CN107229061B

CN107229061B - A kind of star based on low orbit satellite ground difference real-time accurate localization method

Info

Publication number: CN107229061B
Application number: CN201710586437.1A
Authority: CN
Inventors: 王磊; 陈锐志; 李德仁; 潘元进; 李明; 陈亮; 柳景斌
Original assignee: Wuhan University WHU
Current assignee: Wuhan University WHU
Priority date: 2017-07-18
Filing date: 2017-07-18
Publication date: 2019-09-03
Anticipated expiration: 2037-07-18
Also published as: CN107229061A

Abstract

The invention belongs to satellite navigation positioning technical field, difference real-time accurate localization method, this method utilize the observation data and Real-time orbit data of low orbit satellite its spaceborne GNSS receiver to terrestrial broadcasting with disclosing a kind of star based on low orbit satellite；Ground receiver receives and forms double difference observation with local GNSS observation after the difference information that low orbit satellite is broadcast, and carries out the dynamic base station DGNSS positioning based on pseudorange or the dynamic base station RTK positioning based on carrier phase.The present invention can be realized the real-time accurate Differential positioning service in global range, do not depend on the distribution at ground reference station using the mobile low orbit satellite platform in the whole world as reference station.User can be realized difference real-time accurate using single receiver and position, no operatton scope limitation, without considering data link.

Description

A kind of star based on low orbit satellite ground difference real-time accurate localization method

Technical field

The invention belongs to satellite navigation positioning technical field more particularly to a kind of star based on low orbit satellite difference is real-time Precision positioning method.

Background technique

Satellite navigation and location system (GNSS) is capable of providing real-time positioning service in global range, deep change Our life style, and be widely used in many industries of countries in the world.The whole world of mainstream is fixed at present Position system has the Galileo system of GPS, the Russian GLONASS in the U.S., Chinese Beidou system and Europe, India Also in positive Development area navigation system IRNSS.The GNSS location technology precision of standard is about 5-10 meters at present.For precision It is required that higher application, it usually needs use the method for precision positioning.The method that precision positions at present has Differential positioning method And accurate one-point positioning method.Wherein Differential positioning method includes GPS local area differential's method and wide area differential method again.GPS wide area differential GPS Method and Static Precise Point Positioning require the deviation of signal product that various precisions are calculated by the data of ground monitoring net, accurate Then track product etc. is corrected in user terminal using these sophisticated products to improve positioning accuracy.GPS local area differential method master If directly broadcasting the observation data and coordinate of reference station to user, eliminated by way of observation difference respectively in user terminal The influence of kind error, realizes high-precision relative positioning.

In conclusion problem of the existing technology is:

GPS wide area differential GPS is capable of providing large-scale precision positioning service, but due to lacking accurate ionospheric model branch at present It holds, it usually needs convergence in 20-30 minutes could obtain centimeter-level positioning result.In addition the parametric method that GPS wide area differential GPS uses is each It is not identical, it is incompatible between different methods.Format and calculation process are broadcast currently without unified data, is caused different wide Compatible interoperation is unable between the difference product of domain.In addition, it is not deep enough due to understanding at present new GNSS system signal, cause Most of WAAS-Wide Area Augmentation System is only applicable to GPS signal.GPS local area differential's method model is simple, and convergence is fast, and real-time is good, positioning accurate Degree is high, long currently based on the real-time dynamic positioning (RTK) of carrier phase and the difference GNSS technology (DGNSS) based on pseudo range difference It has all been the most widely used precision positioning technology since phase.But differential technique also has the limitation of its own, i.e. user receives The distance between machine (also referred to as rover station) and reference station are restricted.The distance between receiver user and reference station become long-range guided missile and cause Ionosphere and tropospheric delay correlation between two receivers die down, this influences positioning accuracy and convergence time.In addition base The elongated public visible satellite also resulted between receiver user and reference station of line tails off, thus Long baselines data processing is usual In static data processing.In addition long range relative positioning needs to construct real-time communication link, uses ground real time communication chain Road needs the routing forwarding network of high-power wireless electricity transmitting equipment or complexity, and expensive, communication quality is difficult to ensure.

Summary of the invention

In view of the problems of the existing technology, the present invention provides a kind of star based on low orbit satellite difference real-time accurate Localization method.

The invention is realized in this way a kind of star based on low orbit satellite ground difference real-time accurate localization method, the base In low orbit satellite star difference real-time accurate localization method using low orbit satellite as reference station, with ground or terrestrial space Receiver carries out Differential positioning；The low orbit satellite carries out ground increasing using the forecast navigation satellite ephemeris and clock deviation infused on ground Precise autonomous orbit determination or the observation only received using itself carry out autonomous precise orbit determination on strong star；The low orbit satellite Broadcast the satellite-orbit information of the observation of receiver and corresponding moment on star to the ground at a certain time interval；Ground receiver Machine will receive the difference information and locally received navigation satellite distance measuring signal composition double difference observation that low orbit satellite is broadcast, into The RTK or DGNSS of row mobile base station are positioned.Low orbit satellite as reference station is apparently higher than terrestrial user station, guarantees distance Have when farther out and enough regards satellite altogether.Remote Real Time Communication Problems are solved using communication link between star ground, simplify user's operation. Low orbit satellite can be global mobile, breaches limitation of the GPS local area differential to operation distance.

Further, the mode of the Differential positioning includes:

(1) by establishing real time data chain, LEO track and observation data are broadcast to receiver user, it is poor in real time to carry out Divide location Calculation；

(2) real time data connects with establishing star, records LEO observation data respectively and ground receiver observes data, into It acts post-processing precision LEO orbit determination and Differential positioning calculates；

(3) real time data connects with not establishing star, is passed by number or LEO is regularly observed data down transmission by TTC channel, LEO precise orbit determination is carried out on ground, reuses LEO observation, LEO orbit determination result and ground receiver observation data carry out difference Precision positioning.

Further, the application model of the observation includes: to put down using only the station-keeping mode of Pseudo-range Observations, using phase The station-keeping mode of sliding Pseudo-range Observations, the station-keeping mode for using pseudorange and carrier phase observation data simultaneously.

Further, the mode of the LEO autonomous orbit determination includes:

(1) ground enhancement information is not used, observation on star is used only and carries out entirely autonomous precise orbit determination；

(2) by uplink communication links between star ground, the enhancement information on ground is uploaded to low orbit satellite, carries out ground enhancing Star on precise autonomous orbit determination, ground enhancement information include but is not limited to navigation satellite prediction orbit, forecast clock deviation information；

(3) correction information is broadcast to low orbit satellite by data relay satellite or inter-satellite link, carries out the enhancing of star base Real-time accurate orbit determination, repeater satellite enhancement information include but is not limited to navigation satellite prediction orbit, forecast clock deviation information.

Further, observation data of the method for difference precision positioning based on single satellite navigation system, or it is based on multisystem Observation alignment by union.

The terrestrial user reference coordinate frame of difference precision positioning is included global reference frame or is joined using customized coordinate Examine frame or the global reference frame Jing Guo encryption.

