CN100575878C - A kind of quick retrieval method for satellite attitude - Google Patents

Abstract

A kind of quick retrieval method for satellite attitude, it is out of control to judge that earlier current attitude has or not, and then do not carry out normal attitude control as if out of control, otherwise current topworks stops control; Then according to having or not the gyro observation information to determine wave filter, the current attitude of satellite and angular speed observed upgrading handle the line time of going forward side by side and upgrade forecast; Judge whether to arrive targeted attitude according to the current attitude of satellite, do not judge then to targeted attitude whether current time needs topworks is controlled, if need then utilize current attitude to determine that information and the control of targeted attitude instruction carrying out attitude maneuver finish until the appearance control, otherwise repeat said process.The present invention can save the attitude of near-earth satellite out of control fast, eliminate dependence and work load thereof on the one hand, on the other hand, backup each other with autonomous rescue method on the existing star to land station, improve fast autonomous redemption ability on the star, guarantee the satellite operate as normal life-span.

Description

A kind of quick retrieval method for satellite attitude

Technical field

The present invention relates to a kind of attitude of satellite rescue method out of control, relate in particular to the fast autonomous rescue method of a kind of near-earth attitude of satellite out of control, this class satellite possesses three axis magnetometer, appearance control computer and the motor-driven control executing mechanism of operate as normal at least.

Background technology

Attitude is out of control to be situation more common in the satellite failure.More typical case has: there is defective in the design of the posture control system of No. 1 A star of wind and cloud, enter the orbit 35 days promptly forever out of control; No. 1 B star of wind and cloud in operate as normal after 165 days owing to be subjected to the radiation of space high energy particle, computing machine produces frequent saltus step on the star, causes satellite out of hand; 1999,4 in 6 control gyroscopes of Hubble Space Telescope broke down, cause it out of control.According to the health status of the satellite posture control system equipment after out of control, appearance controlling fault degree can be divided into following a few class: 1) CASE1: temporary fault (still can recover) perhaps appears in the equipment non-fault; 2) CASE2: permanent fault appears in equipment component, but system still can carry out certain reconstruct; 3) CASE3: permanent fault appears in some key equipment.Posture control system is to ensure one of critical system that the satellite spatial detection mission successfully is carried out, the attitude of satellite out of control even might jeopardize and put in order the star unexpected termination in serviceable life, and for this reason, research satellite attitude redemption scheme is necessary.

The speed that the attitude of satellite is saved speed is for guaranteeing that satellite serviceable life of flying normally is vital.After the attitude of satellite was out of control, solar array possibly can't charge normal, thereby causes the electric energy of accumulator to deplete, at this moment, will be difficult in a short time the attitude of satellite be saved back the operate as normal attitude, even satellite can't be saved forever, therefore, research rapid posture redemption scheme is very important.

Traditional is to be controlled to be auxilliary redemption mode on master, the star can't reach rapidity requirement usually based on ground station control.In this manner, land station receives attitude of satellite telemetry, handles the back and forms telecommand by land station, is up on the star and carries out.Yet, satellite to set up under the attitude out-of-control condition and ground control station between contact channel need considerable time probably, especially lay under the limited situation in land station.As if some the attitude remedial measure that does not possess under the out-of-control condition, may save the attitude of satellite to land station and cause sizable difficulty on the star.Such as, No. 1 B star of wind and cloud is after satellite is out of control, do not judge attitude of satellite state out of control, do not close the control operation of topworks, cause on the star air storage to exhaust and air storage disorderly spray make satellite high speed rotating (10 rev/mins), so have to utilize environmental torque (gravity gradient stabilization moment and magnetic control moment) to expend the time several months attitude of satellite is saved.Therefore, be necessary to take to be controlled to be on the star that main, land station gets involved is that the attitude of auxilliary (or staying out of) is saved mode.Yet, based on the classic method of ground station control, under the situation that rescue method is invalid on some key equipment breaks down (CASE3) or star, remain last redemption scheme, such as, the emergent safe mode after the appearance control computer fault; The emission space shuttle is repaired Hubble Telescope, or the like.To only consider that fault may take place some equipment on the star though it is pointed out that the present invention, can not cause the thorough infeasible situation of redemption scheme on the star (such as, CASE1 or CASE2).

Can improve the speed that attitude is saved greatly based on attitude rescue method autonomous on the star.This method has greatly reduced the necessity of communicating by letter with land station, can independently arrange attitude to save operation immediately.Current, an important autonomous attitude rescue method of earth observation satellite is to the effect that: when attitude deviation operate as normal attitude, after the appearance control computer is judged as the attitude unusual condition, carrying out attitude reacquisition immediately handles: at first make the windsurfing stall, utilize jet control to finish fast velocity damping (may need), the sun-earth acquisition (the perhaps earth-earth acquisition) then.When moonscope to day, during azimuth information, this attitude redemption stage finishes, and independently changes the normal foundation and the Steady-State Control stage of attitude over the ground afterwards over to.Be not difficult to find out that there is more defective in existing attitude rescue method: 1) scope of application is less.The normal sun sensor of these scheme needs, earth sensor and rate gyro can not be saturated, and catching the solar time can only be in the sun photograph district of track; Adopt jet topworks to carry out attitude maneuver, and jet consume is irremediable.2) the invalid power consumption of motor-driven control is big.This method is utilized unconscious celestial body attitude maneuver operation, capture day, light source, thereby make satellite can determine three-axis attitude, the motor-driven control of three-axis attitude under the three-axis attitude known case relatively, the energy consuming ratio of invalid motor-driven control is more, and jet power consumption can't remedy.Therefore, independently save the ability of attitude, reduce invalid power consumption, be necessary to study other rapid posture rescue method in order to improve satellite.

Current, most satellites all are near-earth satellites, and near-earth satellite utilizes the attitude of terrestrial magnetic field to determine to control security, the reliability of algorithm with control for improving the attitude of satellite, has important effect, and the price of geomagnetic observation equipment and control executing mechanism is low, in light weight, little power consumption and because no rotating mechanism, has goodish reliability, therefore, most of near-earth satellites have all been installed this kind equipment.Wherein, three axis magnetometer is the geomagnetic observation equipment of a main flow.Because the terrestrial magnetic field has been full of whole terrestrial space, therefore, there is not the problem of no observation data in three axis magnetometer.On the contrary, satellite (attitude of satellite may wide-angle depart from normal operation position) when attitude is out of control, can't observe probably day, etc. observable azimuth information constantly under the operate as normal attitude, therefore be difficult to determine the attitude of satellite.In fact,, utilize nonlinear filtering algorithm, can determine the satellite three-axis attitude based on the ground magnetic vector observed quantity sequence on a period of time interval.At present, this algorithm is greatly developed, and includes earth magnetism attitude under the gyro situation and determines that the earth magnetism attitude under algorithm and no gyro (perhaps gyro the is saturated) situation determines algorithm, and the attitude that has successfully solved satellite stable state mission phase sets the tasks.Along with the development of nonlinear filtering algorithm, these attitudes determine that algorithm can finish satellite wide-angle quick rotation, and the attitude under initial attitude evaluated error very big (or the initial attitude the unknown) situation sets the tasks.This is the broad theory basis that the present invention saves satellite out of control.

