CN106441283B - It is a kind of from principal mark compared with the more visual field star sensor data fusion methods of star vector grade - Google Patents
It is a kind of from principal mark compared with the more visual field star sensor data fusion methods of star vector grade Download PDFInfo
- Publication number
- CN106441283B CN106441283B CN201610891255.0A CN201610891255A CN106441283B CN 106441283 B CN106441283 B CN 106441283B CN 201610891255 A CN201610891255 A CN 201610891255A CN 106441283 B CN106441283 B CN 106441283B
- Authority
- CN
- China
- Prior art keywords
- star
- head
- installation matrix
- matrix
- fusion
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/02—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 by astronomical means
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C25/00—Manufacturing, calibrating, cleaning, or repairing instruments or devices referred to in the other groups of this subclass
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- General Physics & Mathematics (AREA)
- Manufacturing & Machinery (AREA)
- Astronomy & Astrophysics (AREA)
- Automation & Control Theory (AREA)
- Navigation (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
It is a kind of from principal mark compared with the more visual field star sensor data fusion methods of star vector grade, applied to more visual field star sensors, more visual field star sensors configure 2~3 optical heads, data fusion is carried out using the information that optical head is detected, the data fusion method comprises the steps of: that S1, multiple optical heads carry out autonomous whole day identification；S2, selection benchmark head, if having loaded installation matrix, jump procedure S3, if thening follow the steps S4 without the installation matrix loaded；S3, star Vector Fusion is carried out using cured installation matrix, exports posture, terminated；S4, carry out installation matrix from principal mark compared with, planet Vector Fusion of going forward side by side, export posture, terminate.Its advantage is that: the fusion method of star vector grade avoids the limitation of corner dimension between the head of star chart fusion method, the problem of fusion of posture grade lacks bring fusion attitude accuracy decline due to single head probe satellite is overcome, the attitude accuracy and posture for effectively promoting star sensor export robustness.
Description
Technical field
The present invention relates to aerospace measurement control technology fields, and in particular to it is a kind of from principal mark compared with the more visual field stars of star vector grade
Sensor data fusion method.
Background technique
A kind of sensitive device of the star sensor as high-precision, high reliability is that the space flight such as satellite, deep space probe fly
Row device provides attitude measurement information.However, the three unbalanced problem of axis precision existing for traditional monoscopic star sensor, affects
It, which develops, uses.In recent years, it is continued to bring out by more visual field star sensors of representative of French HYDRA, merges skill using information
Art comprehensively utilizes the detection information of multiple visual fields, efficiently solves the problems, such as that three axis precision are inconsistent, while star greatly improved
The precision of sensor becomes one of research and development direction of very Rotating Platform for High Precision Star Sensor.
The data fusion mode of current more visual field star sensors is broadly divided into two major classes.The first kind is that the front end of star chart grade is melted
It closes, document [1] (triangle star map recognition method [J] computer measurement of Xiong Xue, the Wang Qing based on more visual field star sensors and control
Make, 2014,22 (1): 225-228.) star chart of different visual fields is spliced by the big visual field of a width using the method for radial characteristic matching
Star chart, then fused star chart is identified.Document [2] (Zhang Hua, Sang Hongshi, Shen Xubang.A
double-FOV star sensor for high dynamic spacecraft navigation[C].Proceedings
Of the SPIE, MIPPR 2009, Yichang, China, 2009,7494:74940X-1~7.) then differed using optical axis
The optical head of 0.1 ° of two visual fields of size, big visual field are used to obtain the rough posture of spacecraft, and small field of view is then used to obtain
High-precision posture, while output posture is improved using the weighted average method of image co-registration or two width star chart nautical star positions
Precision, such can the detection information effectively to different visual fields merged, due to different visual field star pattern matchings and splicing need
Comprising wanting identical information, the setting angle between head is limited, constrains the promotion of optical axis direction precision.Second class is posture grade
Rear end fusion, AA-STR star sensor that Italian SELEX company releases and with the ASF series star of Byelorussia is quick provides
Attitude data fusion function.The posture for calculating different head first merges the attitude data of optical head in fusion output end,
Such method is quick and easy, but when some head leads to posture validation less because of detection star number, fusion attitude accuracy will drop significantly
Low or even posture validation can not embody the advantage of more visual field star sensors to the full extent.The HYDRA of the SODERN company of France
Star sensor merges fixed star vector data, and posture output accuracy is up to 1 ", efficiently solve the head of front-end convergence
Setting angle restricted problem, while improving the robustness of fusion posture output.Installation relation in HYDRA between head is necessarily
Face mark is compared with can not independently optimize, however star sensor is in use, due to the influence of the factors such as vibration, thermal deformation, head
Between installation matrix inevitably change, not can guarantee fusion posture precision and accuracy.
