CN110926480B

CN110926480B - Autonomous aggregation method for remote sensing satellite imaging tasks

Info

Publication number: CN110926480B
Application number: CN201911295970.8A
Authority: CN
Inventors: 袁利; 朱琦; 张聪; 李勇; 瞿涵; 王云鹏; 李志辉
Original assignee: Beijing Institute of Control Engineering
Current assignee: Beijing Institute of Control Engineering
Priority date: 2019-12-16
Filing date: 2019-12-16
Publication date: 2020-11-10
Anticipated expiration: 2039-12-16
Also published as: CN110926480A

Abstract

An autonomous aggregation method for remote sensing satellite imaging tasks includes such steps as recognizing the intensity of tasks in local planning window, further judging whether the candidate task set to be observed needs autonomous aggregation of remote sensing satellite imaging tasks, when the candidate task set to be observed needs autonomous aggregation, dividing the candidate task set to be observed to obtain a plurality of intensive task subsets, then respectively calculating the circumscribed polygons of the dense task subsets, determining the task aggregation mode of the dense task subsets according to the circumscribed polygons of the dense task subsets, then dividing each intensive task subset into a multi-band observation subtask set and a regional observation subtask set, finally generating a multi-band observation aggregated task according to the multi-band observation subtask set, and performing push-broom band division, and generating a regional observation aggregation task according to the plurality of regional observation subtask sets, and dividing push-broom strips to finish autonomous aggregation of the remote sensing satellite imaging task.

Description

Autonomous aggregation method for remote sensing satellite imaging tasks

Technical Field

The invention relates to an autonomous aggregation method for remote sensing satellite imaging tasks, and belongs to the technical field of autonomous task planning of spacecrafts.

Background

The remote sensing satellite with the autonomous task planning capability needs to face the multi-source complex task input problem, and task sources comprise an imaging task provided by a multi-source terminal user, a high-priority imaging task of ground emergency injection, a cooperative guiding task sent by a preamble satellite through an inter-satellite link in a multi-satellite cooperative scene, an imaging task autonomously discovered and generated by an on-satellite perception decision module and the like. Because the multi-source imaging task is not subjected to overall processing of the ground management and control system, the ground target to be observed in the local area is densely distributed. The existing remote sensing satellite autonomous task planning method does not consider the influence of multi-source task input on satellite autonomous task planning, does not relate to the problems of autonomous aggregation and task optimization of targets to be observed on dense ground before satellite autonomous task planning, generally considers the targets to be observed on the ground as independent targets to be observed, so that in the region with dense ground targets to be observed, the attitude maneuvering capability of the remote sensing satellite can not meet the time requirement of load camera pointing adjustment between adjacent ground targets to be observed with compact geographic distance, the problems that the imaging tasks of the targets to be observed on the ground are less executed in unit time, the imaging tasks with high timeliness requirement are difficult to respond in time and the like can not be fully exerted, therefore, the existing remote sensing satellite autonomous task planning method cannot be directly applied to the scene of local intensive tasks in a complex environment.

Disclosure of Invention

The technical problem solved by the invention is as follows: the method overcomes the defects of the prior art, provides the autonomous aggregation method of the imaging tasks of the remote sensing satellite, solves the problem of low task efficiency of the remote sensing satellite in unit time caused by the fact that the remote sensing satellite receives earth observation tasks from multiple terminals and the local areas are densely distributed, is suitable for the agile remote sensing satellite with autonomous task capability, and lays a foundation for realizing the efficient task planning on the satellite.

The technical solution of the invention is as follows: an autonomous polymerization method for remote sensing satellite imaging tasks comprises the following steps:

(1) identifying the task intensity degree in the range of a local planning window, further judging whether a candidate task set to be observed needs autonomous aggregation of remote sensing satellite imaging tasks, if so, turning to the step (2), otherwise, ending the current method, wherein the short period of autonomous task planning of the remote sensing satellite comprises a plurality of local time windows;

(2) dividing the candidate task set to be observed to obtain a plurality of intensive task subsets sub₁,…,sub_nNo aggregated independent observation tasks are performed;

(3) computing a circumscribed polygon of the intensive task subset;

(4) determining a task aggregation mode of the dense task subset according to a circumscribed polygon of the dense task subset, and further dividing each dense task subset into a multi-strip observation sub-task set and a regional observation sub-task set;

(5) generating a multi-strip observation aggregation task according to the multi-strip observation subtask set, and dividing push-broom strips;

(6) and generating a region observation aggregation task according to the plurality of region observation subtask sets, and dividing push-broom strips.

The method for identifying the task intensity in the local planning window range comprises the following steps:

(11) for a local time window (T)₁,T₂) Calculating to obtain a local time window (T) according to the track extrapolation information of the remote sensing satellite and the maximum maneuvering capacity in the rolling pitch direction₁,T₂) The corresponding ground target to be observed can observe the regional scope;

(12) screening all the ground targets to be observed in the range of the region where the ground targets to be observed can be observed according to the longitude, latitude and elevation information of each ground target to be observed in the satellite-borne task pool to form a local time window (T)₁,T₂) Corresponding candidate task set to be observed to be optimized

Wherein m is a positive integer, s_r＝(long_r,lat_r,alt_r) Longitude long corresponding to target r to be observed on ground_rLatitude lat_rElevation information alt_r；

(13) Based on candidate task set to be observed

And identifying whether the candidate task set to be observed needs to be subjected to task autonomous aggregation or not according to the relative geographic distance between the targets to be observed on the middle ground.

The method for identifying whether the candidate task set to be observed needs to be subjected to task autonomous aggregation comprises the following steps:

if the distance between the targets to be observed on the adjacent ground in the candidate task set to be observed is less than delta L, the candidate task set to be observed needs to be subjected to task autonomous aggregation;

otherwise, no optimizable ground target to be observed exists in the candidate task set to be observed, and the tasks are not aggregated autonomously.

Said

Wherein eta is₁、η₂0 is greater than or equal to eta as a regulating factor₁<5， 0≤η₂<5, v represents the linear velocity module value of the satellite subsatellite point,

for the maximum roll angle of the satellite imaging,

the length of an arc section between the ground point and the ground subsatellite point is pointed when the satellite images at the maximum rolling angle,

h is orbital height, R is earth radius, Δ T_vIt is possible to set it to 20s,

a threshold value for angular deviation of satellite imaging poses of adjacent targets in the roll direction.

The method for dividing to obtain the dense task subset comprises the following steps:

(21)

and is

s.t.d_j,k＝d(s_j,s_k)≤ΔL

Wherein the function d(s)_j,s_k) For calculating two on the earth's surfaceThe distance between the points;

(22)

s.t.d_j,k>ΔL，s_jno polymerization is carried out;

(23) traversing to obtain a candidate task set to be observed

The concentrated task subset and the independent ground target observation tasks to be observed which are not aggregated are obtained.

The method for computing the circumscribed polygon of the intensive task subset comprises the following steps:

(31) calculating sub_iGathering intensive tasks into a subset sub through the geometric centers of all the targets to be observed on the ground_i＝{s₁,…,s_kThe geometric center of a limited number of target observation points is denoted as "(long _ o)_i,lat_o_i,alt_o_i) Wherein i and k are positive integers,

(32) using omicron as circle center, to_iThe maximum distance of each ground target to be observed is the radius to form sub_iThe circumscribed circle of the target to be observed on each ground;

(33) to sub_iEach ground target to be observed positioned on the circumcircle and part of target points close to the circumcircle are taken as vertexes, and a convex polygon is drawn, so that the convex polygon comprises all ground targets to be observed in the current dense task subset;

(34) b/2 of parallel outward expansion of each side of the convex polygon obtained in the step (33) to obtain a dense task subset sub_iThe circumscribed polygon of (a), wherein b is the width of the remote sensing satellite load camera.

The method for determining the task aggregation mode of the dense task subset according to the circumscribed polygon of the dense task subset comprises the following steps:

(41) will sub_iWhich is divided into n_iA square small grid with the side length of the load camera width b;

(42)

wherein, num (sub)_i) Is sub_iThe number of targets to be observed on the ground;

(43) if u is_i<U, then sub_iThe ground target to be observed is unevenly distributed in the sub-area, and sub-observation is carried out according to a multi-strip observation mode_iAggregating targets to be observed on the middle ground; if u is_iGreater than or equal to U, then sub_iThe ground target to be observed is uniformly distributed in the sub-area, and sub-is carried out according to the area observation mode_iThe aggregation of observation tasks of the targets to be observed on the middle ground, wherein 0<U≤1；

(44) Repeating the above steps to obtain sub₁,…,sub_nPartitioning into a set of multi-banded Observation subtasks S_stripRegional observation subtask set S_area。

Said

Wherein the content of the first and second substances,

representing the maximum mobility of the satellite platform. Rho₁，ρ₂，ρ₃To determine the threshold parameter.

The step (5) of generating the multi-stripe observation aggregated task according to the multi-stripe observation subtask set, and the method for dividing the push-broom stripes comprises the following steps:

(51) for sub_i∈S_stripThe optimization target for aggregating the targets to be observed on the ground into strips is

The constraint conditions comprise point-to-point attitude mobility, passive imaging push-broom linear velocity average value and active imaging even groundThe average value of the speed tracking push-broom angular velocity and the imaging advance time t before the target to be observed points to the ground_prePointing to the imaging delay time t after the target to be observed on the ground_after；

Wherein N is the number of target observation points covered by the current stripe observation aggregation, p_jAs a task s_j∈sub_iCorresponding priority if s_j∈sub_i∈S_stripWhen is lambda _j1, otherwise λ_j＝0，T_{man_add}The sum of the attitude maneuver time between adjacent strips is obtained from the attitude maneuver time T of each adjacent strip_manAre added to obtain_{image_add}For the sum of the sweep times of the individual strips, from each strip sweep time T_imageAre added to obtain₁、α₂To optimize the target adjustment coefficient, α₁、α₂The value range of (A) is 0 to 1;

(52) when the strip formed by polymerization is in the non-tracking direction, the time T required by the posture presetting process before imaging_manTime T required for imaging process_imageIs composed of

T_man＝α/ω₁

T_image＝β/ω₂

Wherein the content of the first and second substances,

presetting the forward roll angle, theta, for attitude_sPresetting front pitch angle psi for attitude_sA front yaw angle is preset for the attitude,

presetting a rear roll angle, theta, for attitude_ePresetting rear pitch angle psi for attitude_ePresetting a rear yaw angle for the attitude;

for sweeping the strip by a forward roll angle, theta_{str_s}To push the strap across the front pitch angle,

for the strip to push-sweep the roll angle, theta_{str_e}Pushing and sweeping a rear pitch angle for the strip;

when the strip formed after polymerization is in the direction of the trace, the time T required by the posture presetting process before imaging_manUsing satellite point-to-point attitude maneuver capability θ for the difference γ between the satellite's current attitude and the imaged attitude₁°/t₁s、θ₂°/t₂s、θ₃°/t₃s is obtained by interpolation calculation, wherein,

for the current attitude roll angle, θ, of the satellite_nowFor the satellite's current attitude pitch angle, psi_nowFor the current attitude yaw angle of the satellite, the attitude roll angle is imaged along the track direction

The attitude pitch angle of the imaging along the track direction is theta_pasThe attitude yaw angle along the track direction is psi_pas；

Time T required for imaging process_imgaeIs composed of

Wherein, the latitude and longitude of the offset flight imaging starting point and the end point are respectively long_s、long_e、lat_s、lat_e；

(53) And (5) generating an optimization model according to the push-broom strips formed in the step (51), generating optimal push-broom strips which meet constraint conditions and enable the sum of attitude maneuver time between adjacent strips to be minimum and the sum of push-broom time of each strip to be minimum, enabling the number N of target observation points covered by the strip observation aggregation to be maximum, completing the generation of a multi-strip observation aggregation task, and dividing the push-broom strips.

The method for generating the regional observation aggregated task according to the plurality of regional observation subtask sets and dividing the push-broom stripe comprises the following steps:

(61) for sub_i∈S_areaThe longest line segment passing through the circumscribed polygon and passing through the geometric center of the circumscribed polygon is taken as

Will pass through the geometric center of the circumscribed polygon, and

the longest line segment perpendicular to and within the polygon is denoted as

(62) The width b of the load camera is taken as the width of the strip so as to

Calculating the sigma% overlap between adjacent bands for the band length, from sub_iAlong one side of the circumscribed polygon

In the direction of (sub)_iIs divided into strips of a circumscribed polygon and flies from a remote sensing satellite through sub_iStarting scanning in the initial direction of the area; wherein the strip covers more than sub_iThe circumscribed polygon part or the part of the target to be observed, which is not provided with the ground at one end of the strip, is not pushed and swept.

Compared with the prior art, the invention has the advantages that:

(1) the method comprises the steps of identifying the task intensity degree in the range of a local planning window, dividing a dense task subset, calculating an external polygon of a ground target to be observed of the dense task subset, judging the aggregation type of imaging tasks, generating a multi-band observation mode task and dividing a push-broom band, generating a region observation mode task and dividing the push-broom band, autonomously classifying and combining the originally independent dense ground target to be observed into a multi-band observation task or a region observation task, and solving the problem of low task efficiency of a remote sensing satellite in unit time caused by the fact that the ground observation tasks from multiple terminals received by the remote sensing satellite are distributed densely in a local region;

(2) compared with the prior art, the method is based on pushing sweep of the agile remote sensing satellite along the flight direction and imaging pushing sweep of the satellite in the non-tracking direction, two imaging modes of passive imaging and active imaging are considered in the imaging process, and the output of the aggregated planning result can be divided into a multi-strip observation task and a regional target observation task according to the distribution uniformity degree of ground observation points in a dense observation region, so that more optimized input can be provided for an on-satellite autonomous task planning method, and the attitude maneuvering capability of the agile remote sensing satellite can be fully exerted;

(3) the method has clear steps, is convenient for operation and implementation of designers and testers, has been successfully applied to ground tests and tests, and verifies the effectiveness and feasibility of the method.

Drawings

FIG. 1 is a flow chart of a method for autonomous aggregation of remote sensing satellite imaging tasks;

FIG. 2 is a set of objects to be observed on the ground within a local time window

Examples are given;

FIG. 3 is sub₁Circumscribed polygon examples;

FIG. 4 is sub₂Circumscribed polygon examples;

FIG. 5 is sub₁The task aggregation result example of the multi-strip observation mode;

FIG. 6 is sub₂The regional observation mode task aggregate result example of (1).

Detailed Description

The autonomous task planning technology can improve the intelligent autonomous operation and task quick response capability of the remote sensing satellite, and obtains greater economic benefit by improving the application efficiency of the remote sensing satellite, so that the method has stronger market demand. However, earth observation tasks from multiple terminals are not managed by ground overall, and if the earth observation tasks are distributed densely in a local area, the task efficiency of the remote sensing satellite in unit time is not high, which provides challenges for effective implementation and efficient operation of on-satellite autonomous task planning.

Aiming at the scene that the prior art can not be applied to local intensive tasks in a complex environment, the invention provides an autonomous aggregation method of remote sensing satellite imaging tasks from the perspective of engineering application and considering the imaging capability in a non-tracing track and the imaging capability of a strip splicing area of an agile remote sensing satellite, after the remote sensing satellite receives imaging tasks input by multiple sources and completes the task screening of a current planning window, the original independent intensive ground to-be-observed targets are autonomously classified and combined into multi-strip observation tasks or area observation tasks through the identification of the task intensity degree in the range of the local planning window, the division of an intensive task subset, the calculation of an external polygon of the intensive subset ground to-be-observed targets, the judgment of the aggregation type of the imaging tasks, the generation and the division of multi-strip observation mode tasks and the division of push-to-be-observed strips, the generation and the division of the push-to-be-observed tasks, the method provides more optimized task input for the on-satellite autonomous task planning algorithm, and fully exerts the complex path attitude maneuvering capability of the agile remote sensing satellite, thereby greatly improving the task execution efficiency of the remote sensing satellite in unit time. The process of the invention is explained and illustrated in more detail below with reference to the drawing.

The invention relates to an autonomous aggregation method of remote sensing satellite imaging tasks, which autonomously aggregates dense ground targets to be observed into a multi-strip observation task or a regional observation task on a satellite before task planning through the following 6 steps, so that the task execution efficiency of the remote sensing satellite in unit time is improved, and the flow of the 6 steps is shown in figure 1.

(1) Task intensity recognition within a local planning window

Autonomous mission gauge on remote sensing satelliteIn the fine planning process of the planned short-period local time windows (a plurality of local time windows are included in the short period of the autonomous mission planning of the remote sensing satellite), aiming at each local time window, for example (489060100s,489060200s), calculating to obtain the observable area range of the ground target to be observed corresponding to the local time window according to the track extrapolation information of the remote sensing satellite and the maximum maneuvering capacity in the rolling and pitching directions; screening all the ground targets to be observed with the geographic positions within the area range according to the longitude and latitude and elevation information of each ground target to be observed in the satellite-borne task pool to form a candidate task set to be optimized

Wherein s is_r＝(long_r,lat_r,alt_r) Longitude long corresponding to target r to be observed on ground_rLatitude lat_rElevation information alt_r. On the basis, based on the candidate task set to be observed

Identifying whether the candidate task set to be observed needs to be subjected to task autonomous aggregation or not by relative geographic distance between the targets to be observed on the middle ground, wherein the basic method comprises the following steps:

1) if the distance between adjacent ground targets to be observed in the area range is smaller than delta L, identifying that optimizable ground targets to be observed exist in the area range (the ground targets to be observed are the ground targets to be observed of which the tasks need to be observed), and starting a task set

Autonomous aggregation of tasks;

2) otherwise, the optimizable ground target to be observed does not exist in the area range, and the task autonomous aggregation is not performed, namely the target to be observed corresponding to the current local time window does not need to be performed the task aggregation.

Wherein, the setting of the delta L mainly considers the deviation of the over-vertex time of the adjacent target satellite and the rolling angle deviation of the satellite when the adjacent target satellite is imagedTwo factors are different. Wherein the satellite over-the-top time refers to the time when the target is imaged with a pitch zero attitude. It is considered that it is meaningful to perform task aggregation on the adjacent targets to be observed only when the difference between the overhead time of the adjacent targets to be observed on the ground is small and the difference between the rolling angles of the satellites imaging the targets is small. Will be Delta T_vSet as a threshold for the difference between the time when the satellite crosses the top of the neighboring target,

and setting a threshold value of the angular deviation of the satellite imaging postures of the adjacent targets in the rolling direction. Converting the above two factors into expressions in terms of distance, a calculation method of Δ L can be obtained:

wherein eta₁、η₂To regulate the factor, η₁＝3，η₂2, the module value v of the linear velocity of the satellite lower point is 7.39 km/s;

a maximum roll angle for satellite imaging;

when the satellite is imaged at the maximum rolling angle, the length of an arc section between the ground point and the ground subsatellite point is pointed;

wherein the orbit height H is 500km, and the earth radius R is 6378.14 km. Suppose that

Then

Can be set to 10 DEG, Delta T_vCan be set to 20s, therefore

ΔL＝3·20·7.39+2·522.2·10/45＝675.4

For example, as shown in fig. 2, there are 16 ground targets to be observed in the local time window range, which is recorded as

The details are shown in the following table

TABLE 1 ground Observation target information

Wherein s is₁And s₂The geographical distance therebetween is less than the decision threshold deltal,

meeting the criteria of task density within a local time window as described above, so a task set needs to be assembled

And performing autonomous task aggregation. Furthermore, if s is to be₁,…,s₁₆The 16 target points carry out on-satellite autonomous task planning according to independent ground targets to be observed, and a planning scheme and expected execution conditions given according to the attitude mobility of the remote sensing satellite are shown in figure 2, and only s can be planned and executed₁,s₃,s₅,s₇,s₉,s₁₄And six targets to be observed on the ground, and the other ten targets to be observed on the ground do not have execution conditions, so that an imaging task is not planned.

(2) Dense task subset partitioning

If the local area is identified to have intensive tasks in the step 1, autonomous aggregation of the earth observation intensive tasks is required, and the condition that the relative distance of the target to be observed on the ground in the local area is smaller than a threshold value is firstly identifiedAnd generating a dense task subset corresponding to each dense sub-region. Candidate task set to be observed screened based on last step

By calculating the relative distance between the geographic positions of the targets to be observed on each ground, the method further comprises the step of calculating the relative distance between the geographic positions of the targets to be observed on each ground

Divided into several dense task subsets sub₁,…,sub_n. The basic method of dense task subset partitioning is as follows:

1)

and is

Wherein the function d(s)_j,s_k) Defined as calculating the distance between two points on the surface of the earth; the meaning of the formula (1) is that the ground is used for observing an object s to be observed_jClassification in dense task subset sub_iIf and only if_iIn presence of and s_jIs smaller than the threshold value delta L. From the formula (1) can be found_iThe number of the targets to be observed on the ground is more than or equal to 2.

2) If for

I.e. the ground object s to be observed_jAnd

other ground to-be-observed target distanceIf the deviation is larger than the threshold value delta L, s is set_jRemain as independent observation tasks and do not aggregate.

The distance between the targets to be observed on the adjacent ground is calculated in a traversal way according to the method and compared with the threshold value delta L, and the task set shown in the figure 2 can be obtained according to the calculation result

Two dense task subsets sub₁＝{s₁,…,s₅}、sub₂＝{s₆,…,s₁₅The ground target s to be observed₁₆And if the geographic distances between the target and the other ground targets to be observed are larger than delta L, keeping the target to be observed as an independent ground target observation task. In FIG. 2

d_1,2,d_2,3,d_3,4,d_4,5≤ΔL，

d_6,8,d_8,9,d_8,11,d_7,10,d_10,12,d_9,10,d_9,13,d_13,14,d_13,15≤ΔL

(3) Computing external polygon of target to be observed on dense subtask set ground

For sub₁,…,sub_nRespectively calculating each intensive task subset sub_iAnd the external polygon of the target to be observed on each ground surface in the observation area. The calculation method comprises the following steps:

1) first calculate sub_iThe geometric center of each target to be observed on the inner ground;

order intensive task subset sub_i＝{s₁,…,s_k-a geometrical centre of a limited number of target observation points is o, o ═ long o_i,lat_o_i,alt_o_i) The method can be obtained according to the longitude and latitude and the elevation information of each target observation point.

2)

Takes the geometric center as the center of a circle and takes the geometric center to_iThe maximum distance between the targets to be observed on each ground is the radius to form sub_iThe eyes of the middle and various ground to be observedA target circumscribed circle;

3) to sub_iEach ground target to be observed positioned on the circumcircle and part of target points close to the circumcircle are taken as vertexes, and a convex polygon is drawn, so that the convex polygon comprises all ground targets to be observed in the current dense task subset;

4) considering that the width of the remote sensing satellite load camera is b, on the basis of the convex polygon obtained in the previous step, respectively enabling each side to be parallel and outwards expanded by b/2 to obtain corresponding straight lines, wherein the polygon surrounded by intersection points of all the straight lines is used as a dense task subset sub_iIs a circumscribed polygon.

Sub in the example given in fig. 2₁、sub₂Respectively, as shown in fig. 3 and 4. O in figure 3₁Is sub₁Geometric center, s, of the target to be observed in each of the interior surfaces₁s₂s₃s₄s₅Enclosed by sub₁The circumscribed circle of each ground target to be observed is an inward convex polygon r₁r₂r₃r₄r₅r₆r₇r₈Enclosed by sub₁Is a circumscribed polygon. O in figure 4₂Is sub₂Geometric center, s, of the target to be observed in each of the interior surfaces₆s₁₄s₁₅s₁₂s₇Enclosed by sub₂The circumscribed circle of each ground target to be observed is an inward convex polygon r₉r₁₀r₁₁r₁₂r₁₃r₁₄r₁₅r₁₆r₁₇r₁₈Enclosed by sub₂Is a circumscribed polygon.

(4) Imaging task aggregation type determination

Each intensive task subset sub obtained by calculation in the last step₁,…,sub_nOn the basis of the circumscribed polygon, for each circumscribed polygon, e.g. sub_iRespectively according to the distribution uniformity u of the ground target to be observed_iDetermination of sub_iThe specific method of the task aggregation mode is

1) Will sub_iWhich is divided into n_iSquare with length of one side being width b of loaded cameraForming small lattices;

2)

wherein num (sub)_i) Is sub_iThe number of targets to be observed on the ground;

3) if u is_i<U, consider as sub_iIf the ground target to be observed is not uniformly distributed in the sub-area, performing sub-observation according to a multi-strip observation mode_iAnd aggregating observation tasks of the targets to be observed on the middle and various surfaces.

4) If u is_iGreater than or equal to U, considered as sub_iIf the ground target to be observed is uniformly distributed in the sub-area, performing sub-observation according to the area observation mode_iAnd aggregating observation tasks of the targets to be observed on the middle and various surfaces.

And the judgment threshold value 0< U < 1, wherein the specific value is determined by the attitude mobility of the remote sensing satellite (when the attitude mobility of the remote sensing satellite is stronger, the value of U is larger).

ρ₁，ρ₂，ρ₃In order to determine the threshold value parameter,

representing maximum satellite platform maneuverability, i.e. 1 second maximum maneuverability

The above step can be₁,…,sub_nPartitioning into a set of multi-banded Observation subtasks S_stripRegional observation subtask set S_area。

Sub given in fig. 3, assuming that U is 0.5₁Number num (sub) of targets to be observed on middle ground₁) A square small grid n with side length of 5 and width of load camera b₁Is 12, therefore u₁＝0.42<U, then sub₁The aggregation task type of the system is a multi-band observation mode;sub given in FIG. 4₂Number num (sub) of targets to be observed on ground in circumscribed polygon₂) 10, square small grid n with side length of load camera width b₂Is 16.4, so u₂＝0.61>U, then sub₂The aggregated task type of (2) is a region observation mode.

(5) Multi-stripe observation aggregation task generation and push-broom stripe division

For sub_i∈S_stripAggregating the ground targets to be observed into the strip targets in an optimization solving mode, wherein the optimization problem is set as follows:

(2) optimizing an objective

(3) Constraint conditions

a) Point-to-point attitude maneuver capability: in-position and settling times for typical angular pointing adjustments, e.g. theta₁°/t₁s＝5°/6s、θ₂°/t₂s＝10°/7.5s、θ₃°/t₃s is 20 °/10s, meaning the satellite attitude maneuver θ_kThe attitude maneuver is in place and stable at t_ks, wherein k is 1,2, 3;

b) attitude preset angular velocity average value omega₁＝1°/s；

c) The average value v of the passive imaging push-broom linear velocity is 7.39 km/s;

d) active imaging uniform ground speed tracking push-broom angular speed average value omega₂＝0.5°/s；

e) Imaging advance time t before pointing to ground target to be observed_pre1.5s, imaging delay time t after the target to be observed points to the ground_after＝1.5s。

Wherein N is the number of target observation points covered by the current stripe observation aggregation, p_jAs a task s_j∈sub_iCorresponding priority when task s_jWhen covered by the generated push-broom stripe partitioning scheme (i.e., s)_j∈sub_i∈S_stripTime) λ _j1, otherwise λ_j＝0，T_{man_add}Between adjacent stripsSum of attitude maneuver times, from the attitude maneuver time T of each adjacent strip_manAre added to obtain_{image_add}For the sum of the sweep times of the individual strips, from each strip sweep time T_imageAdding to obtain the optimized target regulation coefficient alpha₁＝0.3、α₂＝0.5。

According to an optimization target formula, the optimization index J is in direct proportion to the number of target observation points covered by the strip observation aggregation, and the attitude maneuver time among strips is in inverse proportion to the strip push-scan time, so that the principle of designing the optimization target is to cover as many ground imaging tasks as possible by using the strip push-scan time as less as possible.

For the condition that the formed strip after polymerization is in the non-tracking direction, the time T required by the posture presetting process before imaging_manTime T required for imaging process_imgaeThe calculation method is as follows:

make the posture preset forward roll angle as

Attitude preset front pitch angle theta_s10.4 degrees, and preset front yaw angle psi of attitude_sThe preset attitude has a roll angle of-11.8 DEG

The attitude preset rear pitch angle is theta_e7.6 degrees, and preset attitude rear yaw angle psi_eIs equal to-5.6 degrees and has a preset posture angle of

Obtaining the preset time T of the active imaging attitude according to the preset angular velocity average value of the attitude_man＝α/ω₁＝11.0/1＝11.0s。

The rolling angle of the strip before pushing and sweeping is set as

The pitch angle before strip pushing and sweeping is theta_{str_s}7.6 degrees, and the rolling angle after the strip is pushed and swept is

The pitch angle of the strip after pushing and sweeping is theta_{str_e}The maneuvering angle of the sweeping machine is tracked at a uniform ground speed of 3.4 DEG

The imaging time is T when the active imaging is carried out and the ground speed is uniform_image＝β/ω₂＝8.8/0.5＝17.6s；

For the condition that the strip formed after polymerization is in the direction of the trace, the time T required by the posture presetting process before imaging_manTime T required for imaging process_imgaeThe calculation method is as follows:

let the current attitude of the satellite have a roll angle of

The current attitude of the satellite has a pitch angle theta_now3.4 degrees, and the current attitude yaw angle of the satellite is psi_now-5.6 °, and a roll angle of the imaging attitude along the track direction of

The attitude pitch angle of the imaging along the track direction is theta_pas1.2 degrees, and the attitude yaw angle along the track direction is psi_pasIs equal to-2.5 degrees and has a preset posture angle of

The attitude maneuver time can be imaged along the track direction through the difference gamma between the current attitude and the imaged attitude of the satellite and the point-to-point attitude maneuver capability theta of the satellite₁°/t₁s、θ₂°/t₂s、θ₃°/t₃The attitude maneuver time T is obtained by s interpolation calculation_man＝8.1s；

The latitude and longitude of the offset flight imaging starting point and the offset flight imaging ending point are respectively long_s＝31.82°、long_e＝29.75°、 lat_s＝31.52°、lat_e29.63 °, passive imaging offset imaging time of flight is

The push-broom strip generation optimization model formed based on the optimization target and the constraint condition belongs to the NP-hard problem under the condition that the ground to-be-observed target points are more, so that the sum T of attitude maneuver time between adjacent strips, which meets the constraint condition, is generated through an intelligent optimization algorithm such as a BP-neural network_{man_add}Sum of the push-and-sweep time of each strip T_{image_add}And (3) the minimum target observation point number N of the aggregate coverage of the strip observation is the maximum, the push-broom strip of the maximum optimization index J is obtained, and the optimization problem is solved.

Dense task subset sub in the example given for FIG. 2 in accordance with the above method₁And (3) carrying out multi-strip observation task polymerization on the 5 ground targets to be observed to obtain 3 strips str1, str2 and str3, as shown in fig. 5. str2 is parallel to the off-satellite line locus of the remote sensing satellite, str1 and str3 are not parallel to the off-satellite line locus of the remote sensing satellite. The widths of the three strips are all equal to the width b of the load camera, and the lengths of str1, str2 and str3 are omega respectively₂·(t_pre+t_after)·πR/180°+d_1,2＝0.5·(1.5+1.5)·π·500/180°+18.2＝31.3km、 v·(t_pre+t_after)＝7.39·(1.5+1.5)＝22.17km、ω₂·(t_pre+t_after)·πR/180°+d_4,5＝0.5·(1.5+1.5)·π·500/180°+10.8＝23.9km。

(6) Region observation aggregated task generation and push-broom stripe division

For sub_i∈S_areaIn order to reduce the maneuvering times of a satellite platform, reduce the load starting time and save the platform energy, the band division method of the area observation task adopts the strategy of minimum divided bands and shortest imaging time on the premise of covering the target in a given area according to the push-broom band division strategy of the area observation target, and comprises the following steps of:

1) the longest line segment passing through the circumscribed polygon and through the geometric center of the circumscribed polygon is taken as

Will pass through the geometric center of the circumscribed polygon, and

the longest line segment perpendicular to and within the polygon is denoted as

2) The width b of the load camera is taken as the width of the strip so as to

For the strip length, taking into account the sigma% overlap between adjacent strips, from sub_iAlong one side of the circumscribed polygon

In the direction of (sub)_iIs divided into strips of circumscribed polygon, flying from remote sensing satellite through sub_iStarting scanning in the initial direction of the area;

3) greater than sub for strip coverage_iThe external polygon part or the target part to be observed without the ground at one end of the strip is cut correspondingly, and the invalid push-broom path is reduced.

The dense task subset sub in the example given in fig. 2 is given the above method₂The obtained load camera area push-broom paths are as shown by a 5-section dotted line frame in fig. 6, namely str4, str5, str6, str7 and str8, and the positions of imaging time corresponding to the ground targets to be observed in a formed strip are marked by rectangles.

After the imaging task autonomous aggregation method is adopted, the task planning number and the execution number are upgraded from the original 6 ground targets to be observed to 15 ground targets to be observed.

In summary, compared with the existing remote sensing satellite autonomous task planning method, the remote sensing satellite imaging task autonomous aggregation method provided by the invention comprises the following steps: aiming at the fact that a remote sensing satellite with autonomous task planning capability may receive earth observation tasks from different sources, and the local area of the tasks may be densely distributed without ground planning, the existing remote sensing satellite autonomous task planning method generally considers the ground targets to be observed as independent targets to be observed, and the problem that the agile maneuvering and in-motion imaging capabilities of the remote sensing satellite cannot be fully exerted in the obtained task planning scheme is solved, an imaging task autonomous aggregation method is innovatively provided, the dense ground targets to be observed are autonomously aggregated into a multi-strip observation task or a regional observation task based on the attitude maneuvering capability of the remote sensing satellite, and more optimized input is provided for the on-satellite autonomous task planning method, so that the task execution efficiency of the remote sensing satellite in unit time is greatly improved; aiming at the condition that multi-source tasks are densely distributed in a local area, whether existing tasks in a local time window need to be optimized through aggregation or not is identified according to the attitude mobility of the remote sensing satellite; aiming at the task set in the local time window, identifying each dense area and the corresponding task subset in the task set according to a distance threshold determined by the attitude mobility capability, and respectively performing task aggregation on each subset in the follow-up process, thereby improving the task aggregation efficiency; selecting a task aggregation mode according to the geographical distribution uniformity of the ground target to be observed in each dense task set, adopting a multi-strip observation task aggregation mode when the distribution is not uniform, and adopting a regional observation task aggregation mode when the distribution is uniform, so that the invalid push-broom track of the remote sensing satellite can be reduced, and the number of tasks completed in unit time can be increased; in the process of multi-strip observation, regional observation aggregate task generation and push-broom strip division, the highest priority of the ground covering target, the shortest attitude maneuver time and strip push-broom time are used as optimization targets, so that the generated task and strip division scheme meets the optimization in the aspect of attitude maneuver, and the subsequent autonomous task planning process is further optimized.

The method of the invention is oriented to engineering application, aiming at the practical situation that the remote sensing satellite with autonomous task planning capability can receive earth observation tasks from multiple terminals without ground overall planning, and the obvious problems of autonomous planning on the satellite and reduction of task execution efficiency which are caused therewith, on the basis of fully considering the imaging capability in the non-tracking track and the imaging capability of the strip splicing area of the agile motor-driven remote sensing satellite, determining that the intensive tasks are combined into multi-strip observation or regional observation through four steps of identifying the task intensity in a planning window, classifying the intensive tasks, calculating the coverage area of the intensive tasks and judging the aggregation type of the imaging tasks, and further provides a scheme for dividing the combined push-broom strip, which is suitable for agile remote sensing satellites with autonomous mission planning capability, and the method is about to be applied in orbit and can also provide reference for the task management problem of the intelligent autonomous spacecraft.

Those skilled in the art will appreciate that those matters not described in detail in the present specification are well known in the art.

Claims

1. An autonomous polymerization method for remote sensing satellite imaging tasks is characterized by comprising the following steps:

1) identifying the task intensity degree in the range of a local planning window, further judging whether a candidate task set to be observed needs autonomous aggregation of remote sensing satellite imaging tasks, if so, turning to the step 2), otherwise, ending the current method, wherein the short period of autonomous task planning of the remote sensing satellite comprises a plurality of local time windows;

2) dividing the candidate task set to be observed to obtain a plurality of intensive task subsets sub₁,…,sub_nNo aggregated independent observation tasks are performed;

3) computing a circumscribed polygon of the intensive task subset;

4) determining a task aggregation mode of the dense task subset according to a circumscribed polygon of the dense task subset, and further dividing each dense task subset into a multi-strip observation sub-task set and a regional observation sub-task set;

5) generating a multi-strip observation aggregation task according to the multi-strip observation subtask set, and dividing push-broom strips;

6) generating a regional observation aggregation task according to the plurality of regional observation subtask sets, and dividing push-broom strips;

11) for a local time window (T)₁,T₂) Calculating to obtain a local time window (T) according to the track extrapolation information of the remote sensing satellite and the maximum maneuvering capacity in the rolling pitch direction₁,T₂) The corresponding ground target to be observed can observe the regional scope;

12) screening all the ground targets to be observed in the range of the region where the ground targets to be observed can be observed according to the longitude, latitude and elevation information of each ground target to be observed in the satellite-borne task pool to form a local time window (T)₁,T₂) Corresponding candidate task set to be observed to be optimized

13) Based on candidate task set to be observed

Identifying whether the candidate task set to be observed needs task autonomous aggregation or not according to the relative geographic distance between the targets to be observed on the middle ground;

the method for identifying whether the candidate task set to be observed needs to perform task autonomous aggregation comprises the following steps:

otherwise, no optimizable ground target to be observed exists in the candidate task set to be observed, and the tasks are not aggregated autonomously;

said

Wherein eta is₁、η₂0 is greater than or equal to eta as a regulating factor₁<5，0≤η₂<5, v represents the linear velocity module value of the satellite subsatellite point,

for the maximum roll angle of the satellite imaging,

h is orbital height, R is earth radius, Δ T_vSet to 20 s;

a threshold value of angular deviation of the satellite in the rolling direction of the adjacent target imaging attitude;

21)and is

Wherein the function d(s)_j,s_k) Calculating the distance between two points on the surface of the earth;

22)

s_jno polymerization is carried out;

23) traversing to obtain a candidate task set to be observed

The intensive task subset and the independent ground target observation tasks to be observed which are not aggregated are obtained;

31) calculating sub_iGathering intensive tasks into a subset sub through the geometric centers of all the targets to be observed on the ground_i＝{s₁,…,s_kThe geometric center of a limited number of target observation points is denoted as "(long _ o)_i,lat_o_i,alt_o_i) Wherein i and k are positive integers,

32) using omicron as circle center, to_iThe maximum distance of each ground target to be observed is the radius to form sub_iThe circumscribed circle of the target to be observed on each ground;

33) to sub_iEach ground target to be observed positioned on the circumcircle and part of target points close to the circumcircle are taken as vertexes, and a convex polygon is drawn, so that the convex polygon comprises all ground targets to be observed in the current dense task subset;

34) parallel outward expansion b/2 of each side of the convex polygon obtained in the step 33) to obtain a dense task subset sub_iB is the width of the remote sensing satellite load camera;

41) will sub_iWhich is divided into n_iA square small grid with the side length of the load camera width b;

42)

43) if u is_i<U, then sub_iThe ground target to be observed is unevenly distributed in the sub-areaAccording to a multi-strip observation mode_iAggregating targets to be observed on the middle ground; if u is_iGreater than or equal to U, then sub_iThe ground target to be observed is uniformly distributed in the sub-area, and sub-is carried out according to the area observation mode_iThe aggregation of observation tasks of the targets to be observed on the middle ground, wherein 0<U≤1；

44) Repeating the above steps to obtain sub₁,…,sub_nPartitioning into a set of multi-banded Observation subtasks S_stripRegional observation subtask set S_area。

2. The remote sensing satellite imaging task autonomous aggregation method according to claim 1, characterized in that: said

Wherein the content of the first and second substances,

representing maximum mobility of the satellite platform, p₁，ρ₂，ρ₃To determine the threshold parameter.

3. The remote sensing satellite imaging task autonomous aggregation method according to claim 2, characterized in that: the step 5) generates a multi-stripe observation aggregated task according to the multi-stripe observation subtask set, and the method for dividing the push-broom stripes comprises the following steps:

51) for sub_i∈S_stripThe optimization target for aggregating the targets to be observed on the ground into strips is

The constraint conditions comprise point-to-point attitude mobility, passive imaging push-broom linear velocity average value, active imaging uniform ground velocity tracking push-broom angular velocity average value and imaging advance time t before pointing to a target to be observed on the ground_prePointing to the ground behind the target to be observedImaging delay time t_after；

Wherein N is the number of target observation points covered by the current stripe observation aggregation, p_jAs a task s_j∈sub_iCorresponding priority if s_j∈sub_i∈S_stripWhen is lambda_j1, otherwise λ_j＝0，T_{man_add}The sum of the attitude maneuver time between adjacent strips is obtained from the attitude maneuver time T of each adjacent strip_manAre added to obtain_manNamely the time required by the posture presetting process before imaging; t is_{image_add}For the sum of the sweep times of the individual strips, from each strip sweep time T_imageAre added to obtain_imageI.e. the time required for the imaging process; alpha is alpha₁、α₂To optimize the target adjustment coefficient, α₁、α₂The value range of (A) is 0 to 1;

52) when the strip formed by polymerization is in the non-tracking direction, the time T required by the posture presetting process before imaging_manTime T required for imaging process_imageIs composed of

T_man＝α/ω₁

T_image＝β/ω₂

Wherein the content of the first and second substances,

presetting a rear roll angle, theta, for attitude_ePresetting rear pitch angle psi for attitude_ePresetting a rear yaw angle for the attitude; omega₁Presetting angular velocity, omega, for attitude₂Is the imaging angular velocity;

Time T required for imaging process_imageIs composed of

53) Generating an optimization model according to the push-broom strips formed in the step 51), generating optimal push-broom strips which meet constraint conditions and enable the sum of attitude maneuver time between adjacent strips to be minimum and the sum of push-broom time of each strip to be minimum, enabling the number N of target observation points covered by the strip observation aggregation to be maximum, completing the generation of a multi-strip observation aggregation task, and dividing the push-broom strips.

4. The remote sensing satellite imaging task autonomous aggregation method according to claim 3, characterized in that: the method for generating the regional observation aggregated task according to the plurality of regional observation subtask sets and dividing the push-broom stripe comprises the following steps:

61) for sub_i∈S_areaThe longest line segment passing through the circumscribed polygon and passing through the geometric center of the circumscribed polygon is taken as

Will pass through the geometric center of the circumscribed polygon, and

the longest line segment perpendicular to and within the polygon is denoted as

62) The width b of the load camera is taken as the width of the strip so as to

In the direction of (sub)_iIs divided into strips of a circumscribed polygon and flies from a remote sensing satellite through sub_iStarting scanning in the initial direction of the area; wherein the strip covers more than sub_iOf a circumscribed polygonal portion, or of a strip at one end without a groundThe part facing the target to be observed is not pushed and swept.