But attitude determines that the attitude that algorithm provides determines that information (comprising attitude angle and angular speed) might not directly offer controller, is used for the generation of control moment instruction.The work characteristics of wave filter is that the influence of priori initial estimation error is revised in the observed quantity that foundation constantly obtains gradually, constantly approaches true value.When initial attitude estimation error is big (this is common situation when out of control), the attitude of wave filter initial stage determines that error may be very big, controller utilizes the very big attitude of evaluated error to determine that information carries out attitude control, disturb causing unnecessary attitude control, be unfavorable for the accurate quick control attitude of satellite, and cause a large amount of invalid control energy consumptions.Therefore, attitude also needs when saving the attitude of wave filter is determined whether information can with independently judging, so that when decision starts topworks.

Attitude of satellite redemption can be regarded a three-axis attitude as and rotate motor-driven control problem.Rotate motor-driven control in order to make satellite finish wide-angle fast, motor-driven control law should guarantee that motor-driven control route is enough short, the motor-driven control method of modal time optimal is that Euler rotates motor-driven control, the classical PID control law is widely used in low-angle attitude maneuver control task, but rotate in the motor-driven control task in wide-angle, the maximum speed of the uncontrollable celestial body of this control law, and some equipment of satellite and structure to the celestial body rotating speed have maximum speed require (such as, if rotating speed exceeds the scope that effectively tests the speed of gyro, gyro will be in state of saturation).Passing rank-saturated PID control law is a kind of simple remodeling on the PID control law basis, and the introducing amplitude by restriction attitude misalignment feedback information can be limited to the satellite rotating speed in the specified scope.The relative nonlinear control law, the form of this control law is simple, calculated amount is little, have the extensive prospect of using on star.This is the another broad theory basis that the present invention saves satellite out of control.

At present, above-mentioned each theory all independent development get up, still, do not unite the attitude that is used for satellite out of control as yet and save.And, autonomous attitude rescue method must be practical, reliable, safe on the star, therefore, must restart the determination methods on control opportunity and attitude to the determination methods of attitude of satellite runaway condition and the countermeasure under the out-of-control condition, topworks saves the contents such as judgement on action end opportunity and studies.

Summary of the invention

Technology of the present invention is dealt with problems: solve attitude rescue method based on ground station control in the defective aspect the rapidity, overcome the relatively harsher and high shortcoming of invalid control energy consumption of autonomous attitude rescue method application conditions on the existing star, a kind of quick retrieval method for satellite attitude is provided, can save the attitude of near-earth satellite out of control fast, eliminate dependence and work load thereof on the one hand to land station, on the other hand, backup each other with autonomous rescue method on the existing star, improve fast autonomous redemption ability on the star, guarantee the satellite operate as normal life-span.

Technical solution of the present invention: a quick retrieval method for satellite attitude comprises the following steps:

(1) to judge that current attitude has or not out of control for the appearance control computer, if not out of control, then carries out normal attitude control operation; Otherwise the appearance control computer makes current topworks stop control, and runaway condition is fed back to the Star Service central computer closes inessential equipment on the star, enters step (2) then;

(2) judgement has or not the gyro observation information, if the gyro observation information is arranged, then utilize gyro attitude observation information and ground magnetic vector observation and orientation vector observation information to determine wave filter, the current attitude of satellite and gyro are often floated parameter to be observed upgrading and handles, and learn equation and gyro mechanical model time of carrying out that often floats based on attitude motion and upgrade forecast, enter step (3) then; If there is not the gyro observation information, then utilize no gyro attitude observation information and ground magnetic vector observation and orientation vector observation information to determine wave filter, observe renewal handle to the current attitude of satellite and celestial body angular speed, and learn equation and Euler's kinetic model time of carrying out is upgraded forecast based on attitude motion, enter step (3) then;

(3) the appearance control computer judges whether the current attitude of satellite has arrived targeted attitude, if "Yes", then attitude is saved task termination; Otherwise, change step (4) over to;

(4) the appearance control computer judges whether current time needs to start motor-driven control executing mechanism and control, if "Yes", then utilize current attitude to determine information and targeted attitude instruction, produce the control moment instruction according to three-axis attitude fast reserve control law, drive motor-driven control executing mechanism and carry out attitude maneuver control; Otherwise, return step (2).

The determination methods of runaway condition is in the described step (1):

(a) constantly, according to the nominal installation site of nominal satellite operate as normal attitude and sensor, calculate the magnetic vector observation of nominal ground of three axis magnetometer and estimate, as nominal observed quantity in each observation;

(b) calculate the deviation angle of ground magnetic vector observed quantity and nominal observed quantity;

(c) if the deviation angle of current time first greater than judgment threshold G out of control ₁, then starting the statistics process, order characterizes the parameter K of " attitude is out of control " ₁=1, in each observation after this constantly, if both angles are greater than threshold value G ₁, then make parameter K ₁=K ₁+ 1, otherwise, K ₁Remain unchanged; W ₁After the individual moment, calculate C ₁=K ₁/ W ₁, if C ₁＞R ₁, think that then the attitude of satellite is out of control, and with K ₁Zero clearing; Otherwise, think that the attitude of satellite is normal, equally with K ₁Zero clearing; If if current time deviation angle is first smaller or equal to threshold value G out of control ₁, then carry out the judgement in the next moment.

The appearance control computer judges that the method whether current attitude of satellite has arrived targeted attitude is in the described step (3): as if the deviation of the attitude of satellite of current time and targeted attitude first less than threshold value G ₃, then starting the statistics process, order characterizes the parameter K of " arrival targeted attitude " ₃=1, in each observation after this constantly, the deviation of the attitude of satellite and targeted attitude is less than error threshold G ₃, then make K ₃=K ₃+ 1, otherwise, remain unchanged; When carrying out W ₃After the individual moment, calculate C ₃=K ₃/ W ₃, if C ₃＞R ₃, think that then the attitude of satellite arrives near the targeted attitude, and with K ₃The value zero clearing; Otherwise, think the attitude of satellite not yet Be Controlled arrive near the targeted attitude, and with K ₃The value zero clearing; If the deviation of the attitude of satellite of current time and targeted attitude is first more than or equal to threshold value G ₃, then carry out the judgement in the next moment.

The appearance control computer judges whether current time needs to start the method that motor-driven control executing mechanism controls and be in the described step (4): as if the new breath of current time first less than error threshold G ₄, then starting the statistics process, order characterizes the parameter K of " restarting control " ₄=1, each observation updated time after this is if newly cease less than error threshold G ₄, then make parameter K ₄Add 1, otherwise, remain unchanged; Work as W ₄After the individual moment, calculate C ₄=K ₄/ W ₄, if C ₄＞R ₄, then think and arrived the opportunity of restarting control, and with K ₄The value zero clearing, otherwise, think that current attitude determines that error is excessive, should continue to utilize observed quantity to dwindle attitude and determine error, and with K ₄The value zero clearing; If the new breath of current time is first more than or equal to error threshold G ₄, then carry out the judgement in the next moment.

Three-axis attitude fast reserve control law adopts and passs rank-saturated PID control law in the described step (4)

The present invention's advantage compared with prior art is:

(1) the present invention is based on the rapid posture that the wide-angle attitude determines that algorithm and three motor-driven control algolithms of quick rotation propose a kind of satellite out of control and save new method.Existing relatively method, the present invention only need observe the ground magnetic vector can realize that three-axis attitude is definite, and utilize three motor-driven control methods of quick rotation the attitude of satellite out of control to be saved in the controlled range of normal attitude, and existing method can only catch respectively day by attitude maneuver control unconsciously, local to, make day, sensor can observe day, azimuth information, thereby realize that three-axis attitude determines.On the one hand, the observation sensitive equipment that the present invention needs still less, and motor-driven control is except that adopting jet control, other has the topworks of fast reserve ability can also to adopt momentum wheels, control-moment gyro group etc., therefore, even if day, sensor failure, jet exhausting (or not having jet mechanism) still have feasibility, be mutually backup with existing rescue method, improved on the star attitude and saved ability; On the other hand, utilize three-axis attitude to determine that information controls the attitude of satellite consciously, and designed and closed, enable determination methods and the strategy that motor-driven control gear is controlled reliably, help to reduce invalid control energy consumption.In addition, motor-driven control adopts the angular momentum switch to substitute jet control, also helps to reduce jet loss, so that use where necessary.

(2) judge and decision-making technique on some stars that the present invention proposes, having solved satellite utilizes the wide-angle attitude to determine independently to finish a series of realizations and the application problem that attitude is saved with the motor-driven control algolithm of wide-angle quick rotation, making that autonomous attitude is saved on the star becomes practical, reliable, a safe method, and helps to reduce greatly invalid control energy consumption.Wherein, the autonomous determination methods of satellite runaway condition can guarantee that satellite in time enters control state out of control, reduces unnecessary control energy consumption, reduces the difficulty that the attitude of satellite is saved; The determination methods on opportunity is restarted in topworks, gives satellite and judges that current attitude determines that information could be used for the ability of motor-driven control, is convenient to topworks and carries out effective attitude maneuver control in the suitable time, helps to reduce the control energy consumption; The attitude of satellite reaches targeted attitude whether determination methods, gives posture control system and finishes the current attitude redemption stage, opens the ability of next stage appearance control task.Judge on these stars less with the required calculated amount of decision-making, memory requirement is not high, be easy to realize on the star.

(3) the attitude rescue method of the present invention's proposition reduces the work load of land station and required people, material resources cost greatly based on Autonomous Control on the star, avoids the human operational error.Backup each other with existing quick retrieval method, improved the ability that the attitude of satellite is saved greatly, thus the normal serviceable life of guaranteeing satellite.

Description of drawings

Fig. 1 saves schematic flow sheet for rapid posture on the star of the present invention's proposition;

Fig. 2 judges schematic flow sheet for the satellite runaway condition that the present invention proposes;

The general attitude that Fig. 3 adopts for the present invention is determined the sequential recursion estimation schematic flow sheet of wave filter;

The attitude of satellite that Fig. 4 proposes for the present invention is saved task termination and is judged schematic flow sheet;

Fig. 5 enables the judgement schematic flow sheet for the attitude maneuver control executing mechanism that the present invention proposes.

Embodiment

As shown in Figure 1, a kind of rapid posture rescue method of satellite out of control, concrete steps are described below:

(1) to judge that current attitude has or not out of control for the appearance control computer, if not out of control, then carries out normal attitude control operation; Otherwise the appearance control computer makes current topworks stop control, and runaway condition is fed back to the Star Service central computer so that it closes inessential equipment on the star, enters step (2) then;

(2) judgement has or not the gyro observation information, if the gyro observation information is arranged, then utilize gyro attitude observation information and ground magnetic vector observation and orientation vector observation information to determine wave filter, the current attitude of satellite and gyro are often floated parameter to be observed upgrading and handles, and learn equation and gyro mechanical model time of carrying out that often floats based on attitude motion and upgrade forecast, enter step (3) then; If there is not the gyro observation information, then utilize no gyro attitude observation information and ground magnetic vector observation and orientation vector observation information to determine wave filter, observe renewal handle to the current attitude of satellite and celestial body angular speed, and learn equation and Euler's kinetic model time of carrying out is upgraded forecast based on attitude motion, enter step (3) then;

(3) the appearance control computer judges whether the current attitude of satellite has arrived targeted attitude, if "Yes", then attitude is saved task termination; Otherwise, change step (4) over to;

(4) the appearance control computer judges whether current time needs to start motor-driven control executing mechanism and control.If "Yes" then utilizes current attitude to determine information and targeted attitude instruction, produce the control moment instruction according to three-axis attitude fast reserve control law, drive motor-driven control executing mechanism and carry out attitude maneuver control; Otherwise, return step (2).

The judged result of step (1) attitude of satellite runaway condition, according to the deviation ratio of the priori observed quantity under observed quantity of orientation vector sensor and the operate as normal attitude to drawing.Because satellite operate as normal attitude is known, and the installation site of sensor also is known, therefore, can estimate the orientation vector observed quantity of all kinds of sensors according to the installation site of satellite operate as normal attitude and sensor, is called nominal observed quantity.In order accurately to judge whether attitude is out of control, need add up the deviation situation of nominal observed quantity on a period of time interval and concept of reality measurement.A concrete grammar as shown in Figure 2, its content comprises: if the angle of the observed quantity of current time orientation vector and nominal observed quantity is first greater than threshold value G out of control ₁, then starting the statistics process, order characterizes the parameter K of " attitude is out of control " ₁=1, in each observation after this constantly, if both angles are greater than threshold value G ₁, then make parameter K ₁=K ₁+ 1, otherwise, K ₁Remain unchanged.W ₁After the individual moment, calculate C ₁=K ₁/ W ₁, if C ₁＞R ₁(R ₁Expression statistics empirical probit), think that then the attitude of satellite is out of control, and with K ₁Zero clearing; Otherwise, think that the attitude of satellite is normal, equally with K ₁Zero clearing.If if the angle of observed quantity of current time orientation vector and nominal observed quantity is first smaller or equal to threshold value G out of control ₁, then carry out the judgement in the next moment.Method shown in Figure 2 such as the ground magnetic vector, for the purpose of more accurately, can utilize a plurality of orientation vector observed quantities to carry out the judgement of runaway condition at single orientation vector observed quantity.As long as to the conclusion after two orientation vector judgements all is " out of control ", can confirm that then the attitude of satellite is out of control.

Available attitude observed quantity comprises orientation vector observed quantity and ground magnetic vector observed quantity and has or not gyro observed quantity information in the step (2).The ground magnetic vector is constantly considerable, it is the most basic orientation vector observed quantity, if day,, other sensor non-fault of magnitude, then may observe these several azimuth informations once in a while, therefore, satellite constantly can observe orientation vector information more than 2 simultaneously at some, utilizes the determinacy method for determining posture, decide appearance or many vectors optimum attitude such as two vectors and determine that method just can determine the attitude of satellite in this moment, utilize continuous several attitudes observations can be similar to the angular speed of determining satellite.These attitude estimated values and angular speed estimated value are determined the original state prior estimate of filtering algorithm as attitude, can significantly reduce the transient state time of wave filter, realize that attitude is determined fast and accurately.In addition, a plurality of orientation vector observed quantities are used for attitude and determine that the state observation of wave filter upgrades, and also help to improve the ornamental of system, thereby improve state posteriority estimated accuracy.Suppose that satellite equipped rate gyro and unsaturation, the then combination of angular rate measurement value and orientation vector observed quantity can improve stable state estimated accuracy and speed of convergence that the attitude of satellite is determined greatly.

Attitude after satellite is out of control may wide-angle depart from the operate as normal attitude, and hypercomplex number is best overall attitude characterising parameter, is the first-selection that the wide-angle attitude is determined algorithm.According to having/the no gyro information that tests the speed, the attitude problem identificatioin can be described as following state estimation problem:

(a) gyro observed quantity situation is arranged

Quantity of state is:

x(t _k)＝[ q ^T(t _k)β ^T(t _k)] ^T

Wherein, first representation in components hypercomplex number (4 dimension), second representation in components rate gyro often floats vector (3 components).System equation comprises the single order Markov model (time coefficient infinity) that hypercomplex number kinematical equation and gyro often float:

$\left[\begin{array}{c}\underset{\‾}{\overset{\·}{q}}\left(t_k\right)\\ \stackrel{\·}{\mathrm{\β}}\left(t_k\right)\end{array}\right]=\left[\begin{array}{c}\frac{1}{2}\mathrm{\Ω}\left(\mathrm{\ω}\left(t_k\right)\right)\underset{\‾}{q}\left(t_k\right)\\ {\mathrm{\η}}_u\left(t_k\right)\end{array}\right]$

Wherein, true angular speed ω (t _k) the unknown, need according to the gyro observed quantity

And following gyro to measure model is determined:

$\mathrm{\ω}\left(t_k\right)=\widetilde{\mathrm{\ω}}\left(t_k\right)-\mathrm{\β}\left(t_k\right)-{\mathrm{\η}}_{\mathrm{\υ}}\left(t_k\right)$

(b) no gyro observed quantity situation

Quantity of state is

x(t _k)＝[ q ^T(t _k)ω ^T(t _k)] ^T

System equation comprises hypercomplex number kinematical equation and euler dynamical equations:

$\left[\begin{array}{c}\underset{\‾}{\overset{\·}{q}}\left(t_k\right)\\ \stackrel{\·}{\mathrm{\ω}}\left(t_k\right)\end{array}\right]=\left[\begin{array}{c}\frac{1}{2}\mathrm{\Ω}\left(\mathrm{\ω}\left(t_k\right)\right)\underset{\‾}{q}\left(t_k\right)\\ {I^{-1}[T}_c\left(t_k\right)+T_d\left(t_k\right)-{\stackrel{\·}{h}}_w-\mathrm{\ω}\×(\mathrm{I\ω}+h_w)]\end{array}\right]$

Wherein, I represents whole star inertia matrix (comprising the flywheel gear that is in " freezing " state); T _cThe expression control moment; T _dThe expression disturbance torque; h _wExpression flywheel angular momentum.

Above-mentionedly either way adopt following orientation vector observation equation:

$z_k=\left[\begin{array}{c}{\tilde{b}}_k\\ {\tilde{u}}_k\end{array}\right]=\left[\begin{array}{cc}{A}_k\left(\underset{\‾}{q}\right)& 0_{3\×3}\\ 0_{3\×3}& A_k\left(\underset{\‾}{q}\right)\end{array}\right]\left[\begin{array}{c}{b}_{R,k}\\ r_{R,k}\end{array}\right]+v_k$

Wherein, z _kExpression observed quantity vector comprises ground magnetic vector observation

With other is possible (day, etc.) orientation vector observation

b _{R, k}And r _{R, k}The expression of expression corresponding orientation vector in reference frame; a _kThe attitude matrix of expression system relative reference system is with hypercomplex number qThere is relation of equivalence; v _kExpression observation white noise.

Because the single orientation vector observation in the single moment can only help satellite to determine its 2 dimension attitude information, therefore, ground magnetic vector observed quantity can't make that the three-axis attitude in the single moment of satellite is considerable fully, but the ground magnetic vector observed quantity on a period of time interval can make the satellite three-axis attitude considerable fully.Nonlinear filtering algorithm is the algorithms most in use that solves this type of estimation problem, estimate mode owing to adopt sequential recursion, as shown in Figure 3, the initialization of wave filter elder generation is judged to have or not observed quantity, if having then observed quantity is upgraded, carrying out the state time then upgrades, upgrade otherwise directly carry out the state time, the state time upgrades to finish and carries out the observed quantity judgement again, and this mode is fit to the usefulness that real-time attitude is determined very much.The typical non linear filtering algorithm has EKF (Extended Kalman Filter) and UKF (Unscented Kalman Filter), UKF is owing to adopt " the χ point " of one group of weighting to carry out the estimation (being equivalent to the estimation of average and variance) of gauss' condition posterior density, avoided the linearization process of EKF to system equation, thereby, have better nonlinear filtering convergence capabilities (initial convergence territory big, fast convergence rate).The algorithm that this two classes filtering algorithm solves above-mentioned two class problems also grows up, typically there is the gyro attitude to determine that EKF and UKF algorithm are respectively MEKF algorithm and USQUE algorithm, relevant their concrete formula please refer to document: 1) E.J.Lefferts, F.L.Markley, and M.D.Shuster.Kalman Filtering for Spacecraft AttitudeEstimation.Journal of Guidance, Control and Dynamics.1982,5 (5): 417-429; 2) J.L.Crassidis and F.L.Markley.Unscented Filtering for Spacecraft AttitudeEstimation.Journal of Guidance, Control and Dynamics.2003,26 (4): 536-542.If do not exist test the speed information or gyro of gyro saturated, then utilize Euler's kinetic model to replace gyro angular speed information is provided, there is the gyro attitude to determine that algorithm can not had the gyro attitude accordingly and determines algorithm by above-mentioned, typical no gyro attitude is determined the concrete formula of EKF and the UKF algorithm document that sees reference: 1) M.L.Psiaki, F.Martel, and P.K.Pal.Three-Axis AttitudeDetermination via Kalman Filtering of Magnetometer Data.J.of Guidance, Control and Dynamics.1989,13 (3): 506-514; 2) Huang Lin. the application of nonlinear filtering theory in spacecraft attitude is determined. the doctor of Harbin Institute of Technology thesis .2007, the 3rd chapter.Attitude determines that UKF algorithm relative attitude determines that the EKF algorithm has better nonlinear filtering constringency performance, though and its calculated amount increases to some extent than the latter, but still can finish by computer real-time on the star, be the optimal algorithm that solves the problem identificatioin of wide-angle attitude at present.

The startup of filtering algorithm needs an initialization process.When having only ground magnetic vector observed quantity, can only set one at random or fixing original state prior estimate, then according to attitude determine the convergence of filtering algorithm itself eliminate at the beginning of the evaluated error of state of value, make state estimator approach actual value gradually.When there is plural orientation vector observed quantity in synchronization, the attitude of satellite is considerable, can calculate the attitude of satellite by deterministic algorithm, if all observe a plurality of orientation vector in a plurality of continuous moment, then can also utilize the deterministic algorithm approximate treatment to go out the angular speed of celestial body, the filtering original state is selected the estimated value of attitude and angular speed, in order to accelerate speed of convergence, can select a bigger original state priori covariance matrix.

Step (3) appearance control computer judges whether the attitude of satellite has arrived targeted attitude.When the deviation of the attitude of satellite and targeted attitude continues less than certain scope, think that then the attitude of satellite is controlled near the targeted attitude.A concrete grammar is briefly described as follows as shown in Figure 4: if the deviation of definite attitude of current time and targeted attitude is first less than threshold value G ₃, then starting the statistics process, order characterizes the parameter K of " arrival targeted attitude " ₃=1, in each observation after this constantly, the deviation of satellite (determining) attitude and targeted attitude is less than error threshold G ₃, then make K ₃=K ₃+ 1, otherwise, remain unchanged.When carrying out W ₃After the individual moment, calculate C ₃=K ₃/ W ₃, if C ₃＞R ₃(R ₃Expression statistics empirical probit), think that then the attitude of satellite arrives near the targeted attitude, and with K ₃The value zero clearing; Otherwise, think the attitude of satellite not yet Be Controlled arrive near the targeted attitude, and with K ₃The value zero clearing.If the deviation of definite attitude of current time and targeted attitude is first more than or equal to threshold value G ₃, then carry out the judgement in the next moment.Threshold value G ₃Should carry out choose reasonable according to the Steady-State Control error that attitude is saved scheme.

The judged result whether step (4) topworks enables determines that according to attitude the result makes.Attitude maneuver control must utilize the sufficiently high attitude of estimated accuracy to determine the result.If wave filter has accurately been estimated the attitude of satellite, residual error (or the new breath) sequence estimated of a priori magnetic vector observed quantity that provides of magnetic vector observed quantity and wave filter then, it is a white Gaussian noise sequence (noise amplitude priori is known, such as hundreds of nT) that comprises geomagnetic model noise on observation noise and the star.When attitude determined that wave filter approaches true attitude gradually, residual sequence also showed the trend that converges the white Gaussian noise sequence gradually.Therefore, preestablish a threshold value, in residual error enters this threshold range and remain on when interior, think that then arrived the opportunity that reactivates topworks always greater than priori Gaussian noise amplitude.A concrete grammar is briefly described as follows as shown in Figure 5: if the new breath of current time is first less than error threshold G ₄, then starting the statistics process, order characterizes the parameter K of " restarting control " ₄Be " 1 " that each observation updated time after this newly ceases less than error threshold G ₄, then make parameter K ₄Increase " 1 ", otherwise, do not increase.Work as W ₄After the individual moment, calculate C ₄=K ₄/ W ₄, if C ₄＞R ₄(R ₄Expression statistics empirical probit), then think and arrived the opportunity of restarting control, and with K ₄The value zero clearing.Otherwise, think that current attitude determines that error is excessive, should continue to utilize observed quantity to dwindle attitude and determine error, and with K ₄The value zero clearing.If the new breath of current time is first more than or equal to error threshold G ₄, then carry out the judgement in the next moment.As for the attitude control after restarting constantly, a simple rule is adopted in suggestion: utilize each attitude constantly to determine that information carries out attitude maneuver control.

The suggestion of the three-axis attitude fast reserve control law of step (4) adopt as next have a two-layer constraint of saturation condition pass rank-saturated PD control law:

$T_{\mathrm{com}}=-\underset{T_{\max }}{\mathrm{sat}}I[k_p\underset{L_i}{\mathrm{sat}}\left(q_e\right)+k_d{\mathrm{\ω}}_e]$

Wherein, q _eThe vector part of the deviation hypercomplex number of the relative targeted attitude of hypercomplex number (hypercomplex number) is estimated in expression; ω _eExpression deviation angle speed; k _p, k _dExpression P, D controlled variable; About saturation function, it is defined as follows:

$\underset{U}{\mathrm{sat}}\left(x\right)=\left\{\begin{array}{cc}x& \mathrm{if}\left|\right|x\left|\right|<U\\ x\frac{U}{\left|\right|x\left|\right|}& \mathrm{if}\left|\right|x\left|\right|\&GreaterEqual;U\end{array}\right.$

Wherein, x represents n dimension quantity of state; U represents maximum constrained; || || be Modulo-two operation.For above-mentioned rank-saturated PD control law, the T in the outer saturation function of passing _MaxThe maximum output torque of representing motor-driven control executing mechanism; L in the internal layer saturation function _iThe restriction amplitude of expression single shaft attitude misalignment.In order to improve the dynamic response characteristic of motor-driven control procedure, do not exceed the constraint of celestial body maximum rotation angular velocity again, L _iAdopt following variable amplitude setting:

$L_i=\frac{k_d}{k_p}\min \{\sqrt{2a_i\left|\right|q_{e,i}\left|\right|},{\left|\right|{\mathrm{\&omega;}}_i\left|\right|}_{\max }\},$ i＝x，y，z

Wherein, || ω _i|| _MaxThe maximum angle of rotation speed that allows of expression single shaft;

Then design obtains a according to kinematic principle _iExpression maximum angular acceleration and a _i=T _Max/ I _Ii(about the detailed content of above-mentioned control law, please referring to B.Wie, D.Bailey, and C.Heiberg, Rapid mutil-target acquisition andpointing control of agile spacecraft, AIAA-2000-4546,2000)

Case study on implementation

The earth observation moonlet of low accuracy requirement equipment three axis magnetometer in certain, three rate gyros, scan-type infrared earth sensor, sun sensor, pitch axis bias momentum wheel, jet, three magnetic torquers.Suppose that the attitude sensor sample frequency is 1Hz, gyro sampling period is that 10Hz, control bandwidth are 1Hz.Under normal mode of operation, satellite utilizes three axis magnetometer, infrared quick and rate gyro to carry out attitude to determine (at the track sun according to the district, can also utilize sun sensor), and utilizing pitch axis bias momentum wheel and three magnetic torquers to stablize the earth observation attitude, the steady state controling precision under this pattern can reach about 1 degree.Three rate gyros, scan-type is quick infraredly all is vulnerable equipment, when these two kinds of device fails, can utilize remaining appearance control equipment to carry out system reconfiguration: satellite utilizes three axis magnetometer and sun sensor to carry out three-axis attitude to determine, and utilize bias momentum wheel and three magnetic torquers to carry out stable over the ground control, the control accuracy under this pattern is estimated to reach about 5 degree.Suppose quick infraredly and fault has taken place three rate gyros, and before the appearance control computer is judged the two fault, satellite utilizes the gyro attitude that has under the normal condition to determine that the MEKF algorithm can't provide effective attitude and determine the result, causes bias momentum wheel and three magnetic torquer control attitudes of satellite to depart from normal absolute orientation attitude gradually.Because the attitude of satellite wide-angle has departed from the operate as normal attitude, need utilize the attitude rescue method, control near the target absolute orientation attitude the attitude of satellite is motor-driven, so that reconfiguration system can rebulid the absolute orientation operating attitude.Because the quick damage in ground, existing global attitude acquisition method is invalid, utilizes the flow process of attitude rescue method on the star that the present invention proposes as follows:

(1) the appearance control computer judges according to flow process shown in Figure 2 whether the current attitude of satellite is out of control based on ground magnetic vector observed quantity.Wherein, it is as follows a) to calculate the method for name ground magnetic vector observed quantity: with strict absolute orientation attitude (system is to overlap with track) is the operate as normal attitude, produces ground magnetic vector observed quantity b in this coordinate system according to geomagnetic model on the star _b, and according to the three axis magnetometer installation site A of priori _MbBe converted to name ground magnetic vector observed quantity b _m=A _Mbb _bB) calculating name ground magnetic vector observed quantity and concept of reality measures

Angle:

$\mathrm{\&alpha;}={\cos }^{-1}(\frac{b_m}{\left|b_m\right|}\&CenterDot;\frac{{\tilde{b}}_k}{\left|{\tilde{b}}_k\right|})$

C) as α＞G ₁When occurring first (t=0), then make attitude identification parameter K out of control ₁=1, t=1; After this each be (0＜t＜W constantly ₁), if α＞G ₁, then make K ₁=K ₁+ 1, t=t+1; Otherwise, make t=t+1; When having carried out W ₁(t=W after the individual moment ₁), calculate C ₁=K ₁/ N ₁, if C ₁＞R ₁, think that then the attitude of satellite is out of control, and make K ₁=0, t=0; Otherwise, think that the attitude of satellite is normal, and make K ₁=0, t=0.

Selection about Several Parameters.According to the appearance control precision under the normal condition (about 1 degree), make G ₁=5 degree, W ₁=10, R ₁=0.6.

If learn that according to above-mentioned determination methods current attitude is out of control, then the appearance control computer makes all topworkies that (jet mechanism, bias momentum wheel, three magnetic torquers) stop attitude Control work (momenttum wheel still keeps rotary state), and runaway condition fed back to the Star Service central computer so as its close inessential equipment on the star (such as, some useful load), change step (2) then over to;

(2) the appearance control computer is according to three rate gyro Fault Diagnosis results, and decision utilizes no gyro attitude to determine filtering algorithm.Because ground magnetic vector observation exists constantly, and the sun sensor of operate as normal only some the time observation of solar azimuth information just can be provided, therefore, the original state prior estimate error of wave filter may be bigger, in order to realize fast reserve control, be necessary to strengthen the constringency performance that attitude is determined filtering algorithm.Here adopt no gyro attitude to determine the UKF algorithm, be designated as GUKF, this algorithm has adopted the constraint estimation scheme to the estimation of hypercomplex number, and its actual quantity of state (6 dimension) is defined as:

$X_k\&equiv;{\left[\begin{array}{cc}\mathrm{\&delta;}{P}_k^T& {\mathrm{\&omega;}}_k^T\end{array}\right]}^T$

Wherein, δ P _kFor the deviation attitude of GRP parameter (general rodrigue parameters) expression is estimated; ω _kBe the angular speed vector.Shown in the theing contents are as follows of GUKF algorithm:

A) wave filter initialization (k=0)

Initial hypercomplex number and angular speed Estimation of Mean are respectively

With

Order ${\underset{\&OverBar;}{\overset{^}{q}}}_{0|-1}={\underset{\&OverBar;}{\overset{^}{q}}}_0,$ ${\widehat{\mathrm{\&omega;}}}_{0|-1}={\widehat{\mathrm{\&omega;}}}_0,$ Then original state average and covariance are estimated to be respectively

${\hat{X}}_{0|-1}\&equiv;{\left[\begin{array}{cc}\mathrm{\&delta;}{\hat{P}}_0^T& {\mathrm{\&omega;}}_0^T\end{array}\right]}^T$

$P_{0|-1}^X\&equiv;\left[\begin{array}{cc}{P}_0^{\mathrm{\&delta;P}}& 0_{3\&times;3}\\ 0_{3\&times;3}& P_0^{\mathrm{\&omega;}}\end{array}\right]$

Then, foundation

And P _0|-1 ^XObtain t ₀Point set { the χ of Sigma constantly _0|-1 ⁱ} _I=0 ¹², its corresponding weight value is { w ₀ ^{(m, i)}} _I=0 ¹²{ W ₀ ^{(c, i)}} _I=0 ¹²

Then, utilize transformational relation between hypercomplex number and the GRP parameter:

$\mathrm{\&delta;}{q}_{0|-\mathrm{1,4}}^i=\frac{-a{\left|\right|{\mathrm{\&chi;}}_{0|-1}^{\mathrm{\&delta;P},i}\left|\right|}^2+f\sqrt{f^2+(1-a^2){\left|\right|{\mathrm{\&chi;}}_{0|-1}^{\mathrm{\&delta;P},i}\left|\right|}^2}}{f^2+{\left|\right|{\mathrm{\&chi;}}_{0|-1}^{\mathrm{\&delta;P},i}\left|\right|}^2},$ i＝1，…，12

$\mathrm{\&delta;}{q}_{0|-1}^i=f^{-1}(a+\mathrm{\&delta;}{q}_{0|-\mathrm{1,4}}^i){\mathrm{\&chi;}}_{0|-1}^{\mathrm{\&delta;P},i},$ i＝1，…，12

Obtain priori deviation hypercomplex number Sigma point set ${\left\{\mathrm{\&delta;}{\underset{\&OverBar;}{q}}_{0|-1}^i\right\}}_{i=0}^{12}={\left\{{\left[\begin{array}{cc}{\left(\mathrm{\&delta;}{q}_{0|-1}^i\right)}^T& \mathrm{\&delta;}{q}_{0|-\mathrm{1,4}}^i\end{array}\right]}^T\right\}}_{i=0}^{12},$ And utilize as down conversion

${\underset{\&OverBar;}{q}}_{0|-1}^0={\underset{\&OverBar;}{\overset{^}{q}}}_{0|-1}$

${\underset{\&OverBar;}{q}}_{0|-1}^i=\mathrm{\&delta;}{\underset{\&OverBar;}{q}}_{0|-1}^i\&CircleTimes;{\underset{\&OverBar;}{\overset{^}{q}}}_{0|-1},$ i＝1，…，12

Obtain initial priori hypercomplex number Sigma point set q _0|-1 ⁱ} _I=0 ¹²

B) observation renewal (k=0 ..., N)

At first, according to the sigma point of forecasting observed quantity

$z_k^i=A_k\left({\underset{\&OverBar;}{q}}_{k|k-1}^i\right)b_{R,k},$ i＝0，…，12

CALCULATING PREDICTION observed quantity average

${\hat{z}}_k=\underset{i=0}{\overset{12}{\mathrm{\&Sigma;}}}{W}_k^{(m,i)}{z}_k^i$

Then, utilize ground magnetic vector observation $z_k={\tilde{b}}_k$ Calculating make new advances breath and covariance thereof

${\mathrm{\&upsi;}}_k=z_k-{\hat{z}}_k,$ $P_{k|k-1}^{\mathrm{\&upsi;\&upsi;}}=\underset{i=0}{\overset{12}{\mathrm{\&Sigma;}}}{W}_k^{(c,i)}(z_k^i-{\hat{z}}_k){(z_k^i-{\hat{z}}_k)}^T+R_k$

Wherein, R _kBe geomagnetic observation noise v _{B, k}Covariance.

Relevant covariance matrix and gain matrix are respectively

$P_{k|k-1}^{\mathrm{xz}}=\underset{i=0}{\overset{12}{\mathrm{\&Sigma;}}}{W}_k^{(c,i)}({\mathrm{\&chi;}}_{k|k-1}^i-{\hat{X}}_{k|k-1}){(z_k^i-{\hat{z}}_k)}^T$

$K_k=P_k^{\mathrm{xz}}{\left(P_k^{\mathrm{\&upsi;\&upsi;}}\right)}^{-1}$

State estimation observation is updated to

${\hat{X}}_{k|k}={\hat{X}}_{k|k-1}+K_k{\mathrm{\&upsi;}}_k$

$P_{k|k}^X=P_{k|k-1}^X-K_k{P}_k^{\mathrm{\&upsi;\&upsi;}}{K}_k^T$

Once more by

(be in fact Part) obtaining the deviation hypercomplex number estimates

And finally obtain hypercomplex number observation and upgrade and estimate

${\underset{\&OverBar;}{\overset{^}{q}}}_{k|k}=\mathrm{\&delta;}{\underset{\&OverBar;}{\overset{^}{q}}}_{k|k}\&CircleTimes;{\underset{\&OverBar;}{q}}_{k|k-1}^0$

Note, before next step recursion is estimated,

In the GRP argument section must reset to zero vector.C) time renewal (k=0 ..., N)

At first, foundation

With covariance P _K|k ^XPoint set { the χ of generation state Sigma _K|k ⁱ} _I=0 ¹², its weights are designated as { w _K+1 ^{(m, i)}} _I=0 ¹²{ W _K+1 ^{(c, i)}} _I=0 ¹²

Then, obtain deviation hypercomplex number point set { δ according to the transformational relation between hypercomplex number and the GRP parameter q _K|k ⁱ} _I=0 ¹², and utilize as down conversion

${\underset{\&OverBar;}{q}}_{k|k}^0={\underset{\&OverBar;}{\overset{^}{q}}}_{k|k}$

${\underset{\&OverBar;}{q}}_{k|k}^i=\mathrm{\&delta;}{\underset{\&OverBar;}{q}}_{k|k}^i\&CircleTimes;{\underset{\&OverBar;}{\overset{^}{q}}}_{k|k},$ i＝1，…，12

Obtain the hypercomplex number point set q _K|k ⁱ} _I=0 ¹²

Then, according to q _K|k ⁱ} _I=0 ¹²{ χ _K|k ^{ω, i}} _I=0 ¹²By kinematical equation and euler dynamical equations (3-36) obtain hypercomplex number forecast point set q _K+1|k ⁱ} _I=0 ¹²With angular speed forecast point set { χ _K+1|k ^{ω, i}} _I=0 ¹²

Then, utilize as down conversion

$\mathrm{\&delta;}{\underset{\&OverBar;}{q}}_{k+1|k}^0={\left[\begin{array}{cccc}0& 0& 0& 1\end{array}\right]}^T$

$\mathrm{\&delta;}{\underset{\&OverBar;}{q}}_{k+1|k}^i={\underset{\&OverBar;}{q}}_{k+1|k}^i\&CircleTimes;{\left({\underset{\&OverBar;}{q}}_{k+1|k}^0\right)}^{-1},$ i＝1，…，12

Obtain the point set { δ of deviation hypercomplex number Sigma q _K+1|k ⁱ} _I=0 ¹², be converted to the point set { χ of GRP parameter Sigma at last _K+1|k ^{δ P, i}} _I=0 ¹²

Forecast point set { the x of Sigma _K+1|k ⁱ} _I=0 ¹²Average and covariance be respectively

${\hat{X}}_{k+1|k}=\underset{i=0}{\overset{12}{\mathrm{\&Sigma;}}}{W}_{k+1}^{(m,i)}{x}_{k+1|k}^i$

$P_{k+1|k}^X=\underset{i=0}{\overset{12}{\mathrm{\&Sigma;}}}{W}_{k+1}^{(c,i)}({\mathrm{\&chi;}}_{k+1|k}^i-{\hat{X}}_{k+1|k}){({\mathrm{\&chi;}}_{k+1|k}^i-{\hat{X}}_{k+1|k})}^T+C_{k+1}{Q}_{k+1}{C}_{k+1}^T$

Wherein, C _K+1=[(0 _{3 * 3}) ^TI ^-T] ^TQ _K+1The discrete disturbance torque N of representative _{D, k+1}Covariance.

The introduction of more detailed contents about above-mentioned wave filter is please referring to document: Huang Lin. the application of nonlinear filtering theory in spacecraft attitude is determined. and the doctor of Harbin Institute of Technology thesis .2007, the 3rd chapter.

(3) the appearance control computer judges according to shown in Figure 4 whether the current attitude of satellite has arrived targeted attitude.Wherein, a) calculate the deviation of current definite attitude and targeted attitude: utilize target hypercomplex number q here _ComEstimate with current posteriority hypercomplex number

Between the deviation hypercomplex number $\mathrm{\&Delta;}{\underset{\&OverBar;}{q}}_k={\underset{\&OverBar;}{\overset{^}{q}}}_{k|k}\&CircleTimes;{\left({\underset{\&OverBar;}{q}}_{\mathrm{com}}\right)}^{-1}$ The rotation Eulerian angle as the evaluation index of control deviation, promptly

Δα＝2arccos(Δq ₄)

Wherein, Δ q ₄Expression deviation hypercomplex number Δ q _kScalar component.B) as Δ α＜G ₃When occurring first (t=0), then order characterizes the parameter K of " arrival targeted attitude " ₃=1, t=1; After this each be (0＜t＜W constantly ₃), if Δ α＜G ₃, then make K ₃=K ₃+ 1, t=t+1; Otherwise, make t=t+1; When having carried out W ₃(t=W after the individual moment ₃), calculate C ₃=K ₃/ W ₃, if C ₃＞R ₃, think that then the attitude of satellite has reached targeted attitude, and make K ₃=0, t=0, attitude is saved task termination; Otherwise, think attitude of satellite miss the mark attitude still, and make K ₃=0, t=0.

Selection about Several Parameters.Save the steady state controling precision (such as, 6 degree) of scheme according to attitude, make G ₃=6 degree, W ₃=10, R ₃=0.67.

Learn satellite if reach targeted attitude according to above-mentioned determination methods, then save task termination; Otherwise, change step (4) over to.

(4) the appearance control computer judges currently whether need to start motor-driven control executing mechanism and control according to shown in Figure 5.Wherein, work as υ _k＜G ₄When occurring first (t=0), then order characterizes the parameter K of " restarting control " ₄=1, t=1; After this each be (0＜t＜W constantly ₄), if υ _k＜G ₄, then make parameter K ₄=K ₄+ 1, t=t+1; Otherwise, make t=t+1; When having carried out W ₄(t=W after the individual moment ₄), calculate C ₄=K ₄/ W ₄, if C ₄＞R ₄, think then that satellite is restarted to have arrived the opportunity that topworks controls, make K ₄=0, t=0; Otherwise, think that the current attitude of satellite determines that error is excessive, still can not be used for the usefulness of motor-driven control, and make K ₄=0, t=0.

The selection of above-mentioned parameter should determine that the emulation experience is determined according to the attitude that attitude is saved scheme, such as, make G ₃=1000nT; W ₃=30, R ₃=0.67.

If judge according to said method and to restart control and arrived opportunity, utilize attitude to determine information after carving at this moment

With the targeted attitude director data q _Com, ω _ComThe calculation control deviation information

${\underset{\&OverBar;}{q}}_e={\underset{\&OverBar;}{\overset{^}{q}}}_{k|k}\&CircleTimes;{\left({\underset{\&OverBar;}{q}}_{\mathrm{com}}\right)}^{-1}$

${\mathrm{\&omega;}}_e={\widehat{\mathrm{\&omega;}}}_{k|k}-{\mathrm{\&omega;}}_{\mathrm{com}}$

Then according to the following rank-saturated PD control law of passing

$T_c=-\underset{T_{\max }}{\mathrm{sat}}I[k_p\underset{L_i}{\mathrm{sat}}\left(q_e\right)+k_d{\mathrm{\&omega;}}_e]$

Produce control moment instruction T _cBecause control wave is adopted in jet control, therefore, also the control moment instruction transformation must be become the gating pulse instruction.In order to introduce for simplicity, suppose nozzle along three axial direction mounted in pairs, the moment amplitude of each nozzle is T ₀, then the instruction of the jet pulse of nozzle is $t_{{\mathrm{on}}_i}=T_{\mathrm{ci}}/T_0,$ Yet there is following constraint in jet control: jet time is lower than minimum jet pulse t _MinShi Wufa produces effective control moment, and jet time does not allow to exceed maximum spout gas pulse t yet continuously _MaxConstraint (1s), therefore, adopt following a kind of jet strategy here:

$t_{{\mathrm{on}}_i}=\left\{\begin{array}{cc}0& t_{{\mathrm{on}}_i}<t_{\min }\\ t_{{\mathrm{on}}_i}& t_{\min }<t_{{\mathrm{on}}_i}<t_{\max }\\ T_0& t_{{\mathrm{on}}_i}>t_{\max }\end{array}\right.$

Jet mechanism carries out jet control according to above-mentioned jet strategy, orders about the attitude of satellite and carries out corresponding motor-driven control.

The content that the present invention does not describe in detail is a technology as well known to those skilled in the art.

Claims (5)

1, a kind of quick retrieval method for satellite attitude is characterized in that comprising the following steps:

(1) to judge that current attitude has or not out of control for the appearance control computer, if not out of control, then carries out normal attitude control operation; Otherwise the appearance control computer makes current topworks stop control, and runaway condition is fed back to the Star Service central computer closes inessential equipment on the star, enters step (2) then;

(2) judgement has or not the gyro observation information, if the gyro observation information is arranged, then utilize gyro attitude observation information and ground magnetic vector observation and orientation vector observation information to determine wave filter, the current attitude of satellite and gyro are often floated parameter to be observed upgrading and handles, and learn equation and gyro mechanical model time of carrying out that often floats based on attitude motion and upgrade forecast, enter step (3) then; If there is not the gyro observation information, then utilize no gyro attitude observation information and ground magnetic vector observation and orientation vector observation information to determine wave filter, observe renewal handle to the current attitude of satellite and celestial body angular speed, and learn equation and Euler's kinetic model time of carrying out is upgraded forecast based on attitude motion, enter step (3) then;

(3) the appearance control computer judges whether the current attitude of satellite has arrived targeted attitude, if "Yes", then attitude is saved task termination; Otherwise, change step (4) over to;

(4) the appearance control computer judges whether current time needs to start motor-driven control executing mechanism and control, if "Yes", then utilize current attitude to determine information and targeted attitude instruction, produce the control moment instruction according to three-axis attitude fast reserve control law, drive motor-driven control executing mechanism and carry out attitude maneuver control; Otherwise, return step (2).

2, a kind of quick retrieval method for satellite attitude according to claim 1 is characterized in that: the determination methods of runaway condition is in the described step (1):

(a) constantly, according to the nominal installation site of nominal satellite operate as normal attitude and sensor, calculate the magnetic vector observation of nominal ground of three axis magnetometer and estimate, as nominal observed quantity in each observation;

(b) calculate the deviation angle of ground magnetic vector observed quantity and nominal observed quantity;

(c) if the deviation angle of current time first greater than judgment threshold G out of control ₁, then starting the statistics process, order characterizes the parameter K of " attitude is out of control " ₁=1, in each observation after this constantly, if both angles are greater than threshold value G ₁, then make parameter K ₁=K ₁+ 1, otherwise, K ₁Remain unchanged; W ₁After the individual moment, calculate C ₁=K ₁/ W ₁, if C ₁＞R ₁, think that then the attitude of satellite is out of control, and with K ₁Zero clearing; Otherwise, think that the attitude of satellite is normal, equally with K ₁Zero clearing; If if current time deviation angle is first smaller or equal to threshold value G out of control ₁, then carry out the judgement in the next moment, wherein R ₁Expression statistics empirical probit.

3, a kind of quick retrieval method for satellite attitude according to claim 1 is characterized in that: the appearance control computer judges that the method whether current attitude of satellite has arrived targeted attitude is in the described step (3): the satellite as if current time determines that the deviation of attitude and targeted attitude is first less than threshold value G ₃, then starting the statistics process, order characterizes the parameter K of " arrival targeted attitude " ₃=1, in each observation after this constantly, the deviation of the attitude of satellite and targeted attitude is less than error threshold G ₃, then make K ₃=K ₃+ 1, otherwise, remain unchanged; When carrying out W ₃After the individual moment, calculate C ₃=K ₃/ W ₃, if C ₃＞R ₃, think that then the attitude of satellite arrives near the targeted attitude, and with K ₃The value zero clearing; Otherwise, think the attitude of satellite not yet Be Controlled arrive near the targeted attitude, and with K ₃The value zero clearing; If the satellite of current time determines that the deviation of attitude and targeted attitude is first more than or equal to threshold value G ₃, then carry out the judgement in the next moment, wherein R ₃Expression statistics empirical probit.

4, a kind of quick retrieval method for satellite attitude according to claim 1 is characterized in that: the appearance control computer judges whether current time needs to start the method that motor-driven control executing mechanism controls and be in the described step (4): as if the new breath of current time first less than error threshold G ₄, then starting the statistics process, order characterizes the parameter K of " restarting control " ₄=1, each observation updated time after this is if newly cease less than error threshold G ₄, then make parameter K ₄Add 1, otherwise, remain unchanged; Work as W ₄After the individual moment, calculate C ₄=K ₄/ W ₄, if C ₄＞R ₄, then think and arrived the opportunity of restarting control, and with K ₄The value zero clearing, otherwise, think that current attitude determines that error is excessive, should continue to utilize observed quantity to dwindle attitude and determine error, and with K ₄The value zero clearing; If the new breath of current time is first more than or equal to error threshold G ₄, then carry out the judgement in the next moment, wherein R ₄Expression statistics empirical probit.

5, a kind of quick retrieval method for satellite attitude according to claim 1 is characterized in that: three-axis attitude fast reserve control law adopts and passs rank-saturated PID control law in the described step (4).

Images