CN106153029A - Two frequency machine shaking laser gyroscope shaking signal cancellation devices - Google Patents

Abstract

The present invention relates to a kind of two frequency machine shaking laser gyroscope shaking signal cancellation devices, it is technically characterized by comprising the steps as follows: and includes gyro data acquisition module, calculus of differences module, low order FIR filter module, summation operation module, high frequency clock module, low order iir filter module, IIR high pass filter block, the first digital phase shifter module, shake estimation module, digital dock conversion module, the second digital phase shifter module, band filter module, digital dock management module, the 3rd digital phase shifter module and LMS weight coefficient more new module.The present invention is reasonable in design, it uses adaptive interference cancelling technology based on LMS algorithm to carry out shaking laser gyroscope signal cancellation, offset use just, cosine reference signal Dui You mono-road sampled differential signal obtains through signal phase shift, system input and output response is constant when approximating, can reduce to tremble and offset except time delays, can be widely used for eliminating the mechanical shaking signal in two frequency mechanical shaking lasergyro outputs in real time.

Description

Two frequency machine shaking laser gyroscope shaking signal cancellation devices

Technical field

The invention belongs to two frequency machine laser gyroscope shaking technical fields, especially a kind of two frequency machine shaking laser gyroscope shaking signals Canceller.

Background technology

Quickly following the tracks of occasion, it is desirable to control system real-time output angle rate information.For mechanical shaking laser gyro, its Output signal had both comprised the carrier angular movement information that gyro sensitivity arrives, and also comprised mechanical shaking angular movement.In order to obtain carrier angle Movable information, it is necessary to gyro output signals is demodulated.Typical demodulation scheme mainly has high-speed sampling low-pass filtering and trembles Dynamic counteracting method, high-speed sampling low-pass filtering is generally completed by FIR low pass filter, the FIR filter exponent number that design obtains Higher, filtering delay-time is relatively big, is not suitable for quickly following the tracks of applications, and the feature with zero propagation is offset in shake, can be the most defeated Go out angular rate information, be only applicable to stability contorting etc. and quickly follow the tracks of occasion.

The dither reference signal of existing shake device for stripping is obtained by shake pick-up circuit, before this method needs to be provided with End modulate circuit, also wants additional A/D conversion chip, adds the complexity of hardware.Generally use the pickup shakes such as piezoelectric ceramics Signal, piezoelectric ceramics has lagging characteristics, creep properties and dither nonlinear characteristic, piezoelectric ceramics the signal picked up toward Toward being not real dither signal, show as signal and have non-linear, from frequency spectrum, wherein contain higher harmonic components.

Existing obtaining dither signal from phase-lock technique by gyro, gyro output sampling frequency rate is by fixing sampling clock Completing, it is impossible to just realizing, cosine dither reference signal strictly differs 90 °, therefore, it is poor that shake divests effect.

Existing obtaining dither signal from phase-lock technique by the logical iir filter of band, when gyro, to be operated in output amplitude bigger Occasion time, the band filter exponent number designed is higher, be easily caused IIR output dissipate；Further, since iir filter has Feedback element, it is impossible to realizing full precision cut position during fixed-point calculation, the cut position improper IIR of also resulting in output dissipates, and causes filtering Failure.

Gyro output is directly carried out after calculus of differences by existing shake device for stripping through adaptive jitter device for stripping Shake divests, and LMS algorithm is with feedback element, it is impossible to realize the full precision data cut position of fixed-point calculation, in order to improve system Shake divests effect, it usually needs more word length represents operational data, takies more hardware logic resource, and cut position is not simultaneously When LMS output can be caused to dissipate, shake divests inefficacy.

Existing shake device for stripping all use two-way phase 90 ° just, cosine dither reference signal is for shaking Signal estimation and LMS weight coefficient update, and it is not ideal enough that the shake obtained divests effect.

