Regional remote sensing satellite constellation optimization method based on improved crayfish algorithm

Abstract

The invention is applied to the field of satellite constellation design, and particularly discloses a regional remote sensing satellite constellation optimization method based on an improved crayfish algorithm, which comprises the following steps: S1, selecting an orbit observed in a specific region, the orbit needing to meet the average revisit time and the task requirement; the number of satellites meets the task requirement, the orbit height meets the task requirement, and the remote sensing constellation design objective function is S2: mathematical optimization model, and the orbit control parameters of the satellite constellation can be determined according to the six elements of the satellite orbit. According to the regional remote sensing satellite constellation optimization method based on the improved crayfish algorithm, a population initialization improvement strategy and a refraction reverse learning strategy are set, the convergence speed and optimization precision of the algorithm are improved, initial initialization classification is performed on a population through the population initialization improvement strategy, solution individuals are uniformly distributed, and the optimization efficiency is improved. The subsequent initial optimization efficiency of the population is improved, a reverse learning strategy is refracted, the search range of the algorithm is expanded, and the accuracy of the method is improved.

Description

Technical Field

[0001] The invention relates to the technical field of satellite constellation design, and in particular to a regional remote sensing satellite constellation optimization method based on an improved crayfish algorithm. Background Art

[0002] Analytical methods are often used to solve traditional satellite constellation designs. For regional coverage constellations, due to the diversity and flexibility of their coverage areas, analytical methods are usually not used to solve them. Instead, numerical simulation is used to search for constellation parameters that can provide better coverage performance and shorter revisit time, which has become the mainstream method for optimizing regional coverage constellations. In recent years, with the development of intelligent optimization algorithms, modern intelligent optimization algorithms such as genetics, simulated annealing, and neural networks are used to design satellite orbits, which can search in a wider solution space, speed up the search, and also provide better solutions.

[0003] Meziane-Tani et al. used a multi-objective genetic algorithm to improve satellite coverage, minimize the total number of satellites, and reduce satellite altitude as the objective function. The experiment showed that the constellation size designed using MOGA is significantly smaller than that of the traditional geometric design.

[0004] Shtark et al. proposed a low-orbit satellite constellation optimization design method based on genetic algorithm and gradient optimization, taking total coverage time, revisit time and geometric dilution accuracy percentage as objective functions; Yu Yanjun et al. took the minimum number of satellites and short revisit time as optimization goals, adopted an improved simulated annealing algorithm, combined with an improved equal-area grid point coverage method, and proposed a satellite constellation configuration optimization design method for full coverage of irregular area imaging;

[0005] Tania Savitr et al. optimized the orbital parameters by combining genetic algorithm and semi-analytical method with the optimization objectives of maximizing constellation coverage and minimizing revisit time.

[0006] The 3+4N area coverage constellation model is optimized using a genetic algorithm, but the algorithm takes too long.

[0007] The crayfish optimization algorithm is applied to the solution of five engineering problems and has good search capabilities.

[0008] At present, a lot of research has been done on the modeling and solution of micro-nano satellite constellations. The focus of the research is mainly on the solution of satellite constellations, focusing on solving the coverage and revisit time problems in the solution process. However, there are still some shortcomings in the revisit time and cost issues in the design and optimization of regional remote sensing satellite constellations. Summary of the invention

[0009] The purpose of the present invention is to provide a regional remote sensing satellite constellation optimization method based on an improved crayfish algorithm to solve the problems of revisit time and cost in the design optimization of regional remote sensing satellite constellations proposed in the above background technology.

[0010] To achieve the above object, the present invention provides the following technical solution: a method for optimizing a regional remote sensing satellite constellation based on an improved crayfish algorithm, comprising the following steps;

[0011] S1: Select an orbit for observation in a specific area. The orbit needs to meet the mission requirements in terms of average revisit time, number of satellites, and orbital altitude.