Summary of the invention
The purpose of the present invention is to provide it is a kind of from principal mark compared with the more visual field star sensor data fusion methods of star vector grade,
Firstly, non-referenced head is calculated to the installation matrix between benchmark head using the posture information of optical head under tracing mode, and
Using the weighted filtering algorithm more taken from principal mark compared with the installation matrix between head, while in real time more according to the working condition on head
New mark is compared with mode；It is then calculated according to the data fusion that the installation matrix after marking relatively carries out star vector grade, is finally known using whole
Other star Vector Message carries out quaternary number algorithm for estimating attitude algorithm, obtains the three consistent posture informations of axis precision.By to light
The fixed star vector for learning head is merged, and filtering algorithm is used, compared with the installation matrix between head, can both to pass through increase from principal mark
Setting angle between head promotes fusion accuracy, also can avoid merging attitude accuracy without star or few star bring due to optical head
The problem of decline, enhances robustness.
In order to achieve the above object, the invention is realized by the following technical scheme:
It is a kind of from principal mark compared with the more visual field star sensor data fusion methods of star vector grade, characterized in that this method application
In more visual field star sensors, which is configured at least two optical head, the letter detected using optical head
Breath carries out data fusion, and the data fusion method comprises the steps of:
S1, multiple optical heads carry out autonomous whole day identification；
S2, selection benchmark head, if having loaded installation matrix, jump procedure S3, if without the installation matrix loaded,
Execute step S4；
S3, star Vector Fusion is carried out using cured installation matrix, exports posture, terminated；
S4, carry out installation matrix from principal mark compared with, planet Vector Fusion of going forward side by side, export posture, terminate.
It is above-mentioned from principal mark compared with the more visual field star sensor data fusion methods of star vector grade, wherein the step S1
Specifically include:
S11, more visual field star sensors power on；
S12, multiple optical heads carry out autonomous whole day identification, if whole day identifies successfully, execute step S2, otherwise unidentified
Head continues whole day identification.
It is above-mentioned from principal mark compared with the more visual field star sensor data fusion methods of star vector grade, wherein the step S3
Specifically include:
S31, tracking is entered when one of head completion whole day identifies；
S32, cured installation matrix information is directly utilized, optical axis direction is supplied to other heads, helps other heads
Realize tracking in portion；
S33, judge whether other heads all track successfully, if tracking successfully, then be used directly cured installation matrix
It carries out star Vector Fusion and calculates fusion posture, if tracking failure, thens follow the steps S4, be switched to from principal mark compared with mode.
It is above-mentioned from principal mark compared with the more visual field star sensor data fusion methods of star vector grade, wherein the step S4
Specifically include:
S41, calculate the continuous installation matrix for clapping non-referenced head to benchmark head under stable situation more；
S42, installation matrix is optimized using weighted filtering mode, and the transfer square as star vector that is fixed
Battle array；
Formula is as follows, and calculates the star vector on non-referenced head to the star vector on benchmark head in this, as transfer matrix:
Wherein: T (n) is the n-th installation matrix clapped, and T ' (n+1) is the installation matrix that the (n+1)th bat filtering updates, anFor power
Weight coefficient；Len is the length of filter, and Len=Ne-NS, Len takes 50~100, NeCurrently to clap
Frame number, NsFrame number is clapped for the starting of filter window；
Assuming that the star vector of non-referenced lower identification is wi(i=1,2 ..., n), corresponding star vector matrix are W3×n, non-referenced
The installation matrix on head to benchmark head is T, then the star vector for being transferred to reference head subordinate is TW；
S43, it is calculated according to the data fusion that the installation matrix after marking relatively carries out star vector grade, obtains the perseverance of reference head subordinate
Star information；
S44, quaternary number algorithm for estimating attitude algorithm is carried out using the identified fixed star information of whole of reference head subordinate, obtained
To the posture information of fusion.
