CN102819029A - Supercompact combination satellite navigation receiver - Google Patents

Abstract

The invention discloses a supercompact combination satellite navigation receiver which comprises a radio frequency front end, a signal preprocessor, a base band prefilter, a navigation filter, a micro inertial navigation module, a pseudo-range and pseudo-range rate calculating module and an ephemeris resolving and start selection calculating module. A self-adaptive filter is arranged between the base band prefilter and the navigation filter. An output of the base band prefilter is used as an input to be switched in the self-adaptive filter and then an output of the self-adaptive filter is used as an observed quantity to be input into the navigation filter to carry out data fusion, and thus, the stability and the tracking accuracy of a tracking loop can be obviously improved, so that the adaption capacity of GNSS (Global Navigation Satellite System) positioning on the environment and the motion state is improved. Furthermore, the vector tracking loop is formed by the signal proprocessor, the base band prefilter, the self-adaptive filter, the navigation filter and the pseudo-range and pseudo-range rate calculating module; and the navigation filter adopts a federated filter, so that the stability and the tracking accuracy of the tracking loop can be improved, and thus, the performances of reliability, positioning accuracy and the like of the GNSS positioning are improved.

Description

A kind of hypercompact combined satellite navigation receiver

Technical field

The present invention relates to the satellite navigation technical field, relate in particular to the satellite navigation receiver of a kind of Global Navigation Satellite System (GNSS) and the hypercompact combination of little inertial navigation system (MEMS-IMU).

Background technology

Global Navigation Satellite System (GNSS) be a kind of be the radio navigation system on basis with the satellite, round-the-clock, uninterrupted, high precision, real-time navigation positioning service can be provided for all kinds of carriers of land, sea, air.At present; Using the widest GLONASS is GPS of USA; Be penetrated into the every field of national economy and daily life, like sail, urban traffic control, commercial logistics management, the navigation of boats and ships ocean, when precision receives, geodetic surveying, precision agriculture etc.At present, the construction of the Galileo system that China has also participated in building up in nearly 2 years, and just at independent research Global Positioning System (GPS) Compass (two generations of the Big Dipper), this system provides positioning service in China and surrounding area thereof the end of last year.Therefore, the multimode integrated navigation technology of systems such as research Compass and GPS, Galileo will become the research emphasis in domestic following a period of time.

It is quite faint when gps signal (spread-spectrum signal) arrives ground receiver; Be approximately-130dBmW; Than the low 20～30dB of receiver internal thermal noise, special, in complex environments (being referred to as indoor environment among this paper) such as indoor, city, forest; The GPS received signal to noise ratio is lower, and indoor environment just is one of main environment of mankind's activity.

The GNSS combination receiver can make full use of the satellite-signal resource of each GNSS system, thereby can improve the continuity of system availability and location output.IMU (micro electro mechanical inertia is measured assembly MEMS-IMU, is designated hereinafter simply as IMU) for the autonomous type Position Fixing Navigation System, do not receive the influence of environment basically, and its dynamic property is better.Therefore,, can address the above problem, improve locating accuracy, location continuity and service guarantee ability GNSS combination receiver and IMU system in combination.

GNSS/IMU hypercompact combination navigation Study on Theory is necessary, is one of focus of studying both at home and abroad at present.At present; Though pertinent literature research GPS/IMU (inertial navigation) hypercompact combination navigation mode has been arranged both at home and abroad; Done certain research work at aspects such as combined system implementation, modelings; But hypercompact combination research still exists many problems to further investigate and to solve, and especially based on the hypercompact combination navigation mode of vector tracking, aspects such as the reliability of its work still face many problems and press for solution.In order to improve combined navigation system performance; Can be applied in the higher field reliability requirement in civil aviaton, railway traffic etc.; This patent will provide a kind of GNSS/IMU hypercompact combination navigation vector tracking receiver base-band information fusion treatment scheme, solve above-mentioned key scientific problems.

Summary of the invention

The objective of the invention is in order to overcome the low satellite-signal losing lock that caused with disturbing easily of existing GNSS receiver sensitivity; Thereby can not provide the drawback of positioning result; Through design with the hypercompact assembled scheme of little inertial navigation, improve the vector tracking scheme, take measures to improve GNSS combination receiver tracking accuracy and stability, improve performances such as GNSS reliability of positioning and bearing accuracy.

The invention provides a kind of hypercompact combination navigation DVB, comprise radio-frequency front-end, signal preprocessor; Base band prefilter and Navigation Filter; And little inertial navigation module, pseudorange and pseudorange rates computing module, ephemeris resolves and selects the star computing module; Wherein signal pre-processing module comprises correlator and local signal maker, is provided with sef-adapting filter between said base band prefilter and the Navigation Filter.

Said local signal maker, correlator, the base band prefilter, sef-adapting filter and Navigation Filter, pseudorange and pseudorange rates computing module constitute the vector tracking loop.Pseudorange and pseudorange rates computing module receive information such as clocking error and the GNSS ephemeris parameter that ephemeris resolved and selected the star computing module of output information, the Navigation Filter output of little inertial navigation mechanics layout module; Generate and postpone the adjustment signal; Input local signal maker; Adjust local pseudo-code and carrier wave, this programme can improve track loop stability, can improve the track loop precision again.

Said Navigation Filter is federal wave filter, comprises senior filter and at least one subfilter.The subfilter of said federal wave filter comprises the GNSS wave filter under the pure GNSS pattern, the little inertial navigation wave filter of GNSS/IMU integrated navigation wave filter under the hypercompact integrated mode and pure IMU.As Navigation Filter, introduce the vector tracking loop with federal wave filter, locating information also as the input of vector tracking, is helped the more accurate reinsertion of carrier and pseudo-code.

Further; Said hypercompact combined satellite navigation receiver comprises at least one road navigation signal receiving cable; Each navigation signal receiving cable all is provided with signal preprocessor and base band prefilter, and each roadbed band prefilter is all through inserting Navigation Filter behind the sef-adapting filter.

The state equation of described sef-adapting filter is:

$\frac{d}{\mathrm{dt}}\left[\begin{array}{c}\mathrm{\&delta;\&rho;}\\ \mathrm{\&delta;}\stackrel{\&CenterDot;}{\mathrm{\&rho;}}\\ \mathrm{\&delta;}\stackrel{\&CenterDot;\&CenterDot;}{\mathrm{\&rho;}}\\ \mathrm{\&delta;N}\\ {\mathrm{\&Delta;}}_{\mathrm{ion}}\\ {\stackrel{\&CenterDot;}{\mathrm{\&Delta;}}}_{\mathrm{ion}}\end{array}\right]=\left[\begin{array}{c}\mathrm{\&delta;}\stackrel{\&CenterDot;}{\mathrm{\&rho;}}\\ \mathrm{\&delta;}\stackrel{\&CenterDot;\&CenterDot;}{\mathrm{\&rho;}}\\ 0\\ 0\\ {\stackrel{\&CenterDot;}{\mathrm{\&Delta;}}}_{\mathrm{<figure-callout\; id="0"\; label="ion"\; filenames="HDA00001969162500022.png"\; state="\{\{state\}\}">ion</figure-callout>}}\\ 0\end{array}\right]+\left[\begin{array}{c}0\\ 0\\ w_{\mathrm{<figure-callout\; id="0"\; label="a"\; filenames="HDA00001969162500022.png"\; state="\{\{state\}\}">a</figure-callout>}}\\ 0\\ 0\\ w_{\mathrm{ion}}\end{array}\right]$

Observation equation is:

$\left[\begin{array}{c}\mathrm{\&delta;\&rho;}\\ \mathrm{\&delta;}\stackrel{\&CenterDot;}{\mathrm{\&rho;}}\\ \mathrm{\&delta;}\stackrel{\&CenterDot;\&CenterDot;}{\mathrm{\&rho;}}\\ \mathrm{\&delta;N}\\ {\mathrm{\&Delta;}}_{\mathrm{ion}}\\ {\stackrel{\&CenterDot;}{\mathrm{\&Delta;}}}_{\mathrm{ion}}\end{array}\right]=\left[\begin{array}{c}\mathrm{\&delta;\&rho;}\\ \mathrm{\&delta;}\stackrel{\&CenterDot;}{\mathrm{\&rho;}}\\ \mathrm{\&delta;}\stackrel{\&CenterDot;\&CenterDot;}{\mathrm{\&rho;}}\\ \mathrm{\&delta;N}\\ {\mathrm{\&Delta;}}_{\mathrm{ion}}\\ {\stackrel{\&CenterDot;}{\mathrm{\&Delta;}}}_{\mathrm{ion}}\end{array}\right]+\left[\begin{array}{c}{w}_{\mathrm{\&rho;}}\\ w_{\stackrel{\&CenterDot;}{\mathrm{\&rho;}}}\\ w_{\stackrel{\&CenterDot;\&CenterDot;}{\mathrm{\&rho;}}}\\ w_N\\ w_i\\ w_{\mathrm{io}}\end{array}\right]$

Wherein δ ρ is the pseudorange residual error,

Be the pseudorange rates residual error, Be pseudorange acceleration residual error, δ N is a carrier phase complete cycle residual error, Δ _IonBe the ionosphere time-delay,

Be ionosphere time-delay rate, w _aBe pseudorange acceleration process noise, w _IonBe ionospheric noise, w _ρ,

w _N, W _i, w _IoBe respectively the observation noise of pseudorange, pseudorange rates, pseudorange acceleration, ambiguity of carrier phase, ionosphere time-delay, ionosphere time-delay rate.

Further, said sef-adapting filter is the fuzzy logic sef-adapting filter, comprises fuzzy logic algorithm module and Kalman filter.Said fuzzy logic algorithm module receives pseudorange filtering and newly ceases mean value and mean square deviation, is output as to drive the noise variance weighting coefficient, and the parameter value of the driving noise variance through the said Kalman filter of real-time adjustment is realized auto adapted filtering.Said fuzzy logic algorithm module comprises:

The degree of membership simulation unit is used for confirming that the pseudorange filtering of input newly ceases the membership function of mean value and mean square deviation, and the membership function of the driving noise variance weighting coefficient of output;

The fuzzy inference rule unit is used to be provided with fuzzy inference rule;

The noise variance weighted units is used for calculating driving noise variance weighting coefficient according to said membership function and fuzzy inference rule ambiguity solution.

Fuzzy logic algorithm module of the present invention is according to simulation analysis and actual navigation data analysis to the particular navigation application system; Confirm the fuzzy division and the membership function of fuzzy logic algorithm module I/O amount; The funtcional relationship of I/O amount; And fuzzy logic inference rule; The pseudorange filtering of calculating input newly ceases the membership function value of mean value and mean square deviation, and the membership function value that drives the noise variance weighting coefficient, and utilizes fuzzy logic ordination to calculate driving variance weighting coefficient.Adopt above-mentioned fuzzy logic algorithm to adjust Kalman filter in real time and drive the noise variance weighting coefficient; The filter parameter that makes is adjusted along with the motor-driven situation of carrier; Improve Filtering Estimation precision and filter stability, thereby helped to improve the tracking accuracy and the stability of track loop.

