A single-step predictive avoidance trajectory planning method for spacecraft optical imaging games

Abstract

The present invention discloses a single-step prediction avoidance trajectory planning method for spacecraft optical imaging game, which includes the following steps: constructing a two-layer optimization architecture for the optical imaging game avoidance strategy optimization problem described in the figure. First, in the upper-layer optimization, the avoidance strategy is optimized with the avoidance star angle, and the effective imaging time and orbit constraint penalty term after avoidance are calculated. In addition, the future strategy of the imaging star is predicted in the form of lower-layer optimization, and its speed increment consumption is estimated to calculate the cost-effectiveness of its own avoidance maneuver. Finally, the weighted sum of the effective imaging time, penalty term and cost-effectiveness is calculated to obtain the fitness function, and an iterative optimization calculation is performed using a genetic algorithm to obtain an avoidance strategy based on single-step prediction. While making up for the lack of relevant research on optical imaging avoidance strategies in existing literature, the present invention solves the problem of lack of optimality of unilateral optimization in the long-term game process.

Description

Technical Field

[0001] The present invention belongs to the field of aerospace technology, and in particular relates to a single-step prediction avoidance trajectory planning method for spacecraft optical imaging game. Background Art

[0002] Since the orbital period of geosynchronous orbit satellites is the same as the Earth's rotation period, the spatial geometric relationship between them and the Earth is relatively fixed, and they can provide continuous and long-term services to ground users. Therefore, geosynchronous orbit resources are very valuable and have become a hot spot for space competition among various countries. At present, satellites operating in geosynchronous orbit include communication satellites, data relay satellites, electronic reconnaissance satellites, missile warning satellites, and meteorological satellites. These satellites constitute important nodes of the military information network and have important strategic significance. The literature (Diao Huafei, Zhang Yasheng. Analysis of the Capabilities of US High-orbit Situational Awareness Satellites [J]. Aerospace Electronic Countermeasures, 2019, 35(04): 48-51.) points out that in recent years, the United States and other aerospace powers have vigorously developed various types of high-orbit target situational awareness aircraft, such as high-orbit approach imaging and detection cataloging.

[0003] At present, the pursuit-escape game with relative distance as the game objective is the mainstream research direction. Deep reinforcement learning algorithms are often used to obtain long-term dominant solutions or differential game methods are used to obtain saddle point solutions. Spacecraft optical imaging game is a new scenario in the field of space-based situational awareness game. It not only requires proximity in distance, but also requires the orbit to closely match the changes in sunlight to meet the imaging requirements of the camera. Therefore, multiple imaging constraints need to be considered in the model. The core game point of the pursuit and escape parties is the effective monitoring time to meet the optical imaging constraints. The imaging satellite hopes to extend the monitoring time as much as possible, while the evasion satellite hopes to reduce the other party's monitoring time to zero seconds as much as possible. Prince et al. (Prince ER, Hess JA, Cobb RG, et al. Elliptical Orbit Proximity Operations Differential Games[J]. Journal of Guidance, Control, and Dynamics, 2019, 42(7): 1458-1472.) based on relative orbital dynamics, extended the classic free-time interception game model to models such as capturing sunlight vectors. Although the lighting conditions in the game are taken into account, the lighting is only used as the terminal constraint condition for pursuit, and the influence of more camera imaging factors is not considered. The literature (Liu Xiangchun. Orbit design and mission planning of optical surveillance satellite for space targets [D]. National University of Defense Technology.) designed a space target optical surveillance mission planning method, and conducted research from two aspects: long-range detection and close-range imaging surveillance. The literature (Xiao Yuzhi, Chen Jizheng. Autonomous perception technology of high-orbit targets based on orbital maneuvers [J]. Space Return and Remote Sensing, 2021, 42(01): 1-10.) proposed a flyby and fly-around imaging orbit control method for high-orbit targets based on the analysis of high-orbit lighting reconnaissance needs and combined the target orbit characteristics and lighting characteristics.

[0004] In summary, there are currently few studies on spacecraft optical imaging games, and most of the research is conducted from the perspective of imaging stars. No scholar has yet proposed an optical imaging game avoidance method based on the avoidance star angle and comprehensive consideration of illumination and imaging constraints. The general solution method that uses optimization algorithms to optimize a certain indicator is only a unilateral optimization and cannot consider the impact of the current action on the subsequent game process like the reinforcement learning algorithm. Summary of the Invention

[0005] In view of the above-mentioned deficiencies in the prior art, the purpose of the present invention is to provide a single-step predictive avoidance trajectory planning method for spacecraft optical imaging games. While making up for the lack of relevant research on optical imaging avoidance strategies in the existing literature, it also solves the problem of lack of optimality of unilateral optimization in the long-term game process.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0007] This planning method is a single-step predictive avoidance trajectory planning method for spacecraft optical imaging games, which includes the following steps:

[0008] The planning method includes two parts: upper-level optimization and lower-level optimization. The upper-level optimization is for optimizing the own strategy, while the lower-level optimization is for predicting the target strategy.

[0009] Step 1: Based on the genetic algorithm, the optimization variables and fitness function are defined to solve the optimization problem of the pulse maneuver optical game avoidance strategy of the avoidance star. The optimization variables are selected as the pulse maneuver waiting time and pulse maneuver velocity vector of the avoidance star; the imaging star and the avoidance star are dynamically evolved to the end time of the mission, and the orbit information is obtained by uniformly sampling points. The total time T that meets the optical imaging constraint conditions within the mission time is calculated. s , determine whether the orbit constraint is violated within the avoidance satellite mission time T, and calculate the penalty term ;

[0010] Calculate the cost-effectiveness ratio based on the minimum speed increment consumption ; The total time T s , penalty item and cost-effectiveness The weighted sum is used as the fitness function, and the genetic algorithm is called to perform optimization variable optimization calculation to obtain the single-step prediction avoidance trajectory;

[0011] Step 2: Using a sequential quadratic programming algorithm, solve the lower-level optimization problem, which is the optimization problem for the imaging satellite's pulse maneuver optical game imaging strategy. Define the imaging satellite's optical imaging maneuver mode as flying over the center of the target's optical imaging effective range. Set the transfer time to the center as the optimization variable. Solve the Lambert problem to obtain the imaging satellite's pulse velocity increment vector, and optimize the imaging satellite's minimum fuel consumption as the performance metric. Return the imaging satellite's minimum fuel consumption to step 1.

[0012] Furthermore, the step 1 specifically includes:

[0013] Assume that the initial orbital states of the imaging star and the avoiding star are X p0 =[r p0 , v p0 ] and X e0 = [r e0 , v e0 ], the initial orbit of the imaging star passes through the optical imaging range of the avoidance star;

[0014] X p0 =[r p0 , v p0 ] and X e0= [r e0 , v e0 ] as input, and call the genetic algorithm to obtain the avoidance strategy of the avoidance star, that is, the avoidance time and the pulse maneuver speed increment;

[0015] In the genetic algorithm framework, the avoidance star is set to avoid through a single pulse maneuver, and the optimization variable is defined as the pulse maneuver waiting time of the avoidance star. Velocity vector with pulse maneuvering :

[0016] ;

[0017] in, is the pulse maneuver velocity vector, is the velocity increment amplitude, is the velocity direction angle, then the optimization variable ;

[0018] The dynamic evolution mode is set to pulse maneuver two-body dynamics. According to the initial state and pulse maneuver, the states of the imaging star and the avoiding star are evolved to the end of the mission, and the orbit information is obtained by uniformly sampling points.

[0019] Setting optical imaging constraints consists of two parts:

[0020] Sunlight angle: Set the sunlight angle as the camera imaging line of sight to observe the target along the sun's rays. Within the set range, that is

[0021] ;

[0022] Where, is the lower bound of the sun angle constraint, is the upper bound of the sunlight angle constraint;

[0023] Distance constraint: the distance constraint When the camera images the object, the object distance remains within the set range, that is,

[0024] ;

[0025] Where, is the lower bound of the sight distance constraint, is the upper bound of the sight distance constraint;

[0026] The calculation formula for the time required to meet the optical imaging constraints is:

[0027] ;

[0028] Where, The total time to meet the optical imaging constraints is is the initial moment, T is the mission time, k=1,2 represents the sequence number of the optical imaging constraint, is the switch function corresponding to the sequence constraint:

[0029] ;

[0030] Where, , 、 They are the lower and upper bounds of the sight distance constraint in the distance constraint respectively; , 、 are the lower and upper bounds of the sunlight angle constraint respectively;

