CN113609587A - Active drift multi-objective optimization method and system based on satellite three-dimensional imaging - Google Patents

Active drift multi-objective optimization method and system based on satellite three-dimensional imaging Download PDF

Info

Publication number
CN113609587A
CN113609587A CN202110892841.8A CN202110892841A CN113609587A CN 113609587 A CN113609587 A CN 113609587A CN 202110892841 A CN202110892841 A CN 202110892841A CN 113609587 A CN113609587 A CN 113609587A
Authority
CN
China
Prior art keywords
regression
optimization
orbit
satellite
module
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202110892841.8A
Other languages
Chinese (zh)
Other versions
CN113609587B (en
Inventor
张德新
陈筠力
邵晓巍
曹勇
陈力
鞠潭
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shanghai Jiaotong University
Original Assignee
Shanghai Jiaotong University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shanghai Jiaotong University filed Critical Shanghai Jiaotong University
Priority to CN202110892841.8A priority Critical patent/CN113609587B/en
Publication of CN113609587A publication Critical patent/CN113609587A/en
Application granted granted Critical
Publication of CN113609587B publication Critical patent/CN113609587B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/10—Geometric CAD
    • G06F30/15—Vehicle, aircraft or watercraft design
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/12—Computing arrangements based on biological models using genetic models
    • G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00—Details relating to CAD techniques
    • G06F2111/04—Constraint-based CAD
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00—Details relating to CAD techniques
    • G06F2111/06—Multi-objective optimisation, e.g. Pareto optimisation using simulated annealing [SA], ant colony algorithms or genetic algorithms [GA]

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Geometry (AREA)
  • Biophysics (AREA)
  • General Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Hardware Design (AREA)
  • Software Systems (AREA)
  • Evolutionary Biology (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Mathematical Optimization (AREA)
  • Genetics & Genomics (AREA)
  • Mathematical Analysis (AREA)
  • Computational Mathematics (AREA)
  • Aviation & Aerospace Engineering (AREA)
  • Automation & Control Theory (AREA)
  • Medical Informatics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physiology (AREA)
  • Pure & Applied Mathematics (AREA)
  • Biomedical Technology (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)

Abstract

The invention provides an active drift multi-objective optimization method and system based on satellite three-dimensional imaging, which comprises the following steps: firstly, analyzing the three-dimensional imaging principle of a satellite system; secondly, establishing a regression model of a nominal orbit and a multi-reference transfer orbit by analyzing the dynamic characteristics of the satellite orbit, and establishing an optimization model of fuel, time and regression precision; and finally, performing step-by-step layer-by-layer optimization by taking the optimal fuel as an optimization main line. The method comprises the following steps of firstly, solving a fuel optimal initial value calculation problem meeting time constraints by applying the idea of a pareto optimal solution. And secondly, based on the initial value of the first step, further optimizing the fuel consumption by an improved self-adaptive hybrid cross genetic algorithm under the condition of meeting the requirement of regression precision. The invention solves the problem of fuel optimal orbit design of satellite three-dimensional imaging active drift under the constraint of time and regression precision.

Description

Active drift multi-objective optimization method and system based on satellite three-dimensional imaging
Technical Field
The invention relates to the technical field of spacecraft orbit optimization, in particular to an active drift multi-objective optimization method and system based on satellite three-dimensional imaging.
Background
The three-dimensional imaging technology has huge application potential in geology, forestry, biomass estimation, archaeology, glaciers and detection of underground buried objects. In order to meet the requirements of satellite three-dimensional imaging on periodic observation and illumination conditions of a target area, single-satellite three-dimensional imaging data are usually acquired in a sun synchronous regression orbit mode, frequent orbit transfer control inevitably causes increase of fuel consumption, the fuel consumption is the most concerned problem in a satellite orbit transfer task, the service life of a satellite is directly determined, and the transfer orbit with the optimal fuel consumption is designed due to the fact that fuel brought into space by a spacecraft is limited, so that the method has good practical significance.
Patent document CN105677942A (application number: CN201510997982.0) discloses a method for rapidly simulating repeated orbit satellite-borne natural scene SAR complex image data, which simulates the operation orbit of an SAR satellite, and simulates a ground scene according to the imaging principle of the satellite-borne SAR satellite and a single-view complex image signal model of the SAR satellite, so as to directly obtain multiple repeated orbit SAR complex image data. However, the patent does not calculate the fuel consumption of the satellite operation and is not a good reference for reality.
Disclosure of Invention
Aiming at the defects in the prior art, the invention aims to provide an active drift multi-objective optimization method and system based on satellite three-dimensional imaging.
The active drift multi-objective optimization method based on satellite three-dimensional imaging provided by the invention comprises the following steps:
step 1: establishing a sun synchronous regression orbit parameter table under parameter constraint;
step 2: examining the table and determining all nominal orbit parameters that satisfy the global coverage condition;
and step 3: for each nominal orbit, determining all transition orbit parameters meeting time constraints within a base length range, and obtaining all feasible solutions;
and 4, step 4: calculating the total sampling time and fuel consumption of the feasible solutions to obtain a pareto optimal solution set;
and 5: selecting a fuel preference solution meeting time constraints as an initial value of next optimization;
step 6: establishing a fuel optimization model under regression precision constraint based on the preference solution of the previous step, generating an initial population through a genetic algorithm, and initializing a population algebra;
and 7: calculating individual fitness values by respectively taking regression accuracy and fuel consumption as optimization targets;
and 8: selecting an optimization target according to the difference probability, and calculating an individual fitness value;
and step 9: judging whether the selected individual fitness value reaches an expected value or whether the iteration times reach a termination condition, if so, outputting an optimization result, and otherwise, performing the next operation;
step 10: carrying out genetic algorithm selection operation;
step 11: selecting a cross operation mode according to the probability to carry out cross operation of a genetic algorithm;
step 12: updating the probability of each cross method being selected;
step 13: calculating a fitness value;
step 14: judging whether the fitness value reaches an expected value or whether the iteration times reach a termination condition, if so, outputting a design result, and otherwise, carrying out the next operation;
step 15: and (5) performing genetic algorithm variation operation, returning to the step 7, and performing the next iteration process.
Preferably, the fuel consumption corresponds to the velocity increment of the transfer rail, and the calculation formula is:
Figure BDA0003196642630000021
Figure BDA0003196642630000022
Figure BDA0003196642630000023
Figure BDA0003196642630000024
in the formula (I), the compound is shown in the specification,
Figure BDA0003196642630000025
representing the velocity increment from the kth working track to the nth reference transfer track,
Figure BDA0003196642630000026
denotes the velocity increment, v, from the nth reference transfer track to the (k + 1) th working trackkAnd vk+1The speeds of the kth working track and the kth +1 working track are respectively; Δ vtotalIndicates the amount of change in the total speed; m represents the number of baselines;
Figure BDA0003196642630000027
represents the speed of the 1 st reference transfer track;
Figure BDA0003196642630000031
represents the semimajor axis of the kth nominal track;
Figure BDA0003196642630000032
the semi-major axis of the 1 st transfer track is indicated.
Preferably, the fuel optimization model under the time constraint condition is as follows:
Figure BDA0003196642630000033
wherein, Δ ttotal,ΔvtotalIs the objective function; x is the number of1Is a feasible solution; x is a feasible solution set; minF () represents the optimization model; n is a radical ofnomRepresents the number of days of regression for the nominal orbit; n is a radical ofsRepresents the number of regression days for the data sample;
Figure BDA0003196642630000034
represents the number of regression days for which the nominal orbit is the smallest;
Figure BDA0003196642630000035
the number of regression days for which the nominal orbit was the largest is indicated.
Preferably, the fuel optimization model under the constraint of regression accuracy is as follows:
Figure BDA0003196642630000036
Figure BDA0003196642630000037