Further, the star based on low orbit satellite difference real-time accurate localization method specifically include:

1) when receiver reads the observation time, pseudorange, carrier phase, Doppler, carrier-to-noise ratio C/N0, locking of current epoch Between the corresponding navigation message data of observation value information and current epoch；It is the observation time, pseudorange, carrier phase, how general Le, carrier-to-noise ratio C/N0, the observation value information of locking time and the corresponding navigation message data of current epoch are by receiver The output of portion's PVT algoritic module；The observed quantity for being wherein directly used in positioning is pseudorange and carrier phase observation data, and observation model indicates Are as follows:

P_i=ρ+δ_orb+c(δt^S-δt^R)+I_i+δ_trop+ε_Pi

φ_i=ρ+δ_orb+c(δt^S-δt^R)-I_i+δ_trop+λ_iN_i+ε_φi

In formula: P_iAnd φ_iIndicate the pseudorange and carrier phase observation data (as unit of rice) of i-th of frequency point；

ρ indicates the geometric distance between satellite and receiver, is expressed as Wherein x^S,y^S,z^SIndicate satellite three-dimensional coordinate, x^R,y^R,z^RIndicate receiver antenna phase center three-dimensional coordinate；δt^SWith δ t^RSatellite With the clock jitter of receiver, unit is the second；

I_iAnd δ_tropIndicate the influence in ionosphere and tropospheric delay；

N_iIndicate the integer ambiguity of i-th of frequency point；

ε_PiAnd ε_φiIndicate the receiver observation noise of pseudorange and carrier phase；

C indicates the light velocity in vacuum；

λ_iIndicate the carrier wavelength of i-th of frequency point, unit is rice；

2) track and clock deviation of each navigation satellite of signal emission time are calculated；Signal emission time calculates that mode includes two Kind: (1) the precise signal emission time of each navigation satellite is iteratively solved by standard One-Point Location process and corresponding defended Star orbital road and clock deviation, calculate the track clock deviation of each satellite and the approximate coordinate of ground survey station after successive ignition, and (2) are using pseudo- Away from observation Approximate Calculation signal emission time, Principle representation is calculated are as follows:

T in formula_s,t_rRespectively indicate signal transmitting and the signal time of reception；

3) information of the track clock deviation for the approximate survey station coordinate and each navigation satellite being calculated according to One-Point Location, meter Survey station is calculated to the geometric distance ρ between each satellite, the information such as elevation of satellite；Further rule of thumb ionospheric model, meter Ionosphere is calculated to the influence I of carrier phase and pseudorange_i, rule of thumb Tropospheric Models, calculate troposphere to the shadow of GNSS signal Ring δ_trop, also surplus orbit error δ in the observation after correcting_orb, receiver clock-offsets δ t^RWith integer ambiguity parameter N_iShadow It rings；

4) in the cycle slip of un-differenced observation level detected carrier phase；

5) the difference observation data that the low orbit satellite that processing receives is broadcast；The difference that low orbit satellite is broadcast observes data packet Containing signal reception time, the observation that LEO is received low orbit satellite track corresponding with the moment is received and velocity information；It needs first Difference information is decoded, get necessary information；

6) LEO satellite observation moment corresponding navigation satellite signal emission time track and clock deviation are calculated；

7) the spaceborne receiver observation of LEO is missed according to the track and clock deviation of the track of LEO satellite and navigation satellite Poor source correction；

8) un-differenced observation level carries out data detection, cycle slips detection to the carrier phase observation data of the spaceborne receiver of LEO； The observation for detecting generation cycle slip is marked；

9) after pre-processing respectively to the data of ground receiver and the spaceborne receiver of LEO, double difference observation is initially formed Value；The total view satellite of ground receiver and the spaceborne receiver of LEO is filtered out first, and then each navigation system selects a reference Star is used to form difference observation between star；

10) double difference observation is formed；According to the reference satellite of total view satellite and each system that step 9) determines, double difference is formed Observation, star difference double difference observation expression are as follows:

In formulaIt indicates double difference operator, after executing double difference operation, is not disposed in un-differenced observation pretreatment stage Fractional error source theoretically completely eliminated by way of double difference, these error sources include orbit error δ_orb, satellite clock correction δ t^S, receiver clock-offsets δ t^R；Double difference geometric distance in double difference observation between also surplus station star, double difference ionosphere residual effect, between star The troposphere of difference influences, and carrier phase ambiguity influences and observation noise influences；For remaining after formation double difference observation Error term eliminates the influence of these errors by way of parameter Estimation；

11) cycle slips detection of double difference observation level；Using three poor method cycle slips detections；If to the cycle slip of un-differenced observation Detection has enough confidence, can skip this step, directly carries out in next step；

12) filter temporal updates；Real-time coordinates are estimated, comprising: a) use Sequent least square method and Kalman Filter method；Kalman filter model is related with observation selection, and for Dual Frequency Observation data, there are two types of observation models: being based on nothing The model of ionospheric combination observation and model based on non-combined observation；

A) model based on no ionospheric combination observation eliminates ionosphere effect by Dual-frequency Observations linear combination Single order item, but the observation quantity without ionospheric combination observation halves, observation noise is 3 times of observation before combining；

B) it is directly handled using non-combined observation, every satellite is needed to estimate an Ionospheric Parameters；Star of standing is poor Divide the influence that can weaken ionosphere, according to non-combined model inference；For non-combined model, filter status vector is chosen are as follows:

X=[Δ r, δ_trop,z, I, N]^T；

Wherein Δ r is 3 × 1 increment of coordinate vectors, δ_trop,zFor zenith direction tropospheric delay, each survey station estimates one Parameter；I is each satellite direction of visual lines ionosphere delay, and every visible satellite estimates one, and N is double-differential carrier phase fuzziness ginseng Number, each signal path estimate one；The case where for no ionospheric combination observation, the negligible ionization of filter status vector Layer delay parameter I；Static subscriber is positioned, increment of coordinate Δ r estimates as arbitrary constant；For dynamic positioning, using with The equation of motion of the simulation of machine walk process or additional external constrains increment of coordinate parameter；δ_trop,zUse random walk mould Pattern is quasi-, realizes direction ionosphere delay I or is initialized using prior information, is then simulated using random walk process；Carrier wave Phase Double difference fuzziness is estimated when being continuously tracked using arbitrary constant model, needs parameter to reset when cycle slip occurs；It is sending out When raw reference star changes star, specially treated is done to all fuzziness parameters；

Filter temporal renewal equation indicates are as follows:

In formulaP_t ^-Respectively t moment time updated quantity of state and its variance-covariance matrix, also referred to as forecast Solution；For t-1 moment user location state filtering solution and its variance-covariance matrix；Φ (t, t-1) is t-1 to t moment State-transition matrix；Q (t) is process noise matrix；For random walk parameter and random constant parameter, state-transition matrix It is set as unit matrix；Diagonal matrix is set as n independent random walk process Q (t), it is settable for arbitrary constant parameter Q (t) For null matrix；

13) filter measurement updates；Pseudorange and carrier phase observation according to the analysis of step 10), for double difference positioning Value linearisation indicates are as follows:

E (y)=(J, M, Θ, Λ) (Δ r, δ_trop,z,I,N)^T；

E () is mathematic expectaion operator in formula, to f frequency point, the observation of s visible double difference satellite, for using For the mathematical model of non-combined observation, J be (2*s*f) × 3 Jacobbian matrix, M be (2*s*f) × 1 column to Amount, each element correspond to the troposphere projection function value of respective satellite；Θ is the matrix of (2*s*f) × s；Λ be (2*s*f) × (f*s) matrix；For using the observation model of no ionospheric combination, in addition to ignoring the corresponding design matrix Θ of Ionospheric Parameters Outside, all observation linear combinations also result in observation number and reduce half than non-combined model；After linearisation, this is non-combined The design matrix of observational equation is defined as:

A=(J, M, Θ, Λ)；

Then the measurement updaue equation of Kalman filtering indicates are as follows:

K_t=P_t ^-A(AP_t ^-A^T+R_t)^-1,

K in formula_tFor t moment filtering gain matrix, R_tFor geometry observation variance-covariance matrix；

14) rear residual test is tested；The rear residual error valuation of testing of filtering equations indicates are as follows:

χ is constructed according to rear residual error valuation is tested²Inspected numberObey χ²(n-t) it is distributed, wherein n-t indicates freedom degree；IfThen think to test rear residual error by examining, otherwise it is assumed that testing rear residual test refusal；In formulaIt is significant Property level be α threshold value；

15) output forecast solution；If testing rear residual test not pass through, then it represents that there may be rough errors for observation；Use number It tests according to probe method to rough error and rejects or abandon to use forecast solution instead using filter solutionIt is carried out as final positioning result defeated Out；

16) fuzziness is fixed；If filter solution is by testing rear residual test, special using the integer of carrier phase ambiguity Property obtain centimetre class precision ambiguity fixed solution；The fixed method of fuzziness using least square drop related algorithm LAMBDA or The mode that no ionospheric combination is decomposed into width lane ambiguity is carried out fuzziness step by step and fixed by person；

State vector in step 13) is merged according to real parameters and two class of numeric parameter, then observational equation simplifies table It is shown as E (y)=(B, Λ) (b, N)^T, corresponding filter state amountTwo class parameters corresponding filter solution variance association side Poor matrix is expressed as:

The filter solution of fuzziness parameter is denoted asCommonly referred to as fuzziness real solution, by the real solution of fuzziness parameterAnd its variance-covariance matrix Q_NNIt as input, is scanned for using LAMBDA method, the optimal integer for obtaining fuzziness is solid Fixed solutionWith the integer fixed solution of suboptimum

17) fuzziness is examined；Positioning result is caused to generate flying spot since the fuzziness of mistake is fixed, position error compares floating-point Solution is also big, tests to the fixed result of fuzziness；It is examined using Ratio；

18) float-solution is exported；If the test fails by Ratio, then it is assumed that ambiguity fixed solution is unreliable, needs using filter Wave solutionIt is exported as final positioning result；

19) fixed solution is exported；If Ratio upchecks, ambiguity fixed solution is used to export as positioning result；Mould After paste degree is fixed as integer, corresponding real parametersIt may be updated as the condition solution of integer ambiguity, indicate are as follows:

Corresponding variance-covariance matrix may be expressed as:

If fuzziness is upchecked, useIt is used as last positioning result.

A kind of another object of the present invention is to provide star based on low orbit satellite difference real-time accurate positioning system.

It is fixed that another object of the present invention is to provide a kind of using the above-mentioned star based on low orbit satellite difference real-time accurates The pelagic region positioning system of position system.

It is fixed that another object of the present invention is to provide a kind of using the above-mentioned star based on low orbit satellite difference real-time accurates The positioning system from far-off regions of position system

It is fixed that another object of the present invention is to provide a kind of using the above-mentioned star based on low orbit satellite difference real-time accurates The natural calamity disaster area positioning system of position system.

Advantages of the present invention and good effect are as follows:

It is provided by the invention using low orbit satellite as reference station, realize that difference is fixed with the receiver of ground or terrestrial space The method of position.Low orbit satellite broadcasts the satellite of the observation of receiver and corresponding moment on star to the ground at a certain time interval Orbit information, ground receiver will receive the difference information and locally received navigation satellite distance measuring signal that low orbit satellite is broadcast Double difference observation is formed, realizes RTK or the DGNSS positioning of mobile base station.This method uses low orbit satellite as reference station, real The Differential positioning service of global range is showed, especially pelagic region, remote districts and natural calamity disaster area has been realized quickly Precision positioning is significant.

The method that the present invention uses GPS local area differential, but the utilization of reference station is improved by mobile reference platform Rate so that reference station service range is not limited to around tens kilometers, and is to provide the differential service of global range.If used The constellation of more LEO satellites composition can then provide global range continual station star Differential positioning service as reference base station

The present invention can be to be influenced to cause by natural conditions and economic condition not set up or can not set up earth station Region, such as ocean area, remote districts, the regions such as earthquake-stricken area provide real-time requiring service.

Star of the present invention Differential positioning algorithm model it is simpler than Static Precise Point Positioning (PPP), can be with by the method for double difference The clock deviation and hardware deviation for voluntarily eliminating satellite end and receiver end, are modified without using external signal deviation product.This Invention utilizes low orbit satellite platform, and the low orbit satellite platform of the reference station movement of the fixation on ground is replaced.Structure between star ground At double difference observation also there is certain weakening effect to ionosphere delay.

User terminal of the present invention only needs the receiver that can receive LEO difference information that global model can be realized Enclose interior precision positioning, star the method for difference help user to get rid of the dependence to reference station.Without voluntarily setting up reference station, Without setting up real-time communication link.Application method is just the same with PPP, but can provide precision and convergence time is all better Real time differential precision positioning service.

Detailed description of the invention

Fig. 1 is the star provided in an embodiment of the present invention based on low orbit satellite ground difference real-time accurate localization method flow chart.

Fig. 2 is the schematic diagram of low orbit satellite provided in an embodiment of the present invention with realizing star difference precision positioning.

Specific embodiment

In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to embodiments, to the present invention It is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, it is not used to Limit the present invention.

Application principle of the invention is described in detail with reference to the accompanying drawing.

Star provided in an embodiment of the present invention based on low orbit satellite ground difference real-time accurate localization method, it is described to be based on low rail The star of satellite ground receiver of the difference real-time accurate localization method using low orbit satellite as reference station, with ground or terrestrial space Carry out Differential positioning；The low orbit satellite carries out the star of ground enhancing using the forecast navigation satellite ephemeris and clock deviation infused on ground Upper precise autonomous orbit determination or the observation only received using itself carry out autonomous precise orbit determination；The low orbit satellite is with certain Time interval broadcast the satellite-orbit information of the observation of receiver on star and corresponding moment to the ground；Ground receiver will connect The difference information and locally received navigation satellite distance measuring signal composition double difference observation that low orbit satellite is broadcast are received, is moved The RTK or DGNSS of base station are positioned.

The Differential positioning includes:

(2) real time data connects with establishing star, records LEO observation data respectively and ground receiver observes data, into It acts post-processing LEO precise orbit determination and Differential positioning calculates；

The application model of the observation includes: using only the station-keeping mode of Pseudo-range Observations, using carrier phase smoothed pseudorange The station-keeping mode of observation, the station-keeping mode for using pseudorange and carrier phase observation data simultaneously.

The mode of the LEO autonomous orbit determination includes:

Observation data of the method for difference precision positioning based on single satellite navigation system, or joined based on multisystem observation Close positioning.

The terrestrial user reference coordinate frame of difference precision positioning is included global reference frame or is joined using customized coordinate Examine frame or the global reference frame Jing Guo encryption.Such as ITRF2008, CGCS2000 reference frame.

Star difference precision positioning method be related to navigation satellite constellation and include, but are not limited to the existing whole world/regional satellite Navigation system, such as GPS, GLONASS, Beidou, Galileo, IRNSS, QZSS etc..

Application principle of the invention is further described combined with specific embodiments below.

Star provided in an embodiment of the present invention based on low orbit satellite difference real-time accurate localization method as shown in Figure 1, should Algorithm is the real-time location algorithm for being based on extended BHF approach (EKF), be can be used at static and dynamic data Reason.The algorithm can be realized in the firmware algorithm of the inside of GNSS receiver, also can be implemented as post-processing algorithm.In order to full The demand positioned when full is described below using receiver internal firmware algorithm as reference.Specific implementation steps are as follows:

1) receiver reads the observation time of current epoch, pseudorange, carrier phase, Doppler, carrier-to-noise ratio (C/N0), locking Time etc. observes value information and the corresponding navigation message data of current epoch, these information can be by PVT inside receiver Algoritic module output.The observed quantity for being wherein directly used in positioning is pseudorange and carrier phase observation data, and observation model can be with table It is shown as:

P_i=ρ+δ_orb+c(δt^S-δt^R)+I_i+δ_trop+ε_Pi

φ_i=ρ+δ_orb+c(δt^S-δt^R)-I_i+δ_trop+λ_iN_i+ε_φi

ρ indicates the geometric distance between satellite and receiver, is represented by Wherein x^S,y^S,z^SIndicate satellite three-dimensional coordinate, x^R,y^R,z^RIndicate receiver antenna phase center three-dimensional coordinate；δt^SWith δ t^RSatellite With the clock jitter (in seconds) of receiver；

I_iAnd δ_tropIndicate the influence in ionosphere and tropospheric delay；

N_iIndicate the integer ambiguity of i-th of frequency point；

C indicates the light velocity in vacuum；

λ_iIndicate the carrier wavelength (as unit of rice) of i-th of frequency point；

2) track and clock deviation of each navigation satellite of signal emission time are calculated.In view of the correction of low orbit satellite is broadcast When there is certain data processing and signal to propagate delay, need accurately to calculate signal emission time.In view of middle height The satellite velocities of rail satellite, it is desirable that signal emission time calculates error at least below 1 μ s, just can guarantee that calculating orbit error is less than 1cm.There are two types of signal emission time reckoning modes: (1) iteratively solving each navigation satellite by standard One-Point Location process Precise signal emission time and corresponding satellite orbit and clock deviation, this method need successive ignition to calculate the track clock of each satellite The approximate coordinate of difference and ground survey station, but signal emission time can be calculated to 100ns after restraining；(2) pseudorange is utilized Observation Approximate Calculation signal emission time calculates that principle can be expressed as:

T in formula_s,t_rRespectively indicate signal transmitting and the signal time of reception.Compare pseudorange code P₁Observation model it is found that should Formula does not account for ionosphere, the influence of troposphere and receiver clock-offsets.Wherein the influence of ionosphere and troposphere to ranging is several Ten meters of magnitudes, the influence to the reckoning of signal emission time is in hundred nanosecond orders.Can the influence of receiver clock-offsets, which be ignored, depends on The model of receiver clock-offsets.The model of receiver clock-offsets has amendment in real time and millisecond to correct two kinds of models, and real-time correction model connects Receipts machine clock deviation was stablized in several nanoseconds to the magnitude of several hundred nanoseconds, can satisfy reckoning demand, in millisecond correction model, receiver clock Difference, which has reached 1ms or -1ms, will occur clock jump, must be fixed using single-point if receiver clock uses millisecond correction model Position method iteratively solves navigation satellite signal emission time.Major part receiver uses or can be configured to clock deviation at present repairs in real time Positive model.After solving signal emission time, i.e., the track and clock deviation of each satellite are calculated using broadcast ephemeris.

3) information such as track clock deviation of approximate survey station coordinate and each navigation satellite being calculated according to One-Point Location, meter Survey station is calculated to the geometric distance ρ between each satellite, the information such as elevation of satellite.Further rule of thumb ionospheric model, such as Klobuchar model calculates ionosphere to the influence I of carrier phase and pseudorange_i, rule of thumb Tropospheric Models, such as Hopfield model calculates influence δ of the troposphere to GNSS signal_trop, also surplus orbit error δ in the observation after correcting_orb, Receiver clock-offsets δ t^RWith integer ambiguity parameter N_iInfluence.In addition, being repaired since geometric distance calculating uses approximate coordinate Influence in observation after just also containing geometric distance residual volume, ionospheric convection layer empirical model can only also correct a part It influences, for the residual effect of navigation satellite clock deviation also up to nanosecond order, these error residue amounts can also be retained in revised sight In measured value.

4) in the cycle slip of un-differenced observation level detected carrier phase, the with good grounds carrier lock loop of common method (PLL) export locking time come judge loop whether losing lock.Once loop losing lock, carrier phase will be reset, defeated at this time Carrier phase out can generate step signal, need to be marked.Another situation is exactly signal interference, signal blocks, or The influence of person's multipath causes carrier loop to be shaken, and such case not will lead to loop losing lock generally, but will lead to small week It jumps, such as half cycle slip or complete cycle are jumped.The case where half cycle slip, will affect the bit synchronous and frame synchronization of PVT, but it is logical to do polarity judgement The regular hour often is needed, therefore have several seconds carrier loops and PVT after half cycle slip of generation not judging whether to have occurred half Cycle slip, such case Data processing are usually all handled as cycle slip.Small complete cycle is jumped, needs to sentence using other methods It is disconnected, but the probability is relatively small for ordinary circumstance appearance.Judge that small complete cycle jumps common method and has free-geometry method, if connect The method detection cycle slip of MW combination also can be used in receipts machine pseudorange tracking noise very little.

5) the difference observation data that the low orbit satellite received is broadcast next are handled.The difference observation that low orbit satellite is broadcast Data include signal reception time, the observation that LEO is received low orbit satellite track corresponding with the moment is received and velocity information. It is decoded firstly the need of to difference information, gets necessary information.

6) LEO satellite observation moment corresponding navigation satellite signal emission time track and clock deviation are calculated.Calculation method and That discusses in step 102 is identical, the difference is that the coordinate of LEO satellite is known, sends out if One-Point Location is used to calculate signal Penetrating is that iteration is only needed to update co-ordinates of satellite constantly, and the coordinate of LEO satellite may be considered true value.In view of the fortune of low orbit satellite Dynamic speed is about 7.6km/s (by taking 500Km orbit altitude as an example), when the clock face of LEO observation and the actual signal time of reception it Difference should be less than 1 μ s.If the clock deviation of the spaceborne receiver of LEO uses real-time correction model, when being received with actual signal when clock face For the residual quantity grade at quarter less than hundred nanosecond orders, the influence to LEO track is less than 1cm, can be neglected.The rail that LEO is broadcast to the ground Road information should be by when clock face to guarantee LEO observation of interpolation appropriate or extrapolation process and the difference of the actual signal time of reception Magnitude is less than 1 μ s.

7) the spaceborne receiver observation of LEO is missed according to the track and clock deviation of the track of LEO satellite and navigation satellite Poor source correction.The step implementation procedure is similar with step 3), the difference is that not being that ground connects since LEO belongs to spacecraft Receipts machine, the ionosphere being subject to and tropospheric influencing characterisitic and ground receiver are different.Troposphere influence is often referred to ground To the near-earth atmosphere of 30Km height, therefore it is generally acknowledged that influence of the troposphere to the spaceborne receiver observation of LEO is negligible, place Tropospheric delay is not corrected when reason.In addition ionosphere is smaller than ground to the observation of the spaceborne receiver of LEO, according to Chapman is theoretical, and the ionosphere effect that the spaceborne receiver observation of low orbit satellite is subject to is about the 30%-50% on ground.To star Carry receiver observation ionosphere correction model also usually using ground empirical model multiplied by a certain scale factor as spaceborne Correct empirical model in the ionosphere of receiver observation.Since atmospheric density section curve and temperature profile curve can not be surveyed accurately Fixed, Chapman model is also approximate empirical model, cannot completely eliminate influence of the ionosphere to observation.

8) un-differenced observation level carries out data detection, cycle slips detection to the carrier phase observation data of the spaceborne receiver of LEO. Processing method is identical as step 104, the receiver the difference is that movement velocity of LEO satellite is above the ground level, using some and geometry When detecting cycle slip apart from relevant method, the selection of special consideration threshold value is needed.The observation for detecting generation cycle slip is carried out Label.

9) after pre-processing respectively to the data of ground receiver and the spaceborne receiver of LEO, double difference observation is initially formed Value.Double difference observation usually first carries out difference between survey station, then carries out inter-satellite difference.For star the case where baseline, first The total view satellite of ground receiver and the spaceborne receiver of LEO is filtered out, then each navigation system selects a reference star, is used for shape Difference observation between Cheng Xing.The observation quality of reference star directly affects the quality of all double difference observations, it is therefore desirable to as far as possible The preferable satellite of observation quality is chosen as reference star, usual situation uses the highest satellite of elevation angle or signal carrier-to-noise ratio Best satellite is as reference satellite.On the other hand, reference star changes, and to will lead to all double difference fuzziness parameters discontinuous. When changing reference star, fuzziness parameter needs specially treated to avoid fuzziness parameter from continually resetting, therefore when selection reference star Also reference star can be kept as far as possible constant, to guarantee the continuity of fuzziness parameter.It the case where for multisystem combined positioning, needs Each satellite system selects reference star respectively, since hardware delay has a different between the signal of different navigation system, and this A difference can not be eliminated using the inter-satellite difference of not homologous ray.Deviation can be disappeared by the method for parameter Estimation between this system It removes, more general way is that each navigation system chooses reference satellite respectively and comes cancellation receiver clock deviation (containing deviation between system) It influences.

10) double difference observation is formed；According to the reference satellite of total view satellite and each system that step 9) determines, double difference is formed Observation, star the double difference observation of difference can be expressed as:

In formulaIt indicates double difference operator, after executing double difference operation, is not disposed in un-differenced observation pretreatment stage Fractional error source can theoretically be completely eliminated by way of double difference, these error sources include navigation satellite orbit error δ_orb, navigation satellite clock deviation δ t^S, receiver clock-offsets δ t^R.Double difference geometric distance in double difference observation between also surplus station star, double difference electricity Absciss layer residual effect, the troposphere of difference influences (assuming that LEO observation is not influenced by tropospheric delay) between star, carrier phase mould Paste degree influences and observation noise influences.For error term remaining after formation double difference observation, need to give in function model Consider, the influence of these errors is eliminated by way of parameter Estimation.Consideration star differential signal propagation delay of arriving at a station causes between station When single poor LEO observation moment and earth station's observation moment can not close alignment, this can give double difference observation to introduce time irreversibility Error.Consider star communication bandwidth limitation of arriving at a station, the delay etc. of data processing, star Differencing communication of standing postpone the amount up to the several seconds on star Grade.Time in-synchronization error mainly results in the orbit error of navigation satellite in double difference observation, clock deviation and ionospheric error it is residual Remaining error influence becomes larger.Processing strategie is to be counted respectively in non-difference data processing stage according to LEO and ground receiver observation moment Navigation satellite track and clock deviation are calculated, is then corrected.

11) cycle slips detection of double difference observation level.As the supplement of non-poor level cycle slips detection, double difference observation is formed After can also carry out cycle slips detection in double difference observation level, typical method have three poor method cycle slips detections.The week of double difference level Jumping the not necessary step of detection can skip this step, directly if having enough confidence to the cycle slips detection of un-differenced observation It carries out in next step.

12) filter temporal updates.Complete the data processing of double difference observation, can start parameter estimation process. Real-time coordinates are estimated, Sequent least square method and Kalman filtering method can be used.Kalman filtering method be suitble to it is static and Dynamic data processing, by introducing, parameter eliminates process to Sequent least square method and prior information constrains and also may be implemented and block The effect of Kalman Filtering equivalence.This step illustrates Differential positioning parameter between the star of station so that Kalman filtering solves coordinate parameters as an example The method of estimation.Kalman filter model is related with observation selection, and for Dual Frequency Observation data, there are two types of representative observation moulds Type: the model based on no ionospheric combination observation and the model based on non-combined observation.It is observed based on no ionospheric combination The model of value can eliminate the single order item of ionosphere effect by Dual-frequency Observations linear combination, but observe without ionospheric combination The observation quantity of value halves, and observation noise is about 3 times of observation before combining.Another method is directly to use non-group It closes observation to be handled, but every satellite is needed to estimate an Ionospheric Parameters.Consideration star difference of arriving at a station can weaken electricity The influence of absciss layer, part is according to non-combined model inference below.For non-combined model, filter status vector is chosen are as follows:

X=[Δ r, δ_trop,z, I, N]^T；

Wherein Δ r is 3 × 1 increment of coordinate vectors, δ_trop,zFor zenith direction tropospheric delay, each survey station estimates one Parameter.I is each satellite direction of visual lines ionosphere delay, and every visible satellite estimates that one (in addition to reference satellite), N are carrier wave phase Position double difference fuzziness parameter, each signal path estimate one.The case where for no ionospheric combination observation, filter status Ionosphere delay parameter I can be ignored in vector.Static subscriber is positioned, increment of coordinate Δ r can estimate as arbitrary constant, right In dynamic positioning, random walk process simulation can be used, can also with the equation of motion of additional external to increment of coordinate parameter into Row constraint.δ_trop,zUsually using random walk modeling, it is initial to realize that prior information can be used in ionosphere delay I in direction Change, is then simulated using random walk process.Double-differential carrier phase fuzziness is estimated when being continuously tracked using arbitrary constant model Meter, needs parameter to reset when cycle slip occurs.When generation reference star changes star, need to do special place to all fuzziness parameters Reason avoids fuzziness parameter from resetting.Filter temporal renewal equation may be expressed as:

In formulaP_t ^-Respectively t moment time updated quantity of state and its variance-covariance matrix, also commonly referred to as Forecast solution.For t-1 moment user location state filtering solution and its variance-covariance matrix.Φ (t, t-1) is t-1 to t The state-transition matrix at moment.Q (t) is process noise matrix.For random walk parameter and random constant parameter, state transfer Matrix can be set as unit matrix.Diagonal matrix can be set as n independent random walk process Q (t), for arbitrary constant parameter Q (t) it may be configured as null matrix.

13) filter measurement updates；Pseudorange and carrier phase observation according to the analysis of step 10), for double difference positioning Value can linearize expression are as follows:

E (y)=(J, M, Θ, Λ) (Δ r, δ_trop,z,I,N)^T；

E () is mathematic expectaion operator in formula, to f frequency point, observation (total visible star of s visible double difference satellite Number subtracts reference star number), for using the mathematical model of non-combined observation, J is (2*s*f) × 3 Jacobbian matrix, M are the column vector of (2*s*f) × 1, and each element corresponds to the troposphere projection function value of respective satellite.Θ For the matrix of (2*s*f) × s.Λ is the matrix of (2*s*f) × (f*s).For using the observation model of no ionospheric combination, remove Outside negligible Ionospheric Parameters corresponding design matrix Θ, all observation linear combinations also result in observation number than non-group The model of conjunction reduces half.After linearisation, the design matrix of non-combined model observational equation is defined as:

A=(J, M, Θ, Λ)；

Then the measurement updaue equation of Kalman filtering may be expressed as:

K_t=P_t ^-A(AP_t ^-A^T+R_t)^-1,

P_t ⁺=(I-K_tA)P_t ^-,

K in formula_tFor t moment filtering gain matrix, R_tFor geometry observation variance-covariance matrix.

14) rear residual test is tested；The rear residual error valuation of testing of filtering equations may be expressed as:

χ can be constructed according to rear residual error valuation is tested²Inspected numberObey χ²(n-t) it is distributed, wherein n-t indicates freedom degree. IfThen think to test rear residual error by examining, otherwise it is assumed that testing rear residual test refusal.In formulaIt is aobvious The threshold value that work property level is α.

15) output forecast solution；If testing rear residual test not pass through, then it represents that there may be rough errors for observation.It can be used Data snooping tests rejecting to rough error, can also abandon using forecast solution instead using filter solutionAs final positioning knot Fruit is exported.

16) fuzziness is fixed；If filter solution by testing rear residual test, can be attempted to utilize carrier phase ambiguity The ambiguity fixed solution of integer characteristic acquisition centimetre class precision.Due to star difference baseline group was positioned at double difference observation Journey is equivalent to Long baselines relative positioning.The fixed method of fuzziness can be used least square drop related algorithm (LAMBDA) or The mode that no ionospheric combination is decomposed into width lane ambiguity is carried out fuzziness step by step to fix.By the state in step 13) to Amount merges according to real parameters and two class of numeric parameter, then observational equation, which can simplify, is expressed as E (y)=(B, Λ) (b, N)^T, Corresponding filter state amountThe corresponding filter solution variance-covariance matrix of two class parameters may be expressed as:

The filter solution of fuzziness parameter is denoted asCommonly referred to as fuzziness real solution, by the real solution of fuzziness parameterAnd its variance-covariance matrix Q_NNAs input, scanned for using LAMBDA method, available fuzziness it is optimal whole Number fixed solutionWith the integer fixed solution of suboptimum

17) fuzziness is examined；Since the fixed positioning result that may cause of the fuzziness of mistake generates flying spot, position error ratio Float-solution is also big, therefore to test to the fixed result of fuzziness.It is examined usually using Ratio.Ratio is examined is defined as:

In formulaFor the suboptimal solution of integer ambiguity.For two norm of Euclid.μ is that ratio examines threshold value, is led to Often take 2-5.Think if above formula is set upBy examining, otherwise it is assumed that examining refusal.

18) float-solution is exported；If the test fails by Ratio, then it is assumed that ambiguity fixed solution is unreliable, needs using filter Wave solutionIt is exported as final positioning result.

19) fixed solution is exported；If Ratio upchecks, ambiguity fixed solution is used to export as positioning result.Mould After paste degree is fixed as integer, corresponding real parametersIt may be updated as the condition solution of integer ambiguity, indicate are as follows:

Corresponding variance-covariance matrix may be expressed as:

If fuzziness is upchecked, useIt is used as last positioning result.

Above-mentioned steps can realize that the difference real-time accurate between ground receiver and low orbit satellite positions.

The principle of low orbit satellite provided in an embodiment of the present invention with realizing star difference precision positioning is as shown in Figure 2.In figure GEO/IGSO/MEO refer to it is existing the whole world or region satellite navigation system satellite, including but not limited to GPS, GLONASS, Beidou, Galileo, IRNSS and QZSS navigational satellite system.LEO, which refers to, is equipped with double frequency GNSS signal receiver, has height Precision autonomous orbit determination ability, and there is the low orbit satellite for broadcasting information capability to the ground.Ground reference station, which refers to, is laid in the earth The continuous operation of the reference station on surface, such as monitoring net (IGS) of world GNSS service, more GNSS try nets (MGEX) etc..Ground distributor Analysis center is the observation data for referring to obtain each ground reference station by network, can calculate or forecast in real time that navigation is defended The track and clock deviation of star, and note on the navigation satellite track and clock deviation of forecast can be arrived to the ground centre data of low orbit satellite Processing center.User refers to GNSS signal receive capabilities, and can receive, the difference letter that parsing low orbit satellite is broadcast Breath, determines the receiving device of itself precision coordinate by way of Differential positioning.As shown in Fig. 2, GEO/IGSO/MEO is led in figure Boat satellite it is lasting navigation signal is broadcast with low orbit satellite to the ground, navigation signal is by low orbit satellite, user and ground reference station It receives.The observation received is sent to Ground analysis center by ground reference station, and Ground analysis center is each by COMPREHENSIVE CALCULATING The observation data of a earth station determine the accurate GEO/IGSO/MEO track and clock deviation information in real time and forecast.In view of low rail Satellite TT length of window is limited, it usually needs the track and clock deviation of a few hours is forecast, then by prediction orbit and clock deviation information It carries out data compression and compressed data is uploaded to low orbit satellite in observing and controlling window.Low orbit satellite uses spaceborne receiver Observation data and the ground enhancement information that receives carry out real-time accurate track and determine.Then by the track of itself and with for the moment It carves corresponding observation data to be encoded, broadcast to ground.Terrestrial user uses the observation of the GEO/IGSO/MEO satellite received The difference information that value and low orbit satellite are broadcast, with forming star difference baseline, complete double difference positioning.Although star parallax range compared with It is long, thousands of kilometers are usually had hundreds to, but the double difference observation formed still is able to effectively cancellation receiver clock deviation, satellite Clock deviation weakens ionosphere effect, effectively simplifies the complexity of precision positioning, improves location efficiency and precision.In addition, this articles and opinions The star stated differential system can also be run in autonomous mode, i.e. the mode supported of no Ground analysis center.There is no ground ginseng It examines station and is supported with Ground analysis center, autonomous orbit determination also can achieve the result of 0.5m or so on star.Directly using autonomous fixed The result of rail with carrying out star difference, still can get the Real-Time Positioning better than 1 meter.Ground analysis center and low orbit satellite it Between direct communication link can also indirectly be established by modes such as data relay satellite or inter-satellite links.

The present invention can be improved the utilization rate at ground reference station, and Differential positioning service range is extended to the whole world.

The present invention is capable of providing the Differential positioning service of global range, especially in ocean area, remote districts, earthquake disaster Area etc. sets up the region at ground reference station without condition.

Star of the present invention Differential positioning algorithm model it is simpler than Static Precise Point Positioning (PPP), can be with by the method for double difference The clock deviation and hardware deviation of satellite end and receiver end are voluntarily eliminated, thus is repaired without using external signal deviation product Just.Based on star it is better than Static Precise Point Positioning in terms of the positioning convergence time of differential position system and positioning accuracy.

The present invention is capable of providing the high accuracy positioning service towards active user, and precise single-point positioning technology will rely on a variety of External precision data product corrects observation data, and portioned product has certain delay, this leads to Static Precise Point Positioning Technology is more applied in post-processing and the not high scene of real-time.Method of the invention is suitble to real-time high-precision positioning application.

The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all in essence of the invention Made any modifications, equivalent replacements, and improvements etc., should all be included in the protection scope of the present invention within mind and principle.

Claims

1. a kind of star based on low orbit satellite ground difference real-time accurate localization method, which is characterized in that described to be based on low orbit satellite Star difference real-time accurate localization method using low orbit satellite as reference station, carried out with the receiver of ground or terrestrial space Differential positioning；On the star that the low orbit satellite is enhanced using ground or repeater satellite precise autonomous orbit determination or only using from The observation that body receives carries out autonomous precise orbit determination；The low orbit satellite broadcasts on star to the ground connect at a certain time interval The satellite-orbit information of the observation of receipts machine and corresponding moment；The difference information that ground receiver will receive low orbit satellite and broadcast Double difference observation is formed with locally received navigation satellite distance measuring signal, carries out RTK or the DGNSS positioning of mobile base station；

The mode of the Differential positioning includes:

(1) by establishing real time data chain, LEO track and observation data is broadcast to receiver user, it is fixed to carry out real time differential Position calculates；

(2) real time data connects with establishing star, records LEO observation data and ground receiver observation respectively by ground receiver Data, carry out post-processing LEO precise orbit determination and Differential positioning calculates；

(3) real time data connects with not establishing star, is passed by number or LEO is regularly observed data down transmission by TTC channel, on ground Face carries out LEO precise orbit determination, reuses LEO observation, it is accurate that LEO orbit determination result and ground receiver observation data carry out difference Positioning.

2. the ground of the star based on low orbit satellite difference real-time accurate localization method as described in claim 1, which is characterized in that described The application model of observation includes: using only the station-keeping mode of Pseudo-range Observations, using the positioning of carrier phase smoothed pseudorange observation Mode, the station-keeping mode for using pseudorange and carrier phase observation data simultaneously.

3. the ground of the star based on low orbit satellite difference real-time accurate localization method as described in claim 1, which is characterized in that described The mode of precise autonomous orbit determination includes: on star

(2) by uplink communication links between star ground, the enhancement information on ground is uploaded to low orbit satellite, carries out the star of ground enhancing Upper precise autonomous orbit determination, ground enhancement information include but is not limited to navigation satellite prediction orbit, the information for forecasting clock deviation；

(3) correction information is broadcast to low orbit satellite by data relay satellite or inter-satellite link, carries out the reality of star base enhancing When precise orbit determination, repeater satellite enhancement information include but is not limited to navigation satellite prediction orbit, forecast clock deviation information.

4. the ground of the star based on low orbit satellite difference real-time accurate localization method as described in claim 1, which is characterized in that difference Observation data of the method for precision positioning based on single satellite navigation system, or it is based on multisystem observation alignment by union；

The terrestrial user reference coordinate frame of difference precision positioning includes global reference frame or the customized coordinate reference frame of use Frame or global reference frame Jing Guo encryption.

5. the ground of the star based on low orbit satellite difference real-time accurate localization method as described in claim 1, which is characterized in that described Star based on low orbit satellite difference real-time accurate localization method specifically include:

1) receiver reads the observation time of current epoch, pseudorange, carrier phase, Doppler, carrier-to-noise ratio C/N0, locking time Observe value information and the corresponding navigation message data of current epoch；The observation time, pseudorange, carrier phase, Doppler, load It makes an uproar than C/N0, the observation value information of locking time and the corresponding navigation message data of current epoch by PVT inside receiver Algoritic module output；The observed quantity for being wherein directly used in positioning is pseudorange and carrier phase observation data, and observation model indicates are as follows:

P_i=ρ+δ_orb+c(δt^S-δt^R)+I_i+δ_trop+ε_Pi

φ_i=ρ+δ_orb+c(δt^S-δt^R)-I_i+δ_trop+λ_iN_i+ε_φi

In formula: P_iAnd φ_iIndicate the pseudorange and carrier phase observation data as unit of rice of i-th of frequency point；ρ indicates satellite and connects Geometric distance between receipts machine, is expressed asWherein x^S,y^S,z^SIndicate satellite Three-dimensional coordinate, x^R,y^R,z^RIndicate receiver antenna phase center three-dimensional coordinate；

δt^SWith δ t^RThe clock jitter of satellite and receiver, unit are the second；

I_iAnd δ_tropIndicate the influence in ionosphere and tropospheric delay；

N_iIndicate the integer ambiguity of i-th of frequency point；

C indicates the light velocity in vacuum；

λ_iIndicate the carrier wavelength of i-th of frequency point, unit is rice；

2) track and clock deviation of each navigation satellite of signal emission time are calculated；Signal emission time calculates that mode includes two kinds: (1) by standard One-Point Location process iteratively solve each navigation satellite precise signal emission time and corresponding satellite rail Road and clock deviation, the track clock deviation of each satellite and the approximate coordinate of ground survey station are calculated after successive ignition, and (2) utilize pseudorange to see Measured value Approximate Calculation signal emission time calculates Principle representation are as follows:

3) information of the track clock deviation for the approximate survey station coordinate and each navigation satellite being calculated according to One-Point Location is calculated and is surveyed It stands to the geometric distance ρ between each satellite, satellite altitude angle information；Further rule of thumb ionospheric model, calculating ionize Influence I of the layer to carrier phase and pseudorange_i, rule of thumb Tropospheric Models, calculate influence δ of the troposphere to GNSS signal_trop, Also surplus orbit error δ in observation after amendment_orb, receiver clock-offsets δ t^RWith integer ambiguity parameter N_iInfluence；

5) the difference observation data that the low orbit satellite that processing receives is broadcast；The difference observation data that low orbit satellite is broadcast include letter Number receiving time, the observation that LEO is received low orbit satellite track corresponding with the moment is received and velocity information；Firstly the need of right Difference information is decoded, and gets necessary information；

6) each navigation satellite track of LEO satellite observation moment corresponding signal emission time and clock deviation are calculated；

7) error source is carried out to the spaceborne receiver observation of LEO according to the track and clock deviation of the track of LEO satellite and navigation satellite Correction；

8) un-differenced observation level carries out data detection, cycle slips detection to the carrier phase observation data of the spaceborne receiver of LEO；To inspection The observation for measuring generation cycle slip is marked；

9) after pre-processing respectively to the data of ground receiver and the spaceborne receiver of LEO, double difference observation is initially formed；It is first The total view satellite of ground receiver and the spaceborne receiver of LEO is first filtered out, then each navigation system selects a reference star, is used for Difference observation between formation star；

10) double difference observation is formed；According to the reference satellite of total view satellite and each system that step 9) determines, composition double difference observation Value, star difference double difference observation expression are as follows:

In formulaDouble difference operator is indicated, after executing double difference operation, in the portion that un-differenced observation pretreatment stage is not disposed Error source is divided theoretically to completely eliminate by way of double difference, these error sources include orbit error δ_orb, satellite clock correction δ t^S, connect Receipts machine clock deviation δ t^R；ε_PiAnd ε_φiRespectively indicate the observation noise of pseudorange and carrier phase；In double difference observation between also surplus station star Double difference geometric distance, double difference ionosphere residual effect, the troposphere of difference influences between star, and carrier phase ambiguity influences and observation Influence of noise；For error term remaining after formation double difference observation, the shadow of these errors is eliminated by way of parameter Estimation It rings；

11) cycle slips detection of double difference observation level；Using three poor method cycle slips detections；If to the cycle slips detection of un-differenced observation There is enough confidence, can skip this step；

12) filter temporal updates；Real-time coordinates are estimated, comprising: a) use Sequent least square method and Kalman filtering Method；Kalman filter model is related with observation selection, and for Dual Frequency Observation data, there are two types of observation models: based on no ionization The model and model based on non-combined observation of layer combination observation；

A) model based on no ionospheric combination observation eliminates the single order Xiang Ying in ionosphere by Dual-frequency Observations linear combination It rings, but the observation quantity without ionospheric combination observation halves, observation noise is 3 times of observation before combining；

B) it is directly handled using non-combined observation, every satellite is needed to estimate an Ionospheric Parameters；It stands star difference energy The influence for weakening ionosphere, according to non-combined model inference；For non-combined model, filter status vector is chosen are as follows:

X=[Δ r, δ_trop,z, I, N]^T；

Wherein Δ r is 3 × 1 increment of coordinate vectors, δ_trop,zFor zenith direction tropospheric delay, each survey station estimates a parameter； I is each satellite direction of visual lines ionosphere delay, and every visible satellite estimates one, and N is double-differential carrier phase fuzziness parameter, often A signal path estimates one；The case where for no ionospheric combination observation, filter status vector can be ignored ionosphere and prolong Slow parameter I；Static subscriber is positioned, increment of coordinate Δ r estimates as arbitrary constant；For dynamic positioning, random trip is used The equation of motion for walking process simulation or additional external constrains increment of coordinate parameter；δ_trop,zUse random walk pattern die It is quasi-, it realizes direction ionosphere delay I or is initialized using prior information, then simulated using random walk process；Carrier phase Double difference fuzziness is estimated when being continuously tracked using arbitrary constant model, needs parameter to reset when cycle slip occurs；Joining When examining star and changing star, specially treated is done to all fuzziness parameters；

Filter temporal renewal equation indicates are as follows:

In formulaP_t ^-Respectively t moment time updated quantity of state and its variance-covariance matrix, also referred to as forecast solution；For t-1 moment user location state filtering solution and its variance-covariance matrix；Φ (t, t-1) is that t-1 arrives t moment State-transition matrix；Q (t) is process noise matrix；For random walk parameter and random constant parameter, state-transition matrix is set For unit battle array；Diagonal matrix is set as n independent random walk process Q (t), arbitrary constant parameter Q (t) may be configured as Null matrix；

13) filter measurement updates；According to the analysis of step 10), pseudorange and carrier phase observation data line for double difference positioning Propertyization indicates are as follows:

E (y)=(J, M, Θ, Λ) (Δ r, δ_trop,z,I,N)^T；

E () is mathematic expectaion operator in formula, to f frequency point, the observation of s visible double difference satellite, for using non-group For the mathematical model for closing observation, J is the Jacobbian matrix of (2*s*f) × 3, and M is the column vector of (2*s*f) × 1, often A element corresponds to the troposphere projection function value of respective satellite；Θ is the matrix of (2*s*f) × s；Λ is (2*s*f) × (f*s) Matrix；For using the observation model of no ionospheric combination, other than ignoring the corresponding design matrix Θ of Ionospheric Parameters, institute There is observation linear combination to also result in observation number and reduces half than non-combined model；After linearisation, the non-combined observation The design matrix of equation is defined as:

A=(J, M, Θ, Λ)；

K_t=P_t ^-A(AP_t ^-A^T+R_t)^-1,

P_t ⁺=(I-K_tA)P_t ^-,

K in formula_tFor t moment filtering gain matrix, R_tFor t moment GNSS observation variance-covariance matrix, by empirical value or The method of variance components estimate, which calculates, to be obtained；

χ is constructed according to rear residual error valuation is tested²Inspected numberObey χ²(n-t) it is distributed, wherein n-t indicates freedom degree；IfThen think to test rear residual error by examining, otherwise it is assumed that testing rear residual test refusal；In formulaIt is significant Property level be α threshold value；R is GNSS observation variance-covariance matrix；

15) output forecast solution；If testing rear residual test not pass through, then it represents that there may be rough errors for observation；It is visited using data Survey method, which tests to rough error, to be rejected or abandons to use forecast solution instead using filter solutionIt is exported as final positioning result；

16) fuzziness is fixed；If filter solution by testing rear residual test, is obtained using the integer characteristic of carrier phase ambiguity Obtain the ambiguity fixed solution of centimetre class precision；The fixed method of fuzziness drops related algorithm LAMBDA using least square or will The mode that no ionospheric combination is decomposed into width lane ambiguity carries out fuzziness step by step and fixes；

State vector in step 13) is merged according to real parameters and two class of numeric parameter, then observational equation is simplified shown as E (y)=(B, Λ) (b, N)^T, corresponding filter state amountThe corresponding filter solution variance and covariance square of two class parameters Matrix representation are as follows:

The filter solution of fuzziness parameter is denoted asCommonly referred to as fuzziness real solution, by the real solution of fuzziness parameterAnd Its variance-covariance matrix Q_NNIt as input, is scanned for using LAMBDA method, obtains the optimal integer fixed solution of fuzzinessWith the integer fixed solution of suboptimum

17) fuzziness is examined；Positioning result is caused to generate flying spot since the fuzziness of mistake is fixed, position error is also than float-solution Greatly, it tests to the fixed result of fuzziness；It is examined using Ratio；

18) float-solution is exported；If the test fails by Ratio, then it is assumed that ambiguity fixed solution is unreliable, needs using filter solutionIt is exported as final positioning result；

19) fixed solution is exported；If Ratio upchecks, ambiguity fixed solution is used to export as positioning result；Fuzziness After being fixed as integer, corresponding real parametersIt may be updated as the condition solution of integer ambiguity, indicate are as follows:

Corresponding variance-covariance matrix may be expressed as:

If fuzziness is upchecked, useIt is used as last positioning result.

6. a kind of star based on low orbit satellite as described in claim 1 difference real-time accurate localization method based on low orbit satellite Star ground difference real-time accurate positioning system.

7. a kind of star using low orbit satellite described in claim 6 difference real-time accurate positioning system ocean real-time accurate it is fixed Position system.

8. a kind of star using low orbit satellite described in claim 6 difference real-time accurate positioning system remote districts in real time essence Close positioning system.

9. a kind of star using low orbit satellite as claimed in claim 6 difference real-time accurate positioning system natural calamity disaster area Real-time accurate positioning system.