[0012] The objective function of remote sensing constellation design is:

[0013] Min(ART)

[0014] Min(N)

[0015] 500km≤a≤1200km

[0016] 43°≤i≤50°

[0017] e=0

[0018] S2: Mathematical optimization model, which determines the orbital control parameters of the satellite constellation based on the six numbers of the satellite orbit;

[0019] That is, Sat=(a,e,i,ω,Ω,Μ);

[0020] a is the semi-major axis, e is the eccentricity, i is the orbital inclination, ω is the argument of perigee, Ω is the right ascension of the ascending node, and M is the mean anomaly;

[0021] S3: Improved population initialization strategy, which combines the improved crayfish optimization algorithm with fusion refraction reverse learning.

[0022] From the geometric relationship of the refraction principle, we know that:

[0023]

[0024] Defined by the refractive index have to:

[0025]

[0026] Let δ=L / L', and combine the above formulas to get:

[0027]

[0028] The above formula is the population individual position transformation formula based on the refraction reverse learning strategy, and its population individual positions are more diverse, which can accelerate convergence;

[0029] S4: Crayfish algorithm strategy, which is a convergence factor nonlinear adjustment strategy. During the summer vacation period, the crayfish's goal is to get close to the cave, which is the optimal solution. The standard COA algorithm updates the position formula by adding the convergence factor C to balance the algorithm's global search and local development capabilities. The update formula is as follows:

[0030]

[0031] Preferably, the S2 types need to use the same semi-major axis at the same time, so that the observed position is a low-orbit circular orbit, and the Walker constellation configuration is selected in S2, and the Walker constellation has the same orbital altitude and orbital inclination.

[0032] Preferably, the S1 needs to minimize its average revisit time and the number of satellites.

[0033] Preferably, an improved EMOCOA algorithm is used in S1, and the improved EMOCOA algorithm is called a multi-objective crayfish optimization algorithm.

[0034] Preferably, in the formula of S3, x is the position of the individual in the population, x′ is the position of the individual in the population after conversion by the refraction reverse learning strategy; [m, n] is the search range; the vertical axis is the normal, and its intersection O with the horizontal axis is the midpoint of the [m, n] interval; the length of the incident light is L, the length of the refracted light is L', α is the incident angle, and β is the refraction angle.

[0035] Preferably, in the formula of S4, y is a random number; Gen and MaxGen are the current number of iterations and the maximum number of iterations, respectively.

[0036] Compared with the prior art, the beneficial effects of the present invention are as follows: the regional remote sensing satellite constellation optimization method based on the improved crayfish algorithm:

[0037] 1. The population initialization improvement strategy and the refraction reverse learning strategy are set to improve the convergence speed and optimization accuracy of the algorithm. The population initialization improvement strategy is used to initialize and classify the population, so that the solution individuals are evenly distributed, which improves the subsequent optimization efficiency of the population initialization. The refraction reverse learning strategy is used to expand the search range of the algorithm and improve the accuracy of the method.

[0038] 2. A multi-objective crayfish optimization algorithm is set up to improve the global search and local development capabilities of the algorithm, enhance the local optimal value capability and the global solution capability of the algorithm, ensure the population diversity of the algorithm, and solve complex optimization problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1A schematic diagram of the refraction principle of the refraction reverse learning in the present invention;

[0040] Figure 2 The test results of the improved EMOCOA algorithm of the present invention and the other three algorithms are shown in the ZDT1 function and the standard Pareto frontier diagram;

[0041] Figure 3 The test results of the EMOCOA algorithm of the present invention and the other three algorithms are shown in the ZDT2 function and the standard Pareto frontier diagram;

[0042] Figure 4 The test results of the EMOCOA algorithm of the present invention and the other three algorithms are shown on the ZDT3 function and the standard Pareto frontier diagram;

[0043] Figure 5 The test results of the EMOCOA algorithm of the present invention and the other three algorithms are shown on the ZDT4 function and the standard Pareto frontier diagram;

[0044] Figure 6 The test results of the improved EMOCOA algorithm of the present invention and the other three algorithms are shown in the ZDT6 function and the standard Pareto frontier diagram;

[0045] Figure 7 Schematic diagram of the Pareto frontier of the revisit time and the number of satellites in the present invention. DETAILED DESCRIPTION

[0046] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] See also Figure 1-7 ,The present invention provides a technical solution: a method for optimizing a regional remote sensing satellite constellation based on an improved crayfish algorithm, comprising the following steps;

[0048] S1: Select an orbit for observation in a specific area. The orbit needs to meet the mission requirements in terms of average revisit time, number of satellites, and orbital altitude.