[0031] The avoidance satellite must meet the set orbital constraints while avoiding imaging:

[0032] ;

[0033] in is the change in longitude, is the upper bound of the longitude change; construct the penalty function J p :

[0034] ;

[0035] Based on the orbital information, the sequential quadratic programming algorithm is used to solve the optimization problem and obtain the minimum fuel consumption required for the imaging star to re-image. , thus calculating the cost-effectiveness ratio :

[0036] ;

[0037] T s , and Weighted sum calculation fitness function:

[0038] ;

[0039] Where, is the fitness function value, is the weight coefficient, is the weight coefficient of the penalty term, and the optimization problem of the pulse maneuver optical game avoidance strategy of the avoidance star is modeled as follows:

[0040] ;

[0041] Genetic algorithm is used to solve the above optimization problem and obtain Optimization variable with minimum value , that is, obtaining a single-step predicted avoidance trajectory.

[0042] Furthermore, the step 2 specifically includes:

[0043] The orbital information is used as input and the sequential quadratic programming algorithm is used to solve the optimization problem and make a single-step prediction for the re-imaging strategy of the imaging star:

[0044] The imaging star maneuver mode is set as follows: after the avoidance star maneuver, it will fly over the center of the optical imaging effective range of the avoidance star with a Lambert maneuver and re-perform optical imaging. tran To optimize the variables, the imaging star is obtained from the orbit information at t w Momentary status With the avoidance star in (t w +t tran ) Status at the moment , where X is the state vector; r, v are position and velocity vectors respectively; subscripts e and p refer to the avoidance star and imaging star respectively;

[0045] Calculated in (t w +t tran ) Unit sunlight vector at time , obtain (t w +t tran ) to avoid the center position vector of the star optical imaging , where To avoid the center position vector of star optical imaging;

[0046] Solve the computational Lambert problem:

[0047] ;

[0048] Where, The imaging star is at (t+t tran ) should have a speed, Calculate the function for Lambert;

[0049] The pulse maneuvering speed increment vector is calculated as:

[0050] ;

[0051] Calculate its speed increment consumption ,by In order to minimize the performance index, the sequential quadratic programming method is used to optimize and obtain the predicted minimum fuel consumption of the imaging satellite. , returning the minimum fuel consumption of the imaging star to step 1.

[0052] Beneficial effects of the present invention:

[0053] This method introduces a two-layer optimization architecture to predict the imaging satellite's future strategy from the perspective of the avoidance star. Cost-effectiveness performance indicators are calculated and incorporated into the fitness function for optimization. This method achieves the optimal avoidance strategy that maximizes the imaging satellite's response strategy costs, enabling the avoidance star to successfully avoid optical imaging while maintaining an advantage in future game scenarios. This method addresses the lack of optimality of existing unilateral optimization strategies in multi-long-term game scenarios. Its overall concept is novel and has broad application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 It is the algorithm flow chart of the present invention;

[0055] Figure 2 This is a schematic diagram of the optical imaging game scenario in the present invention;

[0056] Figure 3 is a schematic diagram of the sunlight angle used as an optical imaging constraint in the present invention;

[0057] Figure 4 is the distance switch function image in the present invention;

[0058] Figure 5 is a schematic diagram of the optical imaging strategy of the imaging star in the present invention;

[0059] Figure 6 It is the genetic algorithm iterative optimization image during the optimization of the present invention;

[0060] Figure 7 It is the trajectory image of the inertial system of the center of mass of the avoiding star;

[0061] Figure 8 It is the image of the relative distance change between the two parties;

[0062] Figure 9 It is an image that tracks the changes in the star's sunlight angle;

[0063] Figure 10 It is an image that avoids changes in star longitude. DETAILED DESCRIPTION

[0064] In order to facilitate understanding by those skilled in the art, the present invention will be further described below with reference to embodiments and drawings. The contents mentioned in the embodiments are not intended to limit the present invention.

[0065] The single-step prediction avoidance trajectory planning method for spacecraft optical imaging game provided by the present invention includes the following steps:

[0066] This planning method is a two-level optimization problem, which includes upper-level optimization and lower-level optimization. The upper level is the optimization of its own strategy, and the lower level is the prediction of the target strategy.