It is above-mentioned from principal mark compared with the more visual field star sensor data fusion methods of star vector grade, wherein the step S42
In, specific step is as follows for the filtering of installation matrix weights:
S421, the installation matrix T (n) that non-referenced head to benchmark head is calculated according to output quaternary number, according to formulaThe installation matrix of next bat is calculated, it is still real for calculating the installation matrix of fusion posture at this time
When the T (n) that calculates；
If S422, multiple optical heads are always maintained at tracking mode, according to formula Installation matrix is updated, after the length bat for updating filter, solidification installation matrix, and it is used for star Vector Fusion number
According to calculating；
Matrix is installed between S423, the Attitude Calculation head according to every racket head portion, and is compared with cured installation matrix, is calculated
The difference of the two continues to use cured installation matrix if difference is less than 0.001 °, if it is much more continuous clap calculate in real time
Installation matrix and cured installation matrix differ by more than 0.001 °, then filtering calculates installation matrix again.
It is above-mentioned from principal mark compared with the more visual field star sensor data fusion methods of star vector grade, wherein the step
In S422:
If some non-referenced head continuously more claps tracking failure, the installation matrix weight on the non-referenced head to benchmark head
Newly marked compared with.
It is above-mentioned from principal mark compared with the more visual field star sensor data fusion methods of star vector grade, in which: more visual fields
Star sensor is configured with 3 optical heads.
Compared with the prior art, the present invention has the following advantages:
1, the installation matrix that user's offer can be used in the present invention carries out Vector Fusion, can also relatively install square from host computer mark
Battle array, using flexible；
2, the fusion method of star vector grade avoids the limitation of corner dimension between the head of star chart fusion method, also overcomes
The problem of fusion of posture grade lacks bring fusion attitude accuracy decline due to single head probe satellite, effectively promote star sensor
Attitude accuracy and posture output robustness.
Detailed description of the invention
Fig. 1 is star vector grade data fusion schematic diagram of the invention；
Fig. 2 is installation matrix process for using figure of the invention；
Fig. 3 is the flow chart of fusion method of the invention；
Fig. 4 is posture curve graph of the present invention in experiment simulation under the quadrature condition of head；
Fig. 5 is 1 attitude error curve of head and tracking star number in the emulation experiment of the embodiment of the present invention；
Fig. 6 is 2 attitude error curve of head and tracking star number in the emulation experiment of the embodiment of the present invention；
Fig. 7 is the three-axis attitude error curve in the emulation experiment of the embodiment of the present invention after data fusion.
Specific embodiment
The present invention is further elaborated by the way that a preferable specific embodiment is described in detail below in conjunction with attached drawing.
Precision on traditional monoscopic star sensor optical axis direction is poor, and the present invention proposes a kind of more visual field star sensors
It is then optical axis direction precision to be promoted by merging the fixed star information of multiple visual fields, and then guarantee that the precision of three axis is consistent.Star arrow
The main thought of amount fusion is the fixed star information (three-dimensional coordinate under celestial coordinate system) by each head in tenacious tracking, root
According to the installation relation between head, it is transformed into reference head subordinate, then carries out the estimation of quaternary number using whole effective fixed star information
(QUEST) algorithm obtains fused attitude quaternion.
Based on the above principles, as shown in figure 3, the invention proposes it is a kind of from principal mark compared with the more visual field stars of star vector grade it is sensitive
Device data fusion method, this method are applied to more visual field star sensors, which can configure 2~3 optical heads,
Data fusion is carried out using the information that optical head is detected, the data fusion method comprises the steps of:
S1, multiple optical heads carry out autonomous whole day identification；Specifically, step S1 includes: S11, more visual field stars are sensitive
Device powers on；S12, multiple optical heads carry out autonomous whole day identification, if whole day identifies successfully, execute step S2, otherwise unidentified
Head continues whole day identification.
