阅读说明：本技术 一种水声ofdm接收机中时频域联合抑制ici的方法和系统 (Method and system for inhibiting ICI (inter-carrier interference) by combining time domain and frequency domain in underwater sound OFDM (orthogonal frequency division multiplexing) receiver ) 是由 李渝舟 黄运龙 邱吉慧 于 2020-12-31 设计创作，主要内容包括：本发明公开了一种水声OFDM接收机中时频域联合抑制ICI的方法和系统,属于水声无线通信技术领域。包括：利用分段快速傅里叶变换P-FFT中信号分段的方式,将一个OFDM块持续时间内的基带接收信号分段为多个非重叠短信号；借鉴分数快速傅里叶变换F-FFT在多个频率上进行傅里叶变换的思想,对每个分段后的非重叠短信号在载波频率和以特定基准偏移频率的分数倍偏移载波频率的多个频率上进行傅里叶变换,从而在每个载波处产生多个输出；将每个载波处的多个输出加权合并成一路,用于符号检测,并利用随机梯度算法求解和更新加权合并的权重。本发明通过结合了P-FFT和F-FFT的特点,联合时域和频域抑制ICI,在多普勒因子和载波数都较大的情况下,性能有显著的提升。(The invention discloses a method and a system for inhibiting ICI (inter-carrier interference) by combining time domain and frequency domain in an underwater sound OFDM (orthogonal frequency division multiplexing) receiver, belonging to the technical field of underwater sound wireless communication. The method comprises the following steps: segmenting a baseband receiving signal in the duration of one OFDM block into a plurality of non-overlapping short signals by utilizing a signal segmentation mode in segmented fast Fourier transform (P-FFT); taking into account the idea that fractional fast fourier transform F-FFT performs fourier transform on multiple frequencies, fourier transforming each segmented non-overlapping short signal on a carrier frequency and multiple frequencies that shift the carrier frequency by a fraction of a specific reference shift frequency, thereby producing multiple outputs at each carrier; and combining a plurality of output weights at each carrier into a path for symbol detection, and solving and updating the weights of the weighted combination by using a random gradient algorithm. The invention combines the characteristics of P-FFT and F-FFT, and inhibits ICI by combining the time domain and the frequency domain, and the performance is obviously improved under the condition that both Doppler factors and the number of carriers are large.)

1. A method for jointly suppressing ICI in a time-frequency domain in an underwater acoustic OFDM receiver is characterized by comprising the following steps:

s1, segmenting a baseband receiving signal in the duration of one OFDM block into a plurality of non-overlapping short signals by utilizing a signal segmentation mode in segmented fast Fourier transform (P-FFT);

s2, taking the thought that fractional fast Fourier transform F-FFT carries out Fourier transform on carrier frequency and a plurality of frequencies of the carrier frequency shifted by the fractional times of the carrier interval as reference, carrying out Fourier transform on each segmented non-overlapped short signal on the carrier frequency and a plurality of frequencies of the carrier frequency shifted by the fractional times of the specific reference shift frequency;

and S3, combining a plurality of output weights of the current carrier into one path for symbol detection, updating the weights of the weighted combination of the next carrier, and sequentially carrying out detection until all carriers are detected.

2. The method of claim 1, wherein step S1 is specifically as follows: multiplying the baseband received signal v (t) by a series of non-overlapping rectangular window functions phi_a(t) obtaining A segmented signals v_a(t)＝v(t)φ_a(t)，a＝0，1，...，A-1。

3. The method of claim 2, wherein the rectangular window function φ_a(t) specifically the following:

4. the method of claim 1, wherein step S2 is specifically as follows:

v is to be_a(t) at 2B preset frequenciesFrequency shifting to obtain 2B frequency-shifted signals,where B ± + -1, ± -2, …, ± -B, is compared with the original unshifted signal v_a(t) together form (2B +1) signals v_a，b(t), B ═ 0, ± 1, ± 2, …, ± B, expressed as:

for each v_a，b(t) performing Fourier transform

Wherein f is_eFor the reference offset frequency, B represents the number of single-sided offset frequencies, k represents the carrier number, and Δ f represents the carrier spacing.