Hypercompact combined satellite navigation receiver disclosed by the invention; Through adopting sef-adapting filter; And Navigation Filter adopts measures such as federal wave filter under the vector tracking loop structure; Improve the performances such as reliability and bearing accuracy of GNSS location, and, also can be stable when signal is blocked, exists certain environmental disturbances provide positioning result.

Description of drawings

Fig. 1 is the structural representation of the hypercompact navigational satellite receiver of the present invention;

Fig. 2 is a sef-adapting filter structural representation of the present invention;

Fig. 3 is the structural representation of sef-adapting filter fuzzy logic algorithm module of the present invention;

Fig. 4 is that the filtering of fuzzy logic algorithm input quantity newly ceases mean value membership function synoptic diagram;

Fig. 5 is that the filtering of fuzzy logic algorithm input quantity newly ceases mean square deviation membership function synoptic diagram;

Fig. 6 is a fuzzy logic algorithm output quantity membership function synoptic diagram;

Fig. 7 is federal filter construction synoptic diagram.

Embodiment

Below in conjunction with accompanying drawing and embodiment technical scheme of the present invention is explained further details, following examples do not constitute qualification of the present invention.

As shown in Figure 1 based on the hypercompact combined satellite navigation receiver of the GNSS/IMU of vector tracking; Comprise the radio-frequency front-end 1 that links to each other successively, signal preprocessor 2, base band prefilter 3; Sef-adapting filter 4 and Navigation Filter 5; And little inertial navigation module 6, pseudorange and pseudorange rates computing module 7, ephemeris resolves and selects star computing module 8.

Wherein signal preprocessor 2 comprises correlator 21 and local signal maker 22, and the i/q signal of correlator 21 accumulation outputs is as the observation input of base band prefilter 3, and local signal maker 22 generates local carrier and pseudo-code.

Wherein little inertial navigation module 6 comprises little inertial navigation mechanics layout module 61 and little inertial navigation IMU module 62; Carrier acceleration, the angular velocity information of little inertial navigation IMU62 module output; Be input to little cutter mechanics layout module 61 of being used to, resolve, obtain position, speed and the attitude information of carrier.Little cutter mechanics layout module 61 output terminals of being used to connect pseudorange and pseudorange rates computing module 7 on the one hand, connect navigator fix output terminal as a result on the one hand, output navigator fix result.

As shown in Figure 1, local signal maker 22, correlator 21, base band prefilter 3, sef-adapting filter 4 and Navigation Filter 5, pseudorange and pseudorange rates computing module 7 constitute the vector tracking loop.Pseudorange and pseudorange rates computing module 7 receive the output information of little inertial navigation mechanics layout module 61; The GNSS ephemeris parameter that the receiver clock error of Navigation Filter 5 outputs and ephemeris resolved and selected star computing module 8; Generate and postpone the adjustment signal; Be input to local signal maker 22, adjust local pseudo-code and carrier wave.

Behind the satellite-signal process correlator 21 from satellite RF front end 1; Output homophase (I), quadrature phase (Q) signal; Get into base band prefilter 3 and carry out pre-filtering; Be output as pseudorange error, pseudorange rates sum of errors pseudorange acceleration error, and ionosphere delay error, the ionosphere delay error rate of change information.The effect of base band prefilter 3 is to estimate tracking error and reduce the navigation calculating frequency (to be reduced to 1～10Hz) and calculated amount by 250～1000Hz.

Particularly; Base band prefilter 3 of the present invention receives the pseudo-code phase sum of errors carrier phase error signal of signal preprocessor 2 outputs; For each passage is set up prefilter, through the quantity of state of linear Kalman wave filter estimating system, the prefilter state equation is:

$\frac{d}{\mathrm{dt}}\left[\begin{array}{c}\mathrm{\&delta;\&rho;}\\ \mathrm{\&delta;}\stackrel{\&CenterDot;}{\mathrm{\&rho;}}\\ \mathrm{\&delta;}\stackrel{\&CenterDot;\&CenterDot;}{\mathrm{\&rho;}}\\ \mathrm{\&delta;N}\\ {\mathrm{\&Delta;}}_{\mathrm{ion}}\\ {\stackrel{\&CenterDot;}{\mathrm{\&Delta;}}}_{\mathrm{ion}}\end{array}\right]=\left[\begin{array}{c}\mathrm{\&delta;}\stackrel{\&CenterDot;}{\mathrm{\&rho;}}\\ \mathrm{\&delta;}\stackrel{\&CenterDot;\&CenterDot;}{\mathrm{\&rho;}}\\ 0\\ 0\\ {\stackrel{\&CenterDot;}{\mathrm{\&Delta;}}}_{\mathrm{<figure-callout\; id="0"\; label="ion"\; filenames="HDA00001969162500022.png"\; state="\{\{state\}\}">ion</figure-callout>}}\\ 0\end{array}\right]+\left[\begin{array}{c}0\\ 0\\ w_{\mathrm{<figure-callout\; id="0"\; label="a"\; filenames="HDA00001969162500022.png"\; state="\{\{state\}\}">a</figure-callout>}}\\ 0\\ 0\\ w_{\mathrm{ion}}\end{array}\right]$

Wherein, δ ρ is the pseudorange residual error,

Be the pseudorange rates residual error,

Be pseudorange acceleration residual error, δ N is a carrier phase complete cycle residual error, Δ _IonBe the ionosphere time-delay,

Be ionosphere time-delay rate, w _aBe pseudorange acceleration process noise, w _IonBe ionospheric noise.