[0049] The objective function of remote sensing constellation design is:

[0050] Min(ART)

[0051] Min(N)

[0052] 500km≤a≤1200km

[0053] 43°≤i≤50°

[0054] e=0

[0055] S2: Mathematical optimization model, which determines the orbital control parameters of the satellite constellation based on the six numbers of the satellite orbit;

[0056] That is, Sat=(a,e,i,ω,Ω,Μ);

[0057] a is the semi-major axis, e is the eccentricity, i is the orbital inclination, ω is the argument of perigee, Ω is the right ascension of the ascending node, and M is the mean anomaly;

[0058] S3: Improved population initialization strategy, which combines the improved crayfish optimization algorithm with fusion refraction reverse learning.

[0059] From the geometric relationship of the refraction principle, we know that:

[0060]

[0061] Defined by the refractive index have to:

[0062]

[0063] Let δ=L / L', and combine the above formulas to get:

[0064]

[0065] The above formula is the population individual position transformation formula based on the refraction reverse learning strategy, and its population individual positions are more diverse, which can accelerate convergence;

[0066] S4: Crayfish algorithm strategy, which is a convergence factor nonlinear adjustment strategy. During the summer vacation period, the crayfish's goal is to get close to the cave, which is the optimal solution. The standard COA algorithm updates the position formula by adding the convergence factor C to balance the algorithm's global search and local development capabilities. The update formula is as follows:

[0067]

[0068] In order to ensure the stability and resolution of remote sensing satellite observation, when selecting the orbit, it is necessary to confirm the satellites on the selected orbital plane through the mathematical optimization model and the orbit for observation in the specific area. The same semi-major axis and low-orbit circular orbit with 0 eccentricity are required. At the same time, the launch method and other factors are considered to make the semi-major axis range from 500 to 1200 km. The Walker constellation configuration is selected by considering the stability of the constellation configuration and the launch of multiple satellites on the same orbital plane. At the same time, the Walker constellation configuration code is N / P / F to determine the phase difference between adjacent orbits, where N is the number of satellites, P is the number of orbital planes, and F is the phase factor. The seed satellite of the Walker constellation is the first satellite on the first orbital plane. After confirming that the orbital parameters of the seed satellite are determined, the orbital parameters of the satellite are selected through the formula inside S2. On the basis of S1, the constraints inside S2 are used to select the remote sensing satellite constellation so that the mean revisit time meets the mission requirements, the number of satellites meets the mission requirements, and the orbital height meets the mission requirements. The design objective function of the remote sensing constellation for observation in a specific area is:

[0069] Min(ART)

[0070] Min(N)

[0071] st

[0072] 500km≤a≤1200km

[0073] 43°≤i≤50°

[0074] e=0

[0075] On the basis of mathematical optimization model, the improved EMOCOA algorithm is applied to the optimization design of regional coverage satellite constellation. Due to the limitations of the traditional crayfish algorithm MOCOA, the algorithm is integrated with refraction reverse learning and strategy for population initialization into the internal of MOCOA algorithm to obtain the improved improved crayfish optimization algorithm (EMOCOA). The convergence factor nonlinear adjustment strategy can be adopted in position update to balance the global search and local development capabilities of the algorithm. The population initialization improvement strategy can be used to select and improve the COA algorithm during initialization to expand the search range of the COA algorithm, thereby improving the convergence speed and optimization accuracy of the algorithm. The individual position transformation formula of the population based on the refraction reverse learning strategy can be used according to the formula inside S3, and the individual position of the population is more diverse, which can accelerate convergence. In the summer vacation stage of the existing algorithm, the goal of the crayfish is to get close to the cave, which is the optimal solution. The standard COA algorithm updates the position formula by adding a convergence factor C to balance the global search and local development capabilities of the algorithm. Since the optimal value of the traditional algorithm is poor, the global solution capability of the algorithm is severely weakened. The present invention uses a nonlinear convergence factor to update the formula and change the formula to:

[0076]

[0077] Finally, the algorithm is tested and simulated.

[0078] In order to verify the performance of the improved crayfish algorithm, the classic ZDT series multi-objective test functions (ZDT1, ZDT2, ZDT3, ZDT6) are selected as test functions.