S2, selection benchmark head, if having loaded installation matrix, jump procedure S3, if without the installation matrix loaded,
Execute step S4；The installation matrix that user's offer can be used in this step carries out Vector Fusion, can also relatively pacify from host computer mark
Fill matrix, using flexible；In the present embodiment, as shown in Figure 1, adding by taking the more visual field star sensors for configuring 3 optical heads as an example
With explanation, it is assumed that setting 3 optical heads A, B, C, benchmark head are A, other two heads parts are B and C, and between head
When setting angle is mutually orthogonal, syncretizing effect is best；
S3, star Vector Fusion is carried out using cured installation matrix, exports posture, terminated；Specifically, step S3 has
Body includes: S31, entering tracking when one of head completion whole day identifies；S32, cured installation matrix is directly utilized
Optical axis direction is supplied to other heads, other heads is helped to realize tracking by information；S33, judge other heads whether all with
Track success, if tracking successfully, the cured installation matrix that then be used directly carries out star Vector Fusion and calculates fusion posture, if tracking
Failure, thens follow the steps S4, is switched to from principal mark compared with mode.
S4, carry out installation matrix from principal mark compared with, planet Vector Fusion of going forward side by side, export posture, terminate.
Since the accuracy of installation matrix directly affects the precision of star vector median filters, and then affect the essence of fusion posture
Degree, currently, the installation matrix of more visual field star sensors is mostly ground mark compared with solidifying before emitting in the quick inside of star is in-orbit use.
This method precision is higher, marks relatively complicated compared with process and more demanding for the thermomechanical of star sensor'support, and in-orbit is difficult to
It is relatively corrected from principal mark.On this basis, the invention proposes above-mentioned installs the autonomous of matrix as described in step S3~S4
Mark as shown in connection with fig. 2, is divided into two kinds of situations, situation 1 compared with scheme: if be loaded with correctly installation matrix in star sensor,
The installation matrix for then directlying adopt user's offer completes whole day identification when one of head and enters tracking, then directly utilizes
Matrix information is installed, optical axis direction is supplied to other heads, it is helped to realize tracking, shortens and calculates the time, in the present embodiment,
If the continuous 5 bat whole day identification in the head is unsuccessful, switch to from principal mark compared with mode；Situation 2: if the installation matrix without priori
Information, into from principal mark compared with mode；When head enters under tracing mode, according to the steady-state attitude on each head output calculating head
Between installation matrix, using filtering method update head between installation matrix.
The method of above-mentioned steps S4 culminant star vector grade data fusion is: according to the information of head A, B, C, being independently respectively completed
Identify and calculate the posture on each head；Non-referenced head is calculated to benchmark head according to the tracking output under stable situation
Installation quaternary number or the installation matrix that is directly measured using the quick ground of star；B, all tracking star Vector Messages of C are according to transfer square
Battle array, is converted under the A coordinate system of head.
Again due to from principal mark compared under mode, when the appearance posture bounce for single head occur, the installation matrix of calculating
Also biggish error is had, the precision of fused posture is directly affected.Therefore, it is steady can to calculate continuous more bats for this method proposition
The installation matrix on non-referenced head to benchmark head in the case of state, updates installation matrix by the way of weighted filtering, and by its
The fixed transfer matrix as calculating star vector.
Specifically, the step S4 includes:
S41, calculate the continuous installation matrix for clapping non-referenced head to benchmark head under stable situation more；
S42, installation matrix is optimized using weighted filtering mode, and the transfer square as star vector that is fixed
Battle array；
Formula is as follows, and calculates the star vector on non-referenced head to the star vector on benchmark head in this, as transfer matrix:
Wherein: T (n) is the n-th installation matrix clapped, and T ' (n+1) is the installation matrix that the (n+1)th bat filtering updates, anFor power
Weight coefficient；Len is the length of filter, and Len=Ne-Ns, Len takes 50~100, NeCurrently to clap
Frame number, NSFrame number is clapped for the starting of filter window；
Assuming that the star vector of non-referenced lower identification is wi(i=1,2 ..., n), corresponding star vector matrix are W3×n, non-referenced
The installation matrix on head to benchmark head is T, then the star vector for being transferred to reference head subordinate is TW.
