating a channel estimation value in the LMS scheme in which no weight value is used is given by equation (3). Specifically, the second channel estimation value _C at time _{t 2 (j, t2),} the first channel estimation value _C at time _{t 1} and _{(j, t1)} to step size s and the error information _{R (t1)} It is calculated by adding the product of the series information u ^* .

  A unified step size is used for the method of calculating the channel estimation value according to equation (3). Since a constant addition process is performed on the channel estimation value at the tap where the delay wave does not exist originally, an error is likely to occur.

On the other hand, in channel estimation unit 20 according to the first embodiment, as shown in equation (2), each step size s is multiplied by each weight value w _j to calculate each second channel estimation value. Ru. That is, the step size s at each tap is substantially variable depending on each weight value w _j .

Also, each weight value w _j is calculated based on the time average value of the first channel estimation value for each incoming wave, as shown in equation (1). Since each weight value is calculated based on the power level for each incoming wave, the first channel estimation value corresponding to the CIR where no incoming wave is present is uniformly updated since the updated value is reduced in equation (2). False updates also decrease. Since the error due to updating is reduced, the accuracy of the replica generated by the sequence estimation equalizer 30 is enhanced, and the performance of the equalizer 10 is improved.

(effect)
Summarizing the above, the transmission path estimator 20 in the first embodiment performs transmission for outputting the transmission path estimated value 21 calculated by the least mean square algorithm to the sequence estimation equalizer 30 which equalizes the received signal 1. The path estimator 20 is configured to generate a plurality of first channel estimation values 211 corresponding to each of a plurality of incoming waves forming the reception signal 1 based on the known sequence signal 1a included in the reception signal 1. A plurality of weight values 225 corresponding to the channel estimation unit (initial channel estimation unit 210) and each of the plurality of incoming waves, for weighting the step size 231 in the least mean square algorithm, the plurality of first transmission channels A plurality of weight generation units 220 generated based on the estimated value 211, a plurality of first channel estimation values 211 and a plurality of weight values 225, a plurality of corresponding to each of a plurality of arrival waves Comprises calculating a second channel estimate 212, second channel estimation unit for outputting the sequence estimation equalizer 30 and (LMS channel estimation unit 230), the.

  The transmission channel estimator 20 in the first embodiment does not determine whether to set the transmission channel estimation value to 0 as in the conventional case, but uses the weight value 225 based on the power level of each incoming wave. The step size 231 is adjusted to adaptively realize channel estimation. The equalizer 10 including such a transmission path estimator 20 prevents the extreme performance deterioration due to the determination error as in the prior art and achieves stable equalization performance. Further, in the feedback process in the process of calculating the channel estimation value 21, unnecessary updating of the channel estimation value 21 can be suppressed.

  Also, weight generation unit 220 of transmission channel estimator 20 in the first embodiment receives, for each incoming wave, the plurality of first transmission channel estimation values 211 generated by initial transmission channel estimation unit 210. A channel estimated value storage unit 2210 that stores discriminatively stored for each reception timing of signal 1, and a plurality of first channel estimation values 211 at the past reception timing stored in the channel estimated value storage unit 2210. And a weight calculation unit 2211 that generates each weight value 225 for each wave.

  With such a configuration, the error of the weight value 225 is reduced, and the performance of the transmission path estimator 20 is stabilized.

  In addition, weight calculation section 2211 of channel estimation unit 20 in the first embodiment reads a plurality of first channel estimation values 211 having different reception timings from channel estimation value storage section 2210, and calculates the time for each arrival wave. And a weight calculation unit 2211 that calculates an average value and generates each weight value 225 for each incoming wave based on the time average value.

  With such a configuration, the error of the weight value 225 is reduced, and the performance of the transmission path estimator 20 is stabilized.

  Also, the equalizer 10 according to the first embodiment equalizes the received signal 1 based on the above-described channel estimation unit 20 and the plurality of second channel estimation values 212 output from the LMS channel estimation unit 230. And a sequence estimation equalizer 30.

  With such a configuration, the equalizer 10 realizes adaptive channel estimation for each incoming wave, that is, for each tap. The equalizer 10 can obtain stable performance.

Second Embodiment
A channel estimator according to the second embodiment will be described. Descriptions of configurations and operations similar to those of the first embodiment will be omitted.

  The transmission path estimator according to the second embodiment can set an incoming wave to which the change of the step size 231 by the weight value 225 is applied among the plurality of incoming waves. Therefore, compared with the channel estimator 20 that applies the weight value 225 to all the incoming waves as described in the first embodiment, the channel estimator in the second embodiment has the channel estimation value 21. Accuracy can be improved. As a result, the quality of channel estimation by the equalizer 10 is improved.