[0067] Step 1: Based on the genetic algorithm, define the optimization variables and fitness function to solve the upper optimization problem. The upper optimization is to optimize the pulse maneuver optical game avoidance strategy of the avoidance star. The optimization variables are selected as the pulse maneuver waiting time and pulse maneuver velocity vector of the avoidance star. The imaging star and the avoidance star are dynamically evolved to the end time of the mission, and points are evenly selected to calculate the total time T that meets the optical imaging constraint conditions within the mission time. s ; Determine whether the orbit constraint is violated within the avoidance satellite mission time T and calculate the penalty term .

[0068] The lower layer optimization of step 2 is used to predict the maneuver fuel consumption of the imaging satellite and calculate the cost-effectiveness. . The total time T s , penalty item and The weighted sum is used as the fitness function, and the genetic algorithm is called to perform optimization variable optimization calculation to obtain the single-step prediction avoidance trajectory;

[0069] Step 2: Using a sequential quadratic programming algorithm, solve the underlying optimization problem to optimize the imaging satellite's pulsed maneuver optical game imaging strategy. Define the imaging satellite's optical imaging maneuver mode as flying over the center of the target's effective optical imaging range. Set the transit time to the center as the optimization variable. Solve the Lambert problem to obtain the imaging satellite's pulse velocity increment vector, minimizing the velocity increment cost. Once the solution is complete, return the optimized minimum velocity increment cost to Step 1.

[0070] Furthermore, step one specifically includes:

[0071] The algorithm flow chart is as follows Figure 1 As shown, the input is the initial state of the binary star, the genetic algorithm is called for optimization and solution, the sequential quadratic programming algorithm is called as the lower-level optimization when calculating the fitness function, and the final output is the avoidance strategy of the avoidance star. First, establish the simulation scenario shown in 2, where the avoidance star is initially located on the GEO orbit. The conical area in the figure is the area where the avoidance star can be optically imaged to meet the sunlight angle and distance constraints. The imaging star is near the GEO orbit, and the initial relative distance between the two is about 200km. The orbit of the imaging star will pass through the effective range of the optical imaging of the avoidance star. At this time, the avoidance strategy of the avoidance star is planned. Set the current scene time to 0:0:00 on January 1, 2024, and set the initial orbital states of the imaging star and the avoidance star to X respectively. p0 =[r p0 , v p0 ] and X p0 = [r e0 , v e0], where X is the state vector; r and v are the position and velocity vectors, respectively; the subscripts e and p denote the avoidance and imaging stars, respectively, and 0 represents the initial state. The genetic algorithm population size is set to 150, the maximum number of iterations is set to 60, and parameter initialization is complete.

[0072] The avoidance star is set to avoid through a single pulse maneuver, and the optimization variable is defined as the pulse maneuver waiting time of the avoidance star. Velocity vector with pulse maneuvering :

[0073] ;

[0074] in, is the pulse maneuver velocity vector, is the velocity increment amplitude, is the velocity direction angle. Therefore, the optimized variable .

[0075] Furthermore, the range of optimized variables is determined. Assume that the total mission time is T=120000s and the maximum pulse amplitude of the avoided star is The value range of each variable is , , , .

[0076] Then, the optimization variables are converted from numerical form to binary code form to generate 150 initial populations. Furthermore, the fitness function value corresponding to the initial population code is solved. First, the population code is decoded and restored to the numerical optimization variable, that is, .

[0077] Set the dynamic equations of the spacecraft in the geocentric J2000 inertial coordinate system for:

[0078] ;

[0079] Where, subscript i = “p” represents the imaging star, and subscript i = “e” represents the avoidance star; is the Earth's gravitational constant, x, y and z represent the three-axis position components of the spacecraft in the Earth's center J2000 inertial system, v x , v y and v z They represent the three-axis velocity components of the spacecraft in the Earth's center J2000 inertial system; Δv x , Δv y and Δv z It represents the three-axis velocity components of the spacecraft in the J2000 inertial system at the center of the earth when the spacecraft applies the pulse maneuver velocity increment at the initial moment of each interval; r is the distance from the center of the earth, , R E represents the average radius of the Earth.

[0080] The initial state of the avoidance star is X e0 , and in Always The pulse maneuver is recursively carried out through the above dynamic equations until the mission time ends, that is, at the moment T=120000s, and points are evenly taken in time to obtain the state information within this orbital segment; the initial state of the imaging satellite is X p0 , without maneuvering, the above dynamic equation is recursively deduced to the moment T=120000s, and points are also taken evenly in time.

[0081] Next, we need to calculate the duration for the imaging satellite to meet the optical imaging constraint from the orbital status information of both parties. The optical imaging constraint consists of the following two parts:

[0082] (1) Sunlight angle:

[0083] When the camera forms an image, the line of sight needs to follow the sunlight to observe the target, so it is necessary to ensure the angle between the line of sight and the light, that is, the sunlight angle (like Figure 3 within a certain range, i.e.

[0084] ;

[0085] Where, is the lower bound of the sun angle constraint, The upper bound of the sun angle constraint.

[0086] (2) Distance constraint:

[0087] When the camera images an object, the object distance needs to be kept within an appropriate range, that is,

[0088] ;

[0089] Where, is the lower bound of the sight distance constraint, The upper bound of the sight distance constraint.

[0090] The calculation formula for the time required to meet the optical imaging constraints is:

[0091] ;

[0092] Where, The total time to meet the optical imaging constraints is is the initial moment, T is the task time, k=1,2 represents the sequence number of the two optical imaging constraints, is the switching function corresponding to the sequence constraint (as shown in 4, which is the switching function image of the distance).

[0093] ;

[0094] Where k=1,2 are the subscript numbers of the two optical imaging constraints. , ;

[0095] The avoidance satellite must meet the set orbital constraints while avoiding imaging:

[0096] ;

[0097] in is the change in longitude, is the upper bound of the longitude change. Construct the penalty function J p :

[0098] ;

[0099] Input the orbit information into the lower optimization layer and call the lower optimization layer. The lower optimization layer is described in detail in step 2. The output of the lower optimization layer is the minimum fuel consumption required to predict the imaging star to re-image. , thus calculating the cost-effectiveness ratio :

[0100] ;

[0101] T s , and Weighted sum as fitness function

[0102]

[0103] Where, is the fitness function value, is the weight coefficient, is the weight coefficient of the penalty term. , the evasive star will focus more on the effect of evasive imaging, that is, to make the effective imaging time of the pursuer as short as possible.

[0104] The population's fitness function is thus solved. If the number of iterations does not exceed the maximum number of 60, the population code is replicated, mutated, and crossovered to generate a descendant population. The fitness function is then re-solved, and the number of iterations is increased by one. If the number of iterations exceeds the maximum number of 60, the iteration loop is exited, the population code with the optimal fitness function is output, and a single-step predicted avoidance trajectory is calculated.

[0105] Furthermore, step 2 specifically includes:

[0106] The orbit information calculated by the above optimization is used as input, and the sequential quadratic programming algorithm is used to solve the optimization problem, and a single-step prediction is made for the re-imaging strategy of the imaging star.

[0107] The imaging star's maneuvering mode is set as follows: after the avoidance star maneuver, it will perform a Lambert maneuver to fly over the center of the avoidance star's effective optical imaging range and perform optical imaging again, such as Figure 5 As shown. The transfer time t of Lambert maneuver tran To optimize the variables, the imaging star is obtained from the orbit information at t w Momentary status With the avoidance star in (t w +t tran ) Status at the moment , where X is the state vector; r, v are the position and velocity vectors respectively; subscripts e and p refer to the avoidance star and imaging star respectively. w +t tran ) Unit sunlight vector at time , obtain (t w +t tran ) to avoid the center position vector of the star optical imaging , where To avoid the center position vector of star optical imaging, check the ephemeris to obtain the starting point of the scene (t w +t tran ) in the J2000 coordinate system at the time , calculate the sunlight unit vector:

[0108] ;

[0109] In the formula Sunlight unit vector.

[0110] To calculate (t w +t tran ) to avoid the center position vector of the star optical imaging . Now let’s calculate the Lambert problem:

[0111] ;

[0112] Where, The imaging star is at (t+t tran ) should have a speed, is the Lambert calculation function. Further calculation shows that the pulse maneuvering speed increment vector is

[0113] ;

[0114] And calculate its speed increment consumption .by In order to minimize the performance index, the sequential quadratic programming method is used to optimize and obtain the predicted minimum fuel consumption of the imaging satellite. , returned to the upper layer for optimization.

[0115] According to the above process, the genetic algorithm is used to solve the above optimization problem. The iterative optimization graph of the genetic algorithm is shown in Figure 6. The optimal value gradually converges to a lower value. It can be considered that an approximate optimal solution is obtained, and the penalty constraint is not triggered.

[0116] After strategy planning is completed, the avoidance strategy of the avoidance star is substituted and the scenario is evolved. The resulting trajectory image is shown in Figure 7. It can be seen that the tracking star trajectory fails to pass through the effective optical imaging range of the avoidance star, and the executable imaging duration is 0s. Figure 8 and Figure 9 The relative distance and sunlight angle images, respectively, show that there is no time period in which both constraints are satisfied simultaneously. Therefore, the avoidance star strategy can achieve effective avoidance. As shown in Figure 10, the orbital position of the avoidance star does not exceed the constraint during the entire game.

[0117] On this basis, we add the imaging satellite's response strategy and use the imaging satellite strategy described in step 2 to generate the strategy. The imaging satellite replans the trajectory to bring the target back into the imaging range. The speed increment required is 2.5m / s. At this time, the cost-effectiveness of the pursuit and escape parties is It is 1.5476.

[0118] To further illustrate the effectiveness of the single-step prediction mechanism, a comparative simulation is conducted under the same scenario. In this case, the avoidance star is not predicted in a single step, and only the effective imaging time of the imaging star is minimized during optimization, i.e. At this time, the imaging satellite performs a maneuver imaging strategy again, with a fuel consumption of 1.5 m / s. The fuel consumption ratio between the avoidance star and the imaging satellite is 3.3172.

[0119] In summary, while achieving the same optical avoidance effect (the imaging satellite's original effective imaging duration is reduced to 0 seconds), the avoidance satellite strategy using single-step prediction increases the target satellite's next maneuver speed increment by 1 m / s compared to the strategy without single-step prediction, reducing the fuel consumption ratio between the two targets by 53.35%. This means that while the avoidance satellite consumes the same fuel, the imaging satellite's speed increment required to achieve its target is significantly increased. This shows that introducing a single-step optimization strategy during avoidance optimization effectively increases the difficulty and speed increment of the imaging satellite's reconnaissance trajectory replanning, improves the avoidance satellite's fuel efficiency during avoidance, and ultimately leads to a dominant position in the subsequent game process.

[0120] The present invention has many specific application paths. The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements can be made without departing from the principles of the present invention. These improvements should also be considered as the scope of protection of the present invention.

Claims

1. A single-step prediction avoidance trajectory planning method for spacecraft optical imaging game, characterized by: The steps are as follows: The planning method includes two parts: upper-level optimization and lower-level optimization. The upper-level optimization is for optimizing the own strategy, while the lower-level optimization is for predicting the target strategy. Step 1: Based on the genetic algorithm, the optimization variables and fitness function are defined to solve the optimization problem of the pulse maneuver optical game avoidance strategy of the avoidance star. The optimization variables are selected as the pulse maneuver waiting time and pulse maneuver velocity vector of the avoidance star; the imaging star and the avoidance star are dynamically evolved to the end time of the mission, and the orbit information is obtained by uniformly sampling points. The total time T that meets the optical imaging constraint conditions within the mission time is calculated. s , determine whether the orbit constraint is violated within the avoidance satellite mission time T, and calculate the penalty term J p ; Calculate the cost-effectiveness ratio η by the minimum speed increment consumption; the total time T s , penalty term J p The weighted sum of the cost-effectiveness ratio η is used as the fitness function, and the genetic algorithm is called to perform optimization variable optimization calculation to obtain the single-step predicted avoidance trajectory; Step 2: Based on the sequential quadratic programming algorithm, solve the imaging star's pulse maneuver optical game imaging strategy optimization problem. Define the imaging star's optical imaging maneuver mode as flying over the center of the target's optical imaging effective range. Set the transfer time to the center as the optimization variable. Solve the Lambert problem to obtain the imaging star's pulse velocity increment vector, and use the imaging star's minimum fuel consumption as the minimization performance indicator. Return the imaging star's minimum fuel consumption to step 1. The step 1 specifically includes: Assume that the initial orbital states of the imaging star and the avoiding star are X p0 =[r p0 ,v p0 ] and X e0 =[r e0 ,v e0 ], the initial orbit of the imaging star passes through the optical imaging range of the avoidance star; X p0 =[r p0 ,v p0 ] and X e0 =[r e0 ,v e0 ] as input, and call the genetic algorithm to obtain the avoidance strategy of the avoidance star, that is, the avoidance time and the pulse maneuver speed increment; In the genetic algorithm framework, the avoidance star is assumed to be avoided by a single pulse maneuver, and the optimization variable is defined as the pulse maneuver waiting time t of the avoidance star. w With impulse maneuver velocity vector dv: where △v e is the velocity increment amplitude, α, β are the velocity direction angles, then the optimized variable y=[t w ,dv,α,β]; The dynamic evolution mode is set to pulse maneuver two-body dynamics. According to the initial state and pulse maneuver, the states of the imaging star and the avoiding star are evolved to the end of the mission, and the orbit information is obtained by uniformly sampling points. Setting optical imaging constraints consists of two parts: Sunlight angle: When the camera is imaging, the sight line is along the sun's rays to observe the target. The sunlight angle θ is within the set range, that is, i l ≤θ≤θ u ; Where θ l is the lower bound of the sunlight angle constraint, θ u is the upper bound of the sunlight angle constraint; Distance constraint: The distance constraint r is that when the camera images the object, the object distance remains within the set range, that is, r l ≤r≤r u ； Where r l is the lower bound of the sight distance constraint, r u is the upper bound of the sight distance constraint; The calculation formula for the time required to meet the optical imaging constraints is: Where, T s is the total time to meet the optical imaging constraint, t0 is the initial time, T is the task time, k=1,2 represents the sequence number of the optical imaging constraint, S k is the switch function corresponding to the sequence constraint: Where c1 = r, c 1,l 、c 1,u are the lower and upper bounds of the sight distance constraint in the distance constraint respectively; c2 = θ, c 2,l 、c 2,u are the lower and upper bounds of the sunlight angle constraint respectively; The avoidance satellite must meet the set orbital constraints while avoiding imaging: |△λ|≤△λ u ; Where |△λ| is the change in longitude, λ u is the upper bound of the longitude change; construct the penalty function J p : According to the orbital information, the sequential quadratic programming algorithm is used to solve the optimization problem and obtain the minimum burnup △v required for the imaging star to re-image. pmin , and calculate the cost-effectiveness ratio η: T s , J p The fitness function is calculated by weighted summation of and η: F=ω1η+ω2T s +ω3J p Where F is the fitness function value, ω1 and ω2 are weight coefficients, and ω3 is the weight coefficient of the penalty term. The optimization problem of the pulse maneuver optical game avoidance strategy of avoiding the star is modeled as follows: The genetic algorithm is used to solve the above optimization problem and obtain the optimization variable y=[t w ,dv,α,β], i.e., obtaining a single-step predicted avoidance trajectory; The second step specifically includes: The orbital information is used as input and the sequential quadratic programming algorithm is used to solve the optimization problem and make a single-step prediction for the re-imaging strategy of the imaging star: The imaging star maneuver mode is set as follows: after the avoidance star maneuver, it will fly over the center of the optical imaging effective range of the avoidance star with a Lambert maneuver and re-perform optical imaging. tran To optimize the variables, the imaging star is obtained from the orbit information at t w Time status X p (t w )=[r p (t w ),v p (t w )] and the avoidance star in (t w +t tran ) state X at the moment e (t w +t tran )=[r e (t w +t tran ),v e (t w +t tran )], where X is the state vector; r and v are position and velocity vectors, respectively; subscripts e and p refer to the avoidance star and imaging star, respectively; Calculated in (t w +t tran ) unit sunlight vector r s , obtain (t w +t tran ) to avoid the center position vector r of the star optical imaging c =r e (t w +t tran )-(r l +r u )·r s / 2, where r c To avoid the center position vector of star optical imaging; Solve the computational Lambert problem: V1=Lambert(r e (t w ),r c ,t tran ) Where, V1 is the imaging star at (t+t tran ) is the desired speed, Lambert(·) is the Lambert calculation function; The calculated pulse maneuver speed increment vector is: △v p =V1-v p (t w ) Calculate its speed increment consumption △v p =||△v p ||, with △v p In order to minimize the performance index, the sequential quadratic programming method is used to optimize and obtain the predicted minimum fuel consumption △v of the imaging satellite. pmin , returning the minimum fuel consumption of the imaging star to step 1.

Images