S43, it is calculated according to the data fusion that the installation matrix after marking relatively carries out star vector grade, obtains the perseverance of reference head subordinate
Star information；
S44, quaternary number algorithm for estimating attitude algorithm is carried out using the identified fixed star information of whole of reference head subordinate, obtained
To the posture information of fusion.
In the step S42, specific step is as follows for the filtering of installation matrix weights:
S421, the installation matrix T (n) that non-referenced head to benchmark head is calculated according to output quaternary number, according to formulaThe installation matrix of next bat is calculated, it is still real for calculating the installation matrix of fusion posture at this time
When the T (n) that calculates；
In the present embodiment, it is assumed that the attitude quaternion of head A is Q1=[q10,q11,q12,q13], the posture quaternary of head B
Number is Q2=[q20,q21,q22,q23], the transfer quaternary number of head B to A is Qt=[qt0,qt1,qt2,qt3], head B to head A
Installation matrix be T, then
qt0=q10*q20+q11*q21+q12*q22+q13*q23 (2)
qt1=-q10*q21+q11*q20-q12*q23+q13*q22 (3)
qt2=-q10*q22+q12*q20+q11*q23-q13*q21 (4)
qt3=-q10*q23+q13*q20-q11*q22+q12*q21 (5)
After transfer quaternary number is normalized, installation matrix T is calculated according to formula (6)
If S422, multiple optical heads are always maintained at tracking mode, according to formula Installation matrix is updated, after length (Length) bat for updating filter, solidification installation matrix, and sweared for star
Measure the calculating of fused data；
Matrix is installed between S423, the Attitude Calculation head according to every racket head portion, and is compared with cured installation matrix, is calculated
The difference of the two continues to use cured installation matrix if difference is less than 0.001 °, if it is much more continuous clap calculate in real time
Installation matrix and cured installation matrix differ by more than 0.001 °, then filtering calculates installation matrix again.
In the step S422:
If some non-referenced head continuously more claps tracking failure, the installation matrix weight on the non-referenced head to benchmark head
Newly marked compared with.
A large amount of emulation experiment has been carried out to method of the invention using electric star analogue data.With the angle difference between head
It is illustrated for being 90 °, Fig. 4 gives the output posture curve before and after data fusion, it can be seen from the figure that before number fusion
The variation of the precision of pitch axis and yaw axis less, due to using the blending algorithm of vector grade, is determined appearance star number and is increased therewith, melt afterwards
The attitude accuracy of the axis of rolling has before relatively merging and is significantly promoted after conjunction.
Table 1 gives angle between head, fusion front and back output attitude accuracy
Situation of change.
The output attitude accuracy of the fusion of table 1 front and back
As it can be seen from table 1 fused three axis precision relatively merges when angle is 90 °, 60 ° and 30 ° of angle between head
The preceding promotion for having certain procedures, wherein the precision improvement for rolling axis direction is particularly evident, and with optical axis included angle between head
Reduce, the improvement degree for merging attitude accuracy also reduces therewith.
Further to test the validity for verifying fusion method, a kind of more visual field star sensors are further devised in experiment
Principle prototype meter, and carried out outfield and seen star.Principle prototype include two optical heads, electronic circuit box, high speed communication cables,
Default setting head 1 is benchmark head, wherein optical head uses 14.5 ° of visual fields, and detector resolution is 2048 × 2048,
The calibration result in two title portions is as shown in table 2.
The interference of gas light in order to prevent, the optical axis of two optical heads differs 40 ° and is respectively directed to difference when field observation
Its area.
The calibration result of 2 optical head of table
It chooses certain segment data and carries out precision analysis, the error assembly of noise equivalent angle, high-frequency noise is analyzed.Fig. 5-
Head 1, head 2 and the error curve diagram for merging posture is set forth in Fig. 6.
Wherein Fig. 5 be head 1 attitude error curve and tracking star number, Fig. 6 be head 2 attitude error curve and with
Track star number, Fig. 7 are the three-axis attitude error curves after data fusion, wherein in order to be compared conveniently, by 1 conduct of head
Benchmark head respectively illustrates fused attitude error curve and attitude error curve in figure, consistent with the curve in Fig. 7.
According to precision assessment method, data are calculated, calculated result is as shown in table 3
3 precision statistics table of table
From Fig. 5 and table 3, it can be seen that, the three-axis attitude precision after data fusion has promotion, wherein axis of rolling side
To precision improvement it is the most obvious.Three axis noise equivalent angles (NEA) respectively reach 1.12 rads, 1.01 rads, 1.81 rads.
From table 3 it can be seen that the blending algorithm proposed can significantly improve the precision of three axis.The posture and reference head of fusion
The posture in portion (OH1) is compared, and rolling, the precision improvement in axis direction is the most obvious, and noise equivalent angle is promoted by 12.15 rads
To 1.81 rads, about 6.7 times are improved.In pitching and yaw both direction, precision also while is to a certain degree promoted, with OH2 phase
Than pitching, yaw and rolling dynamic triaxial precision improve 1.15 times, 1.27 and 13 times respectively.
From fig. 5, it can be seen that due to using blending algorithm of the installation from principal mark technology and star vector grade, when single head
When posture is beated, fusion method of the invention can guarantee that fused posture is unaffected, effectively improve fusion appearance
The robustness of state output.
Data are analysis shows data fusion method of the invention can effectively promote the precision of three-axis attitude, in two heads
When portion's angle is 40 °, axis of rolling directional precision improves about 6.7 times.Meanwhile it is effective compared with operational version from principal mark to install matrix
The stability and precision that ensure that fusion posture, when single head pose is beated, fusion posture can also stablize output.
It is discussed in detail although the contents of the present invention have passed through above preferred embodiment, but it should be appreciated that above-mentioned
Description is not considered as limitation of the present invention.After those skilled in the art have read above content, for of the invention
A variety of modifications and substitutions all will be apparent.Therefore, protection scope of the present invention should be limited to the appended claims.
Claims (6)
1. it is a kind of from principal mark compared with the more visual field star sensor data fusion methods of star vector grade, which is characterized in that this method application
In more visual field star sensors, which configures 2~3 optical heads, the information detected using optical head
Data fusion is carried out, the data fusion method comprises the steps of:
S1, multiple optical heads carry out autonomous whole day identification；
S2, selection benchmark head, if having loaded installation matrix, jump procedure S3, if being executed without the installation matrix loaded
Step S4；
S3, star Vector Fusion is carried out using cured installation matrix, exports posture, terminated；
S4, carry out installation matrix from principal mark compared with, planet Vector Fusion of going forward side by side, export posture, terminate；
Wherein, the step S3 specifically includes:
S31, tracking is entered when one of head completion whole day identifies；
S32, cured installation matrix information is directly utilized, optical axis direction is supplied to other heads, helps other heads real
Now track；
S33, judge whether other heads all track successfully, if tracking successfully, the cured installation matrix that then be used directly is carried out
Star Vector Fusion simultaneously calculates fusion posture, if tracking failure, thens follow the steps S4, be switched to from principal mark compared with mode.
2. as described in claim 1 from principal mark compared with the more visual field star sensor data fusion methods of star vector grade, feature exists
In the step S1 specifically includes:
S11, more visual field star sensors power on；
S12, multiple optical heads carry out autonomous whole day identification, if whole day identifies successfully, execute step S2, otherwise unidentified head
Continue whole day identification.
3. as described in claim 1 from principal mark compared with the more visual field star sensor data fusion methods of star vector grade, feature exists
In the step S4 specifically includes:
S41, calculate the continuous installation matrix for clapping non-referenced head to benchmark head under stable situation more；
S42, installation matrix is optimized using weighted filtering mode, and the transfer matrix as star vector that is fixed；
Formula is as follows, and calculates the star vector on non-referenced head to the star vector on benchmark head in this, as transfer matrix:
Wherein: T (n) is the n-th installation matrix clapped, and T ' (n+1) is the installation matrix that the (n+1)th bat filtering updates, anFor weight system
Number；Len is the length of filter, and Len=Ne-Ns, Len takes 50~100, NeFor the frame currently clapped
Number, NsFrame number is clapped for the starting of filter window；
Assuming that the star vector of non-referenced lower identification is wi(i=1,2 ..., n), corresponding star vector matrix are W3×n, non-referenced head
Installation matrix to benchmark head is T, then the star vector for being transferred to reference head subordinate is TW；
S43, it is calculated according to the data fusion that the installation matrix after marking relatively carries out star vector grade, obtains the fixed star letter of reference head subordinate
Breath；
S44, quaternary number algorithm for estimating attitude algorithm is carried out using the identified fixed star information of whole of reference head subordinate, is melted
The posture information of conjunction.
4. as claimed in claim 3 from principal mark compared with the more visual field star sensor data fusion methods of star vector grade, feature exists
In in the step S42, specific step is as follows for the filtering of installation matrix weights:
S421, the installation matrix T (n) that non-referenced head to benchmark head is calculated according to output quaternary number, according to formulaThe installation matrix of next bat is calculated, it is still real for calculating the installation matrix of fusion posture at this time
When the T (n) that calculates；
If S422, multiple optical heads are always maintained at tracking mode, according to formula It updates
Matrix is installed, after the length bat for updating filter, solidification installation matrix, and it is used for the calculating of star Vector Fusion data；
Both matrix is installed between S423, the Attitude Calculation head according to every racket head portion, and is compared with cured installation matrix, calculate
Difference continue to use cured installation matrix if difference is less than 0.001 °, if the much more continuous installations calculated in real time clapped
Matrix and cured installation matrix differ by more than 0.001 °, then filtering calculates installation matrix again.
5. as claimed in claim 4 from principal mark compared with the more visual field star sensor data fusion methods of star vector grade, feature exists
In in the step S422:
If some non-referenced head continuously more claps tracking failure, the installation matrix on the non-referenced head to benchmark head again into
Rower compared with.
6. as described in claim 1 from principal mark compared with the more visual field star sensor data fusion methods of star vector grade, feature exists
In: more visual field star sensors are configured with 3 optical heads.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610891255.0A CN106441283B (en) | 2016-10-12 | 2016-10-12 | It is a kind of from principal mark compared with the more visual field star sensor data fusion methods of star vector grade |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610891255.0A CN106441283B (en) | 2016-10-12 | 2016-10-12 | It is a kind of from principal mark compared with the more visual field star sensor data fusion methods of star vector grade |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106441283A CN106441283A (en) | 2017-02-22 |
| CN106441283B true CN106441283B (en) | 2019-03-26 |
Family
ID=58173762
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610891255.0A Active CN106441283B (en) | 2016-10-12 | 2016-10-12 | It is a kind of from principal mark compared with the more visual field star sensor data fusion methods of star vector grade |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106441283B (en) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107084715B (en) * | 2017-04-10 | 2019-06-18 | 北京控制工程研究所 | A kind of asynchronous multiframe star chart fusion method of star sensor |
| CN107389089B (en) * | 2017-05-31 | 2019-12-20 | 上海航天控制技术研究所 | Satellite-borne multi-probe high-precision star sensor testing method |
| CN107344630B (en) * | 2017-06-09 | 2019-11-12 | 北京空间飞行器总体设计部 | Configuration is laid out the method for determination on a kind of more visual field star sensor stars |
| CN107228683B (en) * | 2017-06-27 | 2020-04-10 | 上海航天控制技术研究所 | Slow-variation error real-time on-orbit correction method among multiple star sensors |
| CN109489657B (en) * | 2018-11-22 | 2022-04-19 | 北京航天计量测试技术研究所 | Double-view-field star sensor attitude fusion measuring device and method |
| CN109945854B (en) * | 2019-03-26 | 2020-12-25 | 上海微小卫星工程中心 | Method for fusing star sensor data based on quaternion difference |
| CN111207772B (en) * | 2020-01-14 | 2021-07-13 | 上海卫星工程研究所 | Method for testing light path and polarity of multi-head star sensor |
| CN114234962B (en) * | 2021-11-10 | 2023-09-12 | 上海航天控制技术研究所 | Multi-star sensor on-orbit thermal deformation correction method, storage medium and electronic equipment |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101592490B (en) * | 2009-07-06 | 2010-12-01 | 北京航空航天大学 | Self-adaptive controlled-array star sensor |
| CN102749089B (en) * | 2012-07-13 | 2015-02-11 | 哈尔滨工业大学 | Method for determining three-probe star sensor gesture |
| CN102865866B (en) * | 2012-10-22 | 2015-01-28 | 哈尔滨工业大学 | Satellite attitude determination method and attitude determination error analytical method based on two star sensors |
| US9073648B2 (en) * | 2013-02-15 | 2015-07-07 | The Boeing Company | Star tracker rate estimation with kalman filter enhancement |
| CN103868510B (en) * | 2014-03-27 | 2016-08-17 | 北京控制工程研究所 | A kind of quickly whole day star chart autonomous fixed star recognition methods |
| CN104406583B (en) * | 2014-12-05 | 2017-02-22 | 中国人民解放军63680部队 | Combined defining method for carrier attitude of double-star sensor |
| CN105318871B (en) * | 2015-11-09 | 2018-03-30 | 中国人民解放军63680部队 | Double star sensor carrier installs matrix dynamic calibrating method |
-
2016
- 2016-10-12 CN CN201610891255.0A patent/CN106441283B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN106441283A (en) | 2017-02-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106441283B (en) | It is a kind of from principal mark compared with the more visual field star sensor data fusion methods of star vector grade | |
| CN107289933B (en) | Double card Kalman Filtering navigation device and method based on MEMS sensor and VLC positioning fusion | |
| CN105974940B (en) | Method for tracking target suitable for aircraft | |
| EP3407294A1 (en) | Information processing method, device, and terminal | |
| CN105021184B (en) | It is a kind of to be used for pose estimating system and method that vision under mobile platform warship navigation | |
| CN104704384B (en) | Specifically for the image processing method of the positioning of the view-based access control model of device | |
| CN103718213B (en) | Automatic scene is calibrated | |
| CN109991636A (en) | Map constructing method and system based on GPS, IMU and binocular vision | |
| US9378559B2 (en) | System and method for motion estimation | |
| CN105737858B (en) | A kind of Airborne Inertial Navigation System attitude parameter calibration method and device | |
| CN109270545A (en) | A kind of positioning true value method of calibration, device, equipment and storage medium | |
| CN107478223A (en) | A kind of human body attitude calculation method based on quaternary number and Kalman filtering | |
| CN106056664A (en) | Real-time three-dimensional scene reconstruction system and method based on inertia and depth vision | |
| JP4956456B2 (en) | Image processing apparatus and image processing method | |
| CN109671105A (en) | A kind of the tracking restoration methods and device of vision navigation system | |
| CN107289910A (en) | A kind of light stream alignment system based on TOF | |
| CN110455301A (en) | A kind of dynamic scene SLAM method based on Inertial Measurement Unit | |
| CN102879010A (en) | Method and system for a self-calibrated multi-magnetometer platform | |
| CN109000633A (en) | Human body attitude motion capture algorithm design based on isomeric data fusion | |
| CN109997150A (en) | System and method for classifying to roadway characteristic | |
| CN108132053B (en) | Pedestrian track construction method and system and inertia measurement device | |
| CN109425348A (en) | A kind of while positioning and the method and apparatus for building figure | |
| CN109931952A (en) | The direct analytic expression coarse alignment method of inertial navigation under the conditions of unknown latitude | |
| CN104913775B (en) | Measurement method, unmanned plane localization method and the device of unmanned plane distance away the ground | |
| KR101755687B1 (en) | System of estimating camaera pose and method thereof |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |