CN106595673A - Space multi-robot autonomous navigation method for geostationary orbit target action - Google Patents

Abstract

The invention discloses a space multi-robot autonomous navigation method for geostationary orbit target action. The method comprises the following steps: designing the formation flight configuration and the orbit parameters of two space robots (set as a primary star and a component star) by adopting a GEO target satellite as an on-orbit service target, and establishing an autonomous navigation system state model according to a relative orbit dynamic model of the satellite in an earth's core inertia coordinate system; providing theoretic illumination conditions and imaging conditions needed by observation of the component star by a primary star sensor; calculating the theoretic azimuth and the theoretic pitch angle of the component star to the primary star, adjusting the direction of a true star sensor optical axis to be consistent with a theoretic direction, truly observing the component star, and establishing an observation equation adopting relative unit direction vector and distance as observations; and establishing the relative position and speed of the primary star by using Unscented Kalman filtering. The method belongs to the technical field of space navigation, provides high-precision navigation information for GEO formation flight of the satellite, and provides references for design of an autonomous navigation system.

Description

In the face of the space multi-robot self-determination air navigation aid of geostationary orbit object run

Technical field

The invention belongs to spacecraft in-orbit service space measurement field, more particularly to it is a kind of in the face of geostationary orbit target The space multi-robot self-determination air navigation aid of operation.

Background technology

Geostationary orbit (Geostationary orbit, GEO) is the unique track resources of the mankind, positioned at this The satellite (abbreviation GEO satellite) of track, area coverage are big, and are static relative to ground, in communication, navigation, early warning, meteorology More and more important effect is increasingly being played just in civil and military field.For some tasks, need multiple Satellite Networkings, Formation constellation, such as DSP (Defense support program) ballistic missile early warning satellite in the U.S., have remained 5 (3 works on GEO tracks Make, 2 are standby) satellite；The high rail section of its space-based infrared system (SBIRS) also includes 4 GEO satellites and 2 highly elliptic orbits Satellite.The dipper system built, is the GPS of China's independent development capability, independent operating, by 5 static rails Road satellite and 30 other types satellite compositions.Compared with developed countries, the on-orbit fault rate of China's satellite is higher, loses in recent years The important GEO satellite of effect includes No. two satellites of prosperous promise (2006), number 04 star of the Big Dipper (2007), Nigeria's star (2007 Year transmitting, fails for 2008), Big Dipper G2 stars (2009) etc., had a strong impact on the development of China's Aerospace Technology.The particularly Big Dipper The failure of G2 stars (one of 5 GEO satellites in Beidou satellite navigation system), have impact on the networking process of whole navigation system, makes China have in 2012 and transmit one substitute star (G2R, also known as G6) fixed point with the position of the failure astrology away from 0.2 ° Put.

To ensure that in-orbit spacecraft is run steadily in the long term, and protect GEO track resources, it is necessary to which development with robot for space is In-orbit service technology for the purpose of means, satellite maintenence and space trash removing.As orbit altitude is high, the 3rd body gravitation can not Ignore, it is also very high for the transmitting of the robot for space itself of GEO services, management and maintenance cost.Therefore need space multimachine Device people system carries out in-orbit maintenance to many GEO satellites in a certain segmental arc, can greatly save the cost of satellite maintenance, improves Rail efficiency of service.Current Space Robot System, the ETS-VII, Orbital Express including emitted and in-orbit demonstration, and The systems such as FREND, the DEOS for carrying out, are that, using single satellite as service object, and service content is more single, it is impossible to full Purpose of the foot in many spacecraft maintainable technology on-orbits of GEO tracks, therefore study the space multirobot towards GEO track in-orbit services System is imperative.

In order to space multirobot carries out formation flight, fixed point and keeps and position control altogether near the GEO satellite, it is necessary to first Position and the attitude information of satellite first can be obtained in real time, and can not produce interference to facing star, as GEO satellite generally exists There is observation signal deficiency of navigating in 36000km height, space multirobot：1. commonly use GNSS navigation mode and there is navigation Signal is weak, the earth blocks the serious problems such as few with visible satellite；2. other independent navigation modes：Earth's magnetic field cannot use, radar Altimeter and celestial navigation cannot provide high accuracy navigation information, all be difficult to meet navigation request, therefore this as observation information Allow for the new observation procedure of research space multirobot and seem urgent important.

The content of the invention

Goal of the invention：The present invention causes the relatively low problem of navigation accuracy for observation information deficiency, proposes a kind of in the face of ground The space multi-robot self-determination air navigation aid of ball stationary orbit object run, using the autonomous Continuous Observation relative direction of star sensor The method of vector, the space multirobot for in-orbit service provide high accuracy relative observation information.

Technical scheme：A kind of space multi-robot self-determination air navigation aid in the face of geostationary orbit object run, step It is as follows：

(1) with GEO target satellites as in-orbit service object, two spaces robot is set to into primary and component, if Meter primary and component formation flight configuration and orbit parameter；

(2) according to satellite relative motion dynamics model under geocentric inertial coordinate system, set up autonomous navigation system state mould Type；

(3) according to the primary and component relative distance for calculating, judge that component is no and meet star sensor observed range requirement, it is full It is sufficient then enter step (4), otherwise into step (12)；

(4) according to the sun, the earth and component three's position relationship for resolving, judge whether component is in solar irradiation area, It is then to enter step (5), otherwise into step (12)；

(5) according to the earth, primary and component three's position relationship for resolving, judge the no entrance star sensor visual field of the earth, It is then to enter step (6), otherwise into step (12)；

(6) according to the component for calculating can the apparent magnitude, judge component can the apparent magnitude whether be less than star sensor Observable threshold value, It is then to enter step (7), otherwise into step (12)；

(7) angle is pointed to star sensor optical axis according to the relative primary direction vector in component for calculating, whether judges component In star sensor field range, it is then to enter step (8), otherwise calculates and pointed to using universal axial adjustment star sensor optical axis Afterwards, continue to judge, be then to enter step (8), otherwise into (12)；

(8) according to the component for calculating in star sensor two dimension image planes battle array coordinate, judge component whether in star sensor two dimension In image planes battle array, it is then to enter step (9), otherwise into step (12)；

(9) theory orientation vector azimuth and the angle of pitch of the component with respect to primary is calculated, into step (10)；

(10) adjust the sensing of star sensor true optical axis consistent with theory orientation, component is truly observed, son is calculated Star relative satellite true directions vector, sets up the observational equation with unit direction vector and distance as observed quantity, into step (11)；

(11) to the state equation set up and observational equation discretization, estimated using Unscented Kalman filtering algorithms Meter satellite position and speed；

(12) terminate observation.

Further, the orbit parameter in the step (1) includes semi-major axis of orbit a, orbital eccentricity e, orbit inclination angle I, right ascension of ascending node Ω, argument of perigee ω, time of perigee passage t_p。

Further, it is as follows autonomous navigation system state model process to be set up in the step (2)：

Under geocentric inertial coordinate system, when primary positional distance is more than component and primary relative distance, foundation is defended Star relative target component dynamics of orbits model

Wherein, δ r⁽¹⁰⁾With δ v⁽¹⁰⁾For component relative satellite direction vector, r⁽⁰⁾And r⁽¹⁾For satellite and component position vector, μ_eFor Gravitational coefficient of the Earth, a_fAffect for perturbative force；

Definition status variable x=[(δ r⁽¹⁰⁾)^T(δv⁽¹⁰⁾)^T]^T, set up autonomous navigation system state model；

Wherein, f [x (t), t] is mission nonlinear continuous state transfer function, and w (t) is state-noise.

Further, judge whether component meets star sensor observed range requirement process in the step (3) as follows：

Primary is calculated with respect to component apart from δ r⁽¹⁰⁾, judge whether which meets condition

L_min≤δr⁽¹⁰⁾≤L_max (3)

Wherein, δ r⁽¹⁰⁾=| δ r⁽¹⁰⁾|=| r⁽¹⁾-r⁽⁰⁾|, r⁽⁰⁾And r⁽¹⁾For primary and component position vector；L_minAnd L_maxFor Minimum and maximum distance needed for observing between star.

Further, judge whether component is in solar irradiation area process in the step (4) as follows：

Analysis earth shaded region and component travel through the critical condition of the shaded region, if component position vector r⁽¹⁾With position of sun vector r^(sun)Angle is ψ, and it is ψ that component enters and leaves the critical angle of earth's shadow scope_cri, then component Being in solar irradiation area needs to meet condition：

ψ ＜ ψ_cri (4)。

Further, judge in the step (5) whether the earth enters star sensor visual field process as follows：

If primary position vector r⁽⁰⁾Component direction vector δ r relative with primary⁽¹⁰⁾Angle be θ, being blocked by the earth causes The excessively weak critical condition of background light is component with respect to primary direction vector δ r⁽¹⁰⁾It is tangent with earth edge, define this critical folder Angle is θ_cri, then the earth be introduced into star sensor viewing conditions and be：

θ ＞ θ_cri (5)。

Further, judge in the step (6) component can the apparent magnitude whether be less than star sensor Observable threshold process It is as follows：

Introduce can the apparent magnitude analyze the observability of component, magnitude value is less, shows that celestial body is brighter；Conversely, celestial body is then darker； If star sensor Observable threshold value is m_thr, component can the apparent magnitude be m, component be observed its can the apparent magnitude need to meet condition

M ＜ m_thr (6)。

Further, judge whether component is as follows in star sensor field range internal procedure in the step (7)：

If component is with respect to primary direction vector δ r⁽¹⁰⁾With star sensor optical axis pointing vectorAngle isStar sensor The angle of visual field is FOV, then direction vector δ r⁽¹⁰⁾In star sensor field range need to meet condition

If relative vector δ r⁽¹⁰⁾Not in field range, calculate and pointed to using universal axial adjustment star sensor optical axis, make Which enters field range, if visual field can not be still entered after rotating, cannot observe.

Further, judge in the step (8) whether component is as follows in star sensor two dimension image planes battle array internal procedure：

According to the relative primary direction vector δ r in component⁽¹⁰⁾It is projected in the geometrical relationship of star sensor two dimension image planes battle array, Xie Qi Coordinate isIf two-dimentional image planes array length degree and width are respectively IP_longthAnd IP_width, then component is in image plane coordinate Need to meet condition

Further, the step (9) operator star of falling into a trap is specially with the angle of pitch with respect to primary direction vector and azimuth：

Component is with respect to primary unit direction vectorBy star sensor obtain, obtain component with respect to primary azimuth angle alpha with bow Elevation angle δ, primary and component relative distance | δ r⁽¹⁰⁾| obtained by inter-satellite link, component is obtained with respect to primary theory orientation vector delta r⁽¹⁰⁾

Wherein,

Component is described by azimuth and the angle of pitch with respect to primary orientation, in satellite body coordinate system o_b-x_by_bz_bIn, definition Azimuth angle alpha is δ r⁽¹⁰⁾In o_b-y_bz_bThe projection of plane and y_bAxle clamp angle, angle of pitch δ are δ r⁽¹⁰⁾With x_bAxle clamp angle, is expressed as

Wherein, It is geocentric inertial coordinate system opposing body's coordinate It is pose transformation matrix.

Further, the step (10) is fallen into a trap operator star relative satellite true directions vector.

According to theory orientation vector azimuth and the angle of pitch of the relative primary of step (9) resulting bottle star, satellite adopts ten thousand Point to axial adjustment star sensor optical axis and match with theory orientation vector, and actual measurement is carried out using star sensor；

Satellite star sensor concept of reality surveys component, exports component relative satellite unit direction vector true measurement Actual measured value between satellite and component is measured by satellite laser rangerSet up component relative satellite observational equation For：

Wherein,

Further, to state model and observation model discretization in the step (11), and Unscented karrs are utilized Graceful filtering algorithm estimates that satellite position and speed are specially：

Discretization is carried out to observation model in state model in step (2) and step (10)

y_k=g (x_k)+v_k (12b)

In formula, k=1,2 ..., f (x_k,u_k) be it is discrete after state transition equation, g (x_k) be it is discrete after observational equation, W (k) and v (k) be respectively it is discrete after system noise and observation noise,

Using Unscented Kalman filtering algorithms, the state model and observation model with reference to described in step is filtered, Corresponding Unscented sampled points can be obtained according to state vector, using System State Model, one-step prediction is carried out to sampled point, And the covariance matrix between the iterative state value obtained with the filtering of a upper moment is drawn, to eliminate model error in state model Affect.

Operation principle：The present invention is the space multi-robot self-determination air navigation aid in the face of geostationary orbit object run, Using the autonomous Continuous Observation component of primary star sensor, component is obtained with respect to primary direction vector and azimuth and the angle of pitch.It is first First with GEO target satellites as in-orbit service object, two spaces robot team formation flight configuration and orbit parameter, Ran Houti are designed Going out primary star sensor observation component needs to meet four kinds of basic illumination conditions：1. primary and component relative distance meet observation away from From requiring；2. component is in solar irradiation area and can be observed completely；3. the earth (or other celestial bodies) is introduced into star sensor visual field； 4. component can the apparent magnitude less than can apparent magnitude threshold value, secondly judge that can primary star sensor observe component：1. whether component In star sensor field range；2. component finally calculates component with respect to primary direction whether in star sensor two dimension image planes battle array Vector azimuth and the angle of pitch, provide data support for the autonomous Continuous Observation component of primary.

Beneficial effect：The present invention can provide high precision position and velocity information, effectively solving satellite for satellite formation flying Institute's relatively low problem of caused navigation accuracy that formation flight observation information is not enough.Relative to prior art, the method have the advantages that In：(1) star sensor is the heavenly body sensor for observing fixed star, and carrying out Inter-satellite relative measure using star sensor needs to meet special Fixed condition, the present invention propose the illumination condition and star sensor observation condition that needs are observed between star, only solve traditional star sensor Energy passive measurement problem, raising independently select star accuracy；(2) realizing between star on the basis of observation, the present invention proposes to calculate in real time Component is with respect to primary orientation vector and azimuth and angle of pitch method, and points to company using universal axial adjustment star sensor optical axis Continuous tracking component, solve conventional observation cannot continuous tracking problem, Continuous Observation efficiency between raising star.

Description of the drawings

Fig. 1 is the inventive method flow chart；

Star sensor observation component flow chart based on Fig. 2

Fig. 3 is primary in the present invention with respect to specific distance range schematic diagram between the star of component；

Fig. 4 is neutron star illumination condition schematic diagram of the present invention；

Fig. 5 is star sensor visual field in the present invention and position of the earth relation schematic diagram；

Fig. 6 is that neutron star of the present invention can apparent magnitude calculating schematic diagram；

Fig. 7 is neutron star of the present invention in star sensor two-dimensional image area array projection schematic diagram；

Fig. 8 is neutron star of the present invention with respect to primary direction vector and azimuth schematic diagram.

Specific embodiment

Below in conjunction with accompanying drawing, the case study on implementation of the present invention is described in detail；

As shown in figure 1, the present invention is a kind of relative observation procedure of space multirobot towards GEO satellite in-orbit service, Towards the GEO satellite in-orbit service stage, space multirobot (being set to primary and component) is independently continuously seen using star sensor The method for surveying relative direction vector is a kind of space multirobot for being very suitable for in-orbit service with respect to observation procedure.Its bag Include step as follows：

(1) designing two spaces robot (being set to primary and component) formation flight configuration and orbit parameter (includes track Semi-major axis a, orbital eccentricity e, orbit inclination angle i, right ascension of ascending node Ω, argument of perigee ω, time of perigee passage t_p), design The optimal installation position of primary star sensor is observing component；

(2) under geocentric inertial coordinate system, when primary positional distance is more than component and primary relative distance, set up satellite Relative target component dynamics of orbits model

Wherein, δ r⁽¹⁰⁾With δ v⁽¹⁰⁾For component relative satellite direction vector, r⁽⁰⁾And r⁽¹⁾For satellite and component position vector, μ_eFor Gravitational coefficient of the Earth, a_fAffect for perturbative force.

Definition status variable x=[(δ r⁽¹⁰⁾)^T(δv⁽¹⁰⁾)^T]^T, set up autonomous navigation system state model；

Wherein, f [x (t), t] is mission nonlinear continuous state transfer function, and w (t) is state-noise.

(3) according to designed two spaces robot track parameter, primary and component relative distance δ r are calculated⁽¹⁰⁾, such as Fig. 2 It is shown, judge whether which meets star sensor observation component needs and meet specific range requirement

L_min≤δr⁽¹⁰⁾≤L_max (15)

Wherein, δ r⁽¹⁰⁾=| δ r⁽¹⁰⁾|=| r⁽¹⁾-r⁽⁰⁾|, r⁽⁰⁾And r⁽¹⁾For primary and component position vector；L_minAnd L_maxFor Minimum and maximum distance needed for observing between star.

(4) when primary observes component, component needs fully to be irradiated by sunlight.When component is in global illumination area, son Star can be fully irradiated by sunlight；Conversely, when component enters earth's shadow area, as the earth is blocked, sunlight cannot irradiate To component, it is therefore desirable to which antithetical phrase star illumination condition is judged.

According to the sun, the earth and component three's geometry site, as shown in figure 3, determining shadow of the sun and component fortune Critical condition of the row track through the shadow region.If sunlight is directional light, component position vector r⁽¹⁾With solar direction vector r^(sun)The angle of formation is

Component enters and leaves the critical angle of earth's shadow scope

Wherein,R_eIt is earth radius.

Thus can obtain that component is in solar irradiation area and shadow region condition is respectively：

Solar irradiation area：ψ ＜ ψ_cri (18a)

Shadow of the sun：ψ≥ψ_cri (18b)

(5) during star sensor observation component, when comparison field light is too strong or excessively weak, which cannot also observe son Star, it is therefore desirable to which analyzing comparison field is affected by celestial body.

Cause to be analyzed as a example by star sensor comparison field is excessively weak by the earth, it is several according to the earth, primary and component three What position relationship, as shown in figure 4, component is with respect to primary direction vector δ r⁽¹⁰⁾With primary direction vector r⁽⁰⁾Angle be

Due to the critical condition that the earth causes background light excessively weak be primary and component line it is tangent with earth edge, then Tangent line with the critical angle of primary position vector is

Thus the condition that star sensor visual field can be obtained is not affected by background light is

θ ＞ θ_cri (21)

The equally applicable judgement component background of the method is blocked by other celestial bodies and causes the too strong situation of light.

(6) magnitude is the concept in astronomy, and it is the physical quantity for weighing celestial body luminosity.Magnitude is generally divided into absolute magnitude With can the apparent magnitude, absolute magnitude referred in the celestial body brightness seen away from 32.6 light-year of celestial body；Can the apparent magnitude refer to the earth Celestial body brightness seen by upper observer.Magnitude value is less, shows that celestial body is brighter；Conversely, celestial body is then darker.Introduce can the apparent magnitude it is general Read the observability that analysis is observed component.

First have to calculate the absolute magnitude of component, absolute magnitude M of component can be calculated by following formula：

Wherein, m_sunBe the sun can the apparent magnitude, its value be -26.73；r_dTo be observed the radius of celestial body；A is celestial body Reflectance；d₀It is the average distance between the earth and the sun, its value is 1.496 × 10¹¹m。

The apparent magnitude m of component can be calculated according to equation below by absolute magnitude M：

Wherein, | r^(sun0)| it is the distance between the sun and component；ξ is relative vector δ r⁽¹⁰⁾Component relative with sun direction Vector r^(sun1)Angle, as shown in figure 5, can be tried to achieve by following formula：

P (ξ) is phase integral, can be tried to achieve by following formula：

It is observed the visual magnitude value of celestial body bigger, its relative star sensor is darker；Conversely, its relative star sensor is brighter. If star sensor Observable threshold value is m_thr, component can the apparent magnitude be m, its can the apparent magnitude need to meet condition

M ＜ m_thr (26)

(7) star sensor optical axis are defined and is pointed to and in body coordinate system direction vector beComponent is calculated with respect to primary side To vector delta r⁽¹⁰⁾With star sensor direction vector it isAngle

Wherein,It is geocentric inertial coordinate system opposing body's coordinate system pose transition matrix.

It is FOV to define the star sensor angle of visual field, judges relative vector δ r⁽¹⁰⁾Whether in star sensor field range

In field range：

Outside field range：

If relative vector δ r⁽¹⁰⁾Not in field range, it is considered to using universal axial adjustment star sensor optical axis director Amount, can be by δ r⁽¹⁰⁾WithIn the plane of composition, directly deflection is equal to or more thanAngle, makes vector delta r⁽¹⁰⁾Enter Enter field range, if visual field can not be still entered after rotating, cannot observe.

(8) according to the relative primary direction vector δ r in component⁽¹⁰⁾The geometrical relationship of star sensor two dimension image planes battle array is projected in, such as Shown in Fig. 6, following (23) formula is resolved, component can be obtained in two-dimentional image planes battle array coordinate

Wherein, f is star sensor focal length

If image plane length and width is respectively IP_longthAnd IP_width, component can be observed to be needed to meet condition

(9) after primary observes component, can be obtained between two satellites apart from δ r by inter-satellite link⁽¹⁰⁾, by star sensor Component can be obtained with respect to primary unit direction vectorAs shown in fig. 7, therefore component can be obtained with respect to primary theory orientation vector being

Wherein,

Component can be described by azimuth and the angle of pitch with respect to primary direction vector, in satellite body coordinate system o_b-x_by_bz_b In, definition azimuth angle alpha is δ r⁽¹⁰⁾In o_b-y_bz_bThe projection of plane and y_bAxle clamp angle, angle of pitch δ are δ r⁽¹⁰⁾With x_bAxle clamp angle, can It is expressed as

Wherein, It is geocentric inertial coordinate system opposing body's coordinate It is pose transformation matrix.

(10) primary star sensor true optical axis are adjusted consistent with theory orientation vector, component is truly observed, is built The vertical observational equation with unit direction vector and distance as observed quantity；

According to theoretical azimuth and the angle of pitch of above-mentioned resulting bottle star relative satellite, satellite adopts universal drive shaft or other machineries Device adjustment star sensor optical axis are pointed to and are matched with the theory orientation, and are truly measured using star sensor, output Star is with respect to the true observation of satellite activity's direction vectorAnd measured between satellite and component using satellite laser ranger True measurementSetting up component relative satellite observational equation is：

Wherein,

(11) to state model and observation model discretization, and satellite is estimated using Unscented Kalman filtering algorithms Position and speed.

Discretization is carried out to observation model in state model in step 2 and step 10

x_k+1=f (x_k,u_k)+w_k (34a)

y_k=g (x_k)+v_k (34b)

Wherein, state vector is x_k∈R^L, input vector is u_k∈Rⁿ, output vector is y_k∈R^M, process noise w_k∈N (0,Q_k), measurement noise：v_k∈N(0,R_k), and w_kAnd v_kIt is uncorrelated.

Using Unscented Kalman filtering algorithms, the state model and observation model with reference to described in step is filtered, Corresponding Unscented sampled points can be obtained according to state vector, using System State Model, one-step prediction is carried out to sampled point, And the covariance matrix between the iterative state value obtained with the filtering of a upper moment is drawn, to eliminate model error in state model Affect.Specific algorithm is as follows

Step 1：For state variable x_k, averageVarianceCarry out Unscented conversion

Step 2：Prediction process

χ_i,k/k-1=f (χ_i,k-1) (36a)

Step 3：Renewal process

Step 4：Return to step 1 carries out the filtering of next cycle.

(12) calculate and terminate.

First with GEO target satellites as in-orbit service object, design two spaces robot (is set to primary and son to the present invention Star) formation flight configuration and orbit parameter, then propose that primary star sensor observation component needs to meet four kinds of basic illumination bars Part：1. primary and component relative distance meet observation specific range and require；2. component is in solar irradiation area and can be observed completely； 3. the earth (or other celestial bodies) is introduced into star sensor visual field；4. component can the apparent magnitude less than can apparent magnitude threshold value, secondly judge Can primary star sensor observe component：1. whether component is in star sensor field range；2. whether component is in star sensor In two-dimentional image planes battle array, component is finally calculated with respect to primary direction vector and azimuth and the angle of pitch, be the autonomous Continuous Observation of primary Component provides data and supports.

First with GEO target satellites as in-orbit service object, design two spaces robot (is set to primary and son to the present invention Star) formation flight configuration and orbit parameter, then according to satellite relative motion dynamics model under geocentric inertial coordinate system, set up Autonomous navigation system state model；Secondly propose theoretical illumination condition and the imaging met needed for primary star sensor observation component Condition.Component is calculated with respect to primary theory azimuth and the angle of pitch, the true star sensor optical axis of adjustment are consistent with theory orientation, right Component is truly observed, and sets up the observational equation with relative unit direction vector and distance as observed quantity；Finally use Unscented Kalman Filter Estimation primary relative positions and speed, the invention belongs to aerospace navigation technical field, not only can High accuracy navigation information is provided in GEO formation flights for satellite, and reference can be provided for its autonomous navigation system design.

Claims (12)

1. a kind of space multi-robot self-determination air navigation aid in the face of geostationary orbit object run, it is characterised in that step It is as follows：

(1) with GEO target satellites as in-orbit service object, two spaces robot is set to into primary and component, design master Star and component formation flight configuration and orbit parameter；

(2) according to satellite relative motion dynamics model under geocentric inertial coordinate system, set up autonomous navigation system state model；

(3) according to the primary and component relative distance for calculating, judge that component is no and meet star sensor observed range requirement, meet then Into step (4), otherwise into step (12)；

(4) according to the sun, the earth and component three's position relationship for resolving, judge whether component is in solar irradiation area, be then Into step (5), otherwise into step (12)；

(5) according to the earth, primary and component three's position relationship for resolving, judge the no entrance star sensor visual field of the earth, be then Into step (6), otherwise into step (12)；

(6) according to the component for calculating can the apparent magnitude, judge component can the apparent magnitude whether be less than star sensor Observable threshold value, be then Into step (7), otherwise into step (12)；

(7) angle is pointed to star sensor optical axis according to the relative primary direction vector in component for calculating, judges component whether in star In sensor field range, it is then to enter step (8), otherwise calculates after being pointed to using universal axial adjustment star sensor optical axis, after It is continuous to judge, it is then to enter step (8), otherwise into (12)；

(8) according to the component for calculating in star sensor two dimension image planes battle array coordinate, judge component whether in star sensor two dimension image planes In battle array, it is then to enter step (9), otherwise into step (12)；

(9) theory orientation vector azimuth and the angle of pitch of the component with respect to primary is calculated, into step (10)；

(10) adjust the sensing of star sensor true optical axis consistent with theory orientation, component is truly observed, component phase is calculated To satellite true directions vector, the observational equation with unit direction vector and distance as observed quantity is set up, into step (11)；

(11) to the state equation set up and observational equation discretization, estimate to defend using Unscented Kalman filtering algorithms Championship is put and speed；

(12) terminate observation.

2. the space multi-robot self-determination air navigation aid in the face of geostationary orbit object run according to claim 1, It is characterized in that：Orbit parameter in the step (1) includes that semi-major axis of orbit a, orbital eccentricity e, orbit inclination angle i, liter are handed over Point right ascension Ω, argument of perigee ω, time of perigee passage t_p。

3. the space multi-robot self-determination air navigation aid in the face of geostationary orbit object run according to claim 1, It is characterized in that：Autonomous navigation system state model process is set up in the step (2) as follows：

Under geocentric inertial coordinate system, when primary positional distance is more than component and primary relative distance, satellite phase is set up To target component dynamics of orbits model

$\mathrm{\δ}{\stackrel{\·}{r}}^{\left(10\right)}={\mathrm{\δv}}^{\left(10\right)}---\left(1a\right)$

$\mathrm{\δ}{\stackrel{\·}{v}}^{\left(10\right)}=-\frac{{\mathrm{\μ}}_e}{{\left|r^{\left(0\right)}\right|}^3}\[{\mathrm{\δr}}^{\left(10\right)}-3\left(\frac{{\left(r^{\left(0\right)}\right)}^T{\mathrm{\δr}}^{\left(10\right)}}{{\left|r^{\left(0\right)}\right|}^2}\right)r^{\left(0\right)}\]+a_f---\left(1b\right)$

Wherein, δ r⁽¹⁰⁾With δ v⁽¹⁰⁾For component relative satellite direction vector, r⁽⁰⁾And r⁽¹⁾For satellite and component position vector, μ_eFor Gravitational coefficient of the Earth, a_fAffect for perturbative force；

Definition status variable x=[(δ r⁽¹⁰⁾)^T (δv⁽¹⁰⁾)^T]^T, set up autonomous navigation system state model；

${\stackrel{\·}{x}}_t=f(x_t,u_t)+w_t---\left(2\right)$

Wherein, f [x (t), t] is mission nonlinear continuous state transfer function, and w (t) is state-noise.

4. the space multi-robot self-determination air navigation aid in the face of geostationary orbit object run according to claim 1, It is characterized in that：Judge whether component meets star sensor observed range requirement process in the step (3) as follows：

Primary is calculated with respect to component apart from δ r⁽¹⁰⁾, judge whether which meets condition

L_min≤δr⁽¹⁰⁾≤L_max (3)

Wherein, δ r⁽¹⁰⁾=| δ r⁽¹⁰⁾|=| r⁽¹⁾-r⁽⁰⁾|, r⁽⁰⁾And r⁽¹⁾For primary and component position vector；L_minAnd L_maxFor between star Minimum and maximum distance needed for observation.

5. the space multi-robot self-determination air navigation aid in the face of geostationary orbit object run according to claim 1, It is characterized in that：Judge whether component is in solar irradiation area process in the step (4) as follows：

Analysis earth shaded region and component travel through the critical condition of the shaded region, if component position vector r⁽¹⁾With Position of sun vector r^(sun)Angle is ψ, and it is ψ that component enters and leaves the critical angle of earth's shadow scope_cri, then component be in Solar irradiation area needs to meet condition：

ψ ＜ ψ_cri (4)。

6. the space multi-robot self-determination air navigation aid in the face of geostationary orbit object run according to claim 1, It is characterized in that：Judge in the step (5) whether the earth enters star sensor visual field process as follows：

If primary position vector r⁽⁰⁾Component direction vector δ r relative with primary⁽¹⁰⁾Angle be θ, being blocked by the earth causes bias light The excessively weak critical condition of line is component with respect to primary direction vector δ r⁽¹⁰⁾Tangent with earth edge, defining this critical angle is θ_cri, then the earth be introduced into star sensor viewing conditions and be：

θ ＞ θ_cri (5)。

7. the space multi-robot self-determination air navigation aid in the face of geostationary orbit object run according to claim 1, It is characterized in that：Judge in the step (6) component can the apparent magnitude whether to be less than star sensor Observable threshold process as follows：

Introduce can the apparent magnitude analyze the observability of component, magnitude value is less, shows that celestial body is brighter；Conversely, celestial body is then darker；If star Sensor Observable threshold value is m_thr, component can the apparent magnitude be m, component be observed its can the apparent magnitude need to meet condition

M ＜ m_thr (6)。

8. the space multi-robot self-determination air navigation aid in the face of geostationary orbit object run according to claim 1, It is characterized in that：Judge whether component is as follows in star sensor field range internal procedure in the step (7)：

If component is with respect to primary direction vector δ r⁽¹⁰⁾With star sensor optical axis pointing vectorAngle isStar sensor visual field Angle is FOV, then direction vector δ r⁽¹⁰⁾In star sensor field range need to meet condition

If relative vector δ r⁽¹⁰⁾Not in field range, calculate and pointed to using universal axial adjustment star sensor optical axis so as to enter Enter field range, if visual field can not be still entered after rotating, cannot observe.

9. the space multi-robot self-determination air navigation aid in the face of geostationary orbit object run according to claim 1, It is characterized in that：Judge in the step (8) whether component is as follows in star sensor two dimension image planes battle array internal procedure：

According to the relative primary direction vector δ r in component⁽¹⁰⁾The geometrical relationship of star sensor two dimension image planes battle array is projected in, solving its coordinate isIf two-dimentional image planes array length degree and width are respectively IP_longthAnd IP_width, then component is full in image plane coordinate needs Sufficient condition

$\left|x_p^{\left(10\right)}\right|<\frac{{\mathrm{IP}}_{longth}}{2}---\left(8a\right)$

$\left|y_p^{\left(10\right)}\right|<\frac{{\mathrm{IP}}_{width}}{2}---\left(8b\right).$

10. the space multi-robot self-determination air navigation aid in the face of geostationary orbit object run according to claim 1, It is characterized in that：The step (9) operator star of falling into a trap is specially with the angle of pitch with respect to primary direction vector and azimuth：

Component is with respect to primary unit direction vectorObtained by star sensor, component is obtained with respect to primary azimuth angle alpha and the angle of pitch δ, primary and component relative distance | δ r⁽¹⁰⁾| obtained by inter-satellite link, component is obtained with respect to primary theory orientation vector delta r⁽¹⁰⁾

${\mathrm{\&delta;r}}^{\left(10\right)}=\mathrm{\&delta;}{\hat{r}}^{\left(10\right)}\&CenterDot;\left|{\mathrm{\&delta;r}}^{\left(10\right)}\right|---\left(9\right)$

Wherein,

Component is described by azimuth and the angle of pitch with respect to primary orientation, in satellite body coordinate system o_b-x_by_bz_bIn, define azimuth α is δ r⁽¹⁰⁾In o_b-y_bz_bThe projection of plane and y_bAxle clamp angle, angle of pitch δ are δ r⁽¹⁰⁾With x_bAxle clamp angle, is expressed as

$\mathrm{\&alpha;}=arctan\left(\frac{{\mathrm{\&delta;r}}_{bz}^{\left(10\right)}}{{\mathrm{\&delta;r}}_{by}^{\left(10\right)}}\right)---\left(10a\right)$

$\mathrm{\&delta;}=arcsin\left(\frac{{\mathrm{\&delta;r}}_{bx}^{\left(10\right)}}{\left|{\mathrm{\&delta;r}}_b^{\left(10\right)}\right|}\right)---\left(10b\right)$

Wherein, It is geocentric inertial coordinate system opposing body's coordinate system appearance State transition matrix.

The 11. space multi-robot self-determination air navigation aids in the face of geostationary orbit object run according to claim 1, It is characterized in that：The step (10) is fallen into a trap operator star relative satellite true directions vector.

According to theory orientation vector azimuth and the angle of pitch of the relative primary of step (9) resulting bottle star, satellite adopts universal drive shaft Adjustment star sensor optical axis are pointed to and are matched with theory orientation vector, and carry out actual measurement using star sensor；

Satellite star sensor concept of reality surveys component, exports component relative satellite unit direction vector true measurementBy satellite Laser range finder measures the actual measured value between satellite and componentSetting up component relative satellite observational equation is：

$y=\left[\begin{array}{c}{y}_1\\ y_2\end{array}\right]=\left[\begin{array}{c}\mathrm{\&delta;}{\hat{r}}_{mes}^{\left(10\right)}\\ {\mathrm{\&delta;r}}_{mes}^{\left(10\right)}\end{array}\right]---\left(11\right)$

Wherein,

The 12. space multi-robot self-determination air navigation aids in the face of geostationary orbit object run according to claim 1, It is characterized in that：To state model and observation model discretization in the step (11), and utilize Unscented Kalman filterings Algorithm estimates that satellite position and speed are specially：

Discretization is carried out to observation model in state model in step (2) and step (10)

${\stackrel{\&CenterDot;}{x}}_{k+1}=f(x_k,u_k)+w_k---\left(12a\right)$

y_k=g (x_k)+v_k (12b)

In formula, k=1,2 ..., f (x_k,u_k) be it is discrete after state transition equation, g (x_k) be it is discrete after observational equation, w (k) With v (k) be respectively it is discrete after system noise and observation noise,

Using Unscented Kalman filtering algorithms, the state model and observation model with reference to described in step is filtered, according to State vector can obtain corresponding Unscented sampled points, using System State Model, one-step prediction is carried out to sampled point, and is obtained The covariance matrix gone out between the iterative state value obtained with the filtering of a upper moment, to eliminate the shadow of model error in state model Ring.