CN103441980A - Sideband-information-free shallow sea underwater acoustic communication pattern selection peak-to-average ratio restraining algorithm based on frequency reversal mirror technology - Google Patents

Abstract

The invention provides a sideband-information-free shallow sea underwater acoustic communication pattern selection peak-to-average ratio restraining algorithm based on a frequency reversal mirror technology. At a data sending end, scrambling is carried out on data to be sent through M sets of different scrambling patterns, meanwhile, M sets of different comb-shaped pilot frequency positions are defined as signs, comb-shaped pilot frequency sequences are added to different pilot frequency positions to obtain candidate OFDM symbols, fast Fourier inversion is carried out on the candidate OFDM symbols to a time domain, and the symbol with the smallest peak-to-average ratio is selected as a transmission symbol; at a receiving end, fast Fourier inversion is carried out on received signals, frequency domain received information corresponding to the different comb-shaped pilot frequency positions is extracted to obtain pilot frequency domain response, matching pursuit is carried out on the pilot frequency domain reesponse, M sets of channel frequency domain response are restored, in frequency domain reversion, the cross correlation coefficients of the pilot frequency domain response and the channel frequency domain response is calculated, the value m corresponding to the maximum value in the M correlation coefficients is obtained, and finally descrambling is carried out through serial number information. The algorithm can guarantee the real-time performance and the reliability of an MIMO-OFDM underwater acoustic communication system.

Description

Select the peak-to-average force ratio Restrainable algorithms based on the frequency time reversal technique without side information shallow-sea underwater acoustic communication pattern

Technical field

What the present invention relates to is a kind of method for restraining peak average power ratio based on the MIMO-OFDM underwater sound communication.

Background technology

Bandwidth resources available in underwater sound communication are very limited, seriously limited quality and the speed of underwater sound communication.Under this background, the orthogonal frequency division multiplexi that high spectrum utilization arranged, in conjunction with not needing to increase available bandwidth or improving transmitting power and can make channel capacity obtain the MIMO technique that essence improves, becomes the new focus of high-speed underwater sound communication in recent years.Shallow Sea Acoustic Channel is looked at as thinning thin channel when slow usually, and in the time of can be with one, constant filter be described.Compressed sensing, as a kind of method for solving that limits the sparse solution of system for linearity, can accurately be recovered sparse signal by a small amount of information, therefore is applicable to the Shallow Sea Acoustic Channel with sparse property is rebuild.Affecting MIMO-OFDM significant obstacle of extensive use in underwater sound communication is its defect that has peak value average power ratio (Peak-to-average-power-ratio, PAPR).This defect not only can limit emission system efficiency, reduces the signal to noise ratio that receives signal, also can cause signal distortion, causes nonlinear distortion, destroy sub-carrier orthogonality, reduces systematic function.Abundant about the document of peak-to-average force ratio Restrainable algorithms in the research radio MIMO-OFDM communication technology both at home and abroad, but due to underwater acoustic channel communication bandwidth resource-constrained, low signal-to-noise ratio, slow time-varying characteristics, simply adopt traditional peak-to-average force ratio Restrainable algorithms can obviously reduce system communication speed.

Selectivity mapping (Selective Mapping, SLM) technology proposes at R.w.Bauml in 1996, and its basic principle is to utilize M scrambling phase sequence

with input signal X=[X (0), X (1) ... X (N-1)] multiply each other, wherein

n=0,1 ... N-1, m=1,2 ... M, after scrambling, data are X ^m=[X ^m(0), X ^m(1) ... X ^m(N-1)].To { X ^m(n) M } different OFDM symbols carry out independently IFFT conversion and obtain x ^m=[x ^m(0), x ^m(1) ... x ^m(N-1)], select the wherein OFDM(OFDM of peak-to-average force ratio minimum) symbol

as transfer sequence.Wherein $\tilde{m}=\underset{m=\mathrm{1,2},...,M}{\arg \min }\left(\underset{n=\mathrm{0,1},...,N-1}{\max }\right|x^m\left(n\right)\left|\right).$

Though it is a bit weaker but there is no the performance of BER loss that the method suppresses the PAPR effect than the method for direct amplitude limit, weak point is to carry out IFFT conversion repeatedly, and computation complexity is higher.And, in order to recover accurately initial data, the supplementary that transmitter need to be adopted is delivered to receiver accurately.Once supplementary makes a mistake, can cause the whole mistakes of information entrained in a symbol, therefore need to carry out special coding to this information, or repeatedly repeat to send to guarantee at receiving terminal restoring signal exactly.So just need a large amount of side information of transmission, cause the reduction of system communication efficiency and the waste of frequency band.

Summary of the invention

The object of the present invention is to provide on a kind of basis not losing PAPR, that can effectively improve band efficiency selects the peak-to-average force ratio Restrainable algorithms based on the frequency time reversal technique without side information shallow-sea underwater acoustic communication pattern.

The object of the present invention is achieved like this:

(1) at data sending terminal, at first by data X to be sent _dadopt M to organize different scrambling pattern P ^mcarry out scrambling, obtain the whole series of data after scrambling

define M simultaneously and organize different Comb Pilot position Po ^mserve as a mark, corresponding different scrambling patterns, at different pilot frequency locations Po ^madd the Comb Pilot sequence X _pobtain candidate OFDM symbol X ^m, candidate OFDM symbol is carried out to inverse fast Fourier transform to time domain, select the symbol of peak-to-average force ratio minimum wherein as transmission symbol;

(2) carry out at first to received signal fast Fourier transform at receiving terminal, extract frequency domain corresponding to different Comb Pilot position and receive information, obtain the pilot tones domain response and it is carried out to match tracing, recover M group channel frequency domain response

in the frequency domain reversion, obtain

calculate the two cross-correlation coefficient, in the M of an acquisition coefficient correlation, the corresponding m value of maximum is the sequence number of scrambling pattern, finally utilizes the serial number information descrambling code.

Described step (1) specific implementation step is as follows:

1. produce random phasic serial signal P ^m=[P ^m(0), P ^m(1) ..., P ^m(N-1)], 0≤m≤M, wherein M is the scramble sequence number.

2. at data sending terminal, by data X to be sent _d=[X _d(0), X _d(1) ..., X _d(N-1)] with all M group scrambling pattern P ^m=[P ^m(0), P ^m(1) ..., P ^m(N-1)], 0≤m≤M multiplies each other, and obtains the whole series of data after scrambling

3. define M and organize different Comb Pilot position Po ^mserve as a mark, corresponding different scrambling patterns.By data sequence after scrambling

with the Comb Pilot sequence X _paccording to correspondence position Po ^mform respectively candidate OFDM symbol X ^m, now in each candidate OFDM symbol, carried scrambling pattern serial number information, be the side information in conventional method.

4. inverse fast Fourier transform: OFDM modulation obtains time-domain signal x ^m.

5. select the wherein minimum symbol of PAPR value

as transmission symbol.

Described step 1. in the degree of correlation of scrambling phase sequence lower, to avoid affecting because the degree of correlation is higher the judgement of scrambling phase sequence.

Described step pilot frequency sequence X 3. _pshould select the sequence without negative effect to the PAPR inhibition, avoid adding the peak value regeneration after pilot tone.

Described step (2) specific implementation step is as follows:

1. fast Fourier transform, the frequency domain data Y (k) obtained.

2. extract respectively the corresponding frequency domain of M group pilot frequency locations and receive information, obtain channel strip frequency domain response estimation information after channel estimating

by the matching pursuit algorithm match channels, recover M group channel frequency domain response

3. by the M group channel frequency domain response obtained

reversion obtains

calculate with

cross-correlation coefficient, in the M of an acquisition coefficient correlation, maximum is corresponding

be the corresponding estimated value of frequency domain of the reality of channel corresponding m value is the sequence number of scrambling pattern.

4. utilize the scrambling pattern obtained to receiving the frequency-region signal descrambling code, and, to its demapping, channel-decoding, can recover to send data, realize selecting the peak-to-average force ratio Restrainable algorithms without the pattern of side information transmission.

The invention provides a kind of MIMO-OFDM under sparse underwater acoustic channel environment pattern system of selection without side information that is applicable to, the method be take the selectivity mapping techniques as basis, in conjunction with the sparse characteristics of underwater acoustic channel, algorithm is improved.Reduce the peak-to-average power ratio of OFDM symbol by the scrambling phase sequence at transmitting terminal, utilize Comb Pilot to carry scramble sequence in the distribution pattern of frequency domain and select information, sparse features at receiving terminal in conjunction with Shallow Water Acoustic Channels, utilize compressive sensing theory to estimate channel, adopt the frequency domain time reversal technique scrambling pattern to be carried out to the autonomous differentiation of high reliability, realized selecting peak-to-average force ratio to suppress without the pattern of side information, under the prerequisite of not losing the PAPR rejection, realized the pattern selection algorithm without the side information transmission, the probability that cause the whole mistakes of whole symbol because of sideband supplementary mistake have been reduced, communication efficiency and the band efficiency of system have also been improved simultaneously.

Major advantage of the present invention is:

1) for the characteristics of OFDM symbol peak value average power ratio, adopt a plurality of scrambling patterns of dot product, therefrom select mode that symbol that PAPR value is less is transmitted to avoid the appearance of peak value, reduce the PAPR value of symbol, the efficiency of raising transmitting transducer.

2) the improvement algorithm that the present invention proposes utilizes Comb Pilot to carry scramble sequence in the distribution pattern of frequency domain and selects information, avoided taking this hard up band resource in underwater sound communication, the band efficiency of raising system, the while has also guaranteed the real-time of communication system.

3) to take compressive sensing theory and frequency time reversal technique be basis to improvement pattern selection algorithm proposed by the invention, utilize the sparse property of Shallow Water Acoustic Channels, by the mode to channel matched and calculated rate reversion cross-correlation coefficient, find the scrambling pattern and select information, saved the necessary side information of conventional pattern selection algorithm, reduced the error code disaster risk caused because of the side information reading mistake, reduced the consumption of system transmitting terminal energy, improved the band system band utilization ratio.

The accompanying drawing explanation

Select to suppress peak-to-average force ratio algorithmic system block diagram without side information shallow-sea underwater acoustic communication pattern in Fig. 1 MIMO-OFDM system.

The pattern detector arrangement block diagram of Fig. 2 based on compressed sensing and frequency inversion technique.

Fig. 3 simulated channel impulse response.

Fig. 4 the present invention and conventional pattern selection mode and scrambled signals CCDF curve comparison figure not.

The system BER performance comparison figure of the pattern selection mode that under Fig. 5 white Gaussian noise, the present invention and traditional approach and side information are known.

The system BER performance comparison figure of the pattern selection mode that under the stable burst noise environment distributed of Fig. 6 Alpha, the present invention and conventional pattern selection mode and side information are known.

The source that the transmitting terminal scrambling pattern that Fig. 7 a-Fig. 7 d is front 65 the OFDM symbols of MIMO_OFDM tank experiments communication experiment is selected and the receiving terminal translation to the scrambling pattern, wherein: Fig. 7 a source; Fig. 7 b hydrophone I translation; Fig. 7 c hydrophone II translation; Translation after Fig. 7 d hydrophone I and hydrophone II merge, after the original error rate 0.011/ decoding<4.25 * 10 ^-6.

Embodiment

Below in conjunction with accompanying drawing and example, the present invention is further described.

Improved pattern is selected the peak-to-average force ratio Restrainable algorithms need to define M and is organized different Comb Pilot positions

corresponding M group different scrambling pattern P ^m=[P ^m(0), P ^m(1) ..., P ^m(N-1)], 0≤m≤M selects one group of pilot frequency sequence X that the PAPR inhibition be there is no to negative effect simultaneously _p=[X _p(0), X _p(1) ..., X _p(N-1)].

Data to be sent are X _d, after scrambling, the signal of data subcarrier position can be expressed as:

$X_D^m=P^{\mathrm{mT}}{X}_D---\left(1\right)$

After now transmitting terminal scrambling, signal can be expressed as at frequency domain:

$x^m\left(k\right)=a{X}_D^m\left(k\right)+b{X}_P\left(k\right)---\left(2\right)$

Wherein:

$\begin{array}{cc}a=1,b=0& k\&NotElement;{\mathrm{Po}}^m\\ a=0,b=1& k\&Element;{\mathrm{Po}}^m\end{array}---\left(3\right)$

The M obtained after all scramblings is organized to the defined formula of time-domain signal according to peak-to-average power ratio:

$\mathrm{PAPR}\left(\mathrm{dB}\right)=10{\mathrm{<figure-callout\; id="10"\; label="log"\; filenames="HDA00003742733700021.png,HDA00003742733700022.png"\; state="\{\{state\}\}">log</figure-callout>}}_{10}\frac{\max {\left|x\right|}^2}{E\left({\left|x\right|}^2\right)}---\left(4\right)$

Calculate respectively its PAPR value, select the sequence of peak-to-average force ratio minimum

as transfer sequence.

At receiving terminal, utilize compressive sensing theory to estimate underwater sound condition of sparse channel, by calculating frequency domain reversion cross-correlation coefficient, determine the scrambling phase sequence of selecting, this is the core that pattern of the present invention is selected the peak-to-average power ratio Restrainable algorithms.

Suppose that frequency domain Y (k) data that obtain at receiving terminal meet formula (5) after Fourier transform:

Y(k)=X(k)H(k)+W(k) （5）

Wherein H (k) is the response of frequency domain k sub-channels, and W (k) is the noise on frequency domain k sub-channels.

Definition for the estimated value of channel H (k), corresponding m group pilot frequency locations Po ^mthe frequency domain reception information of extracting is Y _p ^m, pilot frequency information

now obtain M group channel strip frequency domain response estimation information

${\hat{H}}_P^m=\frac{Y_P}{X_P^m}---\left(6\right)$

By the matching pursuit algorithm match channels, recover M group channel frequency domain response

${\hat{H}}^m=f_{\mathrm{MP}}\left({\hat{H}}_P^m\right)---\left(7\right)$

F wherein _mP() means match tracing.

The M group

there is and only have one group to be channel frequency domain response estimation

actual value, will

reversion obtains

calculate

with

cross-correlation coefficient, in the M of an acquisition coefficient correlation, maximum is corresponding

be the corresponding estimated value of frequency domain of the reality of channel

$\hat{H}=\underset{m=1,...,M}{\max }\left(\mathrm{xcorr}\left({\hat{H}}^m\right)\right)---\left(8\right)$

Corresponding m value is the sequence number of scrambling pattern, can utilize and select

actual value and corresponding scrambling pattern, realize receiving the signal frequency domain response go disturb and balanced.

Pattern detector arrangement based on compressed sensing and frequency inversion technique as shown in Figure 2.

Following emulation experiment is example of the present invention, by this example, can be verified performance of the present invention.

Simulated conditions:

In the Matlab emulation platform, build 22 and receive the MIMO-OFDM underwater sound communication system, FFT length 8192, number of sub carrier wave 1025, communication band 6～12kHz, generate 10 at random ³individual OFDM symbol, QPSK mapping, adopt Space Time Coding and convolution coding, the simulated channel impulse response as shown in Figure 3,4 times of over-samplings.

In order to illustrate that the inventive method does not have lossy characteristics than traditional SLM method PAPR rejection, Fig. 4 has provided the present invention and traditional SLM mode and scrambled signals CCDF curve comparison figure not, as can be seen from the figure, the inventive method overlaps substantially with the CCDF curve of traditional SLM method, show that its PAPR rejection does not lose, simultaneously, than scrambled signals not, the PAPR value of the inventive method is 10 ^-2the about 10dB of decline on level, illustrate that the inventive method can reduce the probability that peak value occurs effectively, has PAPR inhibition preferably.

The inventive method is that a kind of pattern without side information is selected the peak-to-average force ratio Restrainable algorithms, improving band efficiency, reduce the transmitting power loss, and reduce factors such as being subject to noise due to communication system and affect and fail correctly to understand side information and cause the aspects such as risk of error code disaster to there is greater advantage.Traditional pattern is selected the peak-to-average force ratio Restrainable algorithms need in transmitted signal, increase phase sequence and is selected information, and each data symbol needs the side information of bit, be used for determining the sequence number of selected scrambling phase sequence, wherein

what mean is to round downwards, the sum that M is adopted scramble sequence.And need interpolation bulk redundancy (normally coding redundancy) can correctly understand this information to guarantee system receiving terminal, to avoid the whole symbolic information demodulation mistake disaster caused due to side information demodulation mistake.For the system of multiple-input and multiple-output, the shared amount of information of this redundancy is huge.Than the conventional pattern selection algorithm, algorithm of the present invention is transferred to receiving terminal without the sequence number using selected scrambling phase sequence as side information, utilize Comb Pilot to carry scramble sequence in the distribution pattern of frequency domain and select information, by the effective identification to the pilot tone pattern at receiving terminal, realize selecting information effectively to understand to the scrambling pattern.The inventive method has been saved data volume, data transmission efficiency and band efficiency have been improved, the probability that cause the whole mistakes of whole symbol because of sideband supplementary mistake have been reduced, also improved the communication efficiency of system simultaneously, guaranteed the real-time of communication, the cost of paying is the increase of amount of calculation.But, due to underwater sound communication, improve the complexity of system receiving terminal to reduce the consumption of transmitting terminal, thereby raising entire system efficiency and stability are very suitable.

Fig. 5, Fig. 6 have provided respectively the system performance of BER contrast under white Gaussian noise and the stable burst noise environment distributed of Alpha of the known pattern selection mode of the present invention and conventional pattern selection mode and side information, as can be seen from the figure the inventive method is better than traditional SLM method, the method performance of BER known with side information is substantially suitable, can absolutely prove that the inventive method can correctly judge the sequence number of selected phase bit sequence, for receiver provides reliable side information.And, due to the sparse characteristic of shallow sea channel, compression sensing method can be estimated channel frequency domain response more accurately, BER performance that equally can elevator system, make it reach the requirement of reliable communication.The inventive method when 6dB, can guarantee underwater sound communication normally, reliably carry out, be better than the conventional pattern system of selection.From Fig. 6, it can also be seen that, the inventive method has certain anti-burst noise ability, also can reliable and effective deciphering scrambling drawing information in the environment of burst noise, be suitable for existing the underwater sound communication of burst noise situation.

Fig. 7 has provided respectively source that the transmitting terminal scrambling pattern of front 65 the OFDM symbols of certain MIMO_OFDM tank experiments communication experiment selects and the receiving terminal translation to the scrambling pattern.Four patterns of take are example, and in code book, black means selected pattern sequence number, and the code book color is more deeply felt and shown that the probability that is judged as selection pattern sequence number is larger.By contrast, can find out, the inventive method can judge the serial number information of selected pattern accurately, for receiver provides complete, reliable side information, avoided the whole symbolic information demodulation mistake disaster caused due to side information demodulation mistake, improve system communication efficiency, guaranteed carrying out in real time of communication.

In sum, the inventive method is a kind of improvement pattern selection algorithm of inhibition peak-to-average power ratio of applicable shallow-sea underwater acoustic communication.The inventive method has been improved transmission and the judgment mode of side information, has overcome the dependence of traditional SLM method to side information, under the prerequisite of not losing the PAPR rejection, has improved the BER performance of system.Guarantee correctly resolving of receiving terminal scrambling pattern sequence number by the mode that detects channel, improved communication efficiency and the band efficiency of system, guaranteed real-time and the reliability of MIMO-OFDM underwater sound communication system.

The example that above example is only applied for the present invention; content of the present invention should not only be confined in the MIMO-OFDM system; for single-input single-output, singly enter to have more and enter more in the ofdm system that singly goes out to suppress peak-to-average power ratio all to use, therefore all be subject to protection of the present invention.

Claims (1)

1. One kind based on the frequency time reversal technique select the peak-to-average force ratio Restrainable algorithms without side information shallow-sea underwater acoustic communication pattern, it is characterized in that:

   (1) at data sending terminal, at first by data X to be sent _dadopt M to organize different scrambling pattern P ^mcarry out scrambling, obtain the whole series of data after scrambling

   

   define M simultaneously and organize different Comb Pilot position Po ^mserve as a mark, corresponding different scrambling patterns, at different pilot frequency locations Po ^madd the Comb Pilot sequence X _pobtain candidate OFDM symbol X ^m, candidate OFDM symbol is carried out to inverse fast Fourier transform to time domain, select the symbol of peak-to-average force ratio minimum wherein as transmission symbol;

   (2) carry out at first to received signal fast Fourier transform at receiving terminal, extract frequency domain corresponding to different Comb Pilot position and receive information, obtain the pilot tones domain response and it is carried out to match tracing, recover M group channel frequency domain response

   

   in the frequency domain reversion, obtain

   

   calculate the two cross-correlation coefficient, in the M of an acquisition coefficient correlation, the corresponding m value of maximum is the sequence number of scrambling pattern, finally utilizes the serial number information descrambling code.

Images