[0079] Population size: set N = 20;

[0080] External archive capacity: pop = 100;

[0081] Termination criterion: The termination criterion of the run is set to the maximum number of fitness function calculations, and the maximum number of fitness function calculations maxFE is set to 100.

[0082] Evaluation indicators: Select the currently widely used evolutionary algorithm evaluation indicators: IGD, GD, HV and SP (Spacing). Among them, IGD and HV are used to evaluate the convergence and diversity of the algorithm. The smaller the IGD value and the larger the HV value, the better the convergence of the algorithm and the more uniform the distribution. GD is used to evaluate the convergence of the algorithm. The smaller the GD value, the better the convergence of the algorithm. SP is used to evaluate whether the individuals in the obtained Pareto front are evenly distributed. The smaller the SP value, the more uniform the distribution of the Pareto front obtained by the algorithm.

[0083] Depend on Figure 2 , Figure 3 , Figure 4 , Figure 5 and Figure 6 It can be seen that the improved crayfish algorithm EMOCOA performs well in convergence. It is better than NSGA-2, MOPSO, and MSSA algorithms in convergence on four test functions. It is 54.4%, 78.7%, 3.6%, and 27.3% higher than the other three algorithms in terms of IGD, GD, HV, and SP. At the same time, the EMOCOA algorithm has also made great improvements in diversity. The Pareto frontier distribution obtained by the EMOCOA algorithm is more uniform than that of the other three optimization algorithms. Figures 3 to 7 It can be seen that the convergence speed of the improved EMOCOA algorithm is also better than that of the other three optimization algorithms, which proves the effectiveness of the improved EMOCOA algorithm.

[0084] The specific process of using EMOCOA to solve the regional remote sensing micro-nano satellite constellation design problem can be described as follows:

[0085] Step 1: According to the constraints of the constellation model, initialize the crayfish population with a size of N, and set the external archive capacity pop and the number of iterations t;

[0086] Step 2: Use Matlab to call the simulation data of revisit time in the STK coverage module to obtain the fitness of each individual crayfish in the crayfish population;

[0087] Step 3: For each crayfish individual, compare its fitness with the fitness of the best position it has experienced during the summer, competition, and foraging stages, retain the one with better fitness as the current best position and save it in the external archive capacity pop;

[0088] Step 4: For each crayfish individual, compare its fitness with the fitness of the global optimal position, and retain the one with better fitness as the current global best position;

[0089] Step 5: Update the position of each crayfish according to the formula;

[0090] Step6: If the termination condition is met: the maximum number of iterations, output the solution, otherwise return to Step2.

[0091] The selected area is the Beijing-Tianjin-Hebei region, with a longitude range of 113°27′E~119°50′E and a latitude range of 36°05′N~42°40′N, so the orbital inclination i ranges from 43° to 50°. The MATLAB and STK joint simulation method is used, with a simulation period from 12:00:00 on February 20, 2024 to 12:00:00 on February 21, 2024, and a simulation duration of 1 day; the step size of the simulation time is 10 seconds. The sensor field of view is a simple conical field of view with a half cone angle of 50°.

[0092] The parameters of the optimization algorithm are set as follows: the population size N is 50, the external archive capacity pop is 300, the number of iterations t is 500, and the other parameters are the same as those set in Section 3. The simulation results are shown in Figure 7 shown.

[0093] From the simulation results, it can be seen that when the semi-major axis a is 1122km and the orbital inclination i is 43°, three satellites are selected and evenly distributed on three orbital planes, which can achieve an average revisit coverage of the Beijing-Tianjin-Hebei region of 1.54h, providing time guarantee for reconnaissance and detection missions in response to various natural disasters and military conflicts.

[0094] The S2 type also needs to use the same semi-major axis, so that the observed position is a low-orbit circular orbit, and the Walker constellation configuration is selected in S1, and the Walker constellation has the same orbital altitude and orbital inclination;

[0095] In S1, the position of the observed orbit needs to be confirmed. When confirming, the Walker constellation can be locked according to different requirements.

[0096] In S2, the average revisit time needs to be minimized, and the number of satellites needs to be minimized;

[0097] The average revisit time and the number of simultaneous satellites in the area are displayed to facilitate selection

[0098] The improved EMOCOA algorithm is adopted in S2, and the improved EMOCOA algorithm is called a multi-objective crayfish optimization algorithm.

[0099] The improved EMOCOA algorithm combines the population initialization improvement strategy and refraction reverse learning for learning;

[0100] In the formula of S3, x is the position of the individual in the population, x′ is the position of the individual in the population after conversion by the refraction reverse learning strategy; [m, n] is the search range; the vertical axis is the normal, and its intersection O with the horizontal axis is the midpoint of the [m, n] interval; the length of the incident light is L, the length of the refracted light is L′, α is the incident angle, and β is the refraction angle.

[0101] The vertical position of the letters can be substituted into the formula inside to calculate, and the individual positions of the population can be calculated.

[0102] In the formula of S4, y is a random number; Gen and MaxGen are the current number of iterations and the maximum number of iterations respectively;

[0103] Through this formula, the C update formula can be calculated.

[0104] Working principle: When using the regional remote sensing satellite constellation optimization method based on the improved crayfish algorithm, a population initialization improvement strategy and refraction reverse learning are set up to strengthen the algorithm of the device, shorten the calculation time, and increase the overall practicality.

[0105] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A regional remote sensing satellite constellation optimization method based on an improved crayfish algorithm, characterized in that: The steps include: S1: Select an orbit for observing a specific area, which needs to meet the mission requirements with an average revisit time; The number of satellites meets the mission requirements, and the orbital height meets the mission requirements. The objective function of remote sensing constellation design is: Min(ART) Min(N) 500km≤a≤1200km 43°≤i≤50° e＝0 S2: Mathematical optimization model, which determines the orbital control parameters of the satellite constellation based on the six numbers of the satellite orbit; That is, Sat=(a,e,i,ω,Ω,Μ); a is the semi-major axis, e is the eccentricity, i is the orbital inclination, ω is the argument of perigee, Ω is the right ascension of the ascending node, and M is the mean anomaly; S3: Improved population initialization strategy, which combines the improved crayfish optimization algorithm with fusion refraction reverse learning. From the geometric relationship of the refraction principle, we know that: Defined by the refractive index have to: Let δ=L / L', and combine the above formulas to get: The above formula is the population individual position transformation formula based on the refraction reverse learning strategy, and its population individual positions are more diverse, which can accelerate convergence; S4: Crayfish algorithm strategy, which is a convergence factor nonlinear adjustment strategy. During the summer vacation period, the crayfish's goal is to get close to the cave, which is the optimal solution. The standard COA algorithm updates the position formula by adding the convergence factor C to balance the algorithm's global search and local development capabilities. The update formula is as follows: 。 2. The method for optimizing a regional remote sensing satellite constellation based on an improved crayfish algorithm according to claim 1, characterized in that: The S2 types also need to use the same semi-major axis so that the observed position is a low-orbit circular orbit. At the same time, the Walker constellation configuration is selected in S2, and the Walker constellation has the same orbital altitude and orbital inclination.

3. The method for optimizing a regional remote sensing satellite constellation based on an improved crayfish algorithm according to claim 1, characterized in that: The S1 needs to minimize its average revisit time and the number of satellites.

4. The method for optimizing a regional remote sensing satellite constellation based on an improved crayfish algorithm according to claim 1, characterized in that: The improved EMOCOA algorithm is adopted in S1, and the improved EMOCOA algorithm is called a multi-objective crayfish optimization algorithm.

5. The method for optimizing a regional remote sensing satellite constellation based on an improved crayfish algorithm according to claim 1, characterized in that: In the formula of S3, x is the position of the individual in the population, x′ is the position of the individual in the population after conversion by the refraction reverse learning strategy; [m, n] is the search range; the vertical axis is the normal, and its intersection O with the horizontal axis is the midpoint of the [m, n] interval; the length of the incident light is L, the length of the refracted light is L′, α is the incident angle, and β is the refraction angle.

6. The method for optimizing a regional remote sensing satellite constellation based on an improved crayfish algorithm according to claim 1, characterized in that: In the formula of S4, y is a random number; Gen and MaxGen are the current iteration number and the maximum iteration number, respectively.

Images