A clustering-based method for calculating noise variance of underwater acoustic OFDM signals

Abstract

The present invention discloses a method for calculating the noise variance of underwater acoustic OFDM signals based on clustering. At the underwater acoustic OFDM signal modulation transmitting end, the empty carriers are evenly arranged in the OFDM symbol frequency domain to measure the noise value, and at the receiving end, the Doppler effect is used to measure the carrier frequency in the inconsistency of the broadband signal and the residual ICI characteristics of the OFDM underwater acoustic communication signal by using different subcarrier weighting methods, and then the weighted empty carrier frequency measurement values ​​are grouped by clustering, so that the empty carrier frequency measurement values ​​with large errors can be removed, and the accuracy of the final noise variance calculation can be improved.

Description

Technical Field

[0001] The invention relates to the field of underwater acoustic OFDM communication, and in particular to noise variance measurement of underwater acoustic OFDM communication signals, and mainly provides a method for calculating noise variance of underwater acoustic OFDM signals based on clustering. Background Art

[0002] Underwater acoustic OFDM communication is a high-speed underwater information transmission technology. However, due to the complexity of the underwater environment and the particularity of acoustic wave transmission, underwater acoustic OFDM communication faces many challenges, such as inter-symbol interference, inter-carrier interference, etc. In order to improve the performance of underwater acoustic OFDM communication systems, researchers have proposed various signal processing algorithms, such as equalization algorithms, encoding algorithms, decoding algorithms, etc. In particular, soft information decoding has greatly improved the performance of underwater acoustic OFDM communication systems, but soft information decoding requires accurate measurement of received noise, especially under the influence of residual inter-carrier interference. How to accurately measure and calculate the noise variance is a key issue in OFDM underwater acoustic communication technology. Summary of the invention

[0003] In view of the problem that the soft information decoding at the existing underwater acoustic OFDM communication receiving end needs to obtain an accurate symbol signal-to-noise ratio, the present invention provides a method for calculating the noise variance of underwater acoustic OFDM signals based on clustering. Since the underwater acoustic OFDM signal is greatly affected by Doppler, the Doppler estimation compensation algorithm at the receiving end cannot completely eliminate the ICI (inter-carrier interference) caused by the Doppler compensation effect. The clustering method is used to group the noise measured at different carrier positions at the receiving end, which can eliminate the error introduced by ICI, improve the symbol noise measurement accuracy, and further improve the decoding accuracy.

[0004] The objective of the present invention is achieved through the following technical solutions. A method for calculating the noise variance of underwater acoustic OFDM signals based on clustering, in which the OFDM symbols are evenly set with empty subcarriers at the transmitting end, and the value of the empty carrier is extracted when the OFDM demapping is completed at the receiving end, which is the amount of noise in the frequency domain. The clustering algorithm is used to extract noise groups from different empty carrier positions, and the elimination of groups with small samples can improve the accuracy of OFDM symbol noise estimation. The specific steps are as follows:

[0005] Step 1: Setting the empty carrier at the transmitter:

[0006] At the transmitting end of the underwater acoustic OFDM signal modulation, empty carriers are evenly arranged in the OFDM symbol frequency domain to measure the noise value. The number of arranged empty carriers is generally not less than 1 / 64 of the total number of subcarriers and not more than 1 / 16 of the total number of subcarriers.

[0007] Step 2: Extraction and weighting of empty carrier at the receiving end:

[0008] Since the receiving end knows the position of the empty carrier in the symbol, the decoder can extract the noise value distributed on each empty carrier based on the known information, and weight it according to the relative distance between the empty carrier and the center frequency. The larger the weight given to the empty carrier close to the center frequency, the smaller the weight given to the empty carrier far away from the center frequency.

[0009] Step 3: Noise grouping and selection:

[0010] Input the output weighted noise sequence into the cluster grouping selector, and input the number of cluster groups as 2. The cluster grouping selector first selects two cluster centers, then calculates the distance between each noise sample and the two cluster centers, and then regresses each noise sample to the cluster center with the closest distance. Then, the sample mean of the two classes is used as the new cluster center. If the center no longer changes, the iteration ends and the samples of the two classes are output. Compare the number of samples of the two classes, select the class with the larger number for subsequent noise variance calculation. If the number of the two classes is the same, use all samples for noise variance calculation;

[0011] Step 4: Calculation of noise variance:

[0012] The weighted null carrier noise measurement value is extracted after the output grouping, and the square average of all samples is taken to obtain the noise variance of the received OFDM symbol.

[0013] The present invention adopts a clustering-based underwater acoustic OFDM signal noise variance calculation method. Therefore, compared with the existing method of directly using empty carriers for noise estimation, the present invention has the following significant beneficial effects:

[0014] 1. The present invention utilizes the Doppler effect in the inconsistency of broadband signals and the residual ICI characteristics of OFDM underwater acoustic communication signals, and uses different subcarrier weighting methods to measure the carrier frequency, thereby improving the accuracy of carrier frequency measurement;

[0015] 2. The present invention groups the weighted empty carrier frequency measurement values ​​by clustering, thereby removing empty carrier frequency measurement values ​​with large errors and improving the accuracy of the final noise variance calculation. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art or ordinary technicians, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 It is a schematic diagram of the arrangement of empty subcarriers of the present invention.

[0018] Figure 2 It is a schematic diagram of weight allocation according to the position of empty carrier in the present invention. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work belong to the protection scope of the present invention.

[0020] This embodiment illustrates the working mode of a method for calculating noise variance of underwater acoustic OFDM signals based on clustering according to the present invention through specific steps.

[0021] Step 1: Setting the empty carrier at the transmitter:

[0022] In the present invention, the receiver measures the noise variance of the underwater acoustic OFDM symbol by using the empty subcarrier, so the transmitter and the receiver need to share the empty carrier allocation method in advance;

[0023] [1] At the transmitter end of underwater acoustic OFDM signals, it is necessary to first arrange empty carriers in the frequency domain when performing OFDM symbol mapping.

[0024] The carrier wave is not filled with data or pilots, so that the empty carrier position set at the receiving end is only affected by noise;

[0025] [2] The OFDM upper carrier needs to be Figure 1 The method shown is evenly arranged in the effective frequency band, and the number is not less than 1 / 64 of the total subcarriers, so that the measured noise variance is not accurate enough, but not more than 1 / 16 of the total number of subcarriers. Too many empty carriers will reduce the transmission efficiency of OFDM symbols;

[0026] Step 2: Extraction and weighting of empty carrier at the receiving end:

[0027] In the present invention, since the received underwater acoustic OFDM signal is affected by the channel and introduces ICI, the noise at the empty carrier position received at the receiving end is the additive noise of the equivalent ICI and the environmental noise, so a series of pre-processing is required to eliminate the influence of the main ICI on the noise extracted by the receiver. Although the OFDM symbol obtained after a series of processing eliminates the main ICI influence, the residual ICI will also increase the noise error of the measurement, so the extracted noise needs to be weighted;

[0028] [1] First, the received OFDM signal is time synchronized, and then down-converted to baseband for easy processing.

[0029] The baseband OFDM symbols are subjected to Doppler search, Doppler compensation, channel estimation, and channel equalization, and then to FFT (fast

[0030] The main ICI effect can be eliminated after conversion to the frequency domain by fast Fourier transform;

[0031] [2] According to the agreed empty carrier position between the transmitter and the receiver, the empty carrier frequency in the received OFDM symbol is extracted and measured.

[0032] The value is z[m]. Due to the influence of residual ICI, the frequency measurement value of the empty carrier is affected by residual ICI and environmental noise. The farther away from the center frequency position, the more residual ICI there is, and the stronger the impact is. Therefore, it is necessary to perform weighted processing on the frequency measurement value, such as Figure 2 As shown, the highest weight is 1 near the center frequency side, and the lowest weight is 0.8 far from the center frequency side. The weighted null carrier frequency measurement value is γ[m]z[m];

[0033] Step 3: Noise grouping and selection:

[0034] The influence of the channel on OFDM symbols is non-stationary. Although weighted processing is performed in step 2 to reduce the influence of consistent residual ICI, the inconsistent residual ICI caused by the non-stationary process has a greater impact on the noise variance measurement. At the same time, the number of empty carriers measured is small, and a single frequency measurement value with a large error will have a considerable impact on the entire system. Therefore, it is necessary to group and select noise to eliminate frequency measurement values ​​with large errors.

[0035] [1] First, initialize the cluster grouper, initialize the number of cluster groups to 2, and input the number of measurement elements to the selected empty subcarrier

[0036] The number is M, randomly select two elements as the two group centroids of the cluster grouper, and initialize the two groups

[0037] C t =φt=1,2;

[0038] [2] For each weighted empty carrier frequency measurement sample in the data set, calculate their distance to the two new groups of quality.

[0039] Leave where x i is the weighted frequency measurement sample, μ j is the selected centroid, and x i The smallest d ij The corresponding category λ i Update the group at this time

[0040] [3] Recalculate the new centroid for all sample points in the two groups If both centroid vectors are not updated,

[0041] Then proceed to step 3.4, and if the centroid vector is updated, repeat step 3.2;

[0042] [4] Output grouping C = {C1, C2}, select the group with more internal elements as the input group for the next step.

[0043] If the number of group elements is the same, all elements are grouped as inputs for the next step;

[0044] Step 4: Calculation of noise variance:

[0045] Extract the weighted null carrier noise measurement value after output grouping and calculate Where S M is the number of subcarriers in the selected valid group, c[m] is the weighted frequency measurement value in the valid group, and the obtained σ 2 It is the noise variance of the received OFDM symbol.

[0046] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A method for calculating noise variance of underwater acoustic OFDM signals based on clustering, characterized by: The method comprises the following steps: setting empty subcarriers uniformly for OFDM symbols at the transmitting end, extracting the value of the empty carrier when demapping OFDM at the receiving end, which is the amount of noise in the frequency domain, extracting noise groups from different empty carrier positions using a clustering algorithm, and improving the accuracy of OFDM symbol noise estimation by eliminating groups with small samples.

2. The method for calculating noise variance of underwater acoustic OFDM signals based on clustering according to claim 1, characterized in that: The specific steps are as follows: Step 1: Setting the empty carrier at the transmitter: At the transmitting end of the underwater acoustic OFDM signal modulation, the empty carriers are evenly arranged in the OFDM symbol frequency domain to measure the noise value; Step 2: Extraction and weighting of empty carrier at the receiving end: The receiving end knows the position of the empty carrier in the symbol. The decoder extracts the noise value distributed on each empty carrier based on the known information, and weights it according to the relative distance between the empty carrier and the center frequency. The larger the weight given to the empty carrier close to the center frequency, the smaller the weight given to the empty carrier far from the center frequency. Step 3: Noise grouping and selection: The output weighted noise sequence is input into the cluster grouping selector, and the number of cluster groups is 2. The cluster grouping selector first selects two cluster centers, calculates the distance between each noise sample and the two cluster centers, regresses each noise sample to the cluster center with the closest distance, and then uses the sample mean of the two classes as the new cluster center. If the center no longer changes, the iteration ends and the samples of the two classes are output; the number of samples of the two classes is compared, and the class with the larger number is selected for the subsequent noise variance calculation. If the number of the two classes is the same, all samples are used for noise variance calculation; Step 4: Calculation of noise variance: The weighted null carrier noise measurement value is extracted after the output grouping, and the square average of all samples is taken to obtain the noise variance of the received OFDM symbol.

3. The method for calculating noise variance of underwater acoustic OFDM signals based on clustering according to claim 2 is characterized in that: The number of arranged empty carriers shall not be less than 1 / 64 of the total number of subcarriers and shall not exceed 1 / 16 of the total number of subcarriers.

Images