in the formula, x2In order to make a decision on a variable,
Figure BDA0003196642630000038
is the value range of each element in the decision variables, Δ r, Δ vtotalIs an objective function.
Preferably, the selection operation is performed in the form of roulette, the reciprocal of the regression accuracy is used as the fitness value of each individual, the fitness sum of all the individuals is calculated, the relative fitness of each individual is calculated, a preset number of random numbers between 0 and 1 are generated, and the random numbers are respectively compared with the fitness value of each individual to determine the probability of each individual being selected.
The invention provides an active drift multi-objective optimization system based on satellite three-dimensional imaging, which comprises:
module M1: establishing a sun synchronous regression orbit parameter table under parameter constraint;
module M2: examining the table and determining all nominal orbit parameters that satisfy the global coverage condition;
module M3: for each nominal orbit, determining all transition orbit parameters meeting time constraints within a base length range, and obtaining all feasible solutions;
module M4: calculating the total sampling time and fuel consumption of the feasible solutions to obtain a pareto optimal solution set;
module M5: selecting a fuel preference solution meeting time constraints as an initial value of next optimization;
module M6: establishing a fuel optimization model under regression precision constraint based on the preference solution of the module M5, generating an initial population through a genetic algorithm, and initializing a population algebra;
module M7: calculating individual fitness values by respectively taking regression accuracy and fuel consumption as optimization targets;
module M8: selecting an optimization target according to the difference probability, and calculating an individual fitness value;
module M9: judging whether the selected individual fitness value reaches an expected value or whether the iteration times reach a termination condition, if so, outputting an optimization result, otherwise, calling a module M10;
module M10: carrying out genetic algorithm selection operation;
module M11: selecting a cross operation mode according to the probability to carry out cross operation of a genetic algorithm;
module M12: updating the probability of each cross method being selected;
module M13: calculating a fitness value;
module M14: judging whether the fitness value reaches an expected value or whether the iteration times reach a termination condition, if so, outputting a design result, otherwise, calling a module M15;
module M15: and (4) performing genetic algorithm mutation operation, and calling a module M7 to perform the next iteration process.
Preferably, the fuel consumption corresponds to the velocity increment of the transfer rail, and the calculation formula is:
Figure BDA0003196642630000041
Figure BDA0003196642630000042
Figure BDA0003196642630000043
Figure BDA0003196642630000044
in the formula (I), the compound is shown in the specification,
Figure BDA0003196642630000045
representing the velocity increment from the kth working track to the nth reference transfer track,
Figure BDA0003196642630000046
denotes the velocity increment, v, from the nth reference transfer track to the (k + 1) th working trackkAnd vk+1The speeds of the kth working track and the kth +1 working track are respectively; Δ vtotalIndicates the amount of change in the total speed; m represents the number of baselines;
Figure BDA0003196642630000047
represents the speed of the 1 st reference transfer track;
Figure BDA0003196642630000048
represents the semimajor axis of the kth nominal track;
Figure BDA0003196642630000049
the semi-major axis of the 1 st transfer track is indicated.
Preferably, the fuel optimization model under the time constraint condition is as follows:
Figure BDA0003196642630000051
wherein, Δ ttotal,ΔvtotalIs the objective function; x is the number of1Is a feasible solution; x is a feasible solution set; minF () represents the optimization model; n is a radical ofnomRepresents the number of days of regression for the nominal orbit; n is a radical ofsRepresents the number of regression days for the data sample;
Figure BDA0003196642630000052
represents the number of regression days for which the nominal orbit is the smallest;
Figure BDA0003196642630000053
the number of regression days for which the nominal orbit was the largest is indicated.
Preferably, the fuel optimization model under the constraint of regression accuracy is as follows:
Figure BDA0003196642630000054
Figure BDA0003196642630000055
in the formula, x2In order to make a decision on a variable,
Figure BDA0003196642630000056
is the value range of each element in the decision variables, Δ r, Δ vtotalIs an objective function.
Preferably, the selection operation is performed in the form of roulette, the reciprocal of the regression accuracy is used as the fitness value of each individual, the fitness sum of all the individuals is calculated, the relative fitness of each individual is calculated, a preset number of random numbers between 0 and 1 are generated, and the random numbers are respectively compared with the fitness value of each individual to determine the probability of each individual being selected.
Compared with the prior art, the invention has the following beneficial effects:
(1) for the task requirement of single-satellite three-dimensional imaging, the satellite needs to perform multiple orbital maneuvers in the baseline direction, and the proposed transition orbit scheme of the multi-reference transfer orbit can reduce the maneuvering range of the satellite, thereby reducing the fuel consumption;
(2) the two-step optimization strategy adopted by the invention optimizes the fuel consumption, can greatly reduce the fuel consumption and has important significance for prolonging the on-orbit service life of the spacecraft;
(3) the genetic algorithm multi-objective optimization method adopted by the invention dynamically adjusts related parameters according to the fitness of the objective function for multiple times, and selects different genetic cross algorithms according to the difference probability, so that the diversity and the rapid convergence of the population can be kept, and a track meeting the task requirement is designed.
Drawings
Other features, objects and advantages of the invention will become more apparent upon reading of the detailed description of non-limiting embodiments with reference to the following drawings:
FIG. 1 is a flow chart of the method of the present invention;
FIG. 2 is a graph of a relationship between track elevation direction, baseline direction and radar beam pointing;
FIG. 3 is a schematic diagram of a pareto optimal solution;
FIG. 4 is a diagram of the optimization results of the first step;
FIG. 5 is a convergence plot for Sol.3 under the first reference transfer orbit;
FIG. 6 is a convergence plot for Sol.3 at a second reference transfer orbit;
FIG. 7 is a convergence plot for Sol.7 under the first reference transfer orbit;
FIG. 8 is a convergence plot for Sol.7 at the second reference transfer orbit.
Detailed Description
The present invention will be described in detail with reference to specific examples. The following examples will assist those skilled in the art in further understanding the invention, but are not intended to limit the invention in any way. It should be noted that it would be obvious to those skilled in the art that various changes and modifications can be made without departing from the spirit of the invention. All falling within the scope of the present invention.
Example (b):
as shown in fig. 1, the multi-objective optimization design method for a tomosynthesis SAR satellite orbit based on a multi-reference transfer orbit provided by the present invention specifically includes the following steps:
step 1: according to the relation diagram of the orbit altitude direction, the baseline direction and the radar beam direction of the satellite three-dimensional imaging shown in fig. 2, the three-dimensional imaging satellite system adopts a single-satellite multi-navigation mode, the orbit altitude direction is a straight line connecting the earth center and the satellite, the radar beam direction is a beam direction of satellite load work, and an included angle between the radar beam direction straight line and the orbit altitude straight line is defined as a downward viewing angle and is represented by alpha. The base line is vertical to the beam direction of the satellite and points to the upper right by taking the satellite as a starting point. k is 0,1, 2.. multidot.m, which is the number of the working track, i.e. the track where data sampling is performed, and the 0 th working track is defined as the nominal track.
According to the geometrical relationship of the baseline direction in the space, the length of a three-dimensional imaging multi-baseline single segment is set as d, and then based on the projection transformation theory, the quantitative relationship between the baseline and the track variable quantity can be established, wherein the track height direction lifting quantity corresponding to the baseline is as follows:
dislatitude＝d·sinα…………(1)
the offset of the satellite in the direction perpendicular to the orbital altitude (longitude direction) is:
dislongitude＝d·cosα…………(2)
and if the number of the base lines is m, the height-direction aperture length is as follows:
L＝d·m…………(3)
if the rail transfer is performed solely between the nominal rail and the working rail of the lifting section, a large fuel consumption is generated and the task completion time shows a linear increasing relationship with the regression cycle of the nominal rail. On the basis of the nominal orbit, the fuel consumption and the task completion time can be further reduced through the optimized design of a plurality of reference transfer orbits, and the global coverage capability of the nominal orbit can be guaranteed.
Step 2: make (·)nom、
Figure BDA0003196642630000071
Respectively representing the track parameters of a nominal track, a kth working track and a jth reference transfer track;
Figure BDA0003196642630000072
representing a Hoeman ellipse maneuvering from the kth working orbit to the kth transition orbit;
Figure BDA0003196642630000073
representing a huffman ellipse maneuvering from the k-th transition orbit to the k + 1-th working orbit. a. e, i, omega and M respectively represent the semimajor axis, eccentricity, orbit inclination, ascension of ascending intersection point, amplitude angle of perigee and angle of perigee of the satellite orbit, and Deltav represents the pulse velocity increment.
The nominal orbit regression behavior model is as follows:
Figure BDA0003196642630000074
in the formula, NnomAnd RnomIs a relatively prime integer, ωEIs the rate of rotation of the earth, τnomIs the node period. Assuming that the nominal track is a circular track, when anomAnd inomIn such a manner that
Figure BDA0003196642630000075
The nominal orbit is the sun synchronous orbit. Note that the change of the number of tracks in the case of pan is based on linearization J2An item perturbation model.
The longitude and latitude of the satellite point at the initial sampling moment of the satellite are set as
Figure BDA0003196642630000076
The ideal sub-satellite position at the beginning of the kth sample
Figure BDA0003196642630000077
Can be represented by the formulae (5) and (6):
Figure BDA0003196642630000078
Figure BDA0003196642630000079
under the above conditions, in combination with the nominal orbit regression characteristics model described by equation (4), the reference transfer orbit should satisfy the following conditions:
Figure BDA00031966426300000710
in the formula, Δ ω is a drift amount in the longitudinal direction described based on the angle.
The working track and the transition track are both at the same deviation heart rate and the same inclination angle as the nominal track. Consider the Hulman orbital transfer approach. The angular motion amounts of the satellite in the longitude and latitude directions in one sampling period are respectively as follows:
Figure BDA00031966426300000711
Figure BDA00031966426300000712
if every NsOnce per day, and if the longitude drift characteristic described by formula (5) and the latitude regression characteristic described by formula (6) are simultaneously satisfied, the following formula holds:
Figure BDA0003196642630000081
Figure BDA0003196642630000082
in the formula, NsAnd RsAre all integers.
And step 3: in order to avoid simultaneous processing of a multi-dimensional optimization target and face to the problem of active transfer multi-target optimization design of a single-satellite TomosAR track, a step-by-step layer optimization strategy of target decoupling is provided. The three objectives are broken down into two-step optimization, both taking fuel consumption as optimization objective due to its importance in the TomoSAR orbit transfer task. The method comprises the steps of firstly, calculating an initial value of the rail meeting time constraint and fuel optimization by applying a Pareto optimal solution idea; and in the second step, based on the initial value obtained in the first step, the fuel consumption is further optimized under the condition of meeting regression constraint, and the optimal semi-major axes of a plurality of transition tracks are obtained.
In the first step of optimization, in order to reduce the consumption of computing resources, the continuous optimization solving problem of the multi-reference transfer orbit semi-major axis of wide-area continuous time is converted into an initial value computing problem of discrete time regression days, and an optimization model is established as follows:
Figure BDA0003196642630000083
wherein, Δ ttotal,ΔvtotalIs an objective function, x1For feasible solutions, X is a set of feasible solutions.
For the objective function Δ ttotalIn other words, no other solution better than Δ t can be found in the variable spacetotalThen Δ ttotalIs a Pareto optimal solution, and similarly, no other solution superior to delta v can be found in the variable spacetotalThen Δ vtotalIs a Pareto optimal solution.
And 4, step 4: establishing a sun synchronous regression orbit parameter table under parameter constraint; examining the table to determine all nominal orbit parameters that satisfy the global coverage condition; for each nominal orbit, determining all transition orbit parameters meeting time constraints within a base length range, and obtaining all feasible solutions; calculating the total sampling time and fuel consumption of the feasible solutions to obtain a pareto optimal solution set; the fuel preference solution that satisfies the time constraint is selected as the initial value for the next optimization. A schematic diagram of the pareto optimal solution is shown in fig. 3.
And 5: reference transfer orbit semi-major axis obtained by first step optimization
Figure BDA0003196642630000084
A search space for the semi-major axis of the transition track is determined. In the second optimization step, the regression accuracy and fuel consumption are taken as objective functions, and the mathematical model of the optimization problem is as follows:
Figure BDA0003196642630000085
in the formula (I), the compound is shown in the specification,
Figure BDA0003196642630000086
in order to make a decision on a variable,
Figure BDA0003196642630000087
is x2The value range of the variable of each dimension. And selecting a genetic algorithm for optimization solution aiming at the two target optimization problems. And randomly generating an initialization population.
Step 6: calculating the fitness value of each individual by respectively taking the regression precision and the fuel consumption as optimization targets;
and 7: selecting a fitness value calculated by a certain optimization target according to the difference probability;
and 8: judging whether the self-moderate value reaches an expected value or not, or judging whether the iteration times reach a termination condition, if so, outputting a track design result, and if not, carrying out the next operation;
and step 9: a selection operation is performed. At each generation, a difference probability method is used to select a value of the objective function, assuming that regression accuracy is selected. The selection operation is performed by using a roulette method, and the regression accuracy and fuel consumption optimization problems are described in detail as follows:
(1) calculating the sum of fitness of all individuals by taking the reciprocal of the regression precision as the fitness value of the individual;
(2) calculating the relative fitness of each individual;
(3) NP random numbers between 0 and 1 are generated and compared to the fitness value of each individual to determine the probability of each individual being selected.
Step 10: a crossover operation is performed. Crossover operations are the most dominant genetic operations in genetic algorithms, using an improved hybrid crossover strategy. First, the cross probability Pc is selected by an adaptive method.
If f isgIs less than favgThen:
pcg＝(pc3×(favg-fg)+pc2×(fg-fmin))/(favg-fmin),g＝1,...,maxgen…………(14)
otherwise:
pcg＝(pc2×(fmax-fg)+pc1×(fg-favg))/(fmax-favg),g＝1,...,maxgen…………(15)
wherein f isgMaximum regression error of the g-th generation, favg、fminAnd fmaxThe mean regression error, the minimum regression error and the maximum regression error of the previous g generations are respectively, and maxgen is the total evolution generation.
The specific steps of the crossover operation are described as follows:
(1) the pure strategy set is marked as {1,2,3}, and respectively represents an analog binary crossover operator, a non-uniform arithmetic crossover and a guided crossover operator.
(2) Initialization, an initial population of u individuals is generated. For each individual i, an initial probability distribution, ρ, is assigned to the hybrid strategy vectori＝{ρi(1),ρi(2),ρi(3)}。
(3) For each individual i, selecting a cross strategy h according to the probability in the mixed strategy rho, and carrying out cross according to the selected strategy to generate filial generations, wherein the related cross strategies are as follows:
simulated binary crossover operator (SBX): the two genes assumed to be crossed are each k1、k2New values of mating production
Figure BDA0003196642630000091
Comprises the following steps:
Figure BDA0003196642630000101
non-uniform arithmetic intersection:
Figure BDA0003196642630000102
and guiding a crossover operator:
if the regression error of the parent is greater than the regression error of the child, then:
Figure BDA0003196642630000103
otherwise:
Figure BDA0003196642630000104
in the formulae (16), (17), (18) and (19),. beta.cThe calculation method is as follows:
Figure BDA0003196642630000105
in the formula etacFor distributing the control parameters, η is usually takenc＝20。
Updating the child mixing strategy probability according to the following method:
the adopted cross pure strategy is h, h belongs to {1,2,3}, the fitness value of each individual of the parent and the child is compared, if the number of the individuals of the child superior to the parent is more, the pure strategy is strengthened:
Figure BDA0003196642630000106
otherwise, this pure policy is weakened:
Figure BDA0003196642630000107
lambda is used to adjust the rough distribution of the mixing strategy, firstly, the fitness value of an optimization target after a certain body is crossed is selected with equal probability and is compared with the fitness value before the crossing, if the fitness value before the crossing is large, the count value is added with 1, and lambda can be calculated by the following formula:
λ＝(b-a)(1-count/N)+a…………(23)
in the formula, count is the number of individuals superior to the offspring in the parent, N is the size of the population, and lambda belongs to [ a, b ].
Step 11: and calculating the fitness value of the individual after the crossover operation.
Step 12: and (4) judging whether the fitness value of the individual reaches an expected value or whether the iteration times reach a termination condition, if so, outputting a track design result, and otherwise, turning to the step 6 to enter the next iteration process of the genetic algorithm.
The present invention will be described in further detail with reference to examples.
Setting relevant parameters of a tomography task as shown in a table 1; the range of values of the track parameters optimized in the first layer is shown in table 2.
TABLE 1 simulation parameters for three-dimensional imaging tasks
Parameter(s) Numerical value Unit of
L 15 (km)
m 15 -
α 30 (°)
TABLE 2 optimization Range of orbital parameters
Parameter(s) Minimum value Maximum value Unit of
Nnom 16 30 (day)
NS 1 10 (day)
anom 6578.137 7178.137 (km)
The optimization result of the first layer is shown in fig. 4, and the obtained Pareto optimal solution set has 7 solutions (respectively marked as sol.1-sol.7). Selecting Sol.3 and Sol.7 as preference solutions, and the orbit parameters of the two solutions are shown in Table 3.
TABLE 3 orbital parameters of Pareto optimal solution
Figure BDA0003196642630000111
Figure BDA0003196642630000121
In the second layer of optimization, the relevant parameters of the genetic algorithm are shown in table 4.
TABLE 4 simulation parameters of genetic algorithms
Parameter(s) Numerical value Unit of
NP 120 -
NG 180 -
Pm 0.1 -
mu 20 -
mum 20 -
For the multi-reference transfer orbit method, with sol.3 and sol.7 as the input of the second step optimization, the convergence intervals are (7043.565km,7044.665km) and (7045.138km,7046.138km), respectively, and the convergence process of the algorithm is shown in fig. 5, fig. 6, fig. 7 and fig. 8.
The final design values of the transition track are shown in table 5.
TABLE 5 transition track design results
Figure BDA0003196642630000122
In conclusion, the multi-target optimization design method for satellite three-dimensional imaging active drift based on the multi-reference transfer orbit provides the fuel-optimal transfer orbit parameter design under the constraint conditions of regression precision and revisit time through two-step optimization of Pareto optimization and an improved genetic algorithm, and has important significance for saving fuel, prolonging the service life of the satellite and increasing the on-orbit service time.
Those skilled in the art will appreciate that, in addition to implementing the systems, apparatus, and various modules thereof provided by the present invention in purely computer readable program code, the same procedures can be implemented entirely by logically programming method steps such that the systems, apparatus, and various modules thereof are provided in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, embedded microcontrollers and the like. Therefore, the system, the device and the modules thereof provided by the present invention can be considered as a hardware component, and the modules included in the system, the device and the modules thereof for implementing various programs can also be considered as structures in the hardware component; modules for performing various functions may also be considered to be both software programs for performing the methods and structures within hardware components.
The foregoing description of specific embodiments of the present invention has been presented. It is to be understood that the present invention is not limited to the specific embodiments described above, and that various changes or modifications may be made by one skilled in the art within the scope of the appended claims without departing from the spirit of the invention. The embodiments and features of the embodiments of the present application may be combined with each other arbitrarily without conflict.

Claims (10)

1. An active drift multi-objective optimization method based on satellite three-dimensional imaging is characterized by comprising the following steps:
step 1: establishing a sun synchronous regression orbit parameter table under parameter constraint;
step 2: examining the table and determining all nominal orbit parameters that satisfy the global coverage condition;
and step 3: for each nominal orbit, determining all transition orbit parameters meeting time constraints within a base length range, and obtaining all feasible solutions;
and 4, step 4: calculating the total sampling time and fuel consumption of the feasible solutions to obtain a pareto optimal solution set;
and 5: selecting a fuel preference solution meeting time constraints as an initial value of next optimization;
step 6: establishing a fuel optimization model under regression precision constraint based on the preference solution of the previous step, generating an initial population through a genetic algorithm, and initializing a population algebra;
and 7: calculating individual fitness values by respectively taking regression accuracy and fuel consumption as optimization targets;
and 8: selecting an optimization target according to the difference probability, and calculating an individual fitness value;
and step 9: judging whether the selected individual fitness value reaches an expected value or whether the iteration times reach a termination condition, if so, outputting an optimization result, and otherwise, performing the next operation;
step 10: carrying out genetic algorithm selection operation;
step 11: selecting a cross operation mode according to the probability to carry out cross operation of a genetic algorithm;
step 12: updating the probability of each cross method being selected;
step 13: calculating a fitness value;
step 14: judging whether the fitness value reaches an expected value or whether the iteration times reach a termination condition, if so, outputting a design result, and otherwise, carrying out the next operation;
step 15: and (5) performing genetic algorithm variation operation, returning to the step 7, and performing the next iteration process.
2. The active drift multi-objective optimization method based on satellite three-dimensional imaging according to claim 1, wherein fuel consumption corresponds to velocity increment of a transfer orbit, and the calculation formula is as follows:
Figure FDA0003196642620000011
Figure FDA0003196642620000012
Figure FDA0003196642620000013
Figure FDA0003196642620000021
in the formula (I), the compound is shown in the specification,
Figure FDA0003196642620000022
indicating from the kth work trackThe velocity increment to the nth reference transfer track,
Figure FDA0003196642620000023
denotes the velocity increment, v, from the nth reference transfer track to the (k + 1) th working trackkAnd vk+1The speeds of the kth working track and the kth +1 working track are respectively; Δ vtotalIndicates the amount of change in the total speed; m represents the number of baselines;
Figure FDA0003196642620000024
represents the speed of the 1 st reference transfer track;
Figure FDA0003196642620000025
represents the semimajor axis of the kth nominal track;
Figure FDA0003196642620000026
the semi-major axis of the 1 st transfer track is indicated.
3. The active drift multi-objective optimization method based on satellite three-dimensional imaging according to claim 2, wherein a fuel optimization model under a time constraint condition is as follows:
Figure FDA0003196642620000027
wherein, Δ ttotal,ΔvtotalIs the objective function; x is the number of1Is a feasible solution; x is a feasible solution set; minF () represents the optimization model; n is a radical ofnomRepresents the number of days of regression for the nominal orbit; n is a radical ofsRepresents the number of regression days for the data sample;
Figure FDA0003196642620000028
represents the number of regression days for which the nominal orbit is the smallest;
Figure FDA0003196642620000029
indicating signThe maximum number of days of regression on the track is weighed.
4. The active drift multi-objective optimization method based on satellite three-dimensional imaging according to claim 3, wherein a fuel optimization model under regression accuracy constraint is as follows:
Figure FDA00031966426200000210
Figure FDA00031966426200000211
in the formula, x2In order to make a decision on a variable,
Figure FDA00031966426200000212
is the value range of each element in the decision variables, Δ r, Δ vtotalIs an objective function.
5. The active drift multi-objective optimization method based on satellite three-dimensional imaging as claimed in claim 1, wherein the selection operation is performed in the form of roulette, the reciprocal of regression accuracy is used as the fitness value of each individual, the fitness of all individuals is summed, the relative fitness of each individual is calculated, a preset number of random numbers between 0 and 1 are generated, and the random numbers are respectively compared with the fitness value of each individual to determine the probability of each individual being selected.
6. An active drift multi-objective optimization system based on satellite three-dimensional imaging is characterized by comprising the following components:
module M1: establishing a sun synchronous regression orbit parameter table under parameter constraint;
module M2: examining the table and determining all nominal orbit parameters that satisfy the global coverage condition;
module M3: for each nominal orbit, determining all transition orbit parameters meeting time constraints within a base length range, and obtaining all feasible solutions;
module M4: calculating the total sampling time and fuel consumption of the feasible solutions to obtain a pareto optimal solution set;
module M5: selecting a fuel preference solution meeting time constraints as an initial value of next optimization;
module M6: establishing a fuel optimization model under regression precision constraint based on the preference solution of the module M5, generating an initial population through a genetic algorithm, and initializing a population algebra;
module M7: calculating individual fitness values by respectively taking regression accuracy and fuel consumption as optimization targets;
module M8: selecting an optimization target according to the difference probability, and calculating an individual fitness value;
module M9: judging whether the selected individual fitness value reaches an expected value or whether the iteration times reach a termination condition, if so, outputting an optimization result, otherwise, calling a module M10;
module M10: carrying out genetic algorithm selection operation;
module M11: selecting a cross operation mode according to the probability to carry out cross operation of a genetic algorithm;
module M12: updating the probability of each cross method being selected;
module M13: calculating a fitness value;
module M14: judging whether the fitness value reaches an expected value or whether the iteration times reach a termination condition, if so, outputting a design result, otherwise, calling a module M15;
module M15: and (4) performing genetic algorithm mutation operation, and calling a module M7 to perform the next iteration process.
7. The active drift multi-objective optimization system based on satellite three-dimensional imaging according to claim 6, wherein fuel consumption corresponds to the velocity increment of the transfer orbit, and the calculation formula is as follows:
Figure FDA0003196642620000031
Figure FDA0003196642620000032
Figure FDA0003196642620000033
Figure FDA0003196642620000041
in the formula (I), the compound is shown in the specification,
Figure FDA0003196642620000042
representing the velocity increment from the kth working track to the nth reference transfer track,
Figure FDA0003196642620000043
denotes the velocity increment, v, from the nth reference transfer track to the (k + 1) th working trackkAnd vk+1The speeds of the kth working track and the kth +1 working track are respectively; Δ vtotalRepresents the total change in speed; m represents the number of baselines;
Figure FDA0003196642620000044
represents the speed of the 1 st reference transfer track;
Figure FDA0003196642620000045
represents the semimajor axis of the kth nominal track;
Figure FDA0003196642620000046
the semi-major axis of the 1 st transfer track is indicated.
8. The active drift multi-objective optimization system based on satellite three-dimensional imaging according to claim 7, wherein the fuel optimization model under the time constraint condition is as follows:
Figure FDA0003196642620000047
wherein, Δ ttotal,ΔvtotalIs the objective function; x is the number of1Is a feasible solution; x is a feasible solution set; minF () represents the optimization model; n is a radical ofnomRepresents the number of days of regression for the nominal orbit; n is a radical ofsRepresents the number of regression days for the data sample;
Figure FDA0003196642620000048
represents the number of regression days for which the nominal orbit is the smallest;
Figure FDA0003196642620000049
the number of regression days for which the nominal orbit was the largest is indicated.
9. The active drift multi-objective optimization system based on satellite three-dimensional imaging according to claim 8, wherein the fuel optimization model under the constraint of regression accuracy is:
Figure FDA00031966426200000410
Figure FDA00031966426200000411
in the formula, x2In order to make a decision on a variable,
Figure FDA00031966426200000412
is the value range of each element in the decision variables, Δ r, Δ vtotalIs an objective function.
10. The active drift multi-objective optimization system based on satellite three-dimensional imaging as claimed in claim 6, wherein the selection operation is performed in the form of roulette, the reciprocal of regression accuracy is used as the fitness value of each individual, the fitness of all individuals is summed, the relative fitness of each individual is calculated, a preset number of random numbers between 0 and 1 are generated, and the random numbers are respectively compared with the fitness value of each individual to determine the probability of each individual being selected.
CN202110892841.8A 2021-08-04 2021-08-04 Active drift multi-objective optimization method and system based on satellite three-dimensional imaging Active CN113609587B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110892841.8A CN113609587B (en) 2021-08-04 2021-08-04 Active drift multi-objective optimization method and system based on satellite three-dimensional imaging

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110892841.8A CN113609587B (en) 2021-08-04 2021-08-04 Active drift multi-objective optimization method and system based on satellite three-dimensional imaging

Publications (2)

Publication Number Publication Date
CN113609587A true CN113609587A (en) 2021-11-05
CN113609587B CN113609587B (en) 2023-02-17

Family

ID=78339518

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110892841.8A Active CN113609587B (en) 2021-08-04 2021-08-04 Active drift multi-objective optimization method and system based on satellite three-dimensional imaging

Country Status (1)

Country Link
CN (1) CN113609587B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114996842A (en) * 2022-05-26 2022-09-02 哈尔滨工业大学 Track design method for multi-target quick response task

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105677942A (en) * 2015-12-27 2016-06-15 北京航空航天大学 Rapid simulation method of repeat-pass spaceborne natural scene SAR complex image data
CN106647262A (en) * 2016-11-28 2017-05-10 中国人民解放军国防科学技术大学 Differential evolution method facing agile satellite multi-object task planning
CN106845700A (en) * 2017-01-09 2017-06-13 西北工业大学 A kind of many fragments actively remove offline optimal guidance algorithm
US20180357335A1 (en) * 2017-06-08 2018-12-13 Bigwood Technology, Inc. Systems for solving general and user preference-based constrained multi-objective optimization problems

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105677942A (en) * 2015-12-27 2016-06-15 北京航空航天大学 Rapid simulation method of repeat-pass spaceborne natural scene SAR complex image data
CN106647262A (en) * 2016-11-28 2017-05-10 中国人民解放军国防科学技术大学 Differential evolution method facing agile satellite multi-object task planning
CN106845700A (en) * 2017-01-09 2017-06-13 西北工业大学 A kind of many fragments actively remove offline optimal guidance algorithm
US20180357335A1 (en) * 2017-06-08 2018-12-13 Bigwood Technology, Inc. Systems for solving general and user preference-based constrained multi-objective optimization problems

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114996842A (en) * 2022-05-26 2022-09-02 哈尔滨工业大学 Track design method for multi-target quick response task
CN114996842B (en) * 2022-05-26 2022-11-01 哈尔滨工业大学 Track design method for multi-target quick response task

Also Published As

Publication number Publication date
CN113609587B (en) 2023-02-17

Similar Documents

Publication Publication Date Title
CN110046800B (en) Satellite cluster configuration adjustment planning method for space target collaborative observation
CN110378012B (en) Strict regression orbit design method, system and medium considering high-order gravity field
CN113296535B (en) Satellite formation reconstruction algorithm based on stochastic model predictive control
CN113609587B (en) Active drift multi-objective optimization method and system based on satellite three-dimensional imaging
CN113110560B (en) Satellite formation reconstruction model prediction control method based on Chebyshev inequality
Nath et al. Precise halo orbit design and optimal transfer to halo orbits from earth using differential evolution
CN115081343B (en) Space-based passive detection orbit determination method based on neural network combined with genetic algorithm
Judt et al. Development of an automated aircraft subsystem architecture generation and analysis tool
Pan et al. Probabilistic short-term wind power forecasting using sparse Bayesian learning and NWP
CN113093246B (en) Ground multi-target point imaging rapid judging and task parameter calculating method
CN112685841B (en) Finite element modeling and correcting method and system for structure with connection relation
Samsam et al. Nonlinear Model Predictive Control of J2-perturbed impulsive transfer trajectories in long-range rendezvous missions
Ananthasayanam Tuning of the Kalman filter using constant gains
Shirobokov et al. Adaptive neural formation-keeping control for satellites in a low-earth orbit
CN114397818A (en) MAPIO-based spacecraft cluster orbit reconstruction path planning method
CN113609708A (en) Mars global remote sensing track design method and system based on near-fire drift
CN114492175A (en) Multi-target satellite patrol scheme planning method based on intelligent optimization algorithm
Lamassoure et al. Model-based engineering design for trade space exploration throughout the design cycle
Wu et al. Trajectory optimization and maintenance for ascending from the surface of Phobos
Yang et al. A Unified Model for Multi-Satellite Imaging Mission Planning in Various Scenarios
CN116520711B (en) Electric push satellite WALKER constellation networking control planning method
Parsapoor et al. Prediction of solar cycle 24
CN113065693B (en) Traffic flow prediction method based on radial basis function neural network
Liu et al. A Fast Computational Method for the Minimum Duration Transfer Trajectories of Space‐to‐Ground Vehicles
CN118145019A (en) Satellite data processing method, device, equipment and computer storage medium

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant