CN101242199A - Tracking loop for ultra-broadband communication system based on maximal possibility estimation - Google Patents

Abstract

The invention belongs to the wireless communication technology, especially a tracing loop of hyper-broadband communication system based on maximal maximum likelihood estimation. The tracing loop is composed of a low filter, a time error detector, a synchronous sampling circuit, a circular filter and a digital control oscillator. A received signal passes through the lower filter, output signal of which enters the time error detector. Output of the time error detector is sampled by the synchronous sampling circuit. Sampled value enters the loop filter, output of which passes through the digital control oscillator feedback to the synchronous sampling circuit. The tracing loop provided by the invention can trace Gauss impulse of the hyper-broadband communication system with tracing system under concentrated multiple-channel provided by effective IEEE802.153a, which improves error rate performance of the system greatly.

Description

Track loop based on the ultra-wideband communication system of maximal possibility estimation

Technical field

The invention belongs to wireless communication technology field, be specifically related to a kind of track loop that is applied to pulse system ultra wide band (IR-UWB) communication system.

Background technology

In recent years, (Impulse Radio Ultra-wideband, IR-UWB) wireless communication system is subjected to the extensive attention of industrial quarters and academia to pulse system ultra wide band.

According to the ultra broadband definition that Federal Communications Commission (FCC Federal Communications Commission) announces, if the shared instant bandwidth of signal surpasses 500MHz, perhaps relative bandwidth is greater than 0.2, and then this signal is a ultra-broadband signal.Relative bandwidth (or fractional bandwidth, Fractional Bandwidth) is defined as the ratio of energy bandwidth and centre frequency.

But pulse system ultra wide band utilizes these characteristics of burst transmissions information of the super narrow subnanosecond level of base band just, makes also can make very little timing error the track loop losing lock, thereby systematic function is decayed greatly.

Traditional track loop is based on the code tracking loop design of phase-locked loop or digital baseband system, but the super burst pulse cycle of pulse system ultra wide band and super wide bandwidth, make in the pulse system ultra wide band communication system catch and tracking becomes a difficult point, people can not simply regard the waveform that receives square wave as or rise Yu Xuanbo again as in narrowband systems.This just impels our phase-locked loop by research tradition, and delay locked loop road and code tracking loop come to such an extent that make new advances, and is applicable to the track loop of the super narrow Gaussian pulse of pulse system ultra wide band communication system.In many documents, describe in detail in traditional narrow-band digital communication system by sampling and interpolation, realize the interpolation of code tracking and sampled value, simultaneously in the other document, at additive white Gaussian noise channel (AWGN) and IEEE802.15 multipath channel, paired pulses system ultra-wideband track loop is analyzed and is derived, labor the influence of different Gaussian waveform to track loop, but also nobody proposes under the multi-path dense channel track loop effective circuit structure at pulse system ultra wide band.

Summary of the invention

The object of the present invention is to provide and a kind ofly under the multi-path dense channel, can effectively follow the tracks of ultra-wideband pulse, make the error rate of system performance that the track loop of very big raising be arranged.

The present invention adopts better simply circuit structure, even and than also reaching tracking performance preferably under the low signal-to-noise ratio.Prove that by system emulation this invention is to have superior performance and effective track loop at the super narrow Gaussian pulse of pulse system ultra wide band.

The present invention is to provide a kind of track loop of the pulse system ultra wide band based on maximal possibility estimation.This track loop filters little device 4 by low pass filter 1, time error detector 2, synchronized sampling road 3, loop and digital controlled oscillator 5 is formed, received signal is at first passed through low pass filter 1, the output signal entry time error detector 2 of low pass filter 1, the output of time error detector 2 is sampled by synchronized sampling circuit 3, the value of sampling enters loop filter 4, and the output of loop filter 4 feeds back to synchronized sampling circuit 3 through digital controlled oscillator 5.Wherein:

A) 1 pair of input signal of low pass filter carries out low-pass filtering, suppresses the noise signal of channel;

B) time error detector (TED Timing Error Detector) 2, the wave function of prior known received signal, simulation derivative matched filter transfer function is designed to the derivative of wave function, and received signal is by this derivative matched filter, and the signal that obtains promptly is the function of timing error;

C) synchronized sampling circuit 3 is one one bit moduli transducers, estimate the position of the selectivity multipath that according to channel estimating, in the sampling of place that each pulse should finish, income value be actual reception to signal pulse and the value of the channel estimating timing error function of estimating the pulse position that;

D) loop filter 4 adopts second-order low-pass filter.Regulate filter parameter can effectively suppress to sample noise component(s) in the gained timing error functional value, make the timing error functional value that feedbacks more accurate;

E) digital controlled oscillator 5 is converted to time quantum with the timing error functional value, feeds back to the synchronized sampling circuit, realizes the locking of loop; Simultaneously this value is outputed to the demodulation link, adjust the correlation demodulation template position.

Track loop Mathematical Modeling following (seeing shown in Figure 5):

At first time error signal multiply by the loop coefficient by the decision of the algorithm of time error detector, adds behind the noise through transfer function to be that loop filter and the transfer function of F (z) is the digital controlled oscillator of I (z), is fed back to time error signal at last again.

According to track loop of the present invention, the tracing process of signal is as follows: at first, the signal that receives obtains the time error function of received signal and local template through the derivative matched filter, by the synchronized sampling link this time error function signal is sampled then; By a step low-pass loop filter this sampled value is carried out filtering again, suppress noise component(s) wherein; The time error value that at last filtering is obtained converts time quantum to through digital controlled oscillator and feeds back to the synchronized sampling circuit, local template is adjusted in time with the signal that receives alignd, and realizes the locking of loop.In addition, this feedback signal will be exported to the demodulation link, be used for adjusting the position of correlate template in the selectivity Rake receiver branch.

Track loop provided by the invention greatly reduces the requirement of system to sample rate, makes sampling to carry out on the frame aspect of pulse system ultra wide band modulation signal.And this track loop can reduce the error rate of system's demodulation greatly, improves systematic function.This invention proves the effective pulse system ultra wide band track loop that can greatly improve systematic function by emulation.

Description of drawings

The Gaussian pulse ultra broadband that Fig. 1 has provided four bit direct sequence binary phases modulation (DS-BPSK direct spread binary phase shiftkeying) sends signal, and (signal period is T _p).Wherein, (time is long to be T to each bit _s) (time is long to be T by five frames _f) form, a Gaussian pulse is arranged in each frame, each bit is all formed by the modulation of DS sign indicating number, and every frame time is identical, and pulse all appears at the place that frame begins.

Fig. 2 has provided the discrete time impulse response of the defined ultra broadband multipath channel of IEEE802.15.3a.

Fig. 3 has provided the channel estimating that branches is 5 selectivity Rake.When with selectivity (Rake) receiver of 5 branches, after channel estimating is finished, 5 multipaths of the amplitude maximum of selecting, these 5 multipath pulses are duplicating behind transmitted pulse time delay and the amplitude attenuation, 5 pulses need following the tracks of of track loop just.

Fig. 4 has provided the track loop structured flowchart.Received signal was at first advanced a low pass filter, obtained time error signal by the time error detector then, and by carrying out low-pass filtering after the synchronized sampler sampling, filtered signal is fed back to synchronized sampler and outputed to the demodulation link by digital controlled oscillator.

Fig. 5 has provided the Mathematical Modeling of track loop.Wherein, A is an approximate operator that multiplies each other of being derived out by the time error detector, and F (z) is the transfer function of loop filter, I=z ^-1/ (1-z ^-1) be the equivalent function of digital controlled oscillator.

Fig. 6 has provided the S curve chart of time error detector.This figure has characterized the following range of track loop and to the sensitivity of time error.

Fig. 7 has provided the analogous diagram of the variance of track loop time error.The variance of time error has characterized loop to the Noise Suppression performance, and this figure has illustrated that this track loop has good noise suppressed performance.

When Fig. 8 has provided no track loop, the error rate under the different time jitters.This figure emulation during no track loop, under the different time jitter scopes, time jitter is to the error rate of system Effect on Performance, can see that time jitter improves the error rate of system rapidly when shaking to about half pulse period, demodulation performance reduces rapidly.

Fig. 9 has provided as time jitter scope=0.4T _pThe time, the error rate is relatively.By analogous diagram as can be seen, introduce after the track loop, error rate of system probably has the reduction of two orders of magnitude.

Figure 10 has provided as time jitter scope=0.6T _pThe time, the error rate is relatively.By analogous diagram as can be seen, introduce after the track loop, error rate of system probably has the reduction of two orders of magnitude.

Figure 11 has provided as time jitter scope=0.8T _pThe time, the error rate is relatively.By analogous diagram as can be seen, introduce after the track loop, error rate of system has smaller reduction, and this moment, the jitter range of time error exceeded the following range of track loop, so be easy to make the track loop losing lock.

Number in the figure 1 is a low pass filter, and 2 is the time error detector, and 3 is synchronized sampler, and 4 is loop filter, and 5 is digital controlled oscillator.

Embodiment

The present invention is a kind of track loop, and received signal just can be adjusted very little time error by this loop, obtains the exact position of pulse, improves the demodulation performance of system.

In the pulse system ultra wide band wireless communication system, each information bit is transmitted by a symbol, establishes each symbol and comprises N _fFrame, every frame comprise a pulse q (t), and the frame width is T _f, pulse duration is T _dTherefore the width of symbol is Ts:=N _f* T _fFor level and smooth transmitting power spectrum and permission multi-user access, (amplitude of each pulse is by information sequence and DS sign indicating number d for Direct-Sequence, DS) sign indicating number to introduce direct sequence spread spectrum _kProduct modulate, wherein K ∈ [0, N _f-1].The transmission signal p of the symbol lengths of non-modulated _T(t) can be expressed as:

$p_T\left(t\right)={\mathrm{\Σ}}_{k=0}^{N_f-1}{d}_{k}q(t-{\mathrm{kT}}_f)---\left(1\right)$

Suppose that signal transmits under the multi-path dense channel, and noise is that average is zero, one-sided power spectrum density is N ₀Additive white Gaussian noise, simultaneously, intersymbol harasses that harass with interframe can be by selecting T _f〉=τ _{L, p}+ T _pAvoid, wherein τ _{L, p}Be multipath delay.Consider multipath channel to transmitting the influence of signal, the discrete impulse response of channel be expressed as:

$h\left(t\right)=\mathrm{\χ}{\mathrm{\Σ}}_{l=0}^L{\mathrm{\Σ}}_{k=0}^K{\mathrm{\α}}_{k,l}\mathrm{\δ}(t-T_l-{\mathrm{\τ}}_{k,l})---\left(2\right)$

In this formula, χ represents to be the amplitude attenuation of exponential distribution, and channel coefficients is α _{K, l}=ρ _{K, l}. ξ _l. β _{K, l}Wherein, ρ _{K, l}Cause equally distributed stochastic variable between ± 1, multiplier ξ by channel reflection _lAnd β _{K, l}Represent the amplitude of k multipath component in l bunch of multipath and this bunch multipath respectively.Therefore, the signal of the process multipath channel of receiving at receiving terminal is:

$r\left(t\right)={\mathrm{\Σ}}_{m=0}^{N_f-1}{\mathrm{\Σ}}_{k=0}^K{\mathrm{\α}}_{k,l}{\hat{d}}_{m}q(t-k{T}_f-\mathrm{\τ})+n\left(t\right)---\left(3\right)$

Wherein q (t) is the unit energy superwide band pulse wave that receives, and τ is unknown time of delay, and n (t) is an additive white Gaussian noise,

Be under the situation of not considering noise, the energy of the pulse that each receives.Assumes synchronization and channel estimating are finished, and adopt selectivity to refer to (SRake selective rake) receiver more.Then for Gauss and derivative pulse thereof, list of references shows: the multipath parameter is by the multipath effect of IEEE802.15.3a definition, be can be uncared-for in tracking, therefore, for following the tracks of, the ultra broadband multipath channel can be by some through what select, and the multipath component of band amplitude attenuation and time-delay characterizes, as long as followed the tracks of the multipath component of these selections, just realized tracking to the ultra broadband multipath channel.

Suppose that selectivity refers to that the branches of receiver (SRake) is M more, the local reference waveform that is used to be correlated with can followingly be represented:

$s\left(t\right)={\mathrm{\Σ}}_{k=0}^{N_f-1}{\mathrm{\Σ}}_{m=0}^M{d}_{k}g(t-{\mathrm{kT}}_f-{\widehat{\mathrm{\τ}}}_m)---\left(4\right)$

Wherein g (t) is the local unit energy template waveforms that produces, and the energy of this waveform is a unit energy,

The time that is m bar multipath is estimated, so the time error of the multipath that the m bar is selected is $e_m={\mathrm{\τ}}_m-{\widehat{\mathrm{\τ}}}_m.$

For the TED structure that is optimized, consider derivation based on the TED of maximal possibility estimation algorithm.

Suppose that the observation sample interval that this likelihood is estimated is 0≤t≤T ₀, ${\hat{c}}_i\∈\{1,-1\}$ Be that the value of symbol that (decision-directed) obtains is adjudicated in check, establish sample estimates to be:

$\hat{s}\left(t\right)=\underset{i}{\mathrm{\Σ}}{\hat{c}}_{i}g(t-i{T}_f-\widehat{\mathrm{\τ}})---\left(5\right)$

Then by this sample space, be according to obtaining the log-likelihood The Representation Equation in the list of references:

$L\left(r\right|\widehat{\mathrm{\τ}})=\underset{0}{\overset{T_0}{\∫}}r\left(t\right)\hat{s}\left(t\right)\mathrm{dt}-\frac{1}{2}\underset{0}{\overset{T_0}{\∫}}{\hat{s}}^2\left(t\right)\mathrm{dt}---\left(6\right)$

With (5) substitution (6) and differentiate, can obtain following formula:

$L^{\′}\left(r\right|\widehat{\mathrm{\τ}})=-\underset{i}{\mathrm{\Σ}}{\hat{c}}_i\underset{0}{\overset{T_0}{\∫}}r\left(t\right)g^{\′}(t-i{T}_f-\widehat{\mathrm{\τ}})\mathrm{dt}+$

$\underset{i}{\mathrm{\Σ}}\underset{m}{\mathrm{\Σ}}{\hat{c}}_i{\hat{c}}_m\underset{0}{\overset{T_0}{\∫}}{g}^{\′}(t-i{T}_f-\widehat{\mathrm{\τ}})g(T-{\mathrm{mT}}_f-\widehat{\mathrm{\τ}})\mathrm{dt}---\left(7\right)$

Wherein g ' is the derivative of g (t) (t), because T between the area of observation coverage ₀Much larger than frame period (T ₀＞＞T _f), so the unlimited summation to i can be limited in according to interval 0≤i≤L ₀(L ₀Observation sample length for the likelihood estimation).And because the super burst pulse of ultra broadband is that multipath can be distinguished, therefore can ignore interpulse influencing each other, have:

${\∫}_0^{T_0}{g}^{\′}(t-{\mathrm{iT}}_f-\widehat{\mathrm{\τ}})g(t-{\mathrm{mT}}_f-\widehat{\mathrm{\τ}})\mathrm{dt}\≈0$ (i ≠ m),, (8) are expressed as through above two-part approximate:

$L^{\′}\left(r\right|\widehat{\mathrm{\τ}})=\underset{0}{\overset{L_0}{\mathrm{\Σ}}}{\hat{c}}_i{y}^i(i{T}_f+\widehat{\mathrm{\τ}})+\underset{0}{\overset{L_0}{\mathrm{\Σ}}}{\hat{c}}_i^2{h}^{\′}\left(0\right)---\left(8\right)$

In formula (8), y ' be r (t) (t) through derivative matched filter dg (t)/response of dg=-g ' after (t), h ' is the response of signal g (t) through this filter (t), can be expressed as:

$\left\{\begin{array}{c}{y}^{\′}\left(t\right)=-\underset{-\∞}{\overset{\∞}{\∫}}r\left(\mathrm{\ξ}\right)g^{\′}(\mathrm{\ξ}-t)\mathrm{d\ξ}\\ h^{\′}\left(t\right)=g\left(t\right)\⊗[-g^{\′}(-t)]\end{array}\right.---\left(9\right)$

Investigation formula (9), can derive some relatively useful results: because h ' (t) and h (t) :=g (t)  g ' (derivative equivalence t), and the latter is an even function and maximum arranged at zero point, thus h ' (0)=0, and h ' (t)=-h ' (t).

With above substitution as a result (8), and find by the discussion of front

K component, be exactly k time error, the expression formula that then can draw error e (k) is as follows:

$e\left(k\right)={\hat{c}}_k{y}^{\′}({\mathrm{kT}}_f+\widehat{\mathrm{\τ}})---\left(10\right)$

The performance of track loop can be characterized by the variance of S curve and timing error.Wherein the S curve has characterized the size of following range and the time error detector sensitivity to timing error.And the variance of timing error has characterized loop to the Noise Suppression performance.

1.S curve:

By deriving, the expression formula of S curve is as follows:

S(ε)＝-Ch′(e _k) (12)

Wherein $e_k=\mathrm{\τ}-\widehat{\mathrm{\τ}}.$ Because track loop is at e _kWork in the very little scope is used the linear approximation theory: h ' (e near=0 _k)=h ' (0)+h " (0) .e _k, the Mathematical Modeling that obtains this track loop thus as shown in Figure 5, with A=C.h " (0) be defined as an operator that multiplies each other, F (z) is the transfer function of loop filter, I=z ^-1/ (1-z ^-1) be the equivalent function of digital controlled oscillator.

2. the variance of time error:

By mistake shown in Figure 5! Do not find Reference source.Mathematical Modeling can obtain: suppose that input time, error was zero, then variance is determined by loop noise, loop noise R _N(m) comprise mutual incoherent two parts: additive noise and self noise by thermal noise causes are expressed as follows:

R _N(m)＝R _N，A(m)+R _N，S(m) (13)

By mathematical derivation, two parts are respectively as can be known:

${\mathrm{\&sigma;}}_{(N,A)}^2=\frac{2B_{L}T}{{\left({-h}^{\text{'}\text{'}}\left(0\right)\right)}^2{\mathrm{<figure-callout\; id="2"\; label="T"\; filenames="S2008100343256E00021.png,S2008100343256E00041.png"\; state="\{\{state\}\}">T</figure-callout>}}^2}\&CenterDot;\frac{{\mathrm{<figure-callout\; id="0"\; label="N"\; filenames="S2008100343256E00021.png,S2008100343256E00031.png"\; state="\{\{state\}\}">N</figure-callout>}}_0}{2E_S}---\left(14\right)$

${{\mathrm{\&sigma;}}^2}_{(N,S)}=K_F\&CenterDot;\frac{{2B}_{L}T}{{\left({-h}^{\text{'}\text{'}}\left(0\right)\right)}^2{\mathrm{<figure-callout\; id="2"\; label="T"\; filenames="S2008100343256E00021.png,S2008100343256E00041.png"\; state="\{\{state\}\}">T</figure-callout>}}^2}\&CenterDot;\underset{L>0}{\mathrm{\&Sigma;}}{\left(h^{\&prime;}\left(\mathrm{lT}\right)\right)}^2---\left(15\right)$

With (14) and (15) addition, promptly obtain the variance yields of time error, Fig. 7 is normalized analogous diagram, by seeing among the figure, under the situation of low signal-to-noise ratio, when ignoring the error of check judgement, time error variance curve and OK a karaoke club Mi-Lao lower bound are very approaching, so the present invention has good tracking performance as can be known from theory analysis.

3. the design of loop filter:

In the formula (15), K _FBe the value relevant with the closed loop transfer function, of loop, the hypothesis loop filter is a second order: F (z) :=G ₁+ G ₂/ (1-z ^-1), G wherein ₁, G ₂Be filter parameter undetermined.Then closed loop transfer function, is shown below:

$H_{\mathrm{loop}}\left(z\right)=\frac{({\mathrm{<figure-callout\; id="1"\; label="G"\; filenames="S2008100343256E00021.png,S2008100343256E00031.png"\; state="\{\{state\}\}">G</figure-callout>}}_1+G_2)z-G_1}{{\mathrm{<figure-callout\; id="2"\; label="z"\; filenames="S2008100343256E00021.png,S2008100343256E00041.png"\; state="\{\{state\}\}">z</figure-callout>}}^2+[A({\mathrm{<figure-callout\; id="1"\; label="G"\; filenames="S2008100343256E00021.png,S2008100343256E00031.png"\; state="\{\{state\}\}">G</figure-callout>}}_1+G_2)-2]z+{\mathrm{<figure-callout\; id="1"\; label="AG"\; filenames="S2008100343256E00021.png,S2008100343256E00031.png"\; state="\{\{state\}\}">AG</figure-callout>}}_1+1}---\left(16\right)$

Suppose B _LT＜＜1 then has approximate: z=e ^JwT _f≈ 1+jwT _f, show that then formula (16) is expressed as:

$H_{\mathrm{loop}}\left(s\right)=\frac{1}{A}\&CenterDot;\frac{s\left(\frac{T_f({\mathrm{<figure-callout\; id="1"\; label="G"\; filenames="S2008100343256E00021.png,S2008100343256E00031.png"\; state="\{\{state\}\}">G</figure-callout>}}_1+G_2)}{G_1}\right)+1}{s^2\left(\frac{{{\mathrm{<figure-callout\; id="2"\; label="T\; f"\; filenames="S2008100343256E00021.png,S2008100343256E00041.png"\; state="\{\{state\}\}">T</figure-callout>}}_f}^2}{{\mathrm{AG}}_1}\right)+s\left(\frac{T_f({\mathrm{<figure-callout\; id="1"\; label="G"\; filenames="S2008100343256E00021.png,S2008100343256E00031.png"\; state="\{\{state\}\}">G</figure-callout>}}_1+G_2)}{G_1}\right)+1}---\left(17\right)$

Can obtain K _F=2[4 λ ²/ (1+4 λ ²)] ², λ=[A (G wherein ₁+ G ₂)/4G ₁] ^1/2Determined by formula (17).

Beneficial effect of the present invention can do as one likes can analogous diagram embody.

In emulation, be T with normalized pulse duration _dThe Gaussian pulse of=1ns.Each symbolic construction is defined as N _f=5, T _f=40ns.Adopt the modulation system of DS-BPSK, the DS sign indicating number is produced at random by shift-register sequence, and the cycle is 10.Multi-path dense channel under the CM3 situation of channel model employing IEEE 802.15.3a definition.Simultaneously, for obtaining track loop in the influence that has under the situation of timing error system's demodulation performance, each pulse is introduced one and is produced at random, timing error in a certain scope.The scope T of timing error _jRepresent.Receiver adopts has the selectivity of different branches to refer to receiver (SRake) more.By emulation, can obtain the analysis of the BER performance of system under different collection of energy strategies and different timing error scope.

When Fig. 8 has provided no track loop, the error rate under the different time jitters.This figure emulation when not introducing track loop, the bit error rate performance of system.We can find from figure, when timing error at 20%T _pWithin the time, the error rate of system performance is in tolerance interval, still, when timing error further became big, the error rate of system performance reduced rapidly, such as: in Fig. 9, work as T _j=40%T _p, Eb/N ₀During=20dB, the error rate only is 10 ^-2Therefore, when timing error about half pulse period or greater than half pulse period the time, just need to introduce the bit error rate performance that track loop improves system.

Figure 10, Figure 11 can see the effect of track loop in improving the error rate of system performance that this paper proposes, and these figure have described respectively in certain timing error scope, when not having track loop and track loop being arranged, and the contrast of the bit error rate performance of system.Such as in Fig. 9, when adopting 5 SRake of branch, compare the error rate two orders of magnitude that on the Eb/No=15dB point, descended when not having track loop.Equally, in Fig. 7.On the Eb/No=20dB point, the error rate of system performance that adopts 5 branches is compared when not having track loop, and nearly three orders of magnitude have descended.But, as shown in figure 11, work as T _j＞80%T _pThe time, bit error rate performance does not increase significantly after the introducing track loop, and this is that track loop is losing lock very easily, no longer can effectively follow the tracks of timing error because when timing error has exceeded following range.

The above is a preferable embodiment of the present invention; but protection scope of the present invention is not limited thereto; any those of ordinary skill in the art are in the technical scope of the present invention's explanation, and the variation that can expect easily or replacement all should be encompassed within protection scope of the present invention.Therefore, protection scope of the present invention should be as the criterion with the protection range of claims.

Claims (1)

1, a kind of ultra-wideband communication system track loop based on maximal possibility estimation, it is characterized in that this track loop is by low pass filter (1), time error detector (2), synchronized sampling road (3), loop filters little device (4) and digital controlled oscillator (5) is formed, received signal is at first passed through low pass filter (1), the output signal entry time error detector (2) of low pass filter (1), the output of time error detector (2) is sampled by synchronized sampling circuit (3), the value of sampling enters loop filter (4), and the output of loop filter (4) feeds back to synchronized sampling circuit (3) through digital controlled oscillator (5); Wherein:

A) low pass filter (1) carries out low-pass filtering to input signal, suppresses the noise signal of channel;

B) time error detector (2), the wave function of prior known received signal, simulation derivative matched filter transfer function is designed to the derivative of wave function, and received signal is by this derivative matched filter, and the signal that obtains promptly is the function of timing error;

C) synchronized sampling circuit (3) is one one a bit moduli transducer, estimate the position of the selectivity multipath that according to channel estimating, in the sampling of place that each pulse should finish, income value be actual reception to signal pulse and the value of the channel estimating timing error function of estimating the pulse position that;

D) loop filter (4) adopts second-order low-pass filter, regulates filter parameter can effectively suppress to sample noise component(s) in the gained timing error functional value, makes the timing error functional value that feedbacks more accurate;

E) digital controlled oscillator (5) is converted to time quantum with the timing error functional value, feeds back to the synchronized sampling circuit, realizes the locking of loop; Simultaneously this value is outputed to the demodulation link, adjust the correlation demodulation template position.

Images