Summary of the invention

It is an object of the invention to overcome the deficiency of existing design, it is provided that a kind of two frequency machine shaking laser gyroscope shaking signals support Disappear device, solves following problem: 1, existing LMS adaptive jitter divests technology and obtained through band filter by gyro output signals To dither reference signal, in the bigger occasion of angle increment signal amplitude of gyro output, in order to obtain accurate dither reference signal N_d, it is higher that the band designed leads to iir filter exponent number, and the iir filter of high-order is computationally intensive, and when realizing in FPGA, cuts Position is dealt with improperly, and the output of wave filter dissipates, and causes filter normally to work；2, existing LMS adaptive jitter divests Technology shake estimatorPrecision is poor, and the carrier angular movement precision of information after shake divests is poor, it usually needs increase low order The exponent number of low-pass FIR filter improves the precision of gyro output carrier angular movement information, and this invention has redesigned shake and estimated Calculating method, while ensureing Gyro Precision, reduce further the exponent number of low-pass FIR filter, reduces at gyro data The time delay of reason, improves the real-time of system；3, dither reference signal is obtained through gyro, it is not necessary to outside shake pick-up circuit obtains Take dither reference signal, simplify the hardware designs that shake is offset, when gyro obtains dither reference signal, devise low order IIR high pass filter obtains dither reference signal, and devises digital dock transducer (DCC) and realize digital signal and believe to clock Number conversion, to the clock signal obtained 4M frequency multiplication in FPGA, obtain and the sampling clock of same frequency Tong Bu with dither signal Gyro output signals is sampled by signal, obtains phase place and strictly differs the dither reference signal of 90 °, thus improves the effect that shake is offset Really.

The present invention solves it and technical problem is that and take techniques below scheme to realize:

A kind of two frequency machine shaking laser gyroscope shaking signal cancellation devices, including gyro data acquisition module, calculus of differences mould Block, low order FIR filter module, summation operation module, high frequency clock module, low order iir filter module, IIR high-pass filtering Device module, the first digital phase shifter module, shake estimation module, digital dock conversion module, the second digital phase shifter module, Band filter module, digital dock management module, the 3rd digital phase shifter module and LMS weight coefficient more new module；Described top Spiral shell data acquisition module, low order iir filter module, digital dock conversion module, digital dock management module is sequentially connected Connecing, another input of low order iir filter module connects high frequency clock module, and described gyro data acquisition module is according to number Dither signal is sampled by the employing frequency of word clock management module output, and sampled data is the lowest through the output of calculus of differences module Rank FIR filter module, low order FIR filter module outfan connects IIR high pass filter block and summation operation respectively Module, the outfan of this IIR high pass filter block connect respectively the first digital phase shifter module, shake estimation module, second Digital phase shifter module and LMS weight coefficient more new module, the outfan of this first digital phase shifter module connects shake respectively and estimates Meter module and LMS weight coefficient more new module, the outfan of this second digital phase shifter module connect respectively shake estimation module, the Three digital phase shifter modules and LMS weight coefficient more new module, the outfan of the 3rd digital phase shifter module connects shake respectively Estimation module and LMS weight coefficient more new module, the outfan of this shake estimation module connects another input of summation operation module End, the outfan of this summation operation module connects the input of band filter module, and this band filter module connects LMS power Coefficient updating module, this LMS weight coefficient more new module output LMS algorithm updates weight coefficient to shaking estimation module.

Described digital dock conversion module is arranged in FPGA, and the sample clock frequency of its output is 4Mf_s, wherein, M For the integer more than or equal to 1, f_sFor gyro chattering frequency.

Pulse increment Δ N (the n)=Δ N of described calculus of differences module output_b(n)+ΔN_d(n), wherein, Δ N_b(n)、Δ N_dN () represents the gyro base corner motion change in the kth sampling period and mechanical shaking angular movement change respectively；Described low order Angular movement pulse increment Δ N after FIR filter process_lpf(n) be:

ΔN_lpf(n)=Δ N_b′(n)+ΔN_d' (n):

Wherein, Δ N_b' (n) represents the gyro base corner increment of motion through low order FIR filter, Δ N_d' (n) represents warp The mechanical shaking angular movement increment signal of overdamping, K is the exponent number representing FIR filter.

Described IIR high pass filter filters the gyro base corner increment of motion Δ N of low order FIR filter output_bAfter ' (n) Obtain the first dither reference signal Δ N_d1(n): described first digital phase shift module is to Δ N_d1N () obtains the second tunnel after processing and trembles Dynamic reference signal Δ N_d2(n), Δ N_d2(n)=Δ N_d1(n-1)；Described second digital phase shift module is to Δ N_d1N () obtains after processing With Δ N_d1The 3rd tunnel dither reference signal Δ N of (n) phase 90 °_d3(n), this Δ N_d3(n)=Δ N_d1(n-M), M is for being more than Integer equal to 1；Described 3rd digital phase shift module is to the 3rd tunnel dither reference signal Δ N_d3N () obtains the 4th tunnel after processing and trembles Dynamic reference signal Δ N_d4(n), Δ N_d4(n)=Δ N_d3(n-1)。

Described LMS weight coefficient more new module is according to four tunnel dither reference signal Δ N_d1(n)、ΔN_d2(n)、ΔN_d3(n)、Δ N_d4(n) and band filter output valve ε (n) output LMS algorithm renewal weight coefficient:

h₁₁(n+1)=h₁₁(n)+2×μ×ε(n)×ΔN_d2(n)

h₁₂(n+1)=h₁₂(n)+2×μ×ε(n)×ΔN_d1(n)

h₂₁(n+1)=h₂₁(n)+2×μ×ε(n)×ΔN_d4(n)

h₂₂(n+1)=h₂₂(n)+2×μ×ε(n)×ΔN_d3(n)。

Described shake estimation module is according to four tunnel dither reference signal Δ N_d1(n)、ΔN_d2(n)、ΔN_d3(n)、ΔN_d4(n) And four LMS weight coefficients are to Δ N_d' (n) carries out the estimation estimating to obtain shaking angular movement increment

$\mathrm{\Δ}{\hat{N}}_d=h_{11}\left(n\right)\×{\mathrm{\ΔN}}_{d2}\left(n\right)+h_{12}\left(n\right)\×{\mathrm{\ΔN}}_{d1}\left(n\right)+h_{21}\left(n\right)\×{\mathrm{\ΔN}}_{d4}\left(n\right)+h_{22}\left(n\right)\×{\mathrm{\ΔN}}_{d3}\left(n\right)$

Wherein h₁₁(n)、h₁₂(n)、h₂₁(n)、h₂₂N () is the weight coefficient in the n-th sampling period.

Advantages of the present invention and good effect be:

1, the present invention obtains dither reference signal by gyroscope, it is no longer necessary to outside shake pick-up circuit obtains shake ginseng Examine signal, simplify and shake the hardware designs offset, and improve the accuracy of the dither signal of acquisition, thus improve shake The effect offset.

2, after the present invention uses gyro to obtain dither reference signal, during by design low order IIR high pass filter and numeral Clock transducer, obtains and the clock signal of same frequency Tong Bu with dither signal, and the chattering frequency for different gyros is different, can make Sample frequency is the integral multiple of dither signal frequency, shake offset need phase 90 ° just, cosine shake with reference to In FPGA, phase shift obtains, and can improve shake neutralization effect.

3, the iir filter that the present invention uses is IIR filter, there is feedback element, pinpoints in FPGA Cannot ensure full precision cut position during computing, the low order FIR filter that can realize full precision fixed-point calculation in FPGA is defeated to gyro The pulse increment gone out carries out pre-filtering, reduces the amplitude entering high-pass IIR filter pulse increment signal, thus reduces data and exist Expression figure place in FPGA, improves the figure place of significance bit in the case of identical data figure place, improves shake and offsets precision.

4, the dither reference signal of the present invention obtains through high-pass IIR filter, compares the logical iir filter of band and obtains shake ginseng Examine the mode of signal, the exponent number of iir filter can be substantially reduced, thus solve High Order IIR Filter for Fix-Point and easily dissipate, cause shake to be supported Disappear the problem lost, simultaneously the available high-frequency interferencing signal relevant to gyro output medium-high frequency interference, according to adaptive disturbance Principle of cancellation, offsets part High-frequency Interference when shake is offset, and is beneficial to improve shake neutralization effect.

5, the present invention uses three digital phase shifter modules, available four tunnel dither reference signal, by four LMS power is Several estimating shaking interference, signal is estimated in available more accurate shake, improves shake neutralization effect further.

6, the present invention is reasonable in design, and it uses adaptive interference cancelling technology based on LMS algorithm to carry out laser gyro and trembles Dynamic signal cancellation, it is the integral multiple of chattering frequency that sample frequency is offset in shake, offset use just, cosine reference signal is to by one Road sampled differential signal obtains through signal phase shift, constant during system input and output response approximation, thus reduces and tremble counteracting except the time Time delay, can be widely used for eliminating the mechanical shaking signal in two frequency mechanical shaking lasergyro outputs in real time.

Accompanying drawing explanation

Fig. 1 is the circuit theory diagrams of the present invention.

Detailed description of the invention

Below in conjunction with accompanying drawing, the embodiment of the present invention is further described:

A kind of two frequency machine shaking laser gyroscope shaking signal cancellation devices, as it is shown in figure 1, include gyro data acquisition module 1, Calculus of differences module 2, low order FIR filter module 3, summation operation module 4, high frequency clock module 5, low order iir filter mould Block 6, IIR high pass filter block the 7, first digital phase shifter module 8, shake estimation module 9, digital dock conversion module 10, the second digital phase shifter module 11, band filter module 12, digital dock management module the 13, the 3rd digital phase shifter mould Block 14, LMS weight coefficient more new module 15.Described gyro data acquisition module 1, low order iir filter module 6, digital dock turn Parallel operation module 10, digital dock management module 13 is sequentially connected and connects, and another input of low order iir filter module 6 connects High frequency clock module 5, described gyro data acquisition module 1 exports the employing frequency of 4M frequency multiplication according to digital dock management module 13 Sampling, sampled data exports to low order FIR filter module 3, low order FIR filter module 3 through calculus of differences module 2 Outfan connects IIR high pass filter block 7 and summation operation module 4, the outfan of this IIR high pass filter block 7 respectively Connect the first digital phase shifter module 8 respectively, shake estimation module the 9, second digital phase shifter module 11 and LMS weight coefficient updates Module 15, the outfan of this second digital phase shifter module 11 connects shake estimation module the 9, the 3rd digital phase shifter module respectively 14 and LMS weight coefficients more new module 15, the outfan of this first digital phase shifter module 8 connects shake estimation module 9 He respectively LMS weight coefficient more new module 15, the outfan of the 3rd digital phase shifter module 14 connects shake estimation module 9 and LMS respectively Weight coefficient more new module 15, the outfan of shake estimation module 9 connects another input of summation operation module 4, summation operation The outfan of module 4 connects the input of band filter module 12, and this band filter module 12 outfan connects LMS power and is Number more new module 15, this LMS weight coefficient more new module 15 exports LMS algorithm and updates weight coefficient to shake estimation module 9.

The operation principle of the present invention is: count pulse is sampled in laser gyro output, obtain comprising gyro pedestal after difference Angular movement and pulse increment Δ N (n) of mechanical shaking angular movement；Pulse increment Δ N (n) is through low order FIR low pass filter 3 pre-flock Obtain after ripple shaking the gyro base corner motion of pre-decay and the mechanical shaking angular movement pulse increment Δ N of decay_lpf(n)；Machinery Shake angular movement pulse increment Δ N_lpfN () obtains the reference jitter increment signal after pre-decay after high-pass IIR filter 7 ΔN_d1(n)；Reference jitter increment signal Δ N after pre-decay_d1N () obtains and Δ N after digital phase shifter 11_d1N () differs The reference jitter increment signal Δ N of 90 °_d3(n), the reference jitter increment signal Δ N after pre-decay_d1N () is through digital phase shifter 8 After obtain reference jitter increment signal Δ N_d2(n), reference jitter increment signal Δ N_d3N () is joined after digital phase shifter 14 Examine shake increment signal Δ N_d4(n)；Shake estimation module 9 is according to four tunnel dither reference signal Δ N_d1(n)、ΔN_d2(n)、ΔN_d3 (n)、ΔN_d4N () obtains shaking the estimation of angular movement incrementAnd with Δ N_lpfN () is subtracted each other and is obtained inertia angular movement increment Δs(n)；Inertia angular movement increment Delta s (n) passes through LMS weight coefficient more new module 15 and by optimizing after band filter 12 LMS algorithm update weight coefficient h₁₁(n)、h₁₂(n)、h₂₁(n)、h₂₂(n).Laser gyro output counting pulse signal N is through low order Dither signal N is obtained after IIR high pass filter 6_d(n), dither signal N_dN () obtains and shake through digital dock transducer 10 Signal synchronizes and the tune width ripple signal f of same frequency_s, adjust wide ripple signal f_sFrequency signal digital dock manager in FPGA 13 carry out 4M frequency multiplication after obtain the sampling clock Tong Bu with dither reference signal to gyro output signals sample.ΔN_d1N () is through letter Obtain and Δ N after M single sampling delay_d1N () differs the increment signal Δ N of 90 °_d2N (), the shake so obtained is with reference to letter Number Δ N_d1(n)、ΔN_d2N (), close to preferable cosine and sine signal, inputs Δ N_lpfN () offsets the response of output Δ s (n) to shake When being approximately constant, comparatively ideal shake neutralization effect can be obtained.

Below some key points in the present invention are designed and illustrate:

(1) sampling clock design

Gyro output obtains dither reference signal N through low order IIR high pass filter 6_d(n), dither reference signal N_d(n) Can obtain and dither signal Δ N through digital dock conversion module 10_dN () synchronizes and clock signal f of same frequency_s:

y₁(n)=b_l10×(N(n)+b_l11×N(n-1)+b_l12×N(n-2))+a_l11×y₁(n-1)+a_l12×y₁(n-2)

N_d(n)=y₁(n)+b_l21×y₁(n-1)+b_l22×y₁(n-2)+a_l21×N_d(n-1)+a_l22×N_d(n-2)

f_sObtain Tong Bu with dither reference signal after carrying out 4M frequency multiplication by the digital dock management module 13 in FPGA Sampling clock, sample clock frequency is 4Mf_s(M is the integer more than or equal to 1, f_sFor gyro chattering frequency).Shake canceller With 4Mf_sSample frequency to gyro output signals sample.

(2) design of low pass FIR prefilter

Laser gyro output count pulse is obtained by gyro data acquisition module 1, and its sampled value obtains through calculus of differences module 2 To pulse increment:

Δ N (n)=N (n)-N (n-1)=Δ N_b(n)+ΔN_d(n)

Wherein, Δ N_b(n)、ΔN_dN () represents the gyro base corner motion change in the kth sampling period and machinery respectively Shake angular movement change.

Gyro angular movement pulse increment Δ N (n) carries out pre-filtering through low order FIR filter, obtains declining through once shake The angular movement pulse increment Δ N subtracted_lpf(n), Δ N_lpfN desired signal that () offsets as shake:

ΔN_lpf(n)=h₀×ΔN(n)+h₁×ΔN(n-1)+h₂×ΔN(n-2)+…+h_K-1×ΔN(n-K+1)+h_K× Δ N (n-K)=Δ N_b′(n)+ΔN_d′(n)

Wherein, Δ N_b' (n) represents the gyro base corner increment of motion through FIR filter, Δ N_d' (n) represents through declining The mechanical shaking angular movement increment signal subtracted, K is the exponent number representing FIR filter, designs difference according to the actual demand of system The FIR filter of exponent number, the preferred recommendation of K may be selected to be K=35.

(3) acquisition of dither reference signal

Dither reference signal Δ N_d1N () is given by high pass filter 7.ΔN_lpfN () filters gyro base through high pass filter 7 Seat angular movement increment Delta N_b' (n), obtains gyro mechanical shaking angular movement increment Delta N that a road is more satisfactory_d1(n):

order₁(n)=b₁₀×(ΔN_lpf(n)+b₁₁×ΔN_lpf(n-1)+b₁₂×ΔN_lpf(n-2))+a₁₁×order₁ (n-1)+a₁₂×order₁(n-2)

order₂(n)=order₁(n)+b₂₁×order₁(n-1)+b₂₂×order₁(n-2)+a₂₁×order₂(n-1)+ a₂₂×order₂(n-2)

order₃(n)=order₂(n)+b₃₁×order₂(n-1)+b₃₂×order₂(n-2)+a₃₁×order₃(n-1)+ a₃₂×order₃(n-2)

order₄(n)=order₃(n)+b₄₁×order₃(n-1)+b₄₂×order₃(n-2)+a₄₁×order₄(n-1)+ a₄₂×order₄(n-2)

ΔN_d1(n)=order₄(n)+b₅₁×order₄(n-1)+b₅₂×order₄(n-2)+a₅₁×ΔN_d1(n-1)+a₅₂ ×ΔN_d1(n-2)

ΔN_d1N () obtains the second tunnel dither reference signal Δ N after digital phase shift module 8 phase shift_d2(n):

ΔN_d2(n)=Δ N_d1(n-1)

ΔN_d1N () obtains and Δ N after digital phase shift module 11 phase shift_d1The 3rd tunnel shake ginseng of (n) phase 90 ° Examine signal delta N_d3(n):

ΔN_d3(n)=Δ N_d1(n-M)

Owing to sample frequency is 4Mf_s, the most no matter gyro chattering frequency f_sHow to change, Δ N_d1(n)、ΔN_d3Between (n) Phase place differ 90 ° all the time, correspondingly shake offset be input to shake offset output response approximation time constant, shake can be improved Neutralization effect.

ΔN_d3N () obtains the 4th tunnel dither reference signal Δ N after digital phase shift module 14 phase shift_d4(n):

ΔN_d4(n)=Δ N_d3(n-1)

(4) design of algorithm for estimating is shaken

Shake estimation module 9 passes through four tunnel dither reference signal and four LMS weight coefficients to Δ N_d' (n) estimates Obtain

$\mathrm{\Δ}{\hat{N}}_d=h_{11}\left(n\right)\×{\mathrm{\ΔN}}_{d2}\left(n\right)+h_{12}\left(n\right)\×{\mathrm{\ΔN}}_{d1}\left(n\right)+h_{21}\left(n\right)\×{\mathrm{\ΔN}}_{d4}\left(n\right)+h_{22}\left(n\right)\×{\mathrm{\ΔN}}_{d3}\left(n\right)$

Wherein h₁₁(n)、h₁₂(n)、h₂₁(n)、h₂₂N () is the weight coefficient in the n-th sampling period, they are updated by weight coefficient Algoritic module 15 determines.

(5) the weight coefficient update algorithm design of LMS

The weight coefficient update algorithm of LMS passes through four tunnel dither reference signal Δ N_d1(n)、ΔN_d2(n)、ΔN_d3(n)、ΔN_d4 N ε (n) that the output of () and band filter obtains is updated, and its concrete update algorithm is as follows:

h₁₁(n+1)=h₁₁(n)+2×μ×ε(n)×ΔN_d2(n)

h₁₂(n+1)=h₁₂(n)+2×μ×ε(n)×ΔN_d1(n)

h₂₁(n+1)=h₂₁(n)+2×μ×ε(n)×ΔN_d4(n)

h₂₂(n+1)=h₂₂(n)+2×μ×ε(n)×ΔN_d3(n)

(6) gyro base corner increment of motion

Increment Delta N_lpfN () is estimated with shakeObtain shake through summation operation module 4 and offset output Δ s (n), Δ s N () is gyro base corner increment of motion, the gyro output useful signal that i.e. filters solutions is to be obtained, shake is offset output s (n) and is Gyro base corner increment of motion Δ N_bThe approximation of (n).Shake cancellation algorithms is as follows:

$\mathrm{\Δ}s\left(n\right)={\mathrm{\ΔN}}_{lpf}\left(n\right)-\mathrm{\Δ}{\hat{N}}_d\left(n\right)$

(7) design of band filter

Updating and shake neutralization effect for preventing low frequency or the motion of constant gyro base corner from affecting weight coefficient, shake is offset Output s (n) will be used further to weight coefficient after band filter 12 eliminates the motion of gyro base corner and update.Band filter algorithm is such as Under:

ε (n)=-Δ s (n)+Δ s (n-1)+Δ s (n-2)-Δ s (n-3)

8, complete shake cancellation algorithms

The shake cancellation algorithms that the present invention is complete is:

y₁(n)=b_l10×(N(n)+b_l11×N(n-1)+b_l12×N(n-2))+a_l11×y₁(n-1)+a_l12×y₁(n-2)

N_d(n)=y₁(n)+b_l21×y₁(n-1)+b_l22×y₁(n-2)+a_l21×N_d(n-1)+a_l22×N_d(n-2)

ΔN_lpf(n)=h₀×ΔN(n)+h₁×ΔN(n-1)+h₂×ΔN(n-2)+…+h_K-1×ΔN(n-K+1)+h_K× ΔN(n-K)

order₁(n)=b₁₀×(ΔN_lpf(n)+b₁₁×ΔN_lpf(n-1)+b₁₂×ΔN_lpf(n-2))+a₁₁×order₁ (n-1)+a₁₂×order₁(n-2)

order₂(n)=order₁(n)+b₂₁×order₁(n-1)+b₂₂×order₁(n-2)+a₂₁×order₂(n-1)+ a₂₂×order₂(n-2)

order₃(n)=order₂(n)+b₃₁×order₂(n-1)+b₃₂×order₂(n-2)+a₃₁×order₃(n-1)+ a₃₂×order₃(n-2)

order₄(n)=order₃(n)+b₄₁×order₃(n-1)+b₄₂×order₃(n-2)+a₄₁×order₄(n-1)+ a₄₂×order₄(n-2)

ΔN_d1(n)=order₄(n)+b₅₁×order₄(n-1)+b₅₂×order₄(n-2)+a₅₁×ΔN_d1(n-1)+a₅₂ ×ΔN_d1(n-2)

ΔN_d2(n)=Δ N_d1(n-1)

ΔN_d3(n)=Δ N_d1(n-M)

ΔN_d4(n)=Δ N_d3(n-1)

$\mathrm{\Δ}{\hat{N}}_d=h_{11}\left(n\right)\×{\mathrm{\ΔN}}_{d2}\left(n\right)+h_{12}\left(n\right)\×{\mathrm{\ΔN}}_{d1}\left(n\right)+h_{21}\left(n\right)\×{\mathrm{\ΔN}}_{d4}\left(n\right)+h_{22}\left(n\right)\×{\mathrm{\ΔN}}_{d3}\left(n\right)$

$\mathrm{\Δ}s\left(n\right)={\mathrm{\ΔN}}_{lpf}\left(n\right)-\mathrm{\Δ}{\hat{N}}_d\left(n\right)$

ε (n)=-Δ s (n)+Δ s (n-1)+Δ s (n-2)-Δ s (n-3)

h₁₁(n+1)=h₁₁(n)+2×μ×ε(n)×ΔN_d2(n)

h₁₂(n+1)=h₁₂(n)+2×μ×ε(n)×ΔN_d1(n)

h₂₁(n+1)=h₂₁(n)+2×μ×ε(n)×ΔN_d4(n)

h₂₂(n+1)=h₂₂(n)+2×μ×ε(n)×ΔN_d3(n)

It is emphasized that embodiment of the present invention is illustrative rather than determinate, bag the most of the present invention Include the embodiment being not limited to described in detailed description of the invention, every by those skilled in the art according to technical scheme Other embodiments drawn, also belong to the scope of protection of the invention.

Claims (6)

1. a frequency machine shaking laser gyroscope shaking signal cancellation device, it is characterised in that: include gyro data acquisition module, difference Divide computing module, low order FIR filter module, summation operation module, high frequency clock module, low order iir filter module, IIR High pass filter block, the first digital phase shifter module, shake estimation module, digital dock conversion module, the second numeral are moved Phase device module, band filter module, digital dock management module, the 3rd digital phase shifter module and LMS weight coefficient update mould Block；Described gyro data acquisition module, low order iir filter module, digital dock conversion module, digital dock management module Being sequentially connected and connect, another input of low order iir filter module connects high frequency clock module, and described gyro data gathers mould Dither signal is sampled by block according to the employing frequency of digital dock management module output, and sampled data is through calculus of differences module Output to low order FIR filter module, low order FIR filter module outfan connect respectively IIR high pass filter block and Summation operation module, the outfan of this IIR high pass filter block connects the first digital phase shifter module respectively, mould is estimated in shake Block, the second digital phase shifter module and LMS weight coefficient more new module, the outfan of this first digital phase shifter module connects respectively Shake estimation module and LMS weight coefficient more new module, the outfan of this second digital phase shifter module connects shake respectively and estimates Module, the 3rd digital phase shifter module and LMS weight coefficient more new module, the outfan of the 3rd digital phase shifter module connects respectively Connecing shake estimation module and LMS weight coefficient more new module, the outfan of this shake estimation module connects the another of summation operation module One input, the outfan of this summation operation module connects the input of band filter module, and this band filter module is even Connecing LMS weight coefficient more new module, this LMS weight coefficient more new module output LMS algorithm updates weight coefficient to shaking estimation module.

Two frequency machine shaking laser gyroscope shaking signal cancellation devices the most according to claim 1, it is characterised in that: described numeral Clock converter module is arranged in FPGA, and the sample clock frequency of its output is 4Mf_s, wherein, M is whole more than or equal to 1 Number, f_sFor gyro chattering frequency.

Two frequency machine shaking laser gyroscope shaking signal cancellation devices the most according to claim 1, it is characterised in that: described difference Pulse increment Δ N (the n)=Δ N of computing module output_b(n)+ΔN_d(n), wherein, Δ N_b(n)、ΔN_dN () represents kth respectively Gyro base corner motion change in the individual sampling period and mechanical shaking angular movement change；After described low order FIR filter processes Angular movement pulse increment Δ N_lpf(n) be:

ΔN_lpf(n)=Δ N '_b(n)+ΔN′_d(n):

Wherein, Δ N '_bN () represents the gyro base corner increment of motion through low order FIR filter, Δ N '_dN () represents through declining The mechanical shaking angular movement increment signal subtracted, K is the exponent number representing FIR filter.

Two frequency machine shaking laser gyroscope shaking signal cancellation devices the most according to claim 1, it is characterised in that: described IIR High pass filter filters the gyro base corner increment of motion Δ N ' of low order FIR filter output_bThe first shake reference is obtained after (n) Signal delta N_d1(n): described first digital phase shift module is to Δ N_d1N () obtains the second tunnel dither reference signal Δ N after processing_d2 (n), Δ N_d2(n)=Δ N_d1(n-1)；Described second digital phase shift module is to Δ N_d1N () obtains and Δ N after processing_d1(n) phase place Differ the 3rd tunnel dither reference signal Δ N of 90 °_d3(n), this Δ N_d3(n)=Δ N_d1(n-M), M is the integer more than or equal to 1； Described 3rd digital phase shift module is to the 3rd tunnel dither reference signal Δ N_d3N () obtains the 4th tunnel dither reference signal Δ after processing N_d4(n), Δ N_d4(n)=Δ N_d3(n-1)。

Two frequency machine shaking laser gyroscope shaking signal cancellation devices the most according to claim 1, it is characterised in that: described LMS Weight coefficient more new module is according to four tunnel dither reference signal Δ N_d1(n)、ΔN_d2(n)、ΔN_d3(n)、ΔN_d4The logical filter of (n) and band Ripple device output valve ε (n) output LMS algorithm renewal weight coefficient:

h₁₁(n+1)=h₁₁(n)+2×μ×ε(n)×ΔN_d2(n)

h₁₂(n+1)=h₁₂(n)+2×μ×ε(n)×ΔN_d1(n)

h₂₁(n+1)=h₂₁(n)+2×μ×ε(n)×ΔN_d4(n)

h₂₂(n+1)=h₂₂(n)+2×μ×ε(n)×ΔN_d3(n)。

Two frequency machine shaking laser gyroscope shaking signal cancellation devices the most according to claim 1, it is characterised in that: described shake Estimation module is according to four tunnel dither reference signal Δ N_d1(n)、ΔN_d2(n)、ΔN_d3(n)、ΔN_d4(n) and four LMS weight coefficients To Δ N '_dN () carries out the estimation estimating to obtain shaking angular movement increment

$\mathrm{\Δ}{\hat{N}}_d=h_{11}\left(n\right)\×{\mathrm{\ΔN}}_{d2}\left(n\right)+h_{12}\left(n\right)\×{\mathrm{\ΔN}}_{d1}\left(n\right)+h_{21}\left(n\right)\×{\mathrm{\ΔN}}_{d4}\left(n\right)+h_{22}\left(n\right)\×{\mathrm{\ΔN}}_{d3}\left(n\right)$

Wherein h₁₁(n)、h₁₂(n)、h₂₁(n)、h₂₂N () is the weight coefficient in the n-th sampling period.