CN101873152B - Doppler frequency shift estimation method based on optimal bandwidth filtering pretreatment - Google Patents

Abstract

The invention discloses a Doppler frequency shift estimation method based on optimal bandwidth filtering pretreatment, mainly solving the problem that the Doppler frequency shift estimation deviation is bigger under the condition of low signal to noise ratio in the prior art in burst communication. The method comprises the following steps: (1) sampling the received signal; (2) determining the initial bandwidth of a band-pass filter; (3) carrying out optimal bandwidth filtering pretreatment on the sampled signal; (4) carrying out Doppler frequency shift estimation on the signal after optimal bandwidth filtering pretreatment by adopting a level crossing rate method; and (5)feeding the Doppler frequency shift estimation value back to the step (2), carrying out next determination of the band-pass filter bandwidth and estimating the Doppler frequency shift in the step (4). The method can still ensure accurate estimation of Doppler frequency shift under the condition of low signal to noise ratio and can be used for the burst communication with remarkable randomness, burstiness and uncertainty.

Description

Based on the pretreated Doppler frequency shift estimation method of optimal bandwidth filtering

Technical field

The invention belongs to wireless communication technology field, relate to the burst communication technology, is a kind of wireless channel estimation Method, is suitable for the radio communication channel of high complexity, low signal-to-noise ratio and high dynamic characteristic.

Background technology

In burst communication, because the characteristic of channel is discontinuous or the signal transmission is subjected to stronger interference, thereby receiving terminal just receives the signal of burst.Burst communication is a kind of necessary complement of general communication as a kind of important emergency communication means.It adopts burst mode to carry out data and sends, and the transmission time is at random, and the duration is short, has therefore that antijamming capability is strong, reliability is high, the characteristics of good concealment, is widely used in extraordinary commercial communication and military communication at present.Burst communication mainly utilizes ionospheric propagation, and shortwave/ultrashort wave is its important communication mode, and it has that coverage is big, electronic warfare capability is strong, survivability is strong and distinguishing feature such as confidentiality height, is the important component part that the bottom line emergency communication ensures.But the burst communication utilization is not the channel of single rule, but the overlapping and mixing of multiple complicated wireless transmission channel, thereby channel has significant randomness, sudden and uncertain.And radio wave fading ratio in the long-distance communications process is more serious, and noise ratio is bigger, so it is very faint to receive signal, and is usually flooded by various noise easily, even signal power can occur less than noise power, i.e. the situation of low signal-to-noise ratio condition.Simultaneously, the high interference in the short wave channel, decline and multipath phenomenon make the communication performance in these channels be subjected to very big inhibition.

In burst communication, Doppler frequency shift typically refers to because the phase place random fluctuation of the signal that multipath transmisstion and ionospheric inhomogeneities and irregular movement thereof cause.Because this frequency displacement mainly is because the irregular movement of ionospheric irregularities causes, so be referred to as Doppler frequency shift.Existing Doppler effect schematic diagram such as Fig. 1, S are transmitting terminal, and the distance that receiving terminal moves from X-direction Y is d, and v is the relative moving speed of receiving terminal.The distance of supposing receiving terminal and S is far, can think that its angle theta at X place and Y place is identical when the radio wave that receiving terminal receives is discussed, and can think when the length of moving is d, at the path difference of X place and the generation of Y place is:

Δl＝dcosθ＝vΔtcosθ (1)

If λ is wavelength, then the phase difference that brings of path difference can be expressed as thus:

Further can obtain the changing value of the frequency that Doppler effect brings, i.e. Doppler frequency shift:

Obtain the expression formula of Doppler frequency shift from here as can be seen, the value of Doppler frequency shift is with receiving terminal relevant with direction of relative movement transmitting terminal and speed of related movement size.If travelling carriage is towards incident wave direction motion, then Doppler frequency shift is being for just, i.e. receive frequency rising; If travelling carriage incident wave direction motion dorsad, then Doppler frequency shift is for negative, and namely receive frequency descends.Therefore, work as increase frequency, when translational speed was accelerated, Doppler effect was clearly.And in θ=0 o'clock, maximum doppler frequency is arranged:

$f_{d\max }=\frac{v}{\mathrm{\λ}}---\left(4\right)$

Maximum Doppler frequency offset f _DmaxBe one of most important optimization receiver parameters, it can be applied to many aspects such as resource distribution, switch decision and power control.

Doppler effect has a significant impact the frequency spectrum that high-speed mobile unit receives signal, under multi-path environment, can produce different Doppler frequency shifts in the spectrum structure of reception signal, causes Doppler to expand, and influences the correct reception of signal.Therefore in order to guarantee the effectively reliable transmission of signal, require portable terminal should be able to estimate the Doppler frequency shift value real-time and accurately, and come dynamically the Adjustment System parameter and compensate according to Doppler frequency shift, to obtain best receptivity.In short wave channel, because the continuous variation of ionospheric electron density, and frequency range is crowded, the serious interference of decline fast, thus make that receiving signal is flooded by various noise easily, badly influences the accuracy that Doppler frequency shift is estimated.Therefore the reception signal under the utmost point low signal-to-noise ratio situation is carried out Doppler frequency shift and estimate that this technology has earth shaking meaning in short wave communication even wireless communication field.

Existing maximum doppler frequency is estimated to be divided three classes: based on the estimation of level crossing rate (lcr) LCR or zero passage percent of pass ZCR, based on estimation and the maximum likelihood ML estimation of auto-correlation function COV or cross-correlation function ACF.The advantage of LCR algorithm is to realize simply, but noise robustness is relatively poor; Though COV algorithm estimated performance slightly is better than the LCR algorithm, computation complexity is than higher; Maximum likelihood algorithm then requires background noise Gaussian distributed characteristic, and signal to noise ratio information accurately need be provided.And under the low signal-to-noise ratio environment, the noise robustness of said method is not ideal enough, and the deviation between resulting Doppler frequency shift estimated value and the theoretical value is bigger, is unfavorable for the raising of receptivity in the burst communication system and effective transmission of information.Therefore, need further to inquire under the low signal-to-noise ratio situation, be applicable to more effective Doppler frequency shift estimation method reliably of burst communication.

Summary of the invention

The objective of the invention is the problem at above-mentioned method of estimation existence, proposed a kind of based on the pretreated Doppler frequency shift estimation method of optimal bandwidth filtering, to improve the noise robustness of system under the low signal-to-noise ratio situation, further reduce the deviation between Doppler frequency shift estimated value and the theoretical value, guarantee raising and the information transmission effectively reliably of receptivity in the burst communication system.

The technical scheme that realizes the object of the invention is: under the low signal-to-noise ratio situation, the channel characteristic of shortwave/ultra short wave communication, utilize the band pass filter of bandwidth self-adaption at receiving terminal, the optimal bandwidth filtering technology is combined with the level crossing rate (lcr) method of estimation, concrete steps are as follows:

(1) r (t) samples to received signal, the discrete signal r (k) after obtaining sampling;

(2) the initial bandwidth of determining band pass filter is B _m'=2K _of _Dmax, in the formula, Be optimum filtering bandwidth ratio, f _DmaxBe the Doppler frequency shift maximum that occurs in the burst communication system;

(3) the discrete signal r (k) after the sampling is carried out following optimal bandwidth filtering preliminary treatment:

3a) discrete signal r (k) is carried out fast Fourier transform (FFT) and handle, obtain the signal R (K) after FFT handles;

Be B with initial bandwidth 3b) _m' the signal R (K) of band pass filter after FFT is handled carry out filtering, obtain filtered signal

3c) to filtered signal

Carry out inverse fast Fourier transform (IFFT) and handle, obtain the pretreated output signal y of optimal bandwidth filtering (k);

(4) utilize level crossing rate (lcr) (LCR) method that the pretreated output signal y of optimal bandwidth filtering (k) is carried out Doppler frequency shift and estimate, obtain the Doppler frequency shift estimated value

(5) with the Doppler frequency shift estimated value

Feed back to step (2), substitute described f _Dmax, determine the bandwidth of next round band pass filter.

The present invention is owing to carry out the optimal bandwidth filtering preliminary treatment to received signal at receiving terminal, and combine with the level crossing rate (lcr) method of estimation, can effectively improve the noise robustness of burst communication system, reduce the deviation between the Doppler frequency shift estimated value and theoretical value in the burst communication system.In addition, owing to utilized the feedback of Doppler's estimated value in the optimal bandwidth filtering process, so it catches with tracking performance and is improved, and computation complexity has obtained reduction.Particularly under the low signal-to-noise ratio situation, still can guarantee the accurate estimation of Doppler frequency shift, in effective transmission course of the receptivity that improves communication system and the information of guarantee, bring into play important function.

Description of drawings

Fig. 1 is existing Doppler effect schematic diagram;

Fig. 2 is realization flow block diagram of the present invention;

Fig. 3 is the band pass filter schematic diagram that the present invention adopts;

Fig. 4 is the inventive method and existing method simulation performance correlation curve figure.

Embodiment

With reference to Fig. 2, the present invention is based on the pretreated Doppler frequency shift of optimal bandwidth filtering and estimate, may further comprise the steps:

Step 1, r (t) samples to received signal, the discrete signal r (k) after obtaining sampling.

Determine sample frequency f at receiving terminal _s, with sample frequency f _sR (t) samples to received signal, the discrete signal r (k) after obtaining sampling.

Step 2 is determined the initial bandwidth of band pass filter.

2a) determine optimum filtering bandwidth ratio K _o:

At first, under noise circumstance, drawn the evaluated error of LCR algorithm by the expression formula of level crossing rate (lcr):

$e_{\mathrm{LCR}}=\frac{|{\hat{L}}_R-L_R|}{L_R}=|\sqrt{\frac{\mathrm{\γ}+\frac{2}{3}{K}^2}{\mathrm{\γ}+1}}-1|---\left(5\right)$

In the formula, With L _RBe respectively per second by estimated value and the theoretical value of the number of times of particular level threshold values R, γ is for receiving the signal to noise ratio of signal r (t), and K is bandwidth ratio;

Then, formula (5) is carried out Taylor expansion, approximate obtaining:

$e_{\mathrm{LCR}}=|\sqrt{1+\frac{\frac{2}{3}{K}^2-1}{\mathrm{\γ}+1}}-1|=|\frac{1}{2}\frac{\frac{2}{3}{K}^2-1}{\mathrm{\γ}+1}-\frac{1}{8}{\left(\frac{\frac{2}{3}{K}^2-1}{\mathrm{\γ}+1}\right)}^2+\·\·\·|\≈\left|\frac{1}{2}\frac{\frac{2}{3}{K}^2-1}{\mathrm{\γ}+1}\right|;---\left(6\right)$

At last, formula (6) about bandwidth ratio K differentiate, is made that derived function is 0, solve the optimum filtering bandwidth ratio

2b) the initial bandwidth of definition band pass filter is B _m'=2K _of _Dmax, in the formula, f _DmaxBe the Doppler frequency shift maximum that occurs in the burst communication system;

When the optimal bandwidth filtering of signal is handled, think that filtering bandwidth can regulate, utilize this band pass filter that can regulate bandwidth to handle to received signal, can carry out filtering to noise to greatest extent, improve the signal to noise ratio that receives signal, improve the Doppler frequency shift estimation performance.

Step 3 is carried out following optimal bandwidth filtering preliminary treatment to the discrete signal r (k) after the sampling.

3a) discrete signal r (k) is carried out fast Fourier transform (FFT) and handle, obtain the signal R (K) after FFT handles;

3b) with reference to Fig. 3, be B with initial bandwidth _m' the signal R (K) of band pass filter after FFT is handled carry out filtering, obtain filtered signal

3c) to filtered signal

Carry out inverse fast Fourier transform (IFFT) and handle, obtain the pretreated output signal y of optimal bandwidth filtering (k).

Step 4 is utilized level crossing rate (lcr) (LCR) method that the pretreated output signal y of optimal bandwidth filtering (k) is carried out Doppler frequency shift and is estimated.

4a) get the envelope that receives signal, by experiment, statistics obtains envelope per second forward by the estimated value of the number of times of particular level R

4b) according to expression formula ρ=R/R _RmsCalculate ratio ρ, in the formula, R _RmsLocal root mean square level for the pretreated output signal y of optimal bandwidth filtering (k);

4c) obtain the Doppler frequency shift estimated value

Expression formula:

At first, according to level crossing rate (lcr) LCR theory, deriving obtains the estimated value of per second by the particular level number of times and is:

${\hat{L}}_R=\sqrt{\frac{\mathrm{\γ}+\frac{{B\′}_m^2}{6f_d^2}}{\mathrm{\γ}+1}}\sqrt{2\mathrm{\π}}{f}_d\mathrm{\ρ}{e}^{{-\mathrm{\ρ}}^2},---\left(7\right)$

In the formula, B ' _mBandwidth for band pass filter;

Then, definition bandwidth ratio

Obtain the Doppler frequency shift estimated value according to formula (7)

For:

${\hat{f}}_d={\hat{L}}_R/\left(\sqrt{2\mathrm{\π}}{\mathrm{\ρe}}^{{-\mathrm{\ρ}}^2}\sqrt{\frac{1+{\left(\frac{2}{3}{K}^2\mathrm{\γ}\right)}^{-1}}{1+{\mathrm{\γ}}^{-1}}}\right),---\left(8\right)$

In the formula, e is the end of natural logrithm, and value is 2.71828;

At last, with the optimum filtering bandwidth ratio

Substitution following formula (8), then Doppler frequency shift estimated value

Calculation expression can be simplified shown as:

${\hat{f}}_d=\frac{{\hat{L}}_R}{\left(\sqrt{2\mathrm{\π}}{\mathrm{\ρe}}^{{-\mathrm{\ρ}}^2}\right)};---\left(9\right)$

4d) with 4a) in the per second that the obtains estimated value by the particular level number of times

And 4b) calculates the ratio ρ substitution following formula (9) of gained in, can obtain the Doppler frequency shift estimated value

Step 5 is determined the bandwidth of next round band pass filter.

5a) the mean error of definition Doppler frequency shift estimated value For:

$\stackrel{\‾}{e}=\frac{1}{M}\underset{i=1}{\overset{M}{\mathrm{\Σ}}}\frac{|f_d-{\hat{f}}_{d,i}|}{f_d},---\left(10\right)$

In the formula, the number of times of M for estimating to carry out, f _dBe the theoretical value of Doppler frequency shift,

Be the i time Doppler frequency shift f _dEstimated value;

5b) will Feed back to step 2, the bandwidth of band pass filter in the step 2 is updated to ${B\′}_m=K_0{\hat{f}}_d;$

5c) carry out next round sampling and estimation, loop iteration is until the mean error of the defined Doppler frequency shift estimated value of formula (10)

Tend towards stability and reach minimum.

Effect of the present invention can further specify with analysis by following theoretical the derivation:

(A) obtain following three important inferences through derivation and proof:

Inference 1: after the optimal bandwidth filtering preliminary treatment, the signal to noise ratio of signal improves, and noise component(s) has obtained inhibition, and signal energy is enhanced, i.e. signal to noise ratio processing gain G=γ _Out/ γ _In＞1, in the formula, γ _InAnd γ _OutRepresent the input and output signal to noise ratio of optimal bandwidth filtering system respectively.

Inference 2: after the optimal bandwidth filtering preliminary treatment, the Doppler frequency shift expression formula that receives signal can obtain simplifying, and does not need to know in advance the value of signal to noise ratio, workload getparms in the time of can reducing the Doppler frequency shift estimation, improve arithmetic speed, simplify original estimation structure.

Inference 3: after the optimal bandwidth filtering preliminary treatment, system has adaptive ability by interative computation, catches with tracking performance to be improved.

(B) based on the relative estimated bias e of the Doppler frequency shift of level crossing rate (lcr) method _LCRIt is the function about signal to noise ratio γ and bandwidth ratio K.Proof can get following two formulas:

$\underset{\mathrm{\γ}\→+\∞}{\lim }{e}_{\mathrm{LCR}}=0,---\left(11\right)$

$\underset{K\→\sqrt{3/2}}{\lim }{e}_{\mathrm{LCR}}=0,---\left(12\right)$

Owing to adopt optimal bandwidth filtering preliminary treatment structure, the discrete signal r (k) after the sampling is carried out optimal bandwidth filtering handle, noise component(s) is inhibited, and the signal to noise ratio of output signal y (k) is improved significantly.In addition, adopt

Can guarantee the evaluated error minimum as the filtering bandwidth ratio.Therefore, according to following formula (11) and following formula (12), adopt the optimal bandwidth filtering preliminary treatment can effectively reduce the relative estimated bias of Doppler frequency shift;

(C) because this optimal bandwidth filtering preliminary treatment OBF method does not need to obtain extra parameter, only need utilize specific optimal bandwidth filtering bandwidth ratio and corresponding these two parameters of Doppler frequency shift estimated value that signal is operated, therefore this processing method is same being suitable for to above-mentioned multiple method of estimation, and this method can be widely used in and improves Doppler frequency shift estimated performance aspect.

The present invention propose based on optimal bandwidth filtering preliminary treatment OBF method and existing method performance comparison curve chart as shown in Figure 4.As can be seen from Figure 4, the OBF method of the present invention's proposition always is better than existing method 2～4dB.Such as, the OBF method that the present invention proposes is in mean error

Be 10 ^-1The point place compares S﹠amp; The B method exceeds 5.2dB.In carrying OBF method, need carry out FFT and IFFT conversion to received signal, but these operate in the existing filter and can both realize easily, increase operand that therefore can be not extra, its computer computation complexity is compared lower with existing method.

To sum up, proposed by the invention can suppress noise effectively based on the pretreated level crossing rate (lcr) method of optimal bandwidth filtering, has simultaneously also reduced computation complexity to a certain extent, particularly at low signal-to-noise ratio, i.e. and input signal-to-noise ratio γ _InUnder the situation of＜0dB, still can guarantee the accurate estimation of Doppler frequency shift.In addition, algorithm that the present invention carries has good catching and tracking performance, and the receptivity that improves shortwave/ultra short wave communication system has been played important function.

Claims (1)

1. one kind based on the pretreated Doppler frequency shift estimation method of optimal bandwidth filtering, comprises the steps:

(1) r (t) samples to received signal, the discrete signal r (k) after obtaining sampling;

(2) the initial bandwidth of determining band pass filter is B _m'=2K ₀f _Dmax, in the formula,

Be optimum filtering bandwidth ratio, f _DmaxBe the Doppler frequency shift maximum that occurs in the burst communication system;

(3) the discrete signal r (k) after the sampling is carried out following optimal bandwidth filtering preliminary treatment:

3a) discrete signal r (k) is carried out fast Fourier transform FFT and handle, obtain the signal R (K) after FFT handles;

Be B with initial bandwidth 3b) _m' the signal R (K) of band pass filter after FFT is handled carry out filtering, obtain filtered signal

3c) to filtered signal

Carry out inverse fast Fourier transform IFFT and handle, obtain the pretreated output signal y of optimal bandwidth filtering (k);

(4) utilize level crossing rate (lcr) LCR method that the pretreated output signal y of optimal bandwidth filtering (k) is carried out Doppler frequency shift and estimate, obtain the Doppler frequency shift estimated value

4a) by experiment, the statistics pretreated output signal y of optimal bandwidth filtering (k) per second is by the number of times of particular level threshold values R

4b) ratio calculated ρ=R/R _Rms, in the formula, R _RmsLocal root mean square level for the pretreated output signal y of optimal bandwidth filtering (k);

4c) derivation obtains the Doppler frequency shift estimated value

Expression formula:

In the formula, e is the end of natural logrithm, and value is 2.71828;

4d) with 4a) in the estimated value that obtains of experiment statistics

And 4b) Doppler frequency shift estimated value the ratio ρ substitution 4c that calculates in)

Expression formula, calculate the Doppler frequency shift estimated value

(5) with the Doppler frequency shift estimated value

Feed back to step (2), substitute described f _Dmax, determine the bandwidth of next round band pass filter.

Images