The observation equation of prefilter is:

$\left[\begin{array}{c}\mathrm{\&Delta;\&rho;}\\ \mathrm{\&Delta;\&phi;}\end{array}\right]=\left[\begin{array}{c}\mathrm{\&delta;\&rho;}+{\mathrm{\&Delta;}}_{\mathrm{ion}}\\ \mathrm{\&delta;\&rho;}-{\mathrm{\&Delta;}}_{\mathrm{ion}}+\mathrm{\&delta;N}\end{array}\right]$

Wherein, Δ ρ is a pseudorange residual error measured value, and Δ φ is a carrier phase residual error measured value.

The present invention is not 3 outputs of base band prefilter to be directly inputted into Navigation Filter 5 carry out data fusion in order to improve the positioning performance of GNSS system under high current intelligence, but; Carry out auto adapted filtering earlier; 3 outputs of base band prefilter as input, are inserted sef-adapting filter 4, then; The output of auto adapted filtering 4 is inputed to Navigation Filter 5 as observed quantity carry out data fusion; The acute variation of 4 pairs of observation of sef-adapting filter input signal has adaptive ability, can significantly improve track loop stability, improves the adaptive faculty of GNSS location to environment and motion state sudden change thus.

Particularly, sef-adapting filter 4 state equations of each passage are:

$\frac{d}{\mathrm{dt}}\left[\begin{array}{c}\mathrm{\&delta;\&rho;}\\ \mathrm{\&delta;}\stackrel{\&CenterDot;}{\mathrm{\&rho;}}\\ \mathrm{\&delta;}\stackrel{\&CenterDot;\&CenterDot;}{\mathrm{\&rho;}}\\ \mathrm{\&delta;N}\\ {\mathrm{\&Delta;}}_{\mathrm{ion}}\\ {\stackrel{\&CenterDot;}{\mathrm{\&Delta;}}}_{\mathrm{ion}}\end{array}\right]=\left[\begin{array}{c}\mathrm{\&delta;}\stackrel{\&CenterDot;}{\mathrm{\&rho;}}\\ \mathrm{\&delta;}\stackrel{\&CenterDot;\&CenterDot;}{\mathrm{\&rho;}}\\ 0\\ 0\\ {\stackrel{\&CenterDot;}{\mathrm{\&Delta;}}}_{\mathrm{<figure-callout\; id="0"\; label="ion"\; filenames="HDA00001969162500022.png"\; state="\{\{state\}\}">ion</figure-callout>}}\\ 0\end{array}\right]+\left[\begin{array}{c}0\\ 0\\ w_{\mathrm{<figure-callout\; id="0"\; label="a"\; filenames="HDA00001969162500022.png"\; state="\{\{state\}\}">a</figure-callout>}}\\ 0\\ 0\\ w_{\mathrm{ion}}\end{array}\right]$

Wherein, δ ρ is the pseudorange residual error,

Be the pseudorange rates residual error,

Be pseudorange acceleration residual error, δ N is a carrier phase complete cycle residual error, Δ _IonBe the ionosphere time-delay,

Be ionosphere time-delay rate, w _aBe pseudorange acceleration process noise, w _IonBe ionospheric noise.

Observation equation is:

$\left[\begin{array}{c}\mathrm{\&delta;\&rho;}\\ \mathrm{\&delta;}\stackrel{\&CenterDot;}{\mathrm{\&rho;}}\\ \mathrm{\&delta;}\stackrel{\&CenterDot;\&CenterDot;}{\mathrm{\&rho;}}\\ \mathrm{\&delta;N}\\ {\mathrm{\&Delta;}}_{\mathrm{ion}}\\ {\stackrel{\&CenterDot;}{\mathrm{\&Delta;}}}_{\mathrm{ion}}\end{array}\right]=\left[\begin{array}{c}\mathrm{\&delta;\&rho;}\\ \mathrm{\&delta;}\stackrel{\&CenterDot;}{\mathrm{\&rho;}}\\ \mathrm{\&delta;}\stackrel{\&CenterDot;\&CenterDot;}{\mathrm{\&rho;}}\\ \mathrm{\&delta;N}\\ {\mathrm{\&Delta;}}_{\mathrm{ion}}\\ {\stackrel{\&CenterDot;}{\mathrm{\&Delta;}}}_{\mathrm{ion}}\end{array}\right]+\left[\begin{array}{c}{w}_{\mathrm{\&rho;}}\\ w_{\stackrel{\&CenterDot;}{\mathrm{\&rho;}}}\\ w_{\stackrel{\&CenterDot;\&CenterDot;}{\mathrm{\&rho;}}}\\ w_N\\ w_i\\ w_{\mathrm{io}}\end{array}\right]$

Wherein, w _ρ,

w _N, w _i, w _IoBe respectively the observation noise of parameters such as pseudorange, pseudorange rates, pseudorange acceleration, ambiguity of carrier phase, ionosphere time-delay, ionosphere time-delay rate.

Sef-adapting filter 4 of the present invention is the fuzzy logic sef-adapting filter; As shown in Figure 2; Comprise fuzzy logic algorithm module 41 and Kalman filter 42, wherein fuzzy logic algorithm module 41 be input as absolute value and the mean square deviation that pseudorange filtering newly ceases mean value, fuzzy logic algorithm module 41 is output as wave filter and drives the noise variance weighting coefficient; Through the parameter value of real-time adjustment Kalman filter 42 driving noise variances, realize auto adapted filtering.

A large amount of emulation that fuzzy logic algorithm module 41 is passed through system, and through summing up acquisition input fuzzy quantity membership function, fuzzy logic inference rule, finally the process fuzzy logic algorithm calculates and drives the noise variance weighting coefficient.As shown in Figure 3, fuzzy logic algorithm module 41 comprises degree of membership simulation unit 411, fuzzy inference rule unit 412 and noise variance weighted units 413.

Degree of membership simulation unit 411 obtains the membership function that fuzzy logic algorithm I/O domain is divided through simulation analysis; Be used for confirming that the pseudorange filtering of input newly ceases the membership function of mean value and mean square deviation, and the membership function of the driving noise variance weighting coefficient of output.

As shown in Figure 4, particularly, not have when motor-driven, filtering newly ceases the mean value absolute value less than 10 meters degree of membership no better than 1, and maximum near the degree of membership at 0 meter, along with the increase that filtering newly ceases the mean value absolute value, degree of membership reduces gradually.Therefore; The present invention is with [0; 10] rice is divided into an interval; The degree of membership of corresponding fuzzy set ' little '; Be expressed as

0 meter maximum; Along with the increase that filtering newly ceases the mean value absolute value, the degree of membership of fuzzy set ' little '

reduces gradually.

When filtering newly ceases the mean value absolute value greater than 60 meters, explain occurred certainly big motor-driven; When new breath mean value greater than 10 meters and during less than 60 meters, motor-driven greatly greater than 4g still might be taken place; Generation is during greater than 4g big motor-driven, and it is very little less than 10 meters degree of membership that filtering newly ceases the mean value absolute value.Based on this analysis conclusion; The present invention is divided into an interval with [10～150] rice; Field of definition as the degree of membership

of fuzzy set ' greatly '; In interval [10～60] rice; Along with the increase that filtering newly ceases the mean value absolute value, degree of membership

increases gradually; Its degree of membership when filtering newly ceases the mean value absolute value greater than 60 meters

is constantly equal to 1.

The filtering that acceleration and deceleration less than 4 g are motor-driven, circular flight is motor-driven, climb/dive when motor-driven newly ceases the mean value absolute value generally in interval [0～60] rice value; Analysis when flying based on certain fighting captain is navigated can be known it and add/deceleration Chang Keda 12 meter per seconds about 2 that simulation result shows that it is maximum that filtering this moment newly ceases near the degree of membership of mean value absolute value 10 meters, and it is more little to depart from 10 meters degrees of membership more.According to this analysis; The present invention is divided into an interval with [0～60] rice; The degree of membership of corresponding fuzzy set ' in ' when value equals 10 meters, belongs to degree of membership

maximum of ' less '; Along with departing from 10 meters, the degree of membership

that belongs to fuzzy set ' less ' reduces gradually.

Need to prove, among the present invention, adopt the approximate membership function that replaces each linguistic variable of simple function trigonometric function, though with physical presence a certain distance, can not produce considerable influence to operation result.

Similarly available another input fuzzy logic algorithms, filtering new information on the standard deviation of each domain fuzzy set membership

Figure 5;, and fuzzy logic algorithm outputs α membership

Figure 6.

Fuzzy inference rule unit 412 obtains having the fuzzy inference rule of completeness through simulation analysis.

Particularly, fuzzy inference rule expression filtering newly ceases mean value absolute value, mean square deviation and the corresponding relation of wave filter driving noise variance weighting coefficient on value.Fuzzy inference rule is made up of regular former piece (or be called ' condition ') and rule conclusion two parts, and filtering newly ceases mean value absolute value, mean square deviation rule of correspondence former piece, and wave filter drives noise variance weighting coefficient rule of correspondence conclusion part.

The fuzzy inference rule form is:

IF and then

Wherein, (p ₁, p ₂) ∈ (s, and m, b}, s, m, b}) }, k ∈ s, and m, b}, subscript s, m, b represent each fuzzy quantity respectively, relevant each fuzzy quantity is referring to Fig. 4-Fig. 6.The implication of above-mentioned rule is: if input mean value absolute value delta ρ ₁Belong to fuzzy set

(

Be its degree of membership), mean square deviation Δ ρ ₂Belong to fuzzy set

(

Be its degree of membership), so, output (variance Q's) adjustment alpha belongs to fuzzy set B ^(k)(

Be its degree of membership).The rule condition part be (

And

), conclusion part does This rule can be abbreviated as: (p ₁, p ₂) → k.

Illustrating of inference rule: can know with analyzing through emulation; If filtering newly ceases mean value absolute value " bigger ", mean square deviation " bigger than normal "; So, when the equivalent motor-driven generation that has acceleration greater than 2g for the previous period, need to increase wave filter and drive noise mean square deviation Q; Its adjustment alpha value is " little ", and it is following then can to get inference rule:

IF

and

then

is abbreviated as: (b, b) → s

Through simulation analysis, 9 fuzzy inference rules that obtain having completeness are following:

(s，s)→m (m，s)→s (b，s)→s (s，m)→m (m，m)→s (b，m)→s

(s，b)→b (m，b)→s (b，b)→s ?(1)

The fuzzy inference rule that conclusion is identical is merged into 1 rule, and the fuzzy inference rule after the merging is concentrated and to be had 3, and its conclusion part is respectively " little, in, big ".If the former piece degree of membership value of these 3 rules is respectively μ ^(s), μ ^(m), μ ^(b), according to 9 fuzzy inference rule collection, can get:

μ ^(s)＝μ _(m，s)+μ _(m，m)+μ _(m，b)+μ _(b，s)+μ _(b，m)+μ _(b，b)； (2)

μ ^(m)＝μ _(s，s)+μ _(s，m)μ ^(b)＝μ _(s，b)；

Wherein, μ _{(m, s)}Expression rule (m, s) → the former piece degree of membership of s, its computing formula is following:

${\mathrm{\&mu;}}_{(m,s)}={\mathrm{\&mu;}}_{A_1}^m\left({\mathrm{\&Delta;\&rho;}}_1\right)\&times;{\mathrm{\&mu;}}_{A_2}^s\left({\mathrm{\&Delta;\&rho;}}_2\right)---\left(3\right)$

Need to prove that the calculating of the degree of membership of Else Rule former piece is repeated no more at this with reference to following formula.

Noise variance weighted units 413 is used for calculating driving noise variance weighting coefficient α according to membership function and fuzzy inference rule ambiguity solution, and calculation expression is:

$\mathrm{\&alpha;}=\frac{{\mathrm{\&alpha;}}_0\&times;{\mathrm{\&mu;}}^{\left(s\right)}+{\mathrm{\&alpha;}}_1\&times;{\mathrm{\&mu;}}^{\left(m\right)}+{\mathrm{\&alpha;}}_2\&times;{\mathrm{\&mu;}}^{\left(b\right)}}{{\mathrm{\&mu;}}^{\left(s\right)}+{\mathrm{\&mu;}}^{\left(m\right)}+{\mathrm{\&mu;}}^{\left(b\right)}}---\left(4\right)$

α wherein ₀, α ₁, α ₂Being respectively fuzzy logic algorithm output degree of membership is little in fuzzy set, in, the value of corresponding α when degree of membership is maximum under big three kinds of situation; Correspond to Fig. 6; Be respectively 0.7,1 and 1.4, the wave filter that aforementioned calculation result is the final output of fuzzy logic algorithm drives noise variance weighting coefficient α.

An embodiment of fuzzy logic algorithm of the present invention, be input as (5,80) of establishing fuzzy logic algorithm, promptly the mean value absolute value equals 5 meters, and mean square deviation is 80 meters ², calculate and can get by formula (2):

${\mathrm{\&mu;}}^{\left(s\right)}={\mathrm{\&mu;}}_{(m,s)}+{\mathrm{\&mu;}}_{(m,m)}+{\mathrm{\&mu;}}_{(m,b)}+{\mathrm{\&mu;}}_{(b,s)}+{\mathrm{\&mu;}}_{(b,m)}+{\mathrm{\&mu;}}_{(b,b)}={\mathrm{\&mu;}}_{A_1}^m\left(5\right)\&times;{\mathrm{\&mu;}}_{A_2}^s\left(80\right)+{\mathrm{\&mu;}}_{A_1}^m\left(5\right)\&times;{\mathrm{\&mu;}}_{A_2}^m\left(80\right)$

$+{\mathrm{\&mu;}}_{A_1}^m\left(5\right)\&times;{\mathrm{\&mu;}}_{A_2}^b\left(80\right)+{\mathrm{\&mu;}}_{A_1}^b\left(5\right)\&times;{\mathrm{\&mu;}}_{A_2}^s\left(80\right)+{\mathrm{\&mu;}}_{A_1}^b\left(5\right)\&times;{\mathrm{\&mu;}}_{A_2}^m\left(80\right)+{\mathrm{\&mu;}}_{A_1}^b\left(5\right)\&times;{\mathrm{\&mu;}}_{A_2}^b\left(80\right)$

$=0.5\&times;0.79+0.5\&times;0.21+0.5\&times;0+0\&times;0.79+0\&times;0.21+0\&times;0=0.5$

${\mathrm{\&mu;}}^{\left(m\right)}={\mathrm{\&mu;}}_{(s,s)}+{\mathrm{\&mu;}}_{(s,m)}={\mathrm{\&mu;}}_{A_1}^s\left(5\right)\&times;{\mathrm{\&mu;}}_{A_2}^s\left(80\right)+{\mathrm{\&mu;}}_{A_1}^s\left(5\right)\&times;{\mathrm{\&mu;}}_{A_2}^m\left(80\right)$

$=0.5\&times;0.79+0.5\&times;0.21=0.5$

${\mathrm{\&mu;}}^{\left(b\right)}={\mathrm{\&mu;}}_{(s,b)}={\mathrm{\&mu;}}_{A_1}^s\left(5\right)\&times;{\mathrm{\&mu;}}_{A_2}^b\left(80\right)=0.5\&times;0=0$

Then, calculate driving noise variance weighting coefficient α according to formula (4):

$\mathrm{\&alpha;}=\frac{{\mathrm{\&alpha;}}_0\&times;{\mathrm{\&mu;}}^{\left(s\right)}+{\mathrm{\&alpha;}}_1\&times;{\mathrm{\&mu;}}^{\left(m\right)}+{\mathrm{\&alpha;}}_2\&times;{\mathrm{\&mu;}}^{\left(b\right)}}{{\mathrm{\&mu;}}^{\left(s\right)}+{\mathrm{\&mu;}}^{\left(m\right)}+{\mathrm{\&mu;}}^{\left(b\right)}}=\frac{0.7\&times;0.5+1\&times;0.5+1.4\&times;0}{0.5+0.5+0}=0.85$

The present invention adopts above-mentioned fuzzy logic algorithm to adjust Kalman filter in real time and drives the noise variance weighting coefficient; The filter parameter that makes is adjusted along with the motor-driven situation of carrier; Improve Filtering Estimation precision and filter stability, thereby helped to improve the tracking accuracy and the stability of track loop.

The centralized filtering of main both at home and abroad employing EKF is at present carried out modeling to the Navigation Filter of the hypercompact combined vectors track receiver of GNSS; The present invention proposes the Navigation Filter design proposal that adopts federal filtering theory design GNSS vector tracking receiver, improve Navigation Filter fault-tolerant ability and filtering accuracy.

Navigation Filter 5 adopts distributed federal filter structure among the present invention; General layout Plan is as shown in Figure 7; Comprise federal filtering subfilter and federal filtering senior filter; The output of each channel adaptive wave filter 4 is as the observation input of federal filtering subfilter, and federal senior filter carries out information distribution to the state of subfilter.The state estimation output terminal of each subfilter connects the observed quantity input end of federal senior filter, and federal senior filter information distribution coefficient output terminal feedback connects each subfilter estimation error variance.

Particularly, federal wave filter is a kind of two-stage data fusion structure, and subfilter and senior filter concurrent working are as shown in Figure 7, the output X of little inertial navigation system IMU among the figure _kDirectly give federal senior filter on the one hand, it can be defeated by each subfilter as observed reading on the other hand, and the output Z of each passage of GNSS only gives corresponding subfilter, the partial estimation of each subfilter

And covariance matrix P _iThe estimated value of sending into senior filter and senior filter merges to obtain global optimum together to be estimated

In addition, from Fig. 7, can also see, by subfilter and the synthetic overall estimated value of senior filter

And corresponding covariance matrix P _gBe enlarged into

After feed back to the estimated value of subfilter again with the replacement subfilter.

Senior filter comprises time updating block and optimum fusion unit, the prediction error of senior filter

And covariance P _mSend into the optimum fusion unit, simultaneously the senior filter prediction error Covariance P _mAlso can reset to overall covariance matrix

Doubly, be

Be used for the estimated value of senior filter of resetting.

Wherein, β _i(i=1,2 ..., N m) is called the information distribution coefficient, and N is the number of subfilter, and the method for employing dynamic optimal information distribution coefficient is confirmed the information distribution coefficient of each subfilter, different β _iValue can obtain the different structure and the different qualities of federal wave filter.Choosing the information distribution coefficient constantly at k is:

${\mathrm{\&beta;}}_{i\left(k\right)}=\frac{\mathrm{trace}\left[{\left(P_{\mathrm{ii}(k,k-1)}^{-1}\right)}^{*}\right]}{\mathrm{trace}\left[M_i{P}_{g(k,k-1)}^{-1}{M}_i^T\right]}$

Wherein, ${\left(P_{\mathrm{Ii}(k,k-1)}^{-1}\right)}^{*}={[{\mathrm{\&Phi;}}_{k,k-1}{P}_{\mathrm{Ii}(-1)}{\mathrm{\&Phi;}}_{k,k-1}^T+{\mathrm{\&Gamma;}}_{k-1}{Q}_{i(k-1)}{\mathrm{\&Gamma;}}_{K-1}^T]}^{-1},$ M _iBe subsystem state X _IkWith total system state X _kBetween the relational matrix that exists.Φ _{K, k-1}The state-transition matrix of inscribing during for k, Q _{I (k-1)}The driving noise variance matrix of inscribing when being i sub-filters (k-1), P _{Ii (k, k-1)}, P _{Ii (k-1)}Predicated error variance matrix of inscribing when being respectively i sub-filters (k-1) and estimation error variance matrix, Γ is for driving the noise figure matrix.

Federal filtering has 3 sub-filters among the present invention: the GNSS wave filter under the pure GNSS pattern, the GNSS/IMU integrated navigation wave filter under the hypercompact integrated mode, the little inertial navigation wave filter of pure IMU.

Wherein, GNSS wave filter under the pure GNSS pattern adopts the Singer model modeling, and the little inertial navigation wave filter of GNSS/IMU integrated navigation wave filter and pure IMU all adopts the modeling of INS errors propagation model, and these modeling techniques are prior art; Therefore, modeling process is not given unnecessary details at this in detail.

After the hypercompact combined satellite navigation receiver of the present invention is opened navigational system; Initialization procedure comprises little inertial navigation initialization of IMU and pure GNSS navigation initialization; In case the GNSS receiver acquisition is accomplished the initialization alignment procedures to satellite-signal and the little inertial navigation of IMU; System's operation hypercompact combination navigation filtering location algorithm is to improve dynamic property and interference free performance.The little inertial navigation initialization of IMU realizes that mainly inertial equipment powers up the initialization of initialization and course, attitude, location parameter, prepares for INS resolves to resolve with hypercompact combination navigation.The initialization of pure GNSS navigation mode has utilized information such as depositing almanac, ephemeris, realizes catching fast of GNSS receiver.

System mainly relies on the hypercompact combination navigation subfilter under the normal condition; When GNSS signal losing lock can't be located, system mainly relied on the little inertial navigation subfilter of pure IMU; When the little inertial navigation of IMU is broken down or during system initialization, system mainly relies on pure GNSS navigation subfilter.

Above embodiment is only in order to technical scheme of the present invention to be described but not limit it; Under the situation that does not deviate from spirit of the present invention and essence thereof; Those of ordinary skill in the art work as can make various corresponding changes and distortion according to the present invention, but these corresponding changes and distortion all should belong to the protection domain of the appended claim of the present invention.

Claims (8)

1. a hypercompact combined satellite navigation receiver comprises radio-frequency front-end, signal preprocessor; Base band prefilter and Navigation Filter, and little inertial navigation module, pseudorange and pseudorange rates computing module; Ephemeris resolves and selects the star computing module; Wherein signal pre-processing module comprises correlator and local signal maker, it is characterized in that, is provided with sef-adapting filter between said base band prefilter and the Navigation Filter.

