CN103095614A - Joint equalization and frequency offset estimation device in proruption coherent optical fiber communications - Google Patents

Abstract

The invention discloses a joint equalization and frequency offset estimation device in proruption coherent optical fiber communications. Combined utilization of electric channel balance, frequency offset estimation and frequency offset compensation enables iteration step length needed by channel equalization to be reduced under the condition of the same frequency offset, and the condition that iteration is not restrained is avoided. Simultaneously, in the process of self-adaption update of a balance tap weight coefficient, equilibrium value and decision value which removes frequency offset compensation are adopted by an error signal to calculate, therefore, effect of the frequency offset compensation on accuracy of the error signal is removed, and an equalization module can accurately play a role. The joint equalization and frequency offset estimation device not only can estimate and compensate for proruption offset frequency in a certain range, but also can guarantee convergence of electric channel equalization. A proruption coherent optical fiber communication receiving system which is built by the joint equalization and frequency offset estimation device can keep good performance under the condition of big offset frequency.

Description

Joint equalization and frequency deviation estimation device in a kind of coherent fiber communication that happens suddenly

Technical field

The invention belongs to the optical communication technique field, more specifically say, relate to joint equalization and frequency deviation estimation device in a kind of coherent fiber communication that happens suddenly.

Background technology

Along with the progress of high-speed digital signal treatment technology and the people raising to communicating requirement, the high speed coherent optical communication becomes study hotspot.Current a lot of coherent optical communication system all is in continuous duty, and in a single day system sets up, and signal just recurs.But in some communication system, signal is but to be interrupted to occur to occur even suddenly, is called burst communication.This burst mode is a kind of covert communications, has transience, the difficult characteristics of being investigated and disturbing.

the relevant communication system receiver that detects is to utilize a local oscillator laser and the carrier (boc) modulated signals that receives to be concerned with to obtain baseband signal, require in theory the frequency of the frequency of oscillation of local oscillator laser and signal carrier identical, but in fact each laser has a certain amount of frequency of oscillation skew, suppose that the possible vibration frequency deviation region of each laser is [X, X] Hz, the scope of the relative frequency deviation of two lasers just may be [2X, 2X] Hz, so carrier wave frequency deviation is one of principal element that affects high speed coherent optical communication system performance, it will determine accurately demodulation of signal to a great extent, can information full backup.

Fig. 1 is the frequency drift schematic diagram of burst communication.Burst communication is unstable the incipient stage meeting frequency of occurrences, and the unsteadiness of burst frequency can produce a very large impact the judgement of back signal, therefore need to estimate and compensate the burst frequency deviation.Along with the raising of transmission rate, optical fiber dispersion will be more serious on the impact of system on the other hand, thus in high-speed burst coherent fiber communication system except the estimation of the frequency deviation that will happen suddenly and compensation, channel equalization also is absolutely necessary.In existing continuous coherent optical communication system research, estimate and compensation in order to carry out frequency deviation, suppose to have carried out dispersion equilibrium in the light territory, therefore carry out frequency deviation estimate with compensate in do not need to consider the impact of dispersion.

The existing fiber communication system is when carrying out dispersion equilibrium, the radio channel equalizer tap coefficient adopts the self adaptation adjustment, not only can realize dispersion compensation, therefore can also achieve a butt joint the to a certain extent tracking of the phase jitter that causes due to frequency deviation in collecting mail number does not need to carry out in addition yet correct restituted signal of compensate of frequency deviation in the situation that frequency deviation is smaller.But along with the increase of system's frequency deviation, the equalizer iteration step length also needs to increase in order to can trace into the variation of the signal phase that frequency deviation causes thereupon; Excessive iteration step length can cause equalizer tap coefficient unstable, and even convergence, do not cause the seriously deteriorated of systematic function.

Summary of the invention

The object of the invention is to overcome the deficiencies in the prior art, joint equalization and frequency deviation estimation device in a kind of coherent fiber communication that happens suddenly are provided, by use is united in radio channel equilibrium and frequency deviation estimation and compensate of frequency deviation, the balanced iteration step length that needs of frequency deviation the same terms lower channel is reduced, thereby the situation of avoiding iteration not restrain, make balance module can play consistently effect, system can be estimated and the compensate for slower frequency deviation that happens suddenly on a large scale, also can have performance preferably in the situation that frequency deviation is larger.

For achieving the above object, the present invention happen suddenly joint equalization and frequency deviation estimation device in coherent fiber communication is characterized in that comprising:

One balance module is used for receiving the signal r after the AD sample quantization _kCarry out equilibrium treatment, the previous signal r that namely feeds back according to judging module _k-1Decision value

Previous signal r with the frequency deviation estimating modules feedback _k-1Accumulation frequency deviation θ _k-1Carrying out the iteration of balanced tap weight value coefficient upgrades:

$C_k=C_{k-1}-{\mathrm{\Δ\ϵ}}_{k-1}{r}_{k-1}^{*}$

Wherein, C _k, C _k-1It is respectively k and k-1 signal balanced tap weight value coefficient when passing through balance module; r _k-1The reception signal sample of k-1 signal, Be signal r _k-1Conjugate, Δ is iteration step length, is positive number, and ε is set by the user _k-1Be k-1 the error signal that signal obtains, its computing formula is:

${\mathrm{\ϵ}}_{k-1}={\tilde{m}}_{k-1}-{\tilde{m}}_{k-1}\·e^{{\mathrm{j\θ}}_{k-1}}$

Wherein,

The equilibrium value after k-1 signal passes through balance module,

It is judging module pair

The decision value of the signal that obtains after the process compensate of frequency deviation, θ _k-1Represent the accumulation frequency deviation that k-1 signal obtains by frequency deviation estimating modules;

Balance module is according to balanced tap weight value coefficient C _kTo signal r _kCarry out equilibrium, the output equilibrium value

Send judging module and frequency deviation module to;

One judging module receives equilibrium value And the accumulation frequency deviation θ that obtains by frequency deviation estimating modules according to k-1 signal _k-1, the computation balance value

The compensate of frequency deviation signal

And adjudicate the output decision value

Send frequency deviation estimating modules to and feed back to balance module;

One frequency deviation estimating modules receives the equilibrium value that balance module transmits

Decision value with the judging module transmission Equilibrium value

Signal phase is expressed as φ _d,k+ k Δ ω T+ φ _n,k, wherein

The expression phase modulation, k Δ ω T represents the phase error that frequency deviation is introduced,

The phase error that Gaussian noise causes, decision value

Signal phase is

The accumulated phase error estimated value z_tmp of k signal before calculating _k:

$z\_t{\mathrm{mp}}_k=({\tilde{m}}_k\·{{\hat{m}}_k}^{*})$

$={\mathrm{\φ}}_{d,k}+\mathrm{k\Δ\ωT}+{\mathrm{\φ}}_{n,k}-{\mathrm{\φ}}_{d,k}$

$=\mathrm{k\Δ\ωT}+{\mathrm{\φ}}_{n,k}$

Calculate the phase error estimation and phase error value α of k signal _k:

α _k=z_tmp _k·(z_tmp_delay _k-1) ^*

=kΔωT+φ _n,k-[(k-1)ΔωT+φ _n,k-1]

=ΔωT+φ _n,k-φ _n,k-1

Z_tmp_delay wherein _k-1Accumulated phase error estimated value z_tmp by a front k-1 signal _k-1Postponing a bat obtains;

To phase error estimation and phase error value α _kAsk the angle operation to obtain angle value β _kIf, angle value β _kIn [π, π] scope with phase error estimation and phase error value α _kSend into filter and suppress noise phase, otherwise with phase error estimation and phase error value α _kAbandon; Suppress to obtain frequency deviation estimated value Δ ω T after noise phase the accumulation frequency deviation θ of k signal before calculating through filter _k=θ _k-1+ Δ ω T;

Frequency deviation estimating modules will be accumulated frequency deviation θ _kTransmission feeds back to balance module and judging module respectively, is used for equilibrium treatment and the judgement of next signal.