5. The method of claim 4, wherein the step S3 includes the steps of:

all (2B +1) outputs z with the same index a at the same carrier k_k，a，bArranged in a column vector z_k，a＝[z_k，a，-B，...，z_k，a，-1，z_k，a，0，z_k，a，1，...，z_k，a，B]^TThen all A column vectors z at the same carrier k_k，aFurther arranged into column vectorsFinally using a length and z_kEqual column vectors w_kIs multiplied by z_kObtain a combined symbol x_k：

6. The method of claim 5, wherein step S3 further comprises:

carrying out differential coherent detection on the combined signals to obtain detection symbolsPerforming symbol decision on the detected symbol to obtain a decision symbol

7. The method of claim 6, wherein step S3 further comprises:

using detected symbolsAnd decision symbolsThe mean square error between them calculates the original gradient g_kAnd updating the weight at the next carrier by using the scaled gradient.

8. The method of claim 7, wherein the updating the weights at the next carrier using the scaled gradient comprises:

first, the original gradient g is calculated_k，g_kThe calculation expression of (2) is specifically as follows:

wherein the content of the first and second substances,

the original gradient g_kMultiplied by | x_k-1I, zooming to obtain a zoomed gradient

When | e_kI andare respectively smaller than the corresponding preset threshold value e_thAnd g_thUsing scaled gradientsAnd updating the weight w at the next carrier by the preset step size coefficient mu_k+1：

If the threshold condition is not met, the weight is not updated, namely the weight of the next carrier is the same as the weight at the current carrier:

w_k+1＝w_k。

9. the method of claim 5, wherein step S3 further comprises:

coherent detection is carried out on the combined signals to obtain detection symbolsPerforming symbol decision on the detected symbol to obtain a decision symbol

10. A system for jointly suppressing ICI in time-frequency domain in an underwater acoustic OFDM receiver, comprising: a computer-readable storage medium and a processor;

the computer-readable storage medium is used for storing executable instructions;

the processor is configured to read executable instructions stored in the computer-readable storage medium, and perform the method for jointly suppressing ICI in time-frequency domain in the underwater acoustic OFDM receiver according to any one of claims 1 to 9.

Technical Field

The invention belongs to the technical field of underwater sound wireless communication, and particularly relates to a method and a system for jointly inhibiting ICI (inter-carrier interference) in a time-frequency domain in an underwater sound OFDM (orthogonal frequency division multiplexing) receiver.

Background

The ocean occupies 71 percent of the surface area of the earth, contains abundant mineral resources, chemical resources, biological resources and power resources, and is a global consensus for exploring and developing the ocean which is an important space for the future survival and development of human beings. The high-speed underwater acoustic communication technology is used as a key technology for realizing underwater mass information transmission and constructing an intelligent ocean network, and plays a vital role in various military, civil and commercial maritime activities. However, due to the low-speed propagation characteristic of the acoustic wave and the harsh underwater environment, the underwater acoustic channel has the characteristics of low propagation speed, strong multipath delay spread, severe doppler shift and extremely limited available bandwidth compared with a ground radio channel. Thus, designing and deploying high-speed underwater acoustic communication systems faces many challenges.

Orthogonal Frequency Division Multiplexing (OFDM), a multi-carrier technique, converts a high-speed serial transmission data stream into a low-speed parallel transmission data stream, and can effectively combat delay spread caused by multipath effects. On the other hand, the OFDM allows the frequency spectrums among the carriers to be overlapped, and improves the frequency spectrum efficiency. Since OFDM has these technical advantages, it can effectively solve some of the challenges caused by the above-mentioned underwater acoustic channel, and thus is considered as a standard technique for realizing high-speed underwater acoustic communication. To recover the data symbols from the received signal, OFDM systems typically require the use of coherent or differentially coherent detection techniques at the receiver.

For an OFDM system using coherent detection, channel estimation needs to be performed first, so that the influence of a channel on a signal is eliminated according to acquired channel state information, and a data symbol is recovered from a received signal. For OFDM systems using differential coherent detection, the need for channel estimation is eliminated by differentially encoding the data symbols at the transmitter and differentially detecting the data symbols at the receiver using the coherence between adjacent carriers. However, the underwater acoustic channel generally has a strong doppler effect due to the low-speed propagation characteristics of acoustic waves under water. Further, in the case that Inter-Carrier Interference (ICI) caused by doppler shift spreading is not effectively suppressed, orthogonality between OFDM carriers will be seriously destroyed, and the performance of the two detection methods will be greatly reduced, even the two detection methods cannot work normally.

Therefore, an ICI suppression method for an underwater acoustic OFDM system with excellent performance is in need of research.

Disclosure of Invention

In view of the defects and the improved requirements of the prior art, the present invention provides a method and a system for jointly suppressing ICI in time-frequency domain in an underwater acoustic OFDM receiver, which is called segmented-Fast Fourier Transform (PS-FFT), and aims to combine segmented Fast Fourier Transform (P-FFT) and Fractional Fast Fourier Transform (F-FFT) and jointly suppress ICI in time domain and frequency domain. Taking a differential coherent detection system as an example, the invention has obvious performance improvement under the condition of large Doppler factor and carrier number.

To achieve the above object, according to a first aspect of the present invention, there is provided a method for jointly suppressing ICI in a time-frequency domain in an underwater acoustic OFDM receiver, the method including:

s1, segmenting a baseband receiving signal in the duration of one OFDM block into a plurality of non-overlapping short signals by utilizing a signal segmentation mode in segmented fast Fourier transform (P-FFT);

s2, taking the thought that fractional fast Fourier transform F-FFT carries out Fourier transform on carrier frequency and a plurality of frequencies of the carrier frequency shifted by the fractional times of the carrier interval as reference, carrying out Fourier transform on each segmented non-overlapped short signal on the carrier frequency and a plurality of frequencies of the carrier frequency shifted by the fractional times of the specific reference shift frequency;

and S3, combining a plurality of output weights of the current carrier into one path for symbol detection, updating the weights of the weighted combination of the next carrier, and sequentially carrying out detection until all carriers are detected.

The updating method can be a random gradient algorithm or a Newton method and the like. After appropriate weights are obtained by solving through methods such as a random gradient algorithm, the ICI can be effectively inhibited when weighting combination is carried out. Specifically, the processes of weighted combination and weight updating are performed successively on each carrier, the first carrier is weighted and combined by using a preset initialization weight, and then the weight at the second carrier is updated by using a random gradient algorithm and the like. When the second carrier is weighted and combined, the updated weight of the first carrier is used, and then the weight … … of the third carrier is updated by methods such as a random gradient algorithm and the like in sequence until all carriers are detected.

Preferably, step S1 is specifically as follows: multiplying the baseband received signal v (t) by a series of non-overlapping rectangular window functions phi_a(t) obtaining A segmented signals v_a(t)＝v(t)φ_a(t)，a＝0,1,…,A-1。

Preferably, the rectangular window function phi_a(t) specifically the following:

preferably, step S2 is specifically as follows:

v is to be_a(t) at 2B preset frequenciesFrequency shifting is carried out to obtain 2B frequency-shifted signals, wherein B is +/-1, +/-2, …, +/-B, and the original frequency-unshifted signal v_a(t) together form (2B +1) signals v_a,b(t), B ═ 0, ± 1, ± 2, …, ± B, expressed as:

for each v_a,b(t) performing Fourier transform

Wherein f is_eFor the reference offset frequency, B represents the number of single-sided offset frequencies, k represents the carrier number, and Δ f represents the carrier spacing.