2. hypercompact combined satellite navigation receiver as claimed in claim 1 is characterized in that, said local signal maker, and correlator, the base band prefilter, sef-adapting filter and Navigation Filter, pseudorange and pseudorange rates computing module constitute the vector tracking loop.

3. hypercompact combined satellite navigation receiver as claimed in claim 2 is characterized in that said Navigation Filter is federal wave filter, comprises senior filter and at least one subfilter.

4. hypercompact combined satellite navigation receiver as claimed in claim 3; It is characterized in that; The subfilter of said federal wave filter comprises the GNSS wave filter under the pure GNSS pattern, the little inertial navigation wave filter of GNSS/IMU integrated navigation wave filter under the hypercompact integrated mode and pure IMU.

5. like the described hypercompact combined satellite navigation receiver of the arbitrary claim of claim 1-4; It is characterized in that; Said hypercompact combined satellite navigation receiver comprises at least one road navigation signal receiving cable; Each navigation signal receiving cable all is provided with signal preprocessor and base band prefilter, and each roadbed band prefilter is all through inserting Navigation Filter behind the sef-adapting filter.

6. hypercompact combination navigation DVB as claimed in claim 5 is characterized in that, the state equation of described sef-adapting filter is:

$\frac{d}{\mathrm{dt}}\left[\begin{array}{c}\mathrm{\&delta;\&rho;}\\ \mathrm{\&delta;}\stackrel{\&CenterDot;}{\mathrm{\&rho;}}\\ \mathrm{\&delta;}\stackrel{\&CenterDot;\&CenterDot;}{\mathrm{\&rho;}}\\ \mathrm{\&delta;N}\\ {\mathrm{\&Delta;}}_{\mathrm{ion}}\\ {\stackrel{\&CenterDot;}{\mathrm{\&Delta;}}}_{\mathrm{ion}}\end{array}\right]=\left[\begin{array}{c}\mathrm{\&delta;}\stackrel{\&CenterDot;}{\mathrm{\&rho;}}\\ \mathrm{\&delta;}\stackrel{\&CenterDot;\&CenterDot;}{\mathrm{\&rho;}}\\ 0\\ 0\\ {\stackrel{\&CenterDot;}{\mathrm{\&Delta;}}}_{\mathrm{ion}}\\ 0\end{array}\right]+\left[\begin{array}{c}0\\ 0\\ w_a\\ 0\\ 0\\ w_{\mathrm{ion}}\end{array}\right]$

Observation equation is:

$\left[\begin{array}{c}\mathrm{\&delta;\&rho;}\\ \mathrm{\&delta;}\stackrel{\&CenterDot;}{\mathrm{\&rho;}}\\ \mathrm{\&delta;}\stackrel{\&CenterDot;\&CenterDot;}{\mathrm{\&rho;}}\\ \mathrm{\&delta;N}\\ {\mathrm{\&Delta;}}_{\mathrm{ion}}\\ {\stackrel{\&CenterDot;}{\mathrm{\&Delta;}}}_{\mathrm{ion}}\end{array}\right]=\left[\begin{array}{c}\mathrm{\&delta;\&rho;}\\ \mathrm{\&delta;}\stackrel{\&CenterDot;}{\mathrm{\&rho;}}\\ \mathrm{\&delta;}\stackrel{\&CenterDot;\&CenterDot;}{\mathrm{\&rho;}}\\ \mathrm{\&delta;N}\\ {\mathrm{\&Delta;}}_{\mathrm{ion}}\\ {\stackrel{\&CenterDot;}{\mathrm{\&Delta;}}}_{\mathrm{ion}}\end{array}\right]+\left[\begin{array}{c}{w}_{\mathrm{\&rho;}}\\ w_{\stackrel{\&CenterDot;}{\mathrm{\&rho;}}}\\ w_{\stackrel{\&CenterDot;\&CenterDot;}{\mathrm{\&rho;}}}\\ w_N\\ w_i\\ w_{\mathrm{io}}\end{array}\right]$

Wherein δ ρ is the pseudorange residual error,

Be the pseudorange rates residual error,

Be pseudorange acceleration residual error, δ N is a carrier phase complete cycle residual error, Δ _IonBe the ionosphere time-delay,

Be ionosphere time-delay rate, w _aBe pseudorange acceleration process noise, wi _OnBe ionospheric noise, w _ρ,

w _N, w _i, w _IoBe respectively the observation noise of pseudorange, pseudorange rates, pseudorange acceleration, ambiguity of carrier phase, ionosphere time-delay, ionosphere time-delay rate.

7. hypercompact combined satellite navigation receiver as claimed in claim 6; It is characterized in that said sef-adapting filter is the fuzzy logic sef-adapting filter, comprise fuzzy logic algorithm module and Kalman filter; Said fuzzy logic algorithm module receives pseudorange filtering and newly ceases mean value and mean square deviation; Be output as and drive the noise variance weighting coefficient, be used for adjusting in real time the parameter value of the driving noise variance of said Kalman filter, the realization auto adapted filtering.

8. hypercompact combined satellite navigation receiver as claimed in claim 7 is characterized in that, said fuzzy logic algorithm module comprises:

The degree of membership simulation unit is used for confirming that the pseudorange filtering of input newly ceases the membership function of mean value and mean square deviation, and the membership function of the driving noise variance weighting coefficient of output;

The fuzzy inference rule unit is used to be provided with fuzzy inference rule;

The noise variance weighted units is used for calculating driving noise variance weighting coefficient according to said membership function and fuzzy inference rule ambiguity solution.

Images