Wherein, balance module is the adaptive equalizer based on the LMS algorithm, comprises a series of FIR filters.

Further, frequency deviation estimating modules comprises a memory, is used for storage accumulated phase error estimated value z_tmp _k, first take out z_tmp _k-1Postpone a bat processing and obtain z_tmp_delay _k-1Make itself and z_tmp _kSynchronously, then write z_tmp _kSignal.

Wherein, described balanced tap weight value coefficient C _k-1, error signal _k-1, accumulated phase error estimated value z_tmp _k-1, accumulation frequency deviation θ _k-1Initial value carry out initialization by training sequence and obtain, initialization procedure is:

Training sequence pilot _k, length is L, L is arranged by the user, needs the length l greater than the sliding window of frequency deviation estimating modules median filter, and consider the burst frequency deviation disperse the symbol numbers requirement, L needs long enough to guarantee the making balance module convergence; Training sequence pilot _kAfter disturbing processing, simulation obtains burst R _k

Balance module receives burst R _k, balanced tap weight value coefficient is always C _k=1; Signal R ₁Generate the 1st equilibrium value through balance module

Signal R ₂Generate the 2nd equilibrium value through balance module

With z_tmp ₁Postpone one and clap generation z_tmp_delay ₁, phase error estimation and phase error value α ₂=z_tmp ₂(z_tmp_delay ₁) ^*, to α ₂Get angle and obtain the 1st angle value β ₂To burst R _kRepeat and the 2nd processing that signal is identical, until signal R _l+1The time, angle value is accumulate to l, and l angle value enters together filter inhibition noise phase and obtains the 1st frequency deviation estimated values theta _l+1, obtain error signal ${\mathrm{\ϵ}}_{l+1}={\tilde{m}}_{l+1}-{\mathrm{pilot}}_{l+1}\·e^{{\mathrm{j\θ}}_{l+1}};$

Balanced tap weight value coefficient begins iteration and upgrades C _l+2=C _l+1-Δ ε _l+1R _l+1 ^*=1-Δ ε _l+1R _l+1 ^*, signal R _l+2Obtain equilibrium value through balance module

Carry out frequency deviation and estimate to obtain the 1st accumulation frequency deviation θ _l+2Adopt afterwards the balanced tap weight value coefficient of iteration renewal to burst R _kCarry out joint equalization and frequency deviation and estimate, wherein C _k=C _k-1-Δ ε _k-1R _k-1 ^*,

Until training sequence finishes to complete the initialization of balanced tap weight value coefficient, balance module convergence this moment is with the C of this moment _L, ε _L, z_tmp _L, θ _LAs balanced tap weight value coefficient C _k-1, error signal _k-1, accumulated phase error estimated value z_tmp _k-1, accumulation frequency deviation θ _k-1Initial value.

Goal of the invention of the present invention is achieved in that the present invention is in the adaptive updates process of balanced tap weight value coefficient, error signal adopts equilibrium value and the decision value that removes compensate of frequency deviation to calculate, thereby eliminated the impact of compensate of frequency deviation on the error signal accuracy, balance module can be played a role accurately.

Joint equalization and frequency deviation estimation device that the present invention happens suddenly in coherent fiber communication not only can be estimated and compensate the burst frequency deviation in certain limit, and can guarantee the convergence of radio channel equilibrium, allow and adopt the burst coherent fiber communication receiving system that joint equalization of the present invention and frequency deviation estimation device set up in the situation that the burst frequency deviation is arranged or frequency deviation is larger also can keep performance preferably.

Description of drawings

Fig. 1 is the frequency drift schematic diagram of burst communication;

Fig. 2 is the present invention's joint equalization in coherent fiber communication and frequency deviation estimation device be applied to happen suddenly structural representation of coherent fiber communication system that happens suddenly;

Fig. 3 is happen suddenly joint equalization in coherent fiber communication and a kind of embodiment structure chart of frequency deviation estimation device of the present invention;

Fig. 4 is the happen suddenly a kind of burst frequency deviation simulation result figure of embodiment of joint equalization in coherent fiber communication and frequency deviation estimation device of the present invention;

Fig. 5 adopts the present invention happen suddenly joint equalization in coherent fiber communication and the burst coherent fiber communication receiving system error sign ratio simulation result figure of frequency deviation estimation device.

Embodiment

Below in conjunction with accompanying drawing, the specific embodiment of the present invention is described, so that those skilled in the art understands the present invention better.What need to point out especially is that in the following description, when perhaps the detailed description of known function and design can desalinate main contents of the present invention, these were described in here and will be left in the basket.

Fig. 2 is the present invention's joint equalization in coherent fiber communication and frequency deviation estimation device be applied to happen suddenly structural representation of coherent fiber communication system that happens suddenly.As shown in Figure 2, at transmitting terminal, respectively the orthogonal two ways of optical signals in polarization direction is carried out difference quadrature phase shift keying (QPSK) modulation, polarizing coupler is realized the palarization multiplexing to two ways of optical signals, obtain the signal transmission of PDM-QPSK, when signal transmitted at monomode fiber, the effects such as chromatic dispersion, polarization mode dispersion can cause the broadening of light pulse, formed intersymbol interference.

At receiving terminal, polarisation-multiplexed signal carries out independently coherent demodulation through the signal that polarization beam apparatus is divided into two polarization directions.At first 4 road signals of telecommunication after coherent demodulation need to carry out the AD sampling and quantize, more further carry out radio channel equilibrium and compensate of frequency deviation.At last, the phase place judgement recovers the data that send, and carries out the statistics of error sign ratio.

Fig. 3 is happen suddenly joint equalization in coherent fiber communication and a kind of embodiment structure chart of frequency deviation estimation device of the present invention.As shown in Figure 3, the joint equalization and the frequency deviation estimation device that happen suddenly in coherent fiber communication of the present invention comprises balance module 1, judging module 2, frequency deviation estimating modules 3.

Need use decision signal during due to balanced tap weight value coefficient update error signal calculation used, therefore before signal transmission, need balanced tap weight value coefficient C _k-1, error signal _k-1, accumulated phase error estimated value z_tmp _k-1, accumulation frequency deviation θ _k-1Carry out initialization.In present embodiment, adopt one section training sequence to carry out initialization, initialization procedure is:

Training sequence pilot _k, length is L, L is arranged by the user, needs the length l greater than frequency deviation estimating modules 3 median filters 5 sliding windows, and consider the burst frequency deviation disperse the symbol numbers requirement, L needs long enough to guarantee making balance module 1 convergence; Training sequence pilot _kAfter disturbing processing, simulation obtains the burst R that length is L _k

Balance module 1 receives burst R _k, balanced tap weight value coefficient is always C _k=1; Signal R ₁Generate the 1st equilibrium value through balance module 1 Signal R ₂Generate the 2nd equilibrium value through balance module 1

With z_tmp ₁Postpone one and clap generation z_tmp_delay ₁, phase error estimation and phase error value α ₂=z_tmp ₂(z_tmp_delay ₁) ^*, to α ₂Get angle and obtain the 1st angle value β ₂To burst R _kRepeat and the 2nd processing that signal is identical.Because the sliding window length of filter is l, so angle value need be accumulate to l.To signal R _l+1The time, angle value is accumulate to l, and l angle value enters together filter inhibition noise phase and obtains the 1st frequency deviation estimated values theta _l+1, obtain error signal

Error signal in initialization procedure _kCalculating do not use decision value, but use training sequence pilot _kReplace decision value.

Balanced tap weight value coefficient begins iteration and upgrades C _l+2=C _l+1-Δ ε _l+1R _l+1 ^*=1-Δ ε _l+1R _l+1 ^*, signal R _l+2Obtain equilibrium value through balance module 1

Carry out frequency deviation and estimate to obtain the 1st accumulation frequency deviation θ _l+22 pairs of judging module

According to θ _l+1Carry out compensate of frequency deviation and obtain signal

Judgement obtains decision value

Although obtain decision value, do not use in initialization procedure.

Burst R afterwards _kThe balanced tap weight value coefficient that adopts iteration to upgrade carries out joint equalization and frequency deviation is estimated, wherein C _k=C _k-1-Δ ε _k-1R _k-1 ^*, Until training sequence finishes to complete the initialization of balanced tap weight value coefficient, balance module 1 convergence this moment is with the C of this moment _L, ε _L, z_tmp _L, θ _LAs balanced tap weight value coefficient C _k-1, error signal _k-1, accumulated phase error estimated value z_tmp _k-1, accumulation frequency deviation θ _k-1Initial value.

After initialization was completed, joint equalization and frequency deviation estimation device received the signal r that generates after coherent demodulation, AD sample quantization _k, begin to carry out joint equalization and frequency deviation and estimate.

Balance module 1 receives signal r _kCarry out equilibrium treatment, the previous signal r that at first feeds back according to judging module 2 _k-1Decision value

Previous signal r with frequency deviation estimating modules 3 feedbacks _k-1Accumulation frequency deviation θ _k-1Calculate the balanced tap weight value coefficient of this equilibrium:

$C_k=C_{k-1}-{\mathrm{\Δ\ϵ}}_{k-1}{r}_{k-1}^{*}$

As seen, balanced tap weight value coefficient is that iteration is upgraded.Wherein, C _k, C _k-1It is respectively k and k-1 signal balanced tap weight value coefficient when passing through balance module 1; r _k-1The reception signal sample of k-1 signal,

Be its conjugate; Δ is iteration step length, is positive number, is arranged by the user, need be set to enough little of to guarantee that iterative process can restrain; ε _k-1Be k-1 the error signal that signal obtains, its computing formula is:

${\mathrm{\ϵ}}_{k-1}={\tilde{m}}_{k-1}-{\tilde{m}}_{k-1}\·e^{{\mathrm{j\θ}}_{k-1}}$

Wherein,

The equilibrium value after k-1 signal passes through balance module 1,

2 pairs of judging module

Through the signal that obtains after compensate of frequency deviation

Decision value, θ _k-1Represent the accumulation frequency deviation that k-1 signal obtains by frequency deviation estimating modules 3.Owing to having carried out θ before signal decision _k-1Compensate of frequency deviation, and the estimated value after balanced does not have such compensation, therefore the compensate of frequency deviation of decision value need to be removed in error signal, thereby makes error signal _k-1More accurate.

Balance module 1 is according to the balanced tap weight value coefficient C that calculates _kTo signal r _kCarry out equilibrium, the output equilibrium value

Send judging module 2 and frequency deviation estimating modules 3 to.

In present embodiment, balance module 1 is the adaptive equalizer based on the LMS algorithm, is comprised of a series of FIR filters.

Judging module 2 receives the equilibrium value that balance module 1 transmits

And the accumulation frequency deviation θ that obtains by frequency deviation estimating modules according to k-1 signal _k-1, the computation balance value

The compensate of frequency deviation signal

And adjudicate the output decision value

Send frequency deviation estimating modules 3 to, and feed back to balance module 1 for carrying out the renewal of balanced tap weight value coefficient next time.

Frequency deviation estimating modules 3 receives the equilibrium value that balance module 1 transmits

Decision value with judging module 2 transmission

Equilibrium value

Signal phase is expressed as φ _d,k+ k Δ ω T+ φ _n,k, φ wherein _d,kThe expression phase modulation, k Δ ω T represents the phase error that frequency deviation is introduced, φ _n,kThe phase error that Gaussian noise causes, decision value

Signal phase is φ _d,kEquilibrium value

With decision value

Conjugate multiplication, can eliminate the impact of phase modulation, obtain the accumulated phase error estimated value z_tmp of a front k signal _k:

$z\_t{\mathrm{mp}}_k=({\tilde{m}}_k\·{{\hat{m}}_k}^{*})$

$={\mathrm{\φ}}_{d,k}+\mathrm{k\Δ\ωT}+{\mathrm{\φ}}_{n,k}-{\mathrm{\φ}}_{d,k}$

$=\mathrm{k\Δ\ωT}+{\mathrm{\φ}}_{n,k}$

Calculate the phase error estimation and phase error value α of k signal _k:

α _k=z_tmp _k·(z_tmp_delay _k-1) ^*

=kΔωT+φ _n,k-[(k-1)ΔωT+φ _n,k-1]

=ΔωT+φ _n,k-φ _n,k-1

Z_tmp_delay wherein _k-1Accumulated phase error estimated value z_tmp by a front k-1 signal _k-1Postpone a bat and obtain, (z_tmp_delay _k-1) ^*Be its conjugate.As seen, α _kThe frequency deviation that includes only k signal is estimated the phase place change component φ that component Δ ω T and noise cause _n,k-φ _{N, k-1}

Memory in frequency deviation estimating modules 3 is used for storage accumulated phase error estimated value z_tmp _k, first take out z_tmp _k-1Postpone a bat processing and obtain z_tmp_delay _k-1Make itself and z_tmp _kSynchronously, then write z_tmp _kSignal.

The phase component Δ ω T that frequency deviation is estimated can only be between [π, π], otherwise can cause phase ambiguity, makes frequency deviation estimate to have uncertainty, so the present invention limits

Between [π, π].3 couples of phase error estimation and phase error value α of frequency deviation estimating modules _kAsk the angle operation to obtain angle value β _kIf, angle value β _kIn [π, π] scope with phase error estimation and phase error value α _kSend into filter and suppress noise phase, otherwise with phase error estimation and phase error value α _kAbandon; After suppressing noise phase, filter obtains the frequency deviation estimated value Δ ω T of k signal; Because frequency deviation is a cumulative effect, on the accumulation frequency deviation of k-1 signal, calculate the accumulation frequency deviation θ of k signal before Δ ω T being accumulated to _k=θ _k-1+ Δ ω T.

Frequency deviation estimating modules 3 will be accumulated frequency deviation θ _kTransmission feeds back to balance module 1 and judging module 2 respectively, is used for equilibrium treatment and the judgement of next signal.

Embodiment

Joint equalization and frequency deviation estimation device that the present invention is happened suddenly in coherent fiber communication carry out emulation experiment, parameter is set to: the decay factor of monomode fiber is 0.2dB/km, the 2nd order chromatic dispersion coefficient is-20ps^2/km, the Optical Fiber Transmission distance is 100km, receiver bandwidth is set to 50GHz, the length of the frequency deviation estimating modules 3 sliding windows of median filters is 128, and the iteration step length of balance module 1 is the best iteration step length 0.001 of transmission range when being 100 kilometers.

The training sequence length that table 1 adopts when happening suddenly scope for the difference for frequency deviation.

Fig. 4 is the happen suddenly a kind of burst frequency deviation simulation result figure of embodiment of joint equalization in coherent fiber communication and frequency deviation estimation device of the present invention.As shown in Figure 4, curve represents that joint equalization and frequency deviation estimation device that the present invention happens suddenly in coherent fiber communication can estimate and compensate different burst frequency deviation regions, the burst frequency deviation region is larger, the length of the initialization training sequence that needs is longer, could make balance module 1 convergence at the training sequence initial phase like this, make stable the playing a role of balance module 1.Can see in Fig. 4, along with the change of frequency deviation burst scope is large, the optical fiber communication performance of receiving system can be subject to certain impact, but is also acceptable.Even when the burst frequency deviation increase to ± when 6GHz, performance loss neither be very large, illustrate that happen suddenly joint equalization in coherent fiber communication and frequency deviation estimation device of the present invention is applicable to exist the optical fiber communication receiving system of offset frequency situation of happening suddenly.

Fig. 5 adopts the present invention happen suddenly joint equalization in coherent fiber communication and the receiving system error sign ratio simulation result figure of frequency deviation estimation device.As shown in Figure 5, dotted line represents to only have balanced situation, and solid line represents situation balanced and frequency deviation estimation associating use, and the balanced step-length that every Curves is used is optimum stepsize.

Table 2 is that Optical Fiber Transmission distance is 100km, balance module 1 optimum stepsize used under different frequency deviations.

Frequency deviation size (Hz)	1M	10M	100M	500M
					Balanced	0.002	0.05	0.1	－
Balanced and frequency deviation is estimated	0.001	0.001	0.001	0.001

Table 2

As shown in Figure 5, in the situation that only have balancedly, the patient frequency deviation of receiving system institute is very little, and when frequency deviation value was increased to 100M, even high Optical Signal To Noise Ratio OSNR again, the error sign ratio SER of receiving system was very high; And adopt in the situation of joint equalization and frequency deviation estimation device, when frequency deviation was increased to 200M, under best iteration step length, the error sign ratio SER of receiving system obviously reduced, and along with the increase of Optical Signal To Noise Ratio, the error sign ratio SER of receiving system further reduces.As seen joint equalization and frequency deviation estimation device can make burst coherent fiber communication performance of receiving system be greatly improved.

Although the above is described the illustrative embodiment of the present invention; so that those skilled in the art understand the present invention; but should be clear; the invention is not restricted to the scope of embodiment; to those skilled in the art; as long as various variations appended claim limit and the spirit and scope of the present invention determined in, these variations are apparent, all utilize innovation and creation that the present invention conceives all at the row of protection.

Claims (4)

1. joint equalization and the frequency deviation estimation device in the coherent fiber communication that happens suddenly is characterized in that comprising:

One balance module is used for receiving the signal r after the AD sample quantization _kCarry out equilibrium treatment, the previous signal r that namely feeds back according to judging module _k-1Decision value

Previous signal r with the frequency deviation estimating modules feedback _k-1Accumulation frequency deviation θ _k-1Carrying out the iteration of balanced tap weight value coefficient upgrades:

$C_k=C_{k-1}-{\mathrm{\Δ\ϵ}}_{k-1}{r}_{k-1}^{*}$

Wherein, C _k, C _k-1It is respectively k and k-1 signal balanced tap weight value coefficient when passing through balance module; r _k-1Be the reception signal sample of k-1 signal, Δ is iteration step length, is positive number, and ε is set by the user _k-1Be k-1 the error signal that signal obtains, its computing formula is:

${\mathrm{\ϵ}}_{k-1}={\tilde{m}}_{k-1}-{\tilde{m}}_{k-1}\·e^{{\mathrm{j\θ}}_{k-1}}$

Wherein,

The equilibrium value after k-1 signal passes through balance module, It is judging module pair

The decision value of the signal that obtains after the process compensate of frequency deviation, θ _k-1Represent the accumulation frequency deviation that k-1 signal obtains by frequency deviation estimating modules;

Balance module is according to balanced tap weight value coefficient C _kTo signal r _kCarry out equilibrium, the output equilibrium value

Send judging module and frequency deviation module to;

One judging module receives equilibrium value

And the accumulation frequency deviation θ that obtains by frequency deviation estimating modules according to k-1 signal _k-1, the computation balance value

The compensate of frequency deviation signal

And adjudicate the output decision value

Send frequency deviation estimating modules to and feed back to balance module;

One frequency deviation estimating modules receives the equilibrium value that balance module transmits

Decision value with the judging module transmission

Equilibrium value

Signal phase is expressed as φ _d,k+ k Δ ω T+ φ _n,k, wherein

The expression phase modulation, k Δ ω T represents the phase error that frequency deviation is introduced,

The phase error that Gaussian noise causes, decision value

Signal phase is

The accumulated phase error estimated value z_tmp of k signal before calculating _k:

$z\_t{\mathrm{mp}}_k=({\tilde{m}}_k\·{{\hat{m}}_k}^{*})$

$={\mathrm{\φ}}_{d,k}+\mathrm{k\Δ\ωT}+{\mathrm{\φ}}_{n,k}-{\mathrm{\φ}}_{d,k}$

$=\mathrm{k\Δ\ωT}+{\mathrm{\φ}}_{n,k}$

Calculate the phase error estimation and phase error value α of k signal _k:

α _k=z_tmp _k·(z_tmp_delay _k-1) ^*

=kΔωT+φ _n,k-[(k-1)ΔωT+φ _n,k-1]

=ΔωT+φ _n,k-φ _n,k-1

To phase error estimation and phase error value α _kAsk the angle operation to obtain angle value β _kIf, angle value β _kIn [π, π] scope with phase error estimation and phase error value α _kSend into filter and suppress noise phase, otherwise with phase error estimation and phase error value α _kAbandon; Suppress to obtain frequency deviation estimated value Δ ω T after noise phase the accumulation frequency deviation θ of k signal before calculating through filter _k=θ _k-1+ Δ ω T;

Frequency deviation estimating modules will be accumulated frequency deviation θ _kTransmission feeds back to balance module and judging module respectively, is used for equilibrium treatment and the judgement of next signal.

2. joint equalization according to claim 1 and frequency deviation estimation device, is characterized in that, described balance module is the adaptive equalizer based on the LMS algorithm, comprises a series of FIR filters.

3. joint equalization according to claim 1 and frequency deviation estimation device, is characterized in that, described frequency deviation estimating modules comprises a memory, is used for storage accumulated phase error estimated value z_tmp _k, first take out z_tmp _k-1Postpone a bat processing and obtain z_tmp_delay _k-1Make itself and z_tmp _kSynchronously, then write z_tmp _kSignal.

4. joint equalization according to claim 1 and frequency deviation estimation device, is characterized in that, described balanced tap weight value coefficient C _k-1, error signal _k-1, accumulated phase error estimated value z_tmp _k-1, accumulation frequency deviation θ _k-1Initial value carry out initialization by training sequence and obtain, initialization procedure is:

Training sequence pilot _k, length is L, L is arranged by the user, needs the length l greater than the sliding window of frequency deviation estimating modules median filter, and consider the burst frequency deviation disperse the symbol numbers requirement, L needs long enough to guarantee the making balance module convergence; Training sequence pilot _kAfter disturbing processing, simulation obtains burst R _k

Balance module receives burst R _k, balanced tap weight value coefficient is always C _k=1; Signal R ₁Generate the 1st equilibrium value through balance module

Signal R ₂Generate the 2nd equilibrium value through balance module

With z_tmp ₁Postpone one and clap generation z_tmp_delay ₁, phase error estimation and phase error value α ₂=z_tmp ₂(z_tmp_delay ₁) ^*, to α ₂Get angle and obtain the 1st angle value β ₂To burst R _kRepeat and the 2nd processing that signal is identical, until signal R _l+1The time, angle value is accumulate to l, and l angle value enters together filter inhibition noise phase and obtains the 1st frequency deviation estimated values theta _l+1, obtain error signal ${\mathrm{\ϵ}}_{l+1}={\tilde{m}}_{l+1}-{\mathrm{pilot}}_{l+1}\·e^{{\mathrm{j\θ}}_{l+1}};$

Balanced tap weight value coefficient begins iteration and upgrades C _l+2=C _l+1-Δ ε _l+1R _l+1 ^*=1-Δ ε _l+1R _l+1 ^*, signal R _l+2Obtain equilibrium value through balance module

Carry out frequency deviation and estimate to obtain the 1st accumulation frequency deviation θ _l+2Adopt afterwards the balanced tap weight value coefficient of iteration renewal to burst R _kCarry out joint equalization and frequency deviation and estimate, wherein C _k=C _k-1-Δ ε _k-1R _k-1 ^*,

Until training sequence finishes to complete the initialization of balanced tap weight value coefficient, balance module convergence this moment is with the C of this moment _L, ε _L, z_tmp _L, θ _LAs balanced tap weight value coefficient C _k-1, error signal _k-1, accumulated phase error estimated value z_tmp _k-1, accumulation frequency deviation θ _k-1Initial value.

Images