(Constitution)
FIG. 10 is a block diagram showing a configuration of weight generating section 221 included in the transmission path estimator in the second embodiment. The weight generation unit 221 includes a reference arrival wave determination and estimation unit 2212. The reference arrival wave determination / estimation unit 2212 sets a weight value corresponding to the leading wave among the plurality of weight values 225 generated by the weight calculation unit 2211 or a weight value corresponding to the most delayed wave among the plurality of delay waves. Change to a value The leading wave or the most delayed wave is an incoming wave that is a reference of the operation timing of the sequence estimation equalizer 30, that is, the equalization operation. The function of the reference arrival wave determination / estimation unit 2212 is realized by the processing circuit shown in FIG. 6 or 7.

(Operation)
FIG. 11 is a flowchart showing a method of generating weight values in the second embodiment. Steps S310 to S330 are the same as in the first embodiment shown in FIG.

  In step S340, reference incoming wave determination / estimation unit 2212 receives information on a predetermined reference incoming wave indicating the reference of the operation timing of sequence estimation equalizer 30. The predetermined reference arrival wave is, for example, the leading wave or the most delayed wave among a plurality of arrival waves. The information on the predetermined reference arrival wave is stored in advance in, for example, a storage unit (not shown), and the reference arrival wave determination estimation unit 2212 appropriately reads and uses it for the determination.

  In step S350, the reference incoming wave determination / estimation unit 2212 changes the weight value corresponding to the predetermined reference incoming wave to a constant value. For example, when the predetermined reference arrival wave is a leading wave, the reference arrival wave determination and estimation unit 2212 changes the weight value corresponding to the leading wave to one. In addition, the reference arrival wave determination / estimation unit 2212 may also change the weight value corresponding to the most delayed wave to one. Furthermore, the reference arrival wave determination / estimation unit 2212 may also change the weight value corresponding to the delayed wave of the arrival time one before the redelayed wave to one.

  Alternatively, for example, when the predetermined reference arrival wave is the most delayed wave, the reference arrival wave determination and estimation unit 2212 changes the weight value corresponding to the most delayed wave to one. In addition, the reference arrival wave determination / estimation unit 2212 may also change the weight value corresponding to the preceding wave to one. Furthermore, the reference arrival wave determination / estimation unit 2212 may also change the weight value corresponding to the delayed wave of the arrival time one after the preceding wave to one.

(effect)
Summarizing the above, weight generation section 221 of the transmission path estimator in the second embodiment determines a predetermined reference arrival included in a plurality of arrival waves among a plurality of weight values 225 generated by weight calculation section 2211. It further includes a reference incoming wave determination and estimation unit that changes the weight value corresponding to the wave to a constant value. The predetermined reference arrival wave is the leading wave or the most delayed wave that is the reference of the operation timing of the sequence estimation equalizer 30.

  With such a configuration, the weight value is excluded from the incoming wave that is the reference of the operation of the sequence estimation equalizer 30, and the channel estimation value is calculated with the original step size. As a result, the channel estimation accuracy of the channel estimator is improved.

Embodiment 3
A channel estimator according to the third embodiment will be described. Descriptions of configurations and operations similar to those of Embodiment 1 or 2 will be omitted.

(Constitution)
FIG. 12 is a block diagram showing a configuration of weight generation section 222 included in the transmission path estimator in the third embodiment. Weight generation unit 222 includes weight determination unit 2213, update weight value storage unit 2214 and weight update unit 2215 in addition to the configuration of the second embodiment.

  The weight determination unit 2213 determines each weight value 225 generated by the weight calculation unit 2211 with a predetermined threshold value. The predetermined threshold is stored in advance in, for example, a storage unit (not shown), and the weight determination unit 2213 reads out from the storage unit and uses it for determination.

  The update weight value storage unit 2214 stores at least one update weight value determined in advance. The update weight value storage unit 2214 stores, for example, a weight value table including the above-described threshold value and the update weight value corresponding to the threshold value.

  The weight update unit 2215 selects an update weight value from the update weight value storage unit 2214 based on the determination result of the weight determination unit 2213. Then, the weight update unit 2215 updates the one weight value determined by the weight determination unit 2213 with the selected update weight value.

  Each function of weight determination unit 2213, update weight value storage unit 2214 and weight update unit 2215 is realized by the processing circuit shown in FIG. 6 or FIG.

(Operation)
FIG. 13 is a flowchart showing a method of generating weight values in the third embodiment. Steps S310 to S350 are the same as in the second embodiment.

  In step S360, the weight determination unit 2213 determines, for example, whether each weight value 225 is less than or equal to a predetermined threshold. If it is determined that the threshold value is equal to or less than the predetermined threshold value, step S370 is performed. If it is determined that the predetermined threshold value is exceeded, the method of generating weight values ends.

  In step S 370, weight update unit 2215 selects one update weight value from update weight value storage unit 2214, and weight determination unit 2213 determines one weight value determined to be less than or equal to a predetermined threshold value as the weight value. Update with one update weight value.

  In these steps, each weight value 225 may be determined for multiple thresholds. In that case, the update weight value storage unit 2214 stores a weight value table including a plurality of threshold values and a plurality of update weight values corresponding to each of the plurality of threshold values. The weight update unit 2215 selects and updates one update weight value corresponding to each threshold value from the weight value table.

(effect)
Summarizing the above, weight generation section 222 of the transmission path estimator according to the third embodiment includes update weight value storage section 2214 storing update weight values, and weight values 225 generated by weight calculation section 2211. The weight determination unit 2213 selects the update weight value stored in the update weight value storage unit 2214 based on the determination result of the weight determination unit 2213 and the weight determination unit 2213 that determine It further includes a weight update unit 2215 that updates the one weight value with the update weight value.

  With such a configuration, the channel estimator can update abnormal weight values and stabilize the performance of the equalizer.

  Also, the weight update unit 2215 of the transmission channel estimator according to the third embodiment selects an update weight value corresponding to a predetermined threshold from the weight value table stored in the update weight value storage unit 2214.

  According to such a configuration, the transmission path estimator can select the update weight value according to the magnitude of the predetermined threshold which is the determination reference.

  In the present invention, within the scope of the invention, each embodiment can be freely combined, or each embodiment can be appropriately modified or omitted. Although the present invention has been described in detail, the above description is an exemplification in all aspects, and the present invention is not limited thereto. It is understood that countless variations not illustrated are conceivable without departing from the scope of the present invention.

  Reference Signs List 1 received signal, 1a known sequence signal, 10 equalizers, 20 channel estimator, 210 initial channel estimator, 211 first channel estimate, 212 second channel estimate, 220 weight generator, 221 weights Generation unit 222 Weight generation unit 225 Weight value 2210 Transmission path estimated value storage unit 2211 Weight calculation unit 2212 Reference arrival wave judgment estimation unit 2213 Weight judgment unit 2214 Update weight value storage unit 2215 Weight update unit 230 LMS channel estimation unit, 231 step size, 30 sequence estimation equalizer.

Claims (7)

A channel estimator for outputting a channel estimation value calculated by a least mean square algorithm to a sequence estimation equalizer that equalizes a received signal, comprising:
A first channel estimation unit that generates a plurality of first channel estimation values corresponding to each of a plurality of incoming waves forming the received signal based on the known sequence signal included in the received signal;
A weight generation unit that generates, based on the plurality of first channel estimation values, a plurality of weight values corresponding to each of the plurality of incoming waves and weighting the step size in the least mean square algorithm;
A plurality of second channel estimation values corresponding to each of the plurality of incoming waves are calculated based on the plurality of first channel estimation values and the plurality of weight values, and are output to the sequence estimation equalizer. And a second channel estimation unit. The weight generation unit
A channel estimation value storage unit that stores the plurality of first channel estimation values generated by the first channel estimation unit discriminatively for each incoming wave and for each reception timing of the reception signal. When,
A weight calculating unit for generating each weight value for each incoming wave based on the plurality of first channel estimation values at the past reception timing stored in the channel estimation value storage unit; The channel estimator according to 1. The weight calculation unit
The plurality of first channel estimation values at different reception timings are read out from the channel estimation value storage unit, a time average value is calculated for each of the arrival waves, and each of the arrival waves is calculated based on the time average value. The channel estimator according to claim 2, wherein each of the weight values is generated. The weight generation unit
A reference arrival wave determination and estimation unit that changes a weight value corresponding to a predetermined reference arrival wave included in the plurality of arrival waves among the plurality of weight values generated by the weight calculation unit to a constant value; Including
The transmission path estimator according to claim 2 or 3, wherein the predetermined reference incoming wave is a leading wave or a delayed wave which is a reference of operation timing of the sequence estimation equalizer. The weight generation unit
An update weight value storage unit in which the update weight value is stored;
A weight determination unit that determines each of the weight values generated by the weight calculation unit with a predetermined threshold value;
The update weight value stored in the update weight value storage unit is selected based on the determination result in the weight determination unit, and one weight value determined by the weight determination unit is updated with the update weight value. The transmission path estimator according to any one of claims 2 to 4, further comprising: a weight updater. The weight updating unit
The transmission path estimator according to claim 5, wherein the update weight value corresponding to the predetermined threshold is selected from a weight value table stored in the update weight value storage unit. A channel estimator according to any one of claims 1 to 6,
The sequence estimation equalizer which equalizes the received signal based on the plurality of second channel estimation values output from the second channel estimation unit.

Images