A multi-platform space target observation collaborative scheduling method and terminal equipment

Abstract

The present invention discloses a multi-platform space target observation collaborative scheduling method and terminal equipment, the method includes determining the observation target and the observation platform to be used, the observation platform includes at least two of the ground-based fixed observation platform, the ground-based mobile observation platform and the space-based observation platform; obtaining a set of available observation windows of the observation platform for the observation target; using a weighted multi-objective optimization algorithm to filter out observation windows that meet the conditions from the set of available observation windows, and scheduling them in sequence. Aiming at the practical problems of platform diversity, single observation uncertainty, task hierarchy priority, etc. of space target observation, the present invention uniformly considers space-based, ground-based fixed, and ground-based mobile observation platforms from the perspective of multi-platform task collaborative allocation, establishes a heterogeneous platform pre-scheduling model for space target emergency observation, and verifies it based on STK and improved NSGA‑Ⅱ algorithms, providing a useful reference for the establishment of a space-ground integrated space target observation system.

Description

Technical Field

[0001] The present application relates to a multi-platform space target observation collaborative scheduling method and terminal equipment, belonging to the field of space target observation technology. Background Art

[0002] At present, space target observation platforms are divided into three categories: one is the observation station deployed at a fixed point, the second is the observation platform that can be deployed flexibly at a fixed observation position, and the third is the space-based observation platform deployed in orbit. The characteristics of space target observation tasks are: first, the randomness of the task, the source of emergency task requirements is random, such as when a satellite has an abnormality, it is necessary to quickly identify the cause of the failure; second, the time and space constraints, there is a limited visible time window between the observation platform and the target; third, the logical constraints, the observation tasks generally have a relatively fixed logical order, space target search, coarse tracking, precise tracking, target identification, photography, multi-source data fusion; fourth, the timeliness of completion, according to the dynamic changes of the task, the newly generated or newly arrived tasks have obvious timeliness and need to be completed within a limited time, requiring a convenient and fast algorithm to realize dynamic task scheduling; fifth, the environmental constraints, which are greatly constrained by the weather, such as cloud cover, not allowing the shooting of visible light images, etc., target brightness and other characteristics.

[0003] The target requirements for space target observation systems put forward by foreign countries are: to produce the most accurate products, provide users with timely technology, and give owners maximum flexibility in every operation link of the space stage. To achieve this goal, it is necessary to develop a space target observation system composed of space-based, ground-based, and ground-based mobile platforms. The optical ground-based measurement system fixedly distributed on the surface of the earth has a limited coverage range and is restricted by weather, the rotation of the earth and the revolution period. For example, for targets in near-circular orbits at an altitude of 400 to 2000 km, the average angular velocity is 11.3 to 15.6 rev / day. A single ground-based observation platform cannot observe the target in time and needs to be supplemented by other observation platforms. The space-based observation system can effectively make up for the shortcomings of the ground-based observation network, and the ground-based mobile system is also conducive to improving the flexibility of space target observation. The coordinated observation of space targets by space and ground has become a development trend.

[0004] At present, there are many studies on collaborative scheduling such as multi-satellite earth observation and satellite-ground integrated earth observation. In terms of collaborative scheduling structure, a variety of scheduling architectures have been proposed, such as agent-based centralized-distributed architecture, hierarchical task network architecture, and limited centralized satellite-ground integrated collaborative structure. In terms of dynamic scheduling, Du Yonghao established a unified constraint satisfaction model for satellite task scheduling and gave a variety of collaborative solution frameworks; Zhang Ming considered the influence of dynamic uncertainty factors such as satellite resource failure and emergency task joining, and constructed a reactive scheduling multi-objective optimization model; Peng Shuang studied the problem of satellite dynamic scheduling, and proposed a heuristic satellite action sequence optimization adjustment method based on the characteristics of emergency measurement and control and on-board emergency action adjustment timing. In terms of scheduling algorithms, Huang Xiaojun proposed a hybrid scheduling algorithm combining multi-objective evolutionary algorithm with variable neighborhood search to solve the hybrid scheduling problem of joint reconnaissance of electronic reconnaissance satellites; Wu Haiyan proposed a rule-based heuristic solution algorithm for the scheduling problem of astronomical satellite opportunity target tasks; Yan Qingqing studied the rescheduling problem of ground-based monitoring of space targets based on ant colony algorithm. The above studies mostly focus on the scheduling problems of homogeneous platforms and the rescheduling problems with uncertainties such as task failure and the addition of new tasks, but do not fully consider the task execution success rate in the pre-scheduling scheme. Summary of the invention

[0005] The purpose of this application is to provide a multi-platform space target observation collaborative scheduling method to solve the technical problems raised in the above background technology.

[0006] The first embodiment of the present invention provides a multi-platform space target observation coordinated scheduling method, comprising:

[0007] Step 1: Determine the observation target and the observation platform to be used, wherein the observation platform includes at least two of a ground-based fixed observation platform, a ground-based mobile observation platform and a space-based observation platform;

[0008] Step 2: obtaining a set of available observation windows of the observation platform for the observation target;

[0009] Step 3: Filter out observation windows that meet the conditions from the set of available observation windows using a weighted multi-objective optimization algorithm, and schedule them in sequence.

[0010] Preferably, the weighted multi-objective optimization algorithm in step 3 includes an objective function, and the objective function includes: a task completion benefit function, a task completion concentration function, a ground-based movement total consumption function and a platform average load function.

[0011] Preferably, the task completion benefit function is determined according to a first formula, which is:

[0012]

[0013] Where N C is the number of observation tasks successfully scheduled, pr k is the observation task priority, φ j is the scheduling task success rate of observation target j, x i is the decision variable, w i is the i-th observation window, C(x i w i ) is consumed by the i-th observation window; among them, pr k ,φ j and x i They are:

[0014]

[0015]

[0016]

[0017] In the formula, pr i is the priority of observation platform i, pr j is the priority of the observed target j, pr1 is the ground-based fixed observation platform, pr2 is the ground-based mobile observation platform, pr3 is the space-based observation platform, α1, α2, α3, α4 represent the priority coefficients respectively, and α1+α2+α3+α4=1, β j represents the user-defined importance, and β j ∈[1,10], t woji represents the duration of the oth observation of observation target j by observation platform i, T is the duration of the entire scheduling interval, count(w j ) is the number of all observation windows for observing target j, m is the number of observation platforms, n is the number of observed targets, q ij is the number of observation windows of observation platform i for observation target j,

[0018] p k It represents the observation success rate of observation platform i for observation target j corresponding to the kth observation window.

[0019] Preferably, the task completion concentration function is determined according to a second formula, which is:

[0020]

[0021] In the formula, ts wi+1 is the start time of the i+1th observation window, te wi is the end time of the ith observation window, ts wk is the start time of the kth window to be scheduled.

[0022] Preferably, the total consumption function of foundation movement is determined according to a third formula, which is:

[0023]

[0024] In the formula, m1, m2, and m3 are the number of space-based observation platforms, ground-based fixed observation platforms, and ground-based mobile observation platforms, respectively. is the kth observation window corresponding to the ground-based mobile platform i, The k+1th observation window corresponding to the ground-based mobile platform i, is the distance between the geographical coordinates of the kth and k+1th observation windows corresponding to the ground-based mobile platform i, x i is the decision variable, x i Specifically:

[0025]

[0026] Preferably, the platform average load function is determined according to a fourth formula, which is:

[0027]

[0028] Where i is the number of the observation platform, m is the total number of the observation platforms, and N i is the number of observation tasks of the observation platform i, each of which contains one observation target, q i is the total number of observation windows of the observation platform i.

[0029] Preferably, the weighted multi-objective optimization algorithm includes a constraint function, the constraint function includes an observation task uniqueness constraint function, the observation task uniqueness constraint function is determined according to a fifth formula, and the fifth formula is:

[0030]

[0031] In the formula, x ijk The decision variable for the observation platform i to perform observation on the kth observation window of the observation target j.

[0032] Preferably, the constraint function further includes an observation platform logic constraint function, and the observation platform logic constraint function is determined according to a sixth formula, and the sixth formula is:

[0033]

[0034] In the formula, te ijk is the end time of the kth observation window of the observation platform i for the observation target j, tp iis the preparation time of the observation platform i, ts ijk′ is the start time of the k′th observation window of the observation platform i for the observation target j, d i,i′ is the distance between the observation station corresponding to the ground-based mobile observation platform i and the next station corresponding to the platform, v is the maneuvering speed of the ground-based mobile observation platform, m1 is the number of space-based observation platforms, and m2 is the number of ground-based fixed observation platforms.

[0035] Preferably, the constraint function further includes an observation platform capability constraint function, and the observation platform capability constraint function is determined according to a seventh formula, which is:

[0036]

[0037] In the formula, Em(s) i is the energy consumption percentage of a single observation by space-based observation platform i, E i is the maximum available energy of a single orbit of the space-based observation platform i, and M num_i is the maximum number of power on and off times in a single day;.

[0038] Preferably, the weighted multi-objective optimization algorithm includes a fitness function, and the fitness function is determined according to the eighth formula, which is:

[0039] F=μ1M1+μ2M2+μ3M3+μ4M4

[0040] Wherein, μ1~μ4 are weights obtained according to the uniform design table, M1 is the task completion benefit function, M2 is the task completion concentration function, M3 is the total consumption function of ground-based movement, and M4 is the platform average load function.

[0041] A second embodiment of the present invention provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above method when executing the computer program.

[0042] Compared with the prior art, the multi-platform space target observation collaborative scheduling method and terminal device of the present invention have the following beneficial effects:

[0043] Aiming at the practical problems of platform diversity, single observation uncertainty, task hierarchy priority and so on in space target observation, the present invention uniformly considers space-based, ground-based fixed and ground-based mobile observation platforms from the perspective of multi-platform task collaborative allocation, and establishes a heterogeneous platform pre-scheduling model for space target emergency observation. It is verified based on STK and improved NSGA-Ⅱ algorithm, which provides a useful reference for building a space-ground integrated space target observation system. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0045] Figure 1 This is a flow chart of the multi-platform space target observation collaborative scheduling method of the present invention;

[0046] Figure 2 It is a mapping relationship diagram of "task-target-platform-window" of the present invention;

[0047] Figure 3 It is a mapping relationship diagram of "task-target-station-platform-window" of the present invention;

[0048] Figure 4 It is a schematic diagram of the observation window conflict of the present invention;

[0049] Figure 5 It is a schematic diagram of the observation task flow of the present invention;

[0050] Figure 6 is a schematic diagram of chromosome coding of the present invention;

[0051] Figure 7 It is a STK simulation schematic diagram of the O1 observation target of the present invention;

[0052] Figure 8 It is a result schematic diagram of a specific embodiment of the present invention;

[0053] Fig. 9 It is a convergence curve diagram of the objective function of the improved NSGA-II algorithm in a specific embodiment of the present invention;

[0054] Fig.10 It is a convergence curve diagram of the average fitness function of the improved NSGA-II algorithm in a specific embodiment of the present invention. DETAILED DESCRIPTION

[0055] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present invention. However, it should be clear to those skilled in the art that the present invention may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present invention.

[0056] In order to illustrate the technical solution of the present invention, an embodiment is provided below for illustration.

[0057] Figure 1 Flowchart of the multi-platform space target observation collaborative scheduling method of the present invention.

[0058] The multi-platform space target observation coordinated scheduling method of the present invention comprises:

[0059] Step 1, determining the observation target and the observation platform to be used, where the observation platform includes at least two of a ground-based fixed observation platform, a ground-based mobile observation platform and a space-based observation platform;

[0060] Step 2: Obtain a set of available observation windows of the observation platform for the observation target;

[0061] Step 3: Use a weighted multi-objective optimization algorithm to select observation windows that meet the conditions from the set of available observation windows and schedule them in sequence;

[0062] The weighted multi-objective optimization algorithm in step 3 includes objective functions and constraint functions. The objective functions include: task completion benefit function, task completion concentration function, total ground-based mobile consumption function and platform average load function; the constraint functions include observation task uniqueness constraint function, observation platform logic constraint function and observation platform capacity constraint function.

[0063] The task completion benefit function in the objective function is determined according to the first formula, which is:

[0064]

[0065] Where N C is the number of observation tasks successfully scheduled, pr k is the observation task priority, φ j is the scheduling task success rate of observation target j, x i is the decision variable, w i is the i-th observation window, C(x i w i ) is the cost consumed by the i-th observation window; where pr k ,φ j and x i They are:

[0066]

[0067]

[0068]

[0069] In the formula, pr i is the priority of observation platform i, pr jis the priority of the observed target j, pr1 is the ground-based fixed observation platform, pr2 is the ground-based mobile observation platform, pr3 is the space-based observation platform, α1, α2, α3, α4 represent the priority coefficients respectively, and α1+α2+α3+α4=1, β j represents the user-defined importance, and β j ∈[1,10], t woji represents the duration of the oth observation of observation platform i on observation target j, T is the duration of the entire scheduling interval, count(w j ) is the number of all observation windows for observing target j, m is the number of observation platforms, n is the number of observed targets, q ij is the number of observation windows of observation platform i for observation target j,

[0070] p k It represents the observation success rate of observation platform i for observation target j corresponding to the kth observation window.

[0071] The task completion concentration function in the objective function is determined according to the second formula, which is:

[0072]

[0073] In the formula, ts wi+1 For the i+1 The start time of an observation window, te wi is the end time of the ith observation window, ts wk is the start time of the kth window to be scheduled.

[0074] The total consumption function of the ground movement in the objective function is determined according to the third formula, which is:

[0075]

[0076] In the formula, m1, m2, and m3 are the number of space-based observation platforms, ground-based fixed observation platforms, and ground-based mobile observation platforms, respectively. is the kth observation window corresponding to the ground-based mobile platform i, The k+1th observation window corresponding to the ground-based mobile platform i, is the distance between the geographical coordinates of the kth and k+1th observation windows corresponding to the ground-based mobile platform i, x i is the decision variable, x i Specifically:

[0077]

[0078] The platform average load function in the objective function is determined according to the fourth formula, which is:

[0079]

[0080] Where i is the number of the observation platform, m is the total number of the observation platforms, and N i is the number of observation tasks of the observation platform i, each of which contains one observation target, q i is the total number of observation windows of the observation platform i.

[0081] The observation task uniqueness constraint function in the constraint function is determined according to the fifth formula, and the fifth formula is:

[0082]

[0083] In the formula, x ijk The decision variable for the observation platform i to perform observation on the kth observation window of the observation target j.

[0084] The observation platform logic constraint function in the constraint function is determined according to the sixth formula:

[0085]

[0086] In the formula, te ijk is the end time of the kth observation window of the observation platform i for the observation target j, tp i is the preparation time of the observation platform i, ts ijk′ is the start time of the k′th observation window of the observation platform i for the observation target j, d i,i′ is the distance between the observation station corresponding to the ground-based mobile observation platform i and the next station corresponding to the platform, v is the maneuvering speed of the ground-based mobile observation platform, m1 is the number of space-based observation platforms, and m2 is the number of ground-based fixed observation platforms.

[0087] The observation platform capability constraint function in the constraint function is determined according to the seventh formula, and the seventh formula is:

[0088]

[0089] In the formula, Em(s) i is the energy consumption percentage of a single observation by space-based observation platform i, E i is the maximum available energy of a single orbit of space-based observation platform i, and M num_i is the maximum number of power on and off times in a single day.

[0090] The weighted multi-objective optimization algorithm of the present application also includes a fitness function, which is determined according to the eighth formula:

[0091] F=μ1M1+μ2M2+μ3M3+μ4M4

[0092] Wherein, μ1~μ4 are weights obtained according to the uniform design table, M1 is the task completion benefit function, M2 is the task completion concentration function, M3 is the total consumption function of ground-based movement, and M4 is the platform average load function.

[0093] The present application also discloses a terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the steps of the above method are implemented when the processor executes the computer program.

[0094] The multi-platform space target observation collaborative scheduling method of the present invention will be described in detail below.

[0095] 1 Problem Description

[0096] The problem studied in this application is that there are several known space-based observation platforms deployed in known orbits; several ground-based fixed observation platforms with known deployment positions; several mobile observation platforms with known initial positions that can be maneuvered to several observation platforms’ fixed observation positions as needed; and several space-based targets with known existing orbital elements that require emergency collaborative observation. Based on the above known conditions, a collaborative observation strategy of single tasks "observation-assessment-observation", "ground-based fixed observation platform priority-ground-based mobile observation platform-space-based observation platform" is adopted to form a pre-allocation plan for observation tasks. First, the following assumptions are made for this problem:

[0097] The location of the ground-based position information is expressed in terms of longitude and latitude on the surface of the earth. i ,latitude i ,altitude i );

[0098] The orbital elements of the observed target are determined, and the orbital parameters are expressed by six elements, namely (a, e, i, Ω, ω, τ);

[0099] The time it takes for the observation platform beam to reach the satellite is short, and the delay is negligible;

[0100] The space-based observation platform does not consider the maneuverability of the space-based observation platform. It only uses the strategy of close observation within the inherent observation range of the payload near the intersection point by adjusting the platform attitude and camera angle. It is also assumed that command injection, data feedback, and effect evaluation can be completed in real time through relay satellites at all times.

[0101] Once an observation mission has started, it is not allowed to be interrupted;

[0102] A ground-based observation position can only be assigned to one mobile observation platform at a time;

[0103] The success rate of observation of a target by each type of platform is the theoretical probability value.

[0104] 2 Model establishment

[0105] 2.1 Definitions of symbols, parameters and related concepts

[0106] For the sake of uniformity, the characters and parameters appearing in this application are defined as shown in Table 1.

[0107] Table 1 Related symbols, subscripts and parameter definitions

[0108]

[0109] Before building the model, we first give the relevant definitions:

[0110] Definition 1: Emergency Observation Window

[0111] According to the existing window calculation formula, under the premise of known observation targets and observation platforms, the set of available observation windows for each observation platform for each observation target can be calculated (windows that do not meet the conditions such as observation angle, observation time, and observation distance have been eliminated). An observation task contains only one observation target, which corresponds to several observation platforms and several observation windows, such as Figure 2 and Figure 3 shown. Figure 2 The mapping relationship is task-target-platform-window. Figure 3 The mapping relationship is “task-target-station-platform-window”.

[0112] Based on the mapping relationship of “task-target-(station)-platform-window”, the above observation tasks can be mapped to the observation platform, which is expressed as T u '=<id u ,P i ,O j ,WT u ,pr j ,p ij >, id u is the task identifier, P i is the corresponding observation platform or station, O j is the corresponding observation target, pr j is the observation task priority, p ij is the observation success rate, w ijk ,ts ijk 、te ijkRespectively represent the corresponding observation window, the earliest start time and the latest end time. The observation window contains the corresponding observation target, platform, station, number of times and other information. The observation window time series set can be obtained by arranging them in order of the earliest start time, that is,

[0113] Definition 2: Scheduling task success rate

[0114] During pre-scheduling, the scheduling of observation tasks has a certain probability, which is independent of the number of observation cycles and is denoted as P k , represents the observation success rate of platform i corresponding to the kth window for the observation target j. In order to ensure the observation success rate, multiple scheduling is required until the user's needs are met. The pre-scheduling number of the jth observation target can be calculated by formula (1), and the user sets an observation threshold φ j (such as 98%), rounded up.

[0115]

[0116] Definition 3: Observation task conflict degree

[0117] From the perspective of observation targets, conflict refers to the situation where the same target passes through the observation range of multiple observation platforms at the same time, resulting in overlapping observation periods. Figure 4 As shown in the figure, during the time period from t1 to t7, there are 4 visible windows for observing target j, and the observation periods overlap, and the number of window conflicts are 3, 3, 3, and 2 respectively.

[0118] The ratio of the number of observation windows with conflicts to the total number of observation windows during the total mission is denoted as the conflict degree ψ j , obtained by formula (2).

[0119]

[0120] Definition 4: Observation Task Priority

[0121] It is determined by the priority of the observation target and the observation platform. The main factors considered in the target priority are: ① total observable time; ② number of time windows; ③ task conflict (observation time overlap); ④ user-defined importance. The platform priority needs to comprehensively consider factors such as the success rate of completing the task, the total platform consumption, and the platform performance. The ground-based fixed observation platform, ground-based mobile observation platform, and space-based observation platform are 0.9, 0.8, and 0.7 respectively. Therