CN113242189B - Adaptive equalization soft information iteration receiving method combined with channel estimation - Google Patents

Adaptive equalization soft information iteration receiving method combined with channel estimation Download PDF

Info

Publication number
CN113242189B
CN113242189B CN202110392636.5A CN202110392636A CN113242189B CN 113242189 B CN113242189 B CN 113242189B CN 202110392636 A CN202110392636 A CN 202110392636A CN 113242189 B CN113242189 B CN 113242189B
Authority
CN
China
Prior art keywords
equalization
filter
channel
output
adaptive
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202110392636.5A
Other languages
Chinese (zh)
Other versions
CN113242189A (en
Inventor
余华
陈奕毅
季飞
陈芳炯
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
South China University of Technology SCUT
Original Assignee
South China University of Technology SCUT
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by South China University of Technology SCUT filed Critical South China University of Technology SCUT
Priority to CN202110392636.5A priority Critical patent/CN113242189B/en
Publication of CN113242189A publication Critical patent/CN113242189A/en
Application granted granted Critical
Publication of CN113242189B publication Critical patent/CN113242189B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L25/00—Baseband systems
    • H04L25/02—Details ; arrangements for supplying electrical power along data transmission lines
    • H04L25/03—Shaping networks in transmitter or receiver, e.g. adaptive shaping networks
    • H04L25/03006—Arrangements for removing intersymbol interference
    • H04L25/03178—Arrangements involving sequence estimation techniques
    • H04L25/03312—Arrangements specific to the provision of output signals
    • H04L25/03318—Provision of soft decisions
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L1/00—Arrangements for detecting or preventing errors in the information received
    • H04L1/0001—Systems modifying transmission characteristics according to link quality, e.g. power backoff
    • H04L1/0009—Systems modifying transmission characteristics according to link quality, e.g. power backoff by adapting the channel coding
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L1/00—Arrangements for detecting or preventing errors in the information received
    • H04L1/004—Arrangements for detecting or preventing errors in the information received by using forward error control
    • H04L1/0045—Arrangements at the receiver end
    • H04L1/0047—Decoding adapted to other signal detection operation
    • H04L1/005—Iterative decoding, including iteration between signal detection and decoding operation
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L1/00—Arrangements for detecting or preventing errors in the information received
    • H04L1/004—Arrangements for detecting or preventing errors in the information received by using forward error control
    • H04L1/0056—Systems characterized by the type of code used
    • H04L1/0071—Use of interleaving
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L25/00—Baseband systems
    • H04L25/02—Details ; arrangements for supplying electrical power along data transmission lines
    • H04L25/0202—Channel estimation
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L25/00—Baseband systems
    • H04L25/02—Details ; arrangements for supplying electrical power along data transmission lines
    • H04L25/03—Shaping networks in transmitter or receiver, e.g. adaptive shaping networks
    • H04L25/03006—Arrangements for removing intersymbol interference
    • H04L25/03012—Arrangements for removing intersymbol interference operating in the time domain
    • H04L25/03019—Arrangements for removing intersymbol interference operating in the time domain adaptive, i.e. capable of adjustment during data reception
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L25/00—Baseband systems
    • H04L25/02—Details ; arrangements for supplying electrical power along data transmission lines
    • H04L25/03—Shaping networks in transmitter or receiver, e.g. adaptive shaping networks
    • H04L25/03006—Arrangements for removing intersymbol interference
    • H04L25/03178—Arrangements involving sequence estimation techniques
    • H04L25/03248—Arrangements for operating in conjunction with other apparatus
    • H04L25/03254—Operation with other circuitry for removing intersymbol interference
    • H04L25/03267—Operation with other circuitry for removing intersymbol interference with decision feedback equalisers
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L25/00—Baseband systems
    • H04L25/02—Details ; arrangements for supplying electrical power along data transmission lines
    • H04L25/03—Shaping networks in transmitter or receiver, e.g. adaptive shaping networks
    • H04L25/03006—Arrangements for removing intersymbol interference
    • H04L25/03178—Arrangements involving sequence estimation techniques
    • H04L25/03248—Arrangements for operating in conjunction with other apparatus
    • H04L25/03292—Arrangements for operating in conjunction with other apparatus with channel estimation circuitry
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L25/00—Baseband systems
    • H04L25/02—Details ; arrangements for supplying electrical power along data transmission lines
    • H04L25/03—Shaping networks in transmitter or receiver, e.g. adaptive shaping networks
    • H04L25/03006—Arrangements for removing intersymbol interference
    • H04L2025/03592—Adaptation methods
    • H04L2025/03598—Algorithms
    • H04L2025/03611—Iterative algorithms

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Power Engineering (AREA)
  • Quality & Reliability (AREA)
  • Filters That Use Time-Delay Elements (AREA)

Abstract

The invention discloses a self-adaptive equalization soft information iteration receiving method combined with channel estimation, which is applicable to a single carrier transmission system in the field of underwater acoustic communication, and is used for estimating a channel in a time domain, performing time domain and frequency domain equalization on a signal according to channel impulse response, weighting and combining preliminary equalization results to serve as an expected signal of a self-adaptive algorithm, and performing direct self-adaptive equalization on the signal, wherein the decoding performance of a direct equalization stage can be obviously improved; meanwhile, a turbo equalization structure is adopted, soft information is iteratively exchanged between the equalizer and the decoder so as to fully extract the error correction gain of channel coding, and the error rate of the system is gradually reduced on the whole. Compared with other equalization methods, the method can effectively eliminate the intersymbol interference caused by the underwater acoustic channel, can better track the change of the channel along with the time, and has the advantage of lower calculation complexity.

Description

Adaptive equalization soft information iteration receiving method combined with channel estimation
Technical Field
The invention relates to the technical field of mobile communication. In particular to a self-adaptive equalization soft information iteration receiving method combined with channel estimation.
Background
Due to the characteristics of severe multipath effect, fast time variation, limited bandwidth and the like of the underwater acoustic channel, the underwater acoustic communication has great challenge. In recent years, the research direction of underwater acoustic communication technology is mainly to improve the frequency band utilization rate and the information transmission rate, and coherent communication technology has advantages over incoherent technology, and meanwhile, in order to ensure that reliable communication performance can be maintained in a low signal-to-noise ratio environment, the complexity of an algorithm is continuously increased, and the real-time performance of communication is affected.
In an underwater acoustic communication system, the process of sound waves from a sending end to a receiving end is influenced by refraction, reflection, scattering and the like, so that multipath fading is caused; meanwhile, due to the relative movement between the two communication parties, the doppler effect is formed, so that the signal processing is more difficult. In a single carrier transmission system, a transmission signal generates serious amplitude and phase distortion under the influence of a time-varying multipath channel, and introduces self intersymbol interference. To eliminate Inter-Symbol Interference, more complex equalization techniques are used to eliminate Inter Symbol Interference (ISI).
In the current communication system, there are two most commonly used equalization techniques, which are respectively a Turbo equalization technique (CE-TEQ) based on Channel Estimation and a Turbo equalization technique (DA-TEQ) based on Direct Adaptation, and the two methods have advantages and disadvantages, but the primary purpose is to eliminate ISI and improve the error performance of the system. In recent years, a turbo equalization structure is widely used in an underwater acoustic communication system, more coding gains are gradually obtained in a soft information iteration mode, the overall performance is improved, but the demodulation period is increased in the iteration mode, so that the real-time performance is affected.
In order to make a compromise between the complexity and the real-time performance of the algorithm, ensure the equalization reliability of a transmission system, eliminate ISI in the signal transmission process and enable the system to obtain a lower error rate. It is necessary to consider a new receiver design method, and it is urgent to provide an iterative receiving method of adaptive equalization soft information in combination with channel estimation.
Disclosure of Invention
The present invention aims to solve the above-mentioned drawbacks in the prior art, and provides an adaptive equalization soft information iterative receiving method combined with channel estimation. The method is a receiver equalization method which compromises the complexity and real-time performance of the implementation method, and aims at a single carrier transmission system under an underwater acoustic channel, a method based on channel estimation is firstly used for carrying out preliminary equalization on signals in a time domain and a frequency domain, an expected signal is provided for a direct equalization stage of self-adaptive equalization, and an iterative equalization stage obtains error correction gain by utilizing a soft information iteration process between an equalizer and a decoder, so that information loss caused by hard decision is avoided; by improving the reliability of the result of the direct equalization stage, the convergence rate of the iterative process is accelerated, and the overall performance of the single carrier transmission system under the underwater acoustic channel is improved.
The purpose of the invention can be achieved by adopting the following technical scheme:
an adaptive equalization soft information iteration receiving method combined with channel estimation, the receiving method comprises the following steps:
s1, the receiver obtains a frame of received data after frame synchronization, Doppler estimation and compensation operation, and records the frame as a received signal frame y, wherein the received signal frame y comprises NmaxA data block and Nmax+1 training sequences, per data block
Figure GDA0003468722820000021
Comprising NdA received symbol, each training sequence
Figure GDA0003468722820000022
Containing NtA received symbol [. ]]TThe method comprises the steps of performing transposition on vectors, starting a received signal frame y with a training sequence, arranging data blocks and the training sequence in sequence, connecting a training sequence to the front and the back of each data block respectively, respectively referring to a forward training sequence and a backward training sequence of the data block, and extracting a forward training sequence t of an nth data block r from the received signal frame ynAnd a backward training sequence tn+1The impulse response estimation is carried out by adopting a channel estimation algorithm to respectively obtain estimated channels
Figure GDA0003468722820000031
And
Figure GDA0003468722820000032
estimating the length of a channel to be L, namely comprising L tap coefficients;
s2, extracting the nth data block r from the received signal frame y, and using the estimated channel
Figure GDA0003468722820000033
And
Figure GDA0003468722820000034
carrying out frequency domain equalization on the data block r by adopting a maximum ratio combining method to obtain a transmitting symbol sequence
Figure GDA0003468722820000035
Figure GDA0003468722820000035
1 st estimation of
Figure GDA0003468722820000036
The transmission symbol sequence x contains NdA symbol;
s3, estimating the channel according to the estimated channel
Figure GDA0003468722820000037
And
Figure GDA0003468722820000038
separately solving the corresponding forward filters w in the time domain1And a backward filter w2Forward filter w1And a backward filter w2Length L, and then respectively performing time domain equalization on the data blocks r to obtain forward time domain equalization output
Figure GDA0003468722820000039
And backward time domain equalization output
Figure GDA00034687228200000310
Weighted combining
Figure GDA00034687228200000311
And
Figure GDA00034687228200000312
then obtaining the 2 nd estimation of the transmitted symbol sequence x
Figure GDA00034687228200000313
S4, 1 st estimation of symbol sequence x is transmitted
Figure GDA00034687228200000314
And 2 nd estimation
Figure GDA00034687228200000315
Performing equal proportion combination, inputting the hard decision signal into an adaptive equalizer as an expected signal, and obtaining forward adaptive equalization output through forward and reverse adaptive equalization
Figure GDA00034687228200000316
And reverse adaptive equalization output
Figure GDA00034687228200000317
Weighted combining
Figure GDA00034687228200000318
And
Figure GDA00034687228200000319
then the 3 rd estimation of the transmitted symbol sequence x is obtained
Figure GDA00034687228200000320
Last pair of
Figure GDA00034687228200000321
And
Figure GDA00034687228200000322
merging to obtain balanced output of data block r
Figure GDA00034687228200000323
S5, outputting the balance
Figure GDA00034687228200000324
Equalized symbol at time k
Figure GDA00034687228200000325
Mapping to obtain the external information of the equalizer
Figure GDA00034687228200000326
Figure GDA00034687228200000327
To represent
Figure GDA00034687228200000328
The jth bit carried, k 0, …, NdThe value range of-1, j depends on the modulation mode of the symbol,
Figure GDA00034687228200000329
after de-interleaving operation, as prior information
Figure GDA00034687228200000330
To a decoder, bmAn mth bit representing a sequence of bits carried by the data block;
s6, decoder according to prior information
Figure GDA00034687228200000331
Extracting error correction gain of channel coding and outputting posterior information LD(bm) Subtracting the prior information from the posterior information to obtain the external information of the decoder
Figure GDA00034687228200000332
S7, and the posterior information LD(bm) Decoding and checking, when decoding is correct or current iteration number Iter reaches maximum iteration number ItermaxIf so, the demodulation of the nth data block is completed, and then n is equal to n +1 and Iter is equal to 0, and the process returns to step S1 to perform the processing of the next data block until the demodulation of all data blocks in the received signal frame y is completed; otherwise, let Iter be Iter +1, execute step S8, and enter the iterative equalization stage;
s8, decoding the extrinsic information
Figure GDA00034687228200000333
Interleaving as a priori information of the equalizer
Figure GDA0003468722820000041
Figure GDA0003468722820000042
To represent
Figure GDA0003468722820000043
The jth bit that is carried over is,
Figure GDA0003468722820000044
is that
Figure GDA0003468722820000045
Mapping the obtained k-th prior symbol, wherein the prior symbol forms a prior input in a vector form
Figure GDA0003468722820000046
Figure GDA0003468722820000047
Is the feedback filter input signal, a priori symbol, of an adaptive equalizer
Figure GDA0003468722820000048
Will be output in balance with
Figure GDA0003468722820000049
Computing together the desired signals of an adaptive algorithm
Figure GDA00034687228200000410
S9, forming the equalizer of the iterative equalization stage by the feedforward filter and the feedback filter, using the adaptive algorithm, and using the equalization output
Figure GDA00034687228200000411
And a desired signal
Figure GDA00034687228200000412
Updating filter coefficient to obtain balanced output
Figure GDA00034687228200000413
Then, the process goes to step S5 to re-enter the decoding stage.
Further, in step S1, the training sequence in the signal frame y is received
Figure GDA00034687228200000414
Is made up of known training sequences of the transmitting end
Figure GDA00034687228200000415
Obtained after passing through a channel and containing NtA forward training sequence t is extracted from the received signal frame ynAccording to the known forward training sequence z of the transmitting endnSolving the forward estimated channel by a channel estimation algorithm
Figure GDA00034687228200000416
Followed by extraction of the backward training sequence tn+1According to the known backward training sequence z of the transmitting endn+1Solving for the backward estimated channel
Figure GDA00034687228200000417
Further, in step S2, the nth data block r is extracted from the received signal frame y, and r is transformed by discrete fourier transform,
Figure GDA00034687228200000418
And
Figure GDA00034687228200000419
transforming from time domain to frequency domain to obtain frequency domain data block R and frequency domain channel response
Figure GDA00034687228200000420
And
Figure GDA00034687228200000421
and (2) performing frequency domain equalization by adopting a maximum ratio combining method represented by the formula (1) according to the frequency domain data block and the frequency domain channel response:
Figure GDA00034687228200000422
wherein J represents the number of frequency domain channel responses, and is output to frequency domain equalization
Figure GDA00034687228200000423
Inverse discrete Fourier transform to obtain the 1 st estimate of the transmitted symbol sequence x
Figure GDA00034687228200000424
Further, in step S3, the estimated channel in the forward direction is used
Figure GDA00034687228200000425
Solving for the forward filter w according to equation (2)1：
Figure GDA0003468722820000051
In the formula (I), the compound is shown in the specification,
Figure GDA0003468722820000052
as a variance of the noise, IMIs an M-order identity matrix, where M ═ N1+N2+1，N1And N2Respectively causal and non-causal parts of the filter, s is the channel convolution matrix H1N of (2)2+ L columns, channel convolution matrix H1By estimating the channel
Figure GDA0003468722820000053
Is constructed by the following steps:
Figure GDA0003468722820000054
wherein the content of the first and second substances,
Figure GDA0003468722820000055
respectively representing estimated channels
Figure GDA0003468722820000056
L tap coefficients of (i)
Figure GDA0003468722820000057
Obtaining a forward filter w1The forward time domain equalization is then represented as:
Figure GDA0003468722820000058
in the formula, rkIs the input signal at the time of the k-th instant,
Figure GDA0003468722820000059
is the forward equalization symbol at the k-th time, and the equalization symbols at all times of the data block r are obtained
Figure GDA00034687228200000510
Forward time domain equalization outputs combined into vector form
Figure GDA00034687228200000511
Then, the forward estimated channel is transmitted
Figure GDA00034687228200000512
Estimation channel converted into backward direction
Figure GDA00034687228200000513
Calculating a backward filter w2Calculating the backward balanced symbol of all the time of the data block r
Figure GDA00034687228200000514
Backward time domain equalization output in the form of a component vector
Figure GDA00034687228200000515
Forward time domain equalization output using equation (4)
Figure GDA00034687228200000516
And backward time domain equalization output
Figure GDA00034687228200000517
And (3) carrying out weighted combination, wherein beta is a weighting coefficient, and calculating to obtain a 2 nd estimation of the transmitted symbol sequence x:
Figure GDA00034687228200000518
further, in step S4, direct equalization is performed, where the direct equalization stage only has a feedforward filter f, the length of the filter is M, and w1And w2Similarly, for the forward adaptive equalization, the equalization is divided into a training stage and a decision stage, and the equalization output formula of the two stages is as follows:
Figure GDA00034687228200000519
in the formula (I), the compound is shown in the specification,
Figure GDA0003468722820000061
is the input signal of the feed-forward filter,
Figure GDA0003468722820000062
for the balance output at the kth moment, an adaptive algorithm is required to be used for updating the filter coefficient, a Normalized Least Mean Square (NLMS) algorithm is adopted, and a coefficient updating formula is as follows:
Figure GDA0003468722820000063
in the formula (f)kIs a feedforward filter at the k-th moment, xi is a convergence factor, epsilon is a normal number with a small value,
Figure GDA0003468722820000064
for the desired signal at the k-th moment, in the training phase
Figure GDA0003468722820000065
For training sequences, in the decision phase
Figure GDA0003468722820000066
According to
Figure GDA0003468722820000067
And
Figure GDA0003468722820000068
and performing hard decision, wherein the calculation method comprises the following steps:
Figure GDA0003468722820000069
in the formula (I), the compound is shown in the specification,
Figure GDA00034687228200000610
and
Figure GDA00034687228200000611
are respectively
Figure GDA00034687228200000612
And
Figure GDA00034687228200000613
the equalization value at the k-th moment, Q (-) operation represents that hard decision is carried out on the equalization symbol;
after the output calculation at all times of the nth data block is completed, k is equal to 0, …, NdEqualized output at time-1
Figure GDA00034687228200000614
Forward adaptive equalization output in the form of component vectors
Figure GDA00034687228200000615
Final feedforward filter with preserving forward adaptive equalization
Figure GDA00034687228200000616
As the initial value of the feedforward filter in the iterative equalization stage;
reverse adaptive equalization using a backward training sequence tn+1Solving for reverse adaptive equalization outputs
Figure GDA00034687228200000617
The difference from the forward adaptive equalization is that the input and output of the equalizer are time reversed, preserving the reverse selfFeed forward filter for adaptive equalization
Figure GDA00034687228200000618
By combining in equal proportions
Figure GDA00034687228200000619
And
Figure GDA00034687228200000620
the combining coefficient γ is 1/2, yielding the 3 rd estimate of the transmitted symbol sequence x:
Figure GDA00034687228200000621
3 estimates of the symbol sequence x to be transmitted
Figure GDA00034687228200000622
And
Figure GDA00034687228200000623
performing weighting combination to obtain balance output of direct balance stage
Figure GDA00034687228200000624
Figure GDA00034687228200000625
In the formula, alpha1，α2And alpha 33 estimates respectively
Figure GDA00034687228200000626
And
Figure GDA00034687228200000627
the weighting coefficient of (2).
Further, in step S5, the equalization is output
Figure GDA0003468722820000071
Is mapped as outerThe information needs to approximate the statistical model parameters mu and delta by adopting a time averaging method2Mu is the scaling factor of the transmitted symbol sequence x, delta2Then the variance of x, and then the probability value of the symbol is calculated
Figure GDA0003468722820000072
aiIs the ith element of the transmitted symbol set, the number of symbols of the transmitted symbol set depends on the modulation mode, and then the extrinsic information output by the equalizer is obtained
Figure GDA0003468722820000073
Figure GDA0003468722820000074
Performing de-interleaving operation to obtain prior information of decoder
Figure GDA0003468722820000075
Further, in step S6, a priori information is obtained at the decoder
Figure GDA0003468722820000076
Under the guidance of (2), the decoder extracts the error correction gain of the channel coding and outputs a posteriori information LD(bm) The external information of the decoder is obtained by deducting the prior information from the posterior information
Figure GDA0003468722820000077
The calculation formula is as follows:
Figure GDA0003468722820000078
further, in step S7, the posterior information L is processedD(bm) Judging decoding, and judging whether the decoding result is correct according to the error detection code; if the decoding is correct, the decoding process of the current data block is exited, a result is output, n is equal to n +1, the iteration number Iter is equal to 0, and the process returns to step S1 to process the next data block until the received signal is processed; in decodingFailure and current iteration number less than the maximum iteration number ItermaxWhen it is determined that the term "Iter" is equal to term +1, the process proceeds to step S8, and an iterative operation is performed.
Further, in step S8, extrinsic information of the decoder
Figure GDA0003468722820000079
Interleaving as a priori information of the equalizer
Figure GDA00034687228200000710
Mapping prior information at the k-th time into prior symbols
Figure GDA00034687228200000711
Constituent prior inputs
Figure GDA00034687228200000712
As the feedback filter input signal for the adaptive equalizer.
Further, in step S9, iterative equalization is performed, and the equalizer includes a feedforward filter f and a feedback filter b, and remains after the previous equalization
Figure GDA00034687228200000713
And
Figure GDA00034687228200000714
initialization is performed such that the feedback filter does not retain coefficients during the direct equalization phase
Figure GDA00034687228200000715
Then, setting the equalizer as a zero vector, and the output of the equalizer in the training stage and the decision stage is:
Figure GDA00034687228200000716
in the formula (f)kA feedforward filter at the k-th instant, bkIs a feedback filter at the k-th time, rkIs the input signal to the feedforward filter at time k,
Figure GDA0003468722820000081
feeding back an input signal of the filter at the k time; for updating the self-adaptive filter coefficient, NLMS algorithm is adopted, and feedback filter bkThe update formula is:
Figure GDA0003468722820000082
in the training phase, a signal is expected
Figure GDA0003468722820000083
For training sequence tnIn the decision phase, by equalizing the symbols
Figure GDA0003468722820000084
And the a priori symbol of the feedback
Figure GDA0003468722820000085
And after merging, carrying out hard decision to obtain:
Figure GDA0003468722820000086
the iterative equalization of the data block is completed, k is equal to 0, …, NdEqualized symbols at-1 time instant
Figure GDA0003468722820000087
Forming a vector to obtain a forward adaptive equalization output
Figure GDA0003468722820000088
Feed forward filter with retention of last update
Figure GDA0003468722820000089
And a feedback filter
Figure GDA00034687228200000810
Feedforward filter f for inverse adaptive equalizationk'and feedback Filter b'kAccording to retention coefficient
Figure GDA00034687228200000811
And
Figure GDA00034687228200000812
initializing, and performing time reversal on input and output of data in the equalization process to obtain reverse self-adaptive equalization output
Figure GDA00034687228200000813
Feed forward filter with retention of last update
Figure GDA00034687228200000814
And a feedback filter
Figure GDA00034687228200000815
Providing an initial value for the next iteration balance;
to the forward direction adaptive equalization output
Figure GDA00034687228200000816
And reverse adaptive equalization output
Figure GDA00034687228200000817
Merging in equal proportion mode to obtain the balanced output of the iterative balance
Figure GDA00034687228200000818
Then, the process returns to step S5 to enter the decoding stage.
Compared with the prior art, the invention has the following advantages and effects:
1. the present invention provides an initial desired signal for an adaptive equalizer through a preliminary equalization based on channel estimation. The desired signal of the DA-TEQ is provided by its own balanced output, which is prone to error propagation. Compared with DA-TEQ, the adaptive equalizer combined with channel estimation can more quickly converge to an optimal solution under the guidance of an expected signal, the error propagation effect caused by the fact that the equalizer provides the expected signal is reduced, more accurate equalization output can be obtained in a direct equalization stage, and the error rate can be more quickly reduced in a subsequent iterative equalization stage;
2. in the process of combining the channel estimation equalization, the invention only uses the channel estimation equalization in the direct equalization stage. Compared with CE-TEQ, the method can avoid complex operation required by updating the filter coefficient in the iteration stage, and can be applied to scenes with high real-time requirement by combining a self-adaptive equalization mode without performing operations such as matrix multiplication, inversion and the like.
3. The invention combines the advantages of the DA-TEQ mode and the CE-TEQ mode, utilizes the characteristic that the self-adaptive algorithm tracks the change of the channel in the equalization process, is more suitable for being applied to a rapid time-varying underwater acoustic channel compared with the CE-TEQ mode, and simultaneously, the initial expected signal provided by the channel estimation equalization can enable the self-adaptive equalization to obtain the error rate lower than that of the DA-TEQ mode under the same iteration times.
Drawings
Fig. 1 is a flowchart of an adaptive equalization soft information iterative receiving method combined with channel estimation according to the present invention;
fig. 2 is a schematic diagram of a transmission frame structure of an adaptive equalization soft information iterative receiving method combined with channel estimation according to the present invention;
fig. 3 is a schematic diagram of a received frame structure of an adaptive equalization soft information iterative receiving method combined with channel estimation according to the present invention;
fig. 4 is a system structure diagram of an adaptive equalization soft information iterative receiving method combined with channel estimation according to the present invention;
fig. 5 is a bit error rate comparison diagram of an adaptive equalization soft information iterative receiving method combined with channel estimation and other equalization methods according to the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Examples
In order to facilitate understanding of the subsequent receiving method, a brief description is first made on a signal model of a communication system. Information bit dkGenerating a transmission signal after coding interleaving, baseband modulation and molding filtering:
Figure GDA0003468722820000101
wherein x isnIs constellation symbol obtained by modulating coded bit by baseband, g (T) is a shaping filter, and the symbol period is T, EsRepresenting the energy of the symbol. The signal s (t) is modulated by a carrier and then transmitted.
At a receiving end, after signals are subjected to synchronization, matched filtering and down-sampling, baseband signals are obtained, an underwater acoustic channel impulse response is modeled into a time-invariant finite impulse response filter, and then the received signals can be expressed as follows:
Figure GDA0003468722820000102
in the above formula, the channel impulse response length is L, and the symbol y is receivednI.e. including the transmitted symbol x at the current timenThere are also symbols at previous times, which indicate that the signal has introduced intersymbol interference (ISI) after passing through the channel, and white gaussian noise vn。
The iterative receiving method for adaptive equalization soft information combined with channel estimation proposed by this embodiment is based on the signal model as described above.
The specific parameters of the system model are as follows: the information bit is coded by using a recursive systematic convolutional code with a code rate of 1/2, the generator polynomial is (5,7), 12kHz is selected as the carrier frequency of the system, the sampling rate is 96kHz, a QPSK modulation mode is adopted for a transmission symbol, the symbol duration is about 166.67us, and the length of the symbol is 16 sampling points.
Referring to fig. 1, fig. 2, fig. 3 and fig. 4, fig. 1 is a flowchart of a method in the present embodiment, fig. 2 is a schematic diagram of a transmission frame structure of a signal in the present embodiment, fig. 3 is a schematic diagram of a reception frame structure in the present embodiment, and fig. 4 is a schematic diagram of a system structure in the present embodiment.
The meanings of the indices are as follows:
n: the data block number indicates the nth data block currently processed, and the initial value of n is 1 in this embodiment.
Nmax: the number of data blocks carried by the signal frame, N in this embodimentmax＝4。
Iter: the initial value of Iter in this embodiment is 0 for the current iteration number.
Itermax: maximum number of iterations, Iter in this examplemax＝3。
Nt: training sequence symbol length, N in this examplet＝256。
Nd: symbol length of data block, N in this embodimentd＝2048。
Ns: sum of training sequence and data block symbol length, Ns＝Nt+NdIn this embodiment, Ns＝2304。
zn：
Figure GDA0003468722820000111
And the nth training sequence of the signal frame at the transmitting end is shown.
xn：
Figure GDA0003468722820000112
The nth data block of the signal frame at the transmitting end is shown, and the baseband symbol modulation uses QPSK modulation in the embodiment.
tn：
Figure GDA0003468722820000113
Indicating the nth training sequence of the signal frame at the receiving end.
rn：
Figure GDA0003468722820000114
Representing the nth data block of the signal frame at the receiving end, r is used instead of r in the description process for simplifying the expressionn。
y: representing a complete received signal frame, containing a plurality of data blocks and a plurality of training sequences.
Figure GDA0003468722820000115
Figure GDA0003468722820000116
In the present embodiment, different channel impulse response estimates are distinguished by the labels at the lower left corner.
L: the length of the channel impulse response, L is 121 in this embodiment.
Figure GDA0003468722820000117
Figure GDA0003468722820000118
And (4) equalized output of the data block symbols.
Figure GDA0003468722820000121
And the prior symbol vector representing the data block is obtained by interleaving and mapping the extrinsic information output by the decoder.
f：
Figure GDA0003468722820000122
Representing a feed forward equalizer, different moments being distinguished by the reference k in the lower right hand corner, the filter length being M, M being N1+N2+1，Wherein N is2For the non-causal part length of the filter, N1For the length of the causal part, N in this example1＝N2＝60。
b：
Figure GDA0003468722820000123
Representing a feedback equalizer, the different times are distinguished by the reference k in the lower right corner, and the filter length is also M-N1+N2+1, N in this example1＝N2＝60。
In the iterative receiving method of adaptive equalization soft information combined with channel estimation according to this embodiment, for a received signal frame y, the received signal frame y is composed of a plurality of data blocks, the data blocks are separated by a training sequence, and the received signal frame y starts and ends with the training sequence, as shown in fig. 2. In this embodiment, in the processing process, all data blocks of the signal frame need to be traversed, the data block r is subjected to balanced decoding by using the forward and backward training sequence of the nth frame, information carried by the received signal frame y is solved, an initial value of n is 1, and as the processing process progresses, the value of n is continuously increased until the signal demodulation is finished.
An implementation flow and a system structure diagram of an adaptive equalization soft information iterative receiving method combined with channel estimation are shown in fig. 1 and fig. 4, and specifically include the following steps:
s1, extracting training sequence t from received signal frame ynI.e. the forward training sequence of the nth data block, the length of the training sequence being Nt. According to the known training sequence z of the transmitting endnCalculating the estimated channel of forward direction by adopting a matching pursuit algorithm
Figure GDA0003468722820000124
The training sequence t is then extracted from the received signal frame yn+1I.e. the backward training sequence of the nth data block, based on the known training sequence z of the transmitting endn+1Calculating the estimated channel of the backward direction
Figure GDA0003468722820000125
S2, extracting the nth data block r from the received signal frame y, estimating the channel by using discrete Fourier transform r
Figure GDA0003468722820000126
And
Figure GDA0003468722820000127
transforming from time domain to frequency domain to obtain R, R,
Figure GDA0003468722820000128
And
Figure GDA0003468722820000129
the frequency domain equalization formula is as follows:
Figure GDA0003468722820000131
where J denotes the number of available channel impulse responses, and J is 2 in this embodiment. For the equalization result
Figure GDA0003468722820000132
Inverse discrete Fourier transform to obtain the 1 st estimate of the transmitted symbol sequence x
Figure GDA0003468722820000133
S3, estimating channel according to step S1 solution
Figure GDA0003468722820000134
Solving for the forward filter w1：
Figure GDA0003468722820000135
In the formula (I), the compound is shown in the specification,
Figure GDA0003468722820000136
for the variance of Gaussian white noise, in this embodiment, the variance of Gaussian white noise is usedEstimation of the guard interval preceding the signal frame, IMIs an identity matrix, M ═ N1+N2+1，N1As a causal part of the filter, N2For the non-causal part of the filter, N1＝N260, 121, s is the channel convolution matrix H1N of (2)2+ 181 columns, channel convolution matrix H1By estimating the channel
Figure GDA0003468722820000137
Is constructed of and is shown as
Figure GDA0003468722820000138
Obtaining a forward filter w1The forward time domain equalization is then represented as:
Figure GDA0003468722820000139
in the formula (I), the compound is shown in the specification,
Figure GDA00034687228200001310
denotes the forward equalized symbol at time k, rkRepresenting the input signal of the filter at the k-th instant, according to a forward filter w12048 balanced symbols at all the time of the nth data block are obtained, and forward balanced output in a vector form is formed
Figure GDA00034687228200001311
Will be the front
Figure GDA00034687228200001312
Estimating channel in reverse direction
Figure GDA00034687228200001313
Solving for the backward filter w according to equation (B)2. Computing an equalized symbol for each time instant of the nth data block
Figure GDA00034687228200001314
Backward balanced output in the form of a recomposition vector
Figure GDA00034687228200001315
The forward time domain is output in an equal proportion mode according to a formula (D)
Figure GDA0003468722820000141
And backward time domain equalization output
Figure GDA0003468722820000142
Combining is performed, let β be 1/2, and the 2 nd estimate of the transmitted symbol sequence x is obtained:
Figure GDA0003468722820000143
s4, the filter of the direct equalization stage is only a feedforward filter fk. For forward adaptive equalization, the equalization process is divided into two phases, a training phase and a decision phase. At the initial moment of the training phase
Figure GDA0003468722820000144
(i.e. k ═ N)tTime) is set to zero vector, NtEqualized symbols of time instants
Figure GDA0003468722820000145
Equal to 0. The balanced output of the training phase and the decision phase is shown as formula (E):
Figure GDA0003468722820000146
in the formula, rkThen it is the input signal to the feedforward filter at time k. Obtaining a k time equalized symbol
Figure GDA0003468722820000147
Then, the filter is updatedCoefficient fkAnd the updating process adopts NLMS algorithm, and the coefficient updating formula is as follows:
Figure GDA0003468722820000148
where xi is the convergence factor, epsilon is a small number of normal, epsilon is 0.001 in this example,
Figure GDA0003468722820000149
the signal is desired at time k. In the training phase
Figure GDA00034687228200001410
For training sequences, in the decision phase
Figure GDA00034687228200001411
And carrying out hard decision according to the balanced output of the frequency domain and the time domain, wherein the calculation method comprises the following steps:
Figure GDA00034687228200001412
in the formula (I), the compound is shown in the specification,
Figure GDA00034687228200001413
and
Figure GDA00034687228200001414
respectively 1 st estimate of the transmitted symbol sequence x
Figure GDA00034687228200001415
And 2 nd estimation
Figure GDA00034687228200001416
The equalized symbol at time k. For DA-TEQ, the desired signal in the decision phase is an equalized symbol of itself
Figure GDA00034687228200001417
Hard decision is made, which is reliableThe degree is limited by the accuracy of the output symbol, because the filter in the direct equalization stage has no available prior information, the equalization output in the stage has a large error, and the desired signal is obtained through the equalization symbol, so that the reliability is greatly reduced, the performance of the subsequent equalizer is easily deteriorated, and an error propagation effect is caused. Therefore, the time domain and frequency domain equalization result is introduced as the expected signal, and the confidence degree of the equalization result is far higher than the equalization output of the adaptive filter, so that the convergence process of the adaptive algorithm can be accelerated, and a lower initial error rate can be obtained in the direct equalization stage.
After equalization, the decision stage k is set to 0, …, NdEqualized symbols at-1 time instant
Figure GDA0003468722820000151
Forward adaptive equalization output in the form of component vectors
Figure GDA0003468722820000152
Feed forward filter retaining the last of the direct equalization stages
Figure GDA0003468722820000153
As an initial value for the feedforward filter during the training phase of the iterative equalization.
Inverse adaptive equalization using a feedforward filter fk' the input signal of the filter needs to be time-reversed, and the equalized symbol at each time is obtained by equalization
Figure GDA0003468722820000154
Setting the decision phase k to 0, …, NdEqualized symbols at-1 time instant
Figure GDA0003468722820000155
Forming a vector form, and performing time reversal to obtain reverse self-adaptive equalization output
Figure GDA0003468722820000156
Last feedforward filter of reverse self-adaptive equalization
Figure GDA0003468722820000157
It also needs to be preserved that the iterative equalization phase will be used for initialization.
By combining in equal proportions
Figure GDA0003468722820000158
And
Figure GDA0003468722820000159
let γ be 1/2, the 3 rd estimate of the adaptively equalized transmitted symbol sequence x is:
Figure GDA00034687228200001510
finally, 3 estimates of the symbol sequence will be transmitted
Figure GDA00034687228200001511
And
Figure GDA00034687228200001512
merging is carried out, and the formula is as follows:
Figure GDA00034687228200001513
in the formula, alpha1＝α2＝α3The results are combined as the equalization output of the direct equalization stage, 1/3.
S5, in the embodiment, assuming that the received symbols obey Gaussian distribution, the parameters of the probability model are obtained by estimation, wherein the key parameter mukAnd
Figure GDA00034687228200001514
respectively representing the scaling factor and the variance of the transmitted symbols. In this embodiment, a time-averaging method is adopted to calculate the model parameters, and the calculation formula is as follows:
Figure GDA00034687228200001515
Figure GDA00034687228200001516
after model parameters are obtained, the conditional probability is calculated
Figure GDA00034687228200001517
And then, calculating the external information of the equalizer according to Bayesian theorem
Figure GDA0003468722820000161
The demapping calculation formula is as follows:
Figure GDA0003468722820000162
external information
Figure GDA0003468722820000163
Deinterleaving is performed and the result is used as a priori information for the decoder
Figure GDA0003468722820000164
S6, in the embodiment, the decoder calculates the posterior information L by adopting the BCJR algorithm based on the maximum posterior probability criterionD(bm) The external information of the decoder is obtained by deducting the prior information from the posterior information
Figure GDA0003468722820000165
The calculation formula is as follows:
Figure GDA0003468722820000166
wherein b ismIndicating the mth bit before the information bits are not interleaved.
S7 posterior information LD(bm) Is the likelihood probability of the information bit, and carries out decision decoding according to the probability value thereof, this embodimentIn the method, the error detection code adopts a cyclic redundancy check code and is used for checking a decoding result.
When the decoding is correct, the demodulation of the nth data block is completed in advance; or when the decoding result is wrong and reaches the maximum iteration number ItermaxThen, the demodulation of the nth data block is stopped. Then, let n be n +1 and the iteration number Iter be 0, and go back to step S1 to perform the processing of the next data block until the received signal frame processing is completed.
When the decoding result is wrong and the current iteration number Iter is smaller than the maximum iteration number ItermaxIf yes, let Iter be Iter +1, and proceed to the iteration stage of step S8, and still process the current data block.
S8, before iterative equalization, mapping the extrinsic information to symbols, feeding back to equalizer, and decoding the extrinsic information
Figure GDA0003468722820000167
Interleaving to obtain prior information of equalizer
Figure GDA0003468722820000168
Mapping the interleaved prior information into prior symbols
Figure GDA0003468722820000169
The present embodiment adopts QPSK modulation, and the mapping method is as follows:
Figure GDA00034687228200001610
in the formula (I), the compound is shown in the specification,
Figure GDA0003468722820000171
and
Figure GDA0003468722820000172
are respectively a priori symbols
Figure GDA0003468722820000173
A priori information corresponding to two bits of (a), all a priori symbols of the current data blockAfter the number mapping is finished, the prior symbol vectors are combined
Figure GDA0003468722820000174
As the feedback filter input signal to the equalizer during the iterative equalization phase.
S9, the filter of the iterative equalization stage is composed of a feedforward filter fkAnd a feedback filter bkComposition, feedforward filter initial value
Figure GDA0003468722820000175
By preserving coefficients in the direct equalization phase or in the last iteration equalization phase
Figure GDA0003468722820000176
Determining an initial value of a feedback filter
Figure GDA0003468722820000177
Then a decision to set to zero vector or reserve
Figure GDA0003468722820000178
Depending on the number of iterations Iter. The equalizer outputs in the training phase and the decision phase are:
Figure GDA0003468722820000179
in the formula (f)kA feedforward filter at the k-th instant, bkIs a feedback filter at the k-th time, rkIn order to feed forward the input signal,
Figure GDA00034687228200001710
is a feedback input signal.
Filter equalization input at time k
Figure GDA00034687228200001711
Then, the present embodiment adopts the NLMS algorithm to update the filter coefficient, and the feedforward filter fkThe update formula is the same as formula (F) of step S4, and feedback filter bkThe update formula is:
Figure GDA00034687228200001712
wherein the desired signal
Figure GDA00034687228200001713
In the training phase, is a forward training sequence tnIn the decision stage, the symbols are equalized
Figure GDA00034687228200001714
And the a priori symbol of the feedback
Figure GDA00034687228200001715
And after merging, carrying out hard decision to obtain:
Figure GDA00034687228200001716
set time k to 0, …, NdEqualized symbol of-1
Figure GDA00034687228200001717
Forming a vector to obtain a forward adaptive equalization output
Figure GDA00034687228200001718
Then retaining the last updated feedforward filter
Figure GDA00034687228200001719
And a feedback filter
Figure GDA00034687228200001720
The coefficients of the inverse equalization filter are respectively fk'and b'kAfter time reversal of the filter input, the equalized symbols at each time are calculated as
Figure GDA00034687228200001721
Then k is equal to 0, …, Nd-1 time instantThe equalized symbols form a vector, time reversal is carried out to obtain reverse self-adaptive equalization output
Figure GDA00034687228200001722
Feed forward filter with retention of last update
Figure GDA00034687228200001723
And a feedback filter
Figure GDA0003468722820000181
The reverse equalization is completed.
Finally merging in an equal proportion mode
Figure GDA0003468722820000182
And
Figure GDA0003468722820000183
obtaining the balanced output of the iterative balance
Figure GDA0003468722820000184
Then, the process returns to step S5 to re-enter the decoding stage.
The implementation mode and specific parameters of the present invention are explained in detail above, and then the performance comparison with two common equalization modes DA-TEQ and CE-TEQ is performed, wherein the channel estimation equalization algorithm uses linear equalization based on the minimum mean square error criterion, and the adaptive equalization algorithm uses the normalized minimum mean square algorithm. According to the preceding description of the system parameters, N max4, that is, each signal frame carries 4 data blocks, each data block includes 2048 QPSK symbols, the code rate adopts 1/2, then the information bits carried by one signal frame is 8192, 8000 bits are selected to carry valid information in consideration of the reserved positions of check codes and the like and the convenience of calculating the error rate, the remaining 192 bits are used for other purposes or reservations, 100 frame data signals are sent to ensure that the data volume is sufficient, the signal-to-noise ratio is set to be 6dB, various algorithms are set in the decoding process to undergo 8 iterations, finally, different methods are adopted to equalize and calculate the error rate, and the result is shown in fig. 5.
As can be seen from FIG. 5, the DA-TEQ has the defect that the error rate is high in the initial stage, but can be continuously reduced in the iterative equalization process, and can even exceed the performance of the CE-TEQ in the later stage; the CE-TEQ has the advantages that a lower error rate can be obtained in the initial stage, but the gain of the iterative process cannot be well extracted, the error rate platform can be reached after the first several iterations, and the error rate cannot be further reduced; the self-adaptive equalization algorithm combined with the channel estimation can obtain the advantages of the self-adaptive equalization algorithm and the channel estimation, has a lower error code starting point in the initial iteration stage as the CE-TEQ algorithm, and quickly reduces the error code rate in the iteration process. In summary, the adaptive equalization soft information iterative receiving method combined with channel estimation provided by this embodiment has better decoding performance.
The above embodiments are preferred embodiments of the present invention, but the present invention is not limited to the above embodiments, and any other changes, modifications, substitutions, combinations, and simplifications which do not depart from the spirit and principle of the present invention should be construed as equivalents thereof, and all such changes, modifications, substitutions, combinations, and simplifications are intended to be included in the scope of the present invention.

Claims (10)

1. An adaptive equalization soft information iteration receiving method combined with channel estimation is characterized by comprising the following steps:
s1, the receiver obtains a frame of received data after frame synchronization, Doppler estimation and compensation operation, and records the frame as a received signal frame y, wherein the received signal frame y comprises NmaxA data block and Nmax+1 training sequences, per data block
Figure FDA0003468722810000011
Comprising NdA received symbol, each training sequence
Figure FDA0003468722810000012
Containing NtA received symbol [. ]]TIndicating that the vector is transposed, the received signal frame y starts with a training sequence, the data blocks and the training sequence are arranged one after the other, each data block being concatenatedExtracting the forward training sequence t of the nth data block r from the received signal frame y after a training sequence, namely a forward training sequence and a backward training sequence of the data block respectivelynAnd a backward training sequence tn+1The impulse response estimation is carried out by adopting a channel estimation algorithm to respectively obtain estimated channels
Figure FDA0003468722810000013
And
Figure FDA0003468722810000014
estimating the length of a channel to be L, namely comprising L tap coefficients;
s2, extracting the nth data block r from the received signal frame y, and using the estimated channel
Figure FDA0003468722810000015
And
Figure FDA0003468722810000016
carrying out frequency domain equalization on the data block r by adopting a maximum ratio combining method to obtain a transmitting symbol sequence
Figure FDA0003468722810000017
1 st estimation of
Figure FDA0003468722810000018
The transmission symbol sequence x contains NdA symbol;
s3, estimating the channel according to the estimated channel
Figure FDA0003468722810000019
And
Figure FDA00034687228100000110
separately solving the corresponding forward filters w in the time domain1And a backward filter w2Forward filter w1And a backward filter w2Length L, and then respectively making time domain equalization on data block r to obtain forward time domain equalization inputGo out
Figure FDA00034687228100000111
And backward time domain equalization output
Figure FDA00034687228100000112
Weighted combining
Figure FDA00034687228100000113
And
Figure FDA00034687228100000114
then obtaining the 2 nd estimation of the transmitted symbol sequence x
Figure FDA00034687228100000115
S4, 1 st estimation of symbol sequence x is transmitted
Figure FDA00034687228100000116
And 2 nd estimation
Figure FDA00034687228100000117
Performing equal proportion combination, inputting the hard decision signal into an adaptive equalizer as an expected signal, and obtaining forward adaptive equalization output through forward and reverse adaptive equalization
Figure FDA00034687228100000118
And reverse adaptive equalization output
Figure FDA00034687228100000119
Weighted combining
Figure FDA00034687228100000120
And
Figure FDA00034687228100000121
then the 3 rd estimation of the transmitted symbol sequence x is obtained
Figure FDA00034687228100000122
Last pair of
Figure FDA00034687228100000123
And
Figure FDA00034687228100000124
merging to obtain balanced output of data block r
Figure FDA0003468722810000021
S5, outputting the balance
Figure FDA0003468722810000022
Equalized symbol at time k
Figure FDA0003468722810000023
Mapping to obtain the external information of the equalizer
Figure FDA0003468722810000024
Figure FDA0003468722810000025
To represent
Figure FDA0003468722810000026
The jth bit carried, k 0, …, NdThe value range of-1, j depends on the modulation mode of the symbol,
Figure FDA0003468722810000027
after de-interleaving operation, as prior information
Figure FDA0003468722810000028
To a decoder, bmAn mth bit representing a sequence of bits carried by the data block;
s6, decoder according to prior information
Figure FDA0003468722810000029
Extracting error correction gain of channel coding and outputting posterior information LD(bm) Subtracting the prior information from the posterior information to obtain the external information of the decoder
Figure FDA00034687228100000210
S7, and the posterior information LD(bm) Decoding and checking, when decoding is correct or current iteration number Iter reaches maximum iteration number ItermaxIf so, the demodulation of the nth data block is completed, and then n is equal to n +1 and Iter is equal to 0, and the process returns to step S1 to perform the processing of the next data block until the demodulation of all data blocks in the received signal frame y is completed; otherwise, let Iter be Iter +1, execute step S8, and enter the iterative equalization stage;
s8, decoding the extrinsic information
Figure FDA00034687228100000211
Interleaving as a priori information of the equalizer
Figure FDA00034687228100000212
Figure FDA00034687228100000213
To represent
Figure FDA00034687228100000214
The jth bit that is carried over is,
Figure FDA00034687228100000215
is that
Figure FDA00034687228100000216
Mapping the obtained k-th prior symbol, wherein the prior symbol forms a prior input in a vector form
Figure FDA00034687228100000217
Figure FDA00034687228100000218
Is the feedback filter input signal, a priori symbol, of an adaptive equalizer
Figure FDA00034687228100000219
Will be output in balance with
Figure FDA00034687228100000220
Computing together the desired signals of an adaptive algorithm
Figure FDA00034687228100000221
S9, forming the equalizer of the iterative equalization stage by the feedforward filter and the feedback filter, using the adaptive algorithm, and using the equalization output
Figure FDA00034687228100000222
And a desired signal
Figure FDA00034687228100000223
Updating filter coefficient to obtain balanced output
Figure FDA00034687228100000224
Then, the process goes to step S5 to re-enter the decoding stage.
2. The iterative receiving method for adaptive equalization soft information combined with channel estimation as claimed in claim 1, wherein in step S1, the training sequence in signal frame y is received
Figure FDA00034687228100000225
Is made up of known training sequences of the transmitting end
Figure FDA00034687228100000226
Obtained after passing through a channel and containing NtA forward training sequence t is extracted from the received signal frame ynAccording to the known forward training sequence z of the transmitting endnSolving the forward estimated channel by a channel estimation algorithm
Figure FDA0003468722810000031
Followed by extraction of the backward training sequence tn+1According to the known backward training sequence z of the transmitting endn+1Solving for the backward estimated channel
Figure FDA0003468722810000032
3. The iterative receiving method for adaptive equalization soft information combined with channel estimation as claimed in claim 1, characterized in that in step S2, the nth data block r is extracted from the received signal frame y, and r is transformed by discrete fourier transform,
Figure FDA0003468722810000033
And
Figure FDA0003468722810000034
transforming from time domain to frequency domain to obtain frequency domain data block R and frequency domain channel response
Figure FDA0003468722810000035
And
Figure FDA0003468722810000036
and (2) performing frequency domain equalization by adopting a maximum ratio combining method represented by the formula (1) according to the frequency domain data block and the frequency domain channel response:
Figure FDA0003468722810000037
where J represents the number of frequency domain channel responses, forFrequency domain equalization output
Figure FDA0003468722810000038
Inverse discrete Fourier transform to obtain the 1 st estimate of the transmitted symbol sequence x
Figure FDA0003468722810000039
4. The iterative receiving method for adaptive equalization soft information combined with channel estimation as claimed in claim 1, wherein in step S3, the estimated channel in forward direction is used
Figure FDA00034687228100000310
Solving for the forward filter w according to equation (2)1：
Figure FDA00034687228100000311
In the formula (I), the compound is shown in the specification,
Figure FDA00034687228100000312
as a variance of the noise, IMIs an M-order identity matrix, where M ═ N1+N2+1，N1And N2Respectively causal and non-causal parts of the filter, s is the channel convolution matrix H1N of (2)2+ L columns, channel convolution matrix H1By estimating the channel
Figure FDA00034687228100000313
Is constructed by the following steps:
Figure FDA00034687228100000314
wherein the content of the first and second substances,
Figure FDA0003468722810000041
respectively representing estimated channels
Figure FDA0003468722810000042
L tap coefficients of (i)
Figure FDA0003468722810000043
Obtaining a forward filter w1The forward time domain equalization is then represented as:
Figure FDA0003468722810000044
in the formula, rkIs the input signal at the time of the k-th instant,
Figure FDA0003468722810000045
is the forward equalization symbol at the k-th time, and the equalization symbols at all times of the data block r are obtained
Figure FDA0003468722810000046
Forward time domain equalization outputs combined into vector form
Figure FDA0003468722810000047
Then, the forward estimated channel is transmitted
Figure FDA0003468722810000048
Estimation channel converted into backward direction
Figure FDA0003468722810000049
Calculating a backward filter w2Calculating the backward balanced symbol of all the time of the data block r
Figure FDA00034687228100000410
Backward time domain equalization output in the form of a component vector
Figure FDA00034687228100000411
Forward time domain equalization output using equation (4)
Figure FDA00034687228100000412
And backward time domain equalization output
Figure FDA00034687228100000413
And (3) carrying out weighted combination, wherein beta is a weighting coefficient, and calculating to obtain a 2 nd estimation of the transmitted symbol sequence x:
Figure FDA00034687228100000414
5. the iterative receiving method of adaptive equalization soft information combined with channel estimation as claimed in claim 1, wherein in step S4, direct equalization is performed, the direct equalization stage only has a feedforward filter f, the length of the filter is M, and w are1And w2Similarly, for the forward adaptive equalization, the equalization is divided into a training stage and a decision stage, and the equalization output formula of the two stages is as follows:
Figure FDA00034687228100000415
in the formula (I), the compound is shown in the specification,
Figure FDA00034687228100000416
is the input signal of the feed-forward filter,
Figure FDA00034687228100000417
for the balance output at the kth moment, a self-adaptive algorithm is needed for updating the filter coefficient, a normalized least mean square algorithm is adopted, and a coefficient updating formula is as follows:
Figure FDA00034687228100000418
in the formula (f)kIs a feedforward filter at the k-th moment, xi is a convergence factor, epsilon is a normal number with a small value,
Figure FDA00034687228100000419
for the desired signal at the k-th moment, in the training phase
Figure FDA00034687228100000420
For training sequences, in the decision phase
Figure FDA00034687228100000421
According to
Figure FDA0003468722810000051
And
Figure FDA0003468722810000052
and performing hard decision, wherein the calculation method comprises the following steps:
Figure FDA0003468722810000053
in the formula (I), the compound is shown in the specification,
Figure FDA0003468722810000054
and
Figure FDA0003468722810000055
are respectively
Figure FDA0003468722810000056
And
Figure FDA0003468722810000057
the equalization value at the k-th moment, Q (-) operation represents that hard decision is carried out on the equalization symbol;
of all instants of the nth data blockAfter the output calculation is finished, k is equal to 0, …, NdEqualized output at time-1
Figure FDA0003468722810000058
Forward adaptive equalization output in the form of component vectors
Figure FDA0003468722810000059
Final feedforward filter with preserving forward adaptive equalization
Figure FDA00034687228100000510
As the initial value of the feedforward filter in the iterative equalization stage;
reverse adaptive equalization using a backward training sequence tn+1Solving for reverse adaptive equalization outputs
Figure FDA00034687228100000511
The difference from the forward adaptive equalization is that the input and output of the equalizer are time reversed, leaving the feedforward filter of the reverse adaptive equalization
Figure FDA00034687228100000512
By combining in equal proportions
Figure FDA00034687228100000513
And
Figure FDA00034687228100000514
the combining coefficient γ is 1/2, yielding the 3 rd estimate of the transmitted symbol sequence x:
Figure FDA00034687228100000515
3 estimates of the symbol sequence x to be transmitted
Figure FDA00034687228100000516
And
Figure FDA00034687228100000517
performing weighting combination to obtain balance output of direct balance stage
Figure FDA00034687228100000518
Figure FDA00034687228100000519
In the formula, alpha1，α2And alpha33 estimates respectively
Figure FDA00034687228100000520
And
Figure FDA00034687228100000521
the weighting coefficient of (2).
6. The iterative receiving method for adaptive equalization soft information combined with channel estimation as claimed in claim 1, wherein in step S5, the equalization output is outputted
Figure FDA00034687228100000522
Mapping as external information, and approximate solving of statistical model parameters mu and delta by adopting a time average method2Mu is the scaling factor of the transmitted symbol sequence x, delta2Then the variance of x, and then the probability value of the symbol is calculated
Figure FDA00034687228100000523
aiIs the ith element of the transmitted symbol set, the number of symbols of the transmitted symbol set depends on the modulation mode, and then the extrinsic information output by the equalizer is obtained
Figure FDA00034687228100000524
Figure FDA00034687228100000525
Performing de-interleaving operation to obtain prior information of decoder
Figure FDA0003468722810000061
7. The iterative receiving method for adaptive equalization soft information combined with channel estimation as claimed in claim 1, characterized in that in step S6, the a priori information at the decoder
Figure FDA0003468722810000062
Under the guidance of (2), the decoder extracts the error correction gain of the channel coding and outputs a posteriori information LD(bm) The external information of the decoder is obtained by deducting the prior information from the posterior information
Figure FDA0003468722810000063
The calculation formula is as follows:
Figure FDA0003468722810000064
8. the iterative receiving method for adaptive equalization soft information combined with channel estimation as claimed in claim 1, wherein in step S7, a posteriori information L is appliedD(bm) Judging decoding, and judging whether the decoding result is correct according to the error detection code; if the decoding is correct, the decoding process of the current data block is exited, a result is output, n is equal to n +1, the iteration number Iter is equal to 0, and the process returns to step S1 to process the next data block until the received signal is processed; in the case of decoding failure and the current iteration number is less than the maximum iteration number ItermaxWhen it is determined that the term "Iter" is equal to term +1, the process proceeds to step S8, and an iterative operation is performed.
9. The adaptive equalization soft information in combination with channel estimation as claimed in claim 1Iterative receiving method, characterized in that in step S8, extrinsic information of the decoder is obtained
Figure FDA0003468722810000065
Interleaving as a priori information of the equalizer
Figure FDA0003468722810000066
Mapping prior information at the k-th time into prior symbols
Figure FDA0003468722810000067
Constituent prior inputs
Figure FDA0003468722810000068
As the feedback filter input signal for the adaptive equalizer.
10. The iterative receiving method for adaptive equalization soft information combined with channel estimation as claimed in claim 1, wherein in step S9, iterative equalization is performed, and the equalizer comprises a feedforward filter f and a feedback filter b, and the equalization is retained in the last time
Figure FDA0003468722810000069
And
Figure FDA00034687228100000610
initialization is performed such that the feedback filter does not retain coefficients during the direct equalization phase
Figure FDA00034687228100000611
Then, setting the equalizer as a zero vector, and the output of the equalizer in the training stage and the decision stage is:
Figure FDA00034687228100000612
in the formula (f)kA feedforward filter at the k-th instant, bkIs a feedback filter at the k-th time, rkIs the input signal to the feedforward filter at time k,
Figure FDA0003468722810000071
feeding back an input signal of the filter at the k time; for updating the adaptive filter coefficient, a normalized least mean square algorithm is adopted, and a feedback filter bkThe update formula is:
Figure FDA0003468722810000072
in the training phase, a signal is expected
Figure FDA0003468722810000073
For training sequence tnIn the decision phase, by equalizing the symbols
Figure FDA0003468722810000074
And the a priori symbol of the feedback
Figure FDA0003468722810000075
And after merging, carrying out hard decision to obtain:
Figure FDA0003468722810000076
the iterative equalization of the data block is completed, k is equal to 0, …, NdEqualized symbols at-1 time instant
Figure FDA0003468722810000077
Forming a vector to obtain a forward adaptive equalization output
Figure FDA0003468722810000078
Feed forward filter with retention of last update
Figure FDA0003468722810000079
And a feedback filter
Figure FDA00034687228100000710
Feed-forward filter f 'for reverse adaptive equalization'kAnd a feedback filter b'kAccording to retention coefficient
Figure FDA00034687228100000711
And
Figure FDA00034687228100000712
initializing, and performing time reversal on input and output of data in the equalization process to obtain reverse self-adaptive equalization output
Figure FDA00034687228100000713
Feed forward filter with retention of last update
Figure FDA00034687228100000714
And a feedback filter
Figure FDA00034687228100000715
Providing an initial value for the next iteration balance;
to the forward direction adaptive equalization output
Figure FDA00034687228100000716
And reverse adaptive equalization output
Figure FDA00034687228100000717
Merging in equal proportion mode to obtain the balanced output of the iterative balance
Figure FDA00034687228100000718
Then, the process returns to step S5 to enter the decoding stage.
CN202110392636.5A 2021-04-13 2021-04-13 Adaptive equalization soft information iteration receiving method combined with channel estimation Active CN113242189B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110392636.5A CN113242189B (en) 2021-04-13 2021-04-13 Adaptive equalization soft information iteration receiving method combined with channel estimation

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110392636.5A CN113242189B (en) 2021-04-13 2021-04-13 Adaptive equalization soft information iteration receiving method combined with channel estimation

Publications (2)

Publication Number Publication Date
CN113242189A CN113242189A (en) 2021-08-10
CN113242189B true CN113242189B (en) 2022-04-22

Family

ID=77128588

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110392636.5A Active CN113242189B (en) 2021-04-13 2021-04-13 Adaptive equalization soft information iteration receiving method combined with channel estimation

Country Status (1)

Country Link
CN (1) CN113242189B (en)

Families Citing this family (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113542165B (en) * 2021-09-15 2021-12-21 深圳金信诺高新技术股份有限公司 Decoding method, decoding device, electronic equipment and storage medium
CN113905198B (en) * 2021-09-28 2024-04-05 上海富瀚微电子股份有限公司 Adaptive channel equalization method and device for coaxial video transmission system and computer storage medium
CN113922868B (en) * 2021-12-13 2022-03-18 南京先锋硕通无线技术有限公司 DVB-RCS2 RL three-turbine iterative receiver and receiving method
CN114389754B (en) * 2021-12-28 2023-09-19 大连理工大学 Frequency domain self-adaptive Turbo equalization method based on FBNLMS algorithm
CN114614835B (en) * 2022-03-23 2024-05-17 浙江大学 Polarization code BP decoding method suitable for underwater acoustic communication
CN115225437B (en) * 2022-06-07 2023-05-12 哈尔滨工业大学（威海） Combined intelligent equalization and decoding method for underwater acoustic cooperative communication
CN115208481B (en) * 2022-06-30 2023-03-24 哈尔滨工程大学 Single carrier frequency domain equalization receiving processing method of underwater acoustic time-varying channel in polar environment
CN115208483B (en) * 2022-06-30 2023-02-10 哈尔滨工程大学 Underwater acoustic communication method under polar impulse interference

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9210012B2 (en) * 2011-12-29 2015-12-08 Intel Corporation Frequency-domain turbo equalization, including multi-mode adaptive linear equalization, adaptive decision-directed channel estimation, adaptive noise variance estimation, and dynamic iteration control
CN105656823B (en) * 2016-01-27 2019-01-18 华南理工大学 Subsurface communication Turbo based on minimum bit-error rate criterion receives system and method
US10637586B2 (en) * 2017-08-25 2020-04-28 National Science Foundation Turbo receivers for multiple-input multiple-output underwater acoustic communications
CN108712353B (en) * 2018-03-29 2021-05-11 江苏中科羿链通信技术有限公司 Soft iteration channel estimation method
CN109246039B (en) * 2018-08-09 2020-11-24 华南理工大学 Soft information iteration receiving method based on bidirectional time domain equalization

Also Published As

Publication number Publication date
CN113242189A (en) 2021-08-10

Similar Documents

Publication Publication Date Title
CN113242189B (en) Adaptive equalization soft information iteration receiving method combined with channel estimation
CN109246039B (en) Soft information iteration receiving method based on bidirectional time domain equalization
US6819630B1 (en) Iterative decision feedback adaptive equalizer
WO2001082488A1 (en) Turbo decoder with decision feedback equalization
CN113242190B (en) Multichannel communication minimum bit error rate Turbo equalization method based on posterior soft symbol
JP4272665B2 (en) Apparatus, method, and computer program for estimating channel of OFDM transmission system
US7974335B2 (en) Apparatus for equalizing decision feedback based on channel properties change and method therefor
CN112866151B (en) Underwater sound MPSK signal blind Turbo equalization method based on channel blind estimation
US7027497B2 (en) Adaptive equalization method and adaptive equalizer
CN109981501B (en) Underwater sound direct self-adaptive MIMO communication method
CN105553903B (en) A kind of adaptive turbo equalization methods and balanced device, underwater sound communication system
US20030115061A1 (en) MPSK equalizer
CN101119177A (en) Bit-symbol signal processing method for coherent communication machine
CN109861764B (en) Compressed sensing channel estimation method for underwater sound OFDM iterative receiver
CN116016061B (en) Short wave double-selection channel double-iteration Turbo equalization method for high maneuvering platform
CN111682924B (en) Bidirectional frequency domain Turbo equalization method adopting expected propagation
CN115225437A (en) Joint intelligent equalization and decoding method for underwater acoustic cooperative communication
CN112039809B (en) Block iterative equalizer based on mixed soft information and bidirectional block iterative equalizer
JP4187968B2 (en) Adaptive equalization method and adaptive equalizer
Otnes et al. Block SISO linear equalizers for turbo equalization in serial-tone HF modems
Iltis Iterative joint decoding and sparse channel estimation for single-carrier modulation
CN115987721A (en) Channel impulse noise estimation method based on variational Bayesian learning
CN116155663A (en) Design method of underwater acoustic communication decision feedback iterative equalization receiver based on symbol-by-symbol posterior information soft decision
CN115996065B (en) Robust adaptive turbo equalization method and apparatus applied to time-varying underwater acoustic channel
JP7234383B2 (en) Receiving device, radio system and receiving method

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant