CN112054974A - Underwater sound channel identification method based on regularization minimum mean square error variable step length algorithm - Google Patents

Underwater sound channel identification method based on regularization minimum mean square error variable step length algorithm Download PDF

Info

Publication number
CN112054974A
CN112054974A CN202010889449.3A CN202010889449A CN112054974A CN 112054974 A CN112054974 A CN 112054974A CN 202010889449 A CN202010889449 A CN 202010889449A CN 112054974 A CN112054974 A CN 112054974A
Authority
CN
China
Prior art keywords
channel
impulse response
underwater acoustic
mean square
square error
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.)
Granted
Application number
CN202010889449.3A
Other languages
Chinese (zh)
Other versions
CN112054974B (en
Inventor
伍飞云
苏本学
杨坤德
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.)
Northwestern Polytechnical University
Original Assignee
Northwestern Polytechnical University
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 Northwestern Polytechnical University filed Critical Northwestern Polytechnical University
Priority to CN202010889449.3A priority Critical patent/CN112054974B/en
Publication of CN112054974A publication Critical patent/CN112054974A/en
Application granted granted Critical
Publication of CN112054974B publication Critical patent/CN112054974B/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
    • H04L25/0202—Channel estimation
    • H04L25/024—Channel estimation channel estimation algorithms
    • H04L25/0256—Channel estimation using minimum mean square error criteria
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B13/00—Transmission systems characterised by the medium used for transmission, not provided for in groups H04B3/00 - H04B11/00
    • H04B13/02—Transmission systems in which the medium consists of the earth or a large mass of water thereon, e.g. earth telegraphy
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B17/00—Monitoring; Testing
    • H04B17/30—Monitoring; Testing of propagation channels
    • H04B17/309—Measuring or estimating channel quality parameters
    • H04B17/336—Signal-to-interference ratio [SIR] or carrier-to-interference ratio [CIR]
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B17/00—Monitoring; Testing
    • H04B17/30—Monitoring; Testing of propagation channels
    • H04B17/391—Modelling the propagation channel

Landscapes

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

Abstract

The invention relates to an underwater sound channel identification method based on a regularization minimum mean square error variable step size algorithm. And (3) designing a regularized minimum mean square error variable step size algorithm to perform iterative optimization to obtain a time domain underwater sound channel impulse response function by adopting a periodic training mode, namely identifying to obtain time domain underwater sound channel information. The method of the invention can estimate the impulse response function of the time-varying underwater acoustic channel. The method of the invention skillfully uses the change strategy of the moving window control step length, so that more accurate channel estimation is obtained in the iteration process compared with the traditional algorithm, which is beneficial to improving the estimation accuracy based on the output of the channel estimation equalizer.

Description

Underwater sound channel identification method based on regularization minimum mean square error variable step length algorithm
Technical Field
The invention belongs to the fields of underwater acoustic communication, underwater acoustic signal processing and the like, relates to an underwater acoustic channel identification method based on a regularization minimum mean square error variable step size algorithm, and relates to an underwater acoustic channel identification method for identifying time domains of received signals and training sequences under impulse interference.
Background
The problems of underwater acoustic channel identification, underwater acoustic communication and the like can be solved into an identification optimization problem of an impulse response function, and the problems are used for identifying the underwater acoustic channel expression based on a training sequence and a received signal. Currently, the identification method for the underwater acoustic channel includes a finite impulse response framework and a time domain block-by-block identification framework. For the details of the finite impulse response framework, see "New spark adaptive basic on the natural gradient and the L0-norm", published in 2013 at the No. 38 of IEEE Journal of scientific Engineering, with a start page 323. The frame of block-by-block identification of the time domain is detailed in "Estimation of vertical time-varying sparse channels", which is published in "IEEE Journal of scientific Engineering" No. 32 in 2007, and the start page number is 927.
Due to the expansion and time-varying characteristics of the underwater acoustic channel, the identification of the impulse response function of the underwater acoustic channel is extremely difficult, and in addition, the received signal may be interfered by the impulse signal, and the traditional adaptive estimation method loses the identification effect on the channel under the condition of large impulse interference. Thus, the time domain of the hydroacoustic channel can be characterized in view of its spread and time-varying properties. The invention is based on the model to identify the underwater sound channel. However, the fixed step length of the adaptive identification method cannot simultaneously guarantee the convergence rate and the convergence accuracy of the algorithm, so that a new step length updating strategy is designed to guarantee the convergence rate and hopefully improve the identification accuracy. Considering the time-varying property of the impulse response function of the underwater acoustic channel in practice, the invention adopts the moving window function to control the change of the step length so as to improve the identification and tracking capability of the underwater acoustic channel.
Disclosure of Invention
Technical problem to be solved
In order to avoid the defects of the prior art, the invention provides an underwater sound channel identification method based on a regularization minimum mean square error variable step size algorithm.
Technical scheme
An underwater acoustic channel identification method based on a regularization minimum mean square error variable step size algorithm is characterized by comprising the following steps:
step 1: setting a parameter L as the impulse response length of an underwater acoustic channel, and setting a window function updating parameter eta of a control step length to be 0.99;
step 2: setting the window function weighting factor C of the control step as 1.483[1+5/(L-1)]Initializing a gradient factor of an impulse response of the underwater acoustic channel to
Figure BDA0002656484750000021
Initializing an underwater acoustic channel impulse response function to h0＝0；
And step 3: given an input training signal x and an output signal y, the following i-th iteration is repeated:
calculating an identification error:
Figure BDA0002656484750000022
wherein y isiIs the discrete value of the received signal at the ith time,
Figure BDA0002656484750000023
for the transposition of the transmitted signal for training at the i-th instant, hi-1Is the channel impulse response function at the i-1 th moment;
calculating the absolute value of the gradient factor:
Figure BDA0002656484750000024
wherein | | | xiI represents solving a Euclidean norm of an input signal;
updating the gradient factor:
Figure BDA0002656484750000025
wherein: min (B)i) Represents from BiSelecting the minimum value;
generating a step updating expression:
Figure BDA0002656484750000026
obtaining channel identification iteration: h isi＝hi-1+μibixi。
The number of iterations is equal to the difference between the length of the processed data and the length of the channel.
Advantageous effects
The invention provides an underwater acoustic channel identification method based on a regularization minimum mean square error variable step size algorithm. And (3) designing a regularized minimum mean square error variable step size algorithm to perform iterative optimization to obtain a time domain underwater sound channel impulse response function by adopting a periodic training mode, namely identifying to obtain time domain underwater sound channel information. The method of the invention can estimate the impulse response function of the time-varying underwater acoustic channel. The method of the invention skillfully uses the change strategy of the moving window control step length, so that more accurate channel estimation is obtained in the iteration process compared with the traditional algorithm, which is beneficial to improving the estimation accuracy based on the output of the channel estimation equalizer.
The beneficial effects are as follows: the method controls the step change based on the sliding window function, effectively improves the convergence speed of the algorithm, and simultaneously can ensure that the method has lower steady-state error.
Drawings
FIG. 1 is a graph showing the comparison result between the method of the present invention (Normalized least-mean-square fraction with variable step-size, NLMSFV) and the conventional regularized least-mean-square error class 1 algorithm (NLMS 1) and class 2 algorithm (NLMS2), regularized least-mean-square fraction algorithm (NLMSF), etc.
Fig. 2 is a comparison graph of the recognition results of the conventional NLMS1, NLMS2, NLMSF method and the present invention method (NLMSFV) for two randomly generated channel traces.
Fig. 3 is a graph comparing the learning effect of testing different convergence rates of NLMS1, NLMS2, NLMSF method and the present invention method (NLMSFV).
Detailed Description
The invention will now be further described with reference to the following examples and drawings:
in order to solve the problem of identifying the time domain underwater acoustic channel, the specific implementation steps are provided as follows:
(1) and setting the parameter L as the impulse response length of the underwater acoustic channel, and setting the window function updating parameter eta of the control step length to be 0.99.
(2) Setting the window function weighting factor C of the control step as 1.483[1+5/(L-1)]Initializing a gradient factor of an impulse response of the underwater acoustic channel to
Figure BDA0002656484750000041
Initializing an underwater acoustic channel impulse response function to h0＝0。
(3) Given an input training signal x and an output signal y, the following i-th iteration is repeated:
calculating an identification error:
Figure BDA0002656484750000042
wherein y isiIs the discrete value of the received signal at the ith time,
Figure BDA0002656484750000043
for the transposition of the transmitted signal for training at the i-th instant, hi-1Is the channel impulse response function at the i-1 th moment;
absolute value of gradient factor:
Figure BDA0002656484750000044
wherein | | | xiI represents solving a Euclidean norm of an input signal;
the past L gradient factor absolute values are made up of a window: b isi＝[bi,bi-1,...,bi-L+1]；
Gradient factor updating:
Figure BDA0002656484750000045
where η represents the updateParameter, C denotes a weighting factor, min (B)i) Represents from BiSelecting the minimum value;
generating a step updating expression:
Figure BDA0002656484750000046
channel identification iteration: h isi＝hi-1+μibixi。
The number of iterations above is equal to the difference between the length of the processed data and the length of the channel.
The invention will be further described with reference to specific embodiments thereof. Referring to fig. 1, setting a channel impulse response function length to be 100, generating a 6000-point random gaussian signal according to a standard normal distribution mode, setting a parameter of a conventional algorithm to be 200, μ to be 0.4, setting a gaussian noise signal-to-noise ratio to be 35dB, setting an impulse interference variance to be 1000 times of a received signal variance, and forming by adopting a shell-and-shell effort distribution, wherein a probability of generating impulse interference is Pr to be 0.5, and an obtained result is shown in fig. 1.
The identification result of the time-varying underwater acoustic channel is shown in fig. 2, the data length is always set at 12000 points, and the channel is suddenly changed at the moment when the algorithm is iterated to 6000 points, so as to evaluate the tracking capability of each algorithm on the channel under the time-varying condition.
To further examine the trade-off between convergence rate and steady state error of different algorithms, different β of the traditional algorithm is set to generate different convergence rate and steady state error, and it can be seen that, when the steady state error is reduced, the convergence rate of the traditional NLMS1, NLMS2, NLMSF methods is reduced, and as the learning curve result of the present invention is shown in fig. 3, it can be seen that the method of the present invention can obtain obvious advantages in terms of both convergence rate and steady state error.

Claims (2)

1. An underwater acoustic channel identification method based on a regularization minimum mean square error variable step size algorithm is characterized by comprising the following steps:
step 1: setting a parameter L as the impulse response length of an underwater acoustic channel, and setting a window function updating parameter eta of a control step length to be 0.99;
step 2: setting the window function weighting factor C of the control step as 1.483[1+5/(L-1)]Initializing a gradient factor of an impulse response of the underwater acoustic channel to
Figure FDA0002656484740000011
Initializing an underwater acoustic channel impulse response function to h0＝0；
And step 3: given an input training signal x and an output signal y, the following i-th iteration is repeated:
calculating an identification error:
Figure FDA0002656484740000012
wherein y isiIs the discrete value of the received signal at the ith time,
Figure FDA0002656484740000013
for the transposition of the transmitted signal for training at the i-th instant, hi-1Is the channel impulse response function at the i-1 th moment;
calculating the absolute value of the gradient factor:
Figure FDA0002656484740000014
wherein | | | xiI represents solving a Euclidean norm of an input signal;
updating the gradient factor:
Figure FDA0002656484740000015
wherein: min (B)i) Represents from BiSelecting the minimum value;
generating a step updating expression:
Figure FDA0002656484740000016
obtaining channel identification iteration: h isi＝hi-1+μibixi。
2. The underwater acoustic channel identification method based on the regularized minimum mean square error variable step size algorithm according to claim 1, wherein: the number of iterations is equal to the difference between the length of the processed data and the length of the channel.
CN202010889449.3A 2020-08-28 2020-08-28 Underwater sound channel identification method based on regularization minimum mean square error variable step length algorithm Active CN112054974B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010889449.3A CN112054974B (en) 2020-08-28 2020-08-28 Underwater sound channel identification method based on regularization minimum mean square error variable step length algorithm

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010889449.3A CN112054974B (en) 2020-08-28 2020-08-28 Underwater sound channel identification method based on regularization minimum mean square error variable step length algorithm

Publications (2)

Publication Number Publication Date
CN112054974A true CN112054974A (en) 2020-12-08
CN112054974B CN112054974B (en) 2021-07-16

Family

ID=73607535

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010889449.3A Active CN112054974B (en) 2020-08-28 2020-08-28 Underwater sound channel identification method based on regularization minimum mean square error variable step length algorithm

Country Status (1)

Country Link
CN (1) CN112054974B (en)

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20180191533A1 (en) * 2016-05-06 2018-07-05 Amlogic (Shanghai) Co., Ltd. Channel estimation method
CN108510996A (en) * 2017-02-27 2018-09-07 上海闻通信息科技有限公司 A kind of iteratively faster adaptive filter method
CN108833312A (en) * 2018-06-08 2018-11-16 西北工业大学 It is a kind of based on delay-Doppler domain when be thinned out underwater acoustic channel estimation method
CN109981501A (en) * 2019-03-08 2019-07-05 哈尔滨工程大学 A kind of direct adaptive MIMO communication means of the underwater sound
CN110162739A (en) * 2019-04-30 2019-08-23 哈尔滨工业大学 Based on the RFFKLMS algorithm right value update optimization method for becoming forgetting factor
CN111106877A (en) * 2019-12-11 2020-05-05 中国科学院声学研究所 Underwater acoustic communication transmission method based on Farrow filtering and code word matching
CN111404634A (en) * 2020-02-16 2020-07-10 西安电子科技大学 Large-scale MIMO detection method, system and application based on variable step length iteration

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20180191533A1 (en) * 2016-05-06 2018-07-05 Amlogic (Shanghai) Co., Ltd. Channel estimation method
CN108510996A (en) * 2017-02-27 2018-09-07 上海闻通信息科技有限公司 A kind of iteratively faster adaptive filter method
CN108833312A (en) * 2018-06-08 2018-11-16 西北工业大学 It is a kind of based on delay-Doppler domain when be thinned out underwater acoustic channel estimation method
CN109981501A (en) * 2019-03-08 2019-07-05 哈尔滨工程大学 A kind of direct adaptive MIMO communication means of the underwater sound
CN110162739A (en) * 2019-04-30 2019-08-23 哈尔滨工业大学 Based on the RFFKLMS algorithm right value update optimization method for becoming forgetting factor
CN111106877A (en) * 2019-12-11 2020-05-05 中国科学院声学研究所 Underwater acoustic communication transmission method based on Farrow filtering and code word matching
CN111404634A (en) * 2020-02-16 2020-07-10 西安电子科技大学 Large-scale MIMO detection method, system and application based on variable step length iteration

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
管四海、李智、黄辉、王哲: "带零吸收项的变步长L0范数归一化最小均方误差算法", 《兵工学报》 *

Also Published As

Publication number Publication date
CN112054974B (en) 2021-07-16

Similar Documents

Publication Publication Date Title
CN111653288B (en) Target person voice enhancement method based on conditional variation self-encoder
CN107703486B (en) Sound source positioning method based on convolutional neural network CNN
CN112418014A (en) Modulation signal identification method based on wavelet transformation and convolution long-short term memory neural network
US8737460B2 (en) Equalizer and detector arrangement employing joint entropy-based calibration
CN113094993B (en) Modulation signal denoising method based on self-coding neural network
CN110162739B (en) RFFKBMS algorithm weight updating and optimizing method based on forgetting factor
CN107124379B (en) Orthogonal wavelet normal-modulus blind equalization method based on improved wolf pack optimization
CN112054973B (en) Minimum mean square error sparse underwater acoustic channel estimation method
CN117296061A (en) Diffusion model with improved accuracy and reduced computing resource consumption
CN114172490A (en) Robust adaptive noise elimination method based on Ekbum norm
CN112887239A (en) Method for rapidly and accurately identifying underwater sound signal modulation mode based on deep hybrid neural network
CN111711584A (en) Sparse underwater acoustic channel estimation method based on proportional normalization minimum mean square error
CN112910812A (en) Modulation mode identification method for deep learning based on space-time feature extraction
CN112115821A (en) Multi-signal intelligent modulation mode identification method based on wavelet approximate coefficient entropy
CN112054974B (en) Underwater sound channel identification method based on regularization minimum mean square error variable step length algorithm
EP2011114A1 (en) Signal analysis method with non-gaussian auto-regressive model
CN107018103B (en) Wavelet constant modulus blind equalization method based on adaptive step size monkey swarm optimization
CN117636894A (en) Voice dereverberation method based on multi-channel blind identification and multi-channel equalization
CN111555994A (en) Cluster sparse channel estimation method based on maximum skip rule algorithm
Sharma et al. A comparative analysis of mean square error adaptive filter algorithms for generation of modified scaling and wavelet function
CN110233808A (en) A kind of FTN system signal detection method
CN113420870B (en) U-Net structure generation countermeasure network and method for underwater sound target recognition
CN109474258B (en) Nuclear parameter optimization method of random Fourier feature kernel LMS (least mean square) based on nuclear polarization strategy
CN112821971A (en) Time-varying channel signal detection method based on countermeasure learning
CN113541649B (en) Variable step length kernel number error self-adaptive filter

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