Preferably, step S3 includes the steps of:

all (2B +1) outputs z with the same index a at the same carrier k_k,a,bArranged in a column vector z_k,a＝[z_k,a,-B,…,z_k,a,-1,z_k,a,0,z_k,a,1,…,z_k,a,B]^TThen all A column vectors z at the same carrier k_k,aFurther arranged into column vectorsFinally using a length and z_kEqual column vectors w_kIs multiplied by z_kObtain a combined symbol x_k：

Preferably, step S3 further includes:

carrying out differential coherent detection on the combined signals to obtain detection symbolsPerforming symbol decision on the detected symbol to obtain a decision symbol

Preferably, the updating the weight of the next carrier by using the scaled gradient specifically includes the following steps:

first, the original gradient g is calculated_k，g_kThe calculation expression of (2) is specifically as follows:

wherein the content of the first and second substances,

the original gradient g_kMultiplied by | x_k-1I, zooming to obtain a zoomed gradient

When | e_kI andare respectively smaller than the corresponding preset threshold value e_thAnd g_thUsing scaled gradientsAnd updating the weight w at the next carrier by the preset step size coefficient mu_k+1：

If the threshold condition is not met, the weight is not updated, namely the weight of the next carrier is the same as the weight at the current carrier:

w_k+1＝w_k。

preferably, step S3 further includes:

coherent detection is carried out on the combined signals to obtain detection symbolsPerforming symbol decision on the detected symbol to obtain a decision symbol

To achieve the above object, according to a second aspect of the present invention, there is provided a system for jointly suppressing ICI in a time-frequency domain in an underwater acoustic OFDM receiver, including: a computer-readable storage medium and a processor;

the computer-readable storage medium is used for storing executable instructions;

the processor is configured to read executable instructions stored in the computer-readable storage medium, and execute the method for jointly suppressing ICI in time-frequency domain in an underwater acoustic OFDM receiver according to the first aspect.

Generally, by the above technical solution conceived by the present invention, the following beneficial effects can be obtained:

the invention designs a new ICI inhibition method called PS-FFT by combining the characteristics of P-FFT and F-FFT. On one hand, a signal segmentation mode in P-FFT is adopted to segment a baseband receiving signal into a plurality of non-overlapped short signals, and because the channel variation is smaller in a shorter time interval after the segmentation than in a longer time interval before the segmentation, the time-varying property of the channel is reduced in a time domain, and the ICI is inhibited from the angle of the time domain; on the other hand, taking the idea that F-FFT performs fourier transform on a carrier frequency and a plurality of frequencies that shift the carrier frequency by a fraction of the carrier spacing as reference, fourier transform is performed on a signal on a carrier frequency and a plurality of frequencies that shift the carrier frequency by a fraction of a specific reference shift frequency, equivalently, the signal is frequency shifted by a plurality of frequencies first, and then fourier transform (demodulation) is performed on the signal on the carrier frequency, and the reference shift frequency F is set reasonably according to the magnitude of the doppler shift of the system_eCan effectively compensate forThe ICI is suppressed from the frequency domain perspective. Combining the two, ICI can be suppressed in both time and frequency domains jointly. The invention has obvious performance improvement under the condition of large Doppler factor and carrier number。

Drawings

Fig. 1 is a block diagram of a receiver according to the present invention;

FIG. 2 is a graph of the path gain of a simulated underwater acoustic channel under test in accordance with the present invention;

FIG. 3 is a diagram of the MSE performance of the present invention and the prior art in a differential coherent detection system as a function of Doppler factor;

fig. 4 is a test result graph of MSE performance varying with the signal-to-noise ratio in a differential coherent detection system according to the present invention and the prior art.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. In addition, the technical features involved in the embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.

The invention provides a method for inhibiting ICI (inter-carrier interference) by combining time domain and frequency domain in an underwater sound OFDM (orthogonal frequency division multiplexing) receiver, which depends on a specific underwater sound OFDM system, and an end-to-end system flow from a transmitter data bit to a receiver data bit specifically comprises the following processing steps:

step one, for a differential coherent detection system, in a transmitter, performing Phase-Shift Keying (PSK) modulation, differential coding, and inverse fourier transform on original data bits in sequence to obtain a baseband time domain transmission signal s (t). For the coherent detection system, the transmitter sequentially performs Q-phase shift keying modulation and inverse fourier transform on the original data bits to obtain a baseband time domain transmission signal s (t).

Specifically, in a differential coherent detection system, an original data symbol b is generated by performing energy-normalized Q-order PSK modulation on original data bits_kBelonging to a set of constellation symbolsWherein, aggregateMiddle element a_q＝e^j2πq/Q,q＝0,1,…,Q-1。

To b is_kCarrying out differential coding according to a coding rule shown in formula (1) to obtain a differential symbol d actually carried on each carrier wave for transmission_k；

Where K is the number of OFDM system carriers.

Further, for d_kPerforming an inverse fourier transform to obtain the baseband time domain transmission signal s (t) involves a transform formula as shown in (2).

Where T is the OFDM block duration, f_k＝f₀+ k Δ f is the frequency of the k-th carrier, f₀Is the lowest carrier frequency and Δ f is the carrier spacing. In a digital system, equation (2) can be implemented by a corresponding Inverse Fast Fourier Transform (IFFT).

In particular, in the coherent detection system, the steps are the same as those in the above-described differential coherent detection system except that there is no differential encoding step shown in formula (1).

And step two, adding a cyclic prefix to the baseband time domain transmission signal s (t) by the transmitter, carrying out up-conversion to obtain a radio frequency transmission signal, transmitting the radio frequency transmission signal to the receiver through an underwater acoustic channel, carrying out resampling on the received radio frequency signal by the receiver, and carrying out down-conversion and cyclic prefix removal to obtain a baseband receiving signal v (t).

Step three, as shown in fig. 1, at the receiver, the baseband received signal v (t) within one OFDM block duration is segmented and then fourier transformed at a plurality of preset frequencies, so that at each carrier, a signal will be generatedGenerating a multi-path demodulated output z_k,a,b。

Specifically, the step of segmenting the baseband received signal v (t) is equivalent to multiplying v (t) by a series of non-overlapping rectangular window functions phi shown in formula (4)_a(t) obtaining A segmented signals v_a(t)：

v_a(t)＝v(t)φ_a(t),a＝0,1,…,A-1 (3)

Wherein the function

Then, the segmented signal v is processed_a(t) the step of performing fourier transform at a plurality of predetermined frequencies is equivalent to:

v is first introduced_a(t) at 2B preset frequenciesB + -1 + -2, … + -B to obtain 2B frequency-shifted signals, and mixing with the original un-frequency-shifted signal v_a(t) together form (2B +1) signals v_a,b(t), expressed as:

for each signal v_a,b(t) Fourier transform to obtain a multi-path demodulated output z_k,a,bTo the transformation formula shown in (6).

The Fourier Transform expression shown in equation (6) is a continuous form, and can be implemented in a digital system by using a corresponding Fast Fourier Transform (FFT).

Step four, at the receiver, outputting the A (2B +1) demodulation outputs z at each carrier wave_k,a,bWeighting and combining into a signal x_k。

Specifically, first, all (2B +1) demodulated outputs z having the same index a at the same carrier k are output_k,a,bArranged in a column vector z_k,a：

z_k,a＝[z_k,a,-B,…,z_k,a,-1,z_k,a,0,z_k,a,1,…,z_k,a,B]^T (7)

Then, all A column vectors z at the same carrier k are added_k,aFurther arranged as a column vector z_k：

Finally, again using a length and z_kEqual column vectors w_kIs multiplied by z_kObtaining a weighted and combined symbol x_k：

Step five, for the differential coherent detection system, at the receiver, the combined symbol x_kAnd carrying out differential coherent detection, symbol judgment and Q-order phase shift keying demodulation in sequence to obtain received data bits. For coherent detection systems, at the receiver, the combined symbols x are combined_kAnd carrying out coherent detection, symbol judgment and Q-order phase shift keying demodulation in sequence to obtain received data bits.

In particular, in a differentially coherent detection system, the current symbol x is utilized_kAnd the previous symbol x_k-1The coherence between them is detected by differential coherence to obtain the detected symbolIs expressed as:

then for the detected symbolIs subjected to symbol decision to obtainBy calculatingAnd point a on the PSK constellation_qThe mean square error between, find a that minimizes the mean square error_qIs the judgment resultThe specific calculation formula is expressed as:

in particular, in a coherent detection system, detection in a differential coherent detection system as shown in (10)The formula of (a) needs to be replaced with a formula in the coherent detection system, and other steps are the same as those in the differential coherent detection system described above.

And step six, solving and updating the weight of the weighted combination at the receiver by using a random gradient algorithm.

Specifically, in the differential coherent detection system, the detection symbol is calculated in the manner shown in formula (12)And decision symbolsMean square error of | e between_k|²：

And one-step approximation is carried out:

then on the mean square error | e_k|²Computing a vector w about weights_kThe solution result is expressed as:

wherein the content of the first and second substances,

taking the inverse number of the derivation result to obtain the original gradient g_k：

For the obtained original gradient g_kMultiplied by | x_k-1I, zooming to obtain a zoomed gradient

Finally, according to the error e_kOf the model and the original gradient vector g_kThe size of the medium data and the test experience reasonably set two thresholds e_thAnd g_thWhen error e_kModulo e of_kI and original gradient g_kInner product of (2)When the conditions shown in the formulae (17) and (18) are satisfied simultaneously, the gradient is appliedUpdating the weight vector w for weighted combination at the next carrier with the preset step size coefficient mu_kAs shown in formula (19):

|e_k|<e_th (17)

if the threshold condition is not met, the weights are not updated, i.e. the weights at the next carrier are the same as the weights at the current carrier:

w_k+1＝w_k (20)

it should be noted that, the algorithm for updating the weights is divided into a training mode and a decision-driven mode. At the transmitter, data symbols on the first several carriers of a frame of data are set as pilots, i.e. these symbols are known to the receiver. The receiver performs symbol-by-symbol detection, the first pilot symbol is detected and placed at the head of the data frame, when the algorithm works in training mode, and when calculating the mean square error shown in equation (12), the algorithm will workSymbol b arranged as transmitter_k. After the detection at the pilot symbols is completed, the algorithm enters a decision driven mode,is a detected symbolAnd (4) judging results.

In particular, in coherent detection systems, due to detectionIs different from the formula (10) in the differential coherent detection system, and accordingly, is determined byResulting mean squared error | e_k|²Gradient g_kThe expression of the result is also different from the results in the formulas (12), (13), etc. The other steps are the same as in the above-described differential coherent detection system.

Fig. 2 is a path gain diagram of a simulated underwater acoustic channel for testing, in which Single-FFT, P-FFT, F-FFT and the method of the present invention are respectively used to perform end-to-end OFDM system simulation testing in the simulated underwater acoustic channel environment shown in fig. 2, so as to show the performance of different methods for suppressing ICI in a differential coherent detection system, and the test results are shown in fig. 3 and fig. 4, where fig. 3 is a test result diagram of MSE performance varying with doppler factor, and fig. 4 is a test result diagram of MSE performance varying with signal-to-noise ratio.

Through simulation experiment results, compared with P-FFT and F-FFT technologies, the PS-FFT technology provided by the invention has the advantage of relatively obvious Mean Square Error (MSE) performance in a differential underwater acoustic OFDM system with a large carrier number and a large Doppler factor.

It will be understood by those skilled in the art that the foregoing is only a preferred embodiment of the present invention, and is not intended to limit the invention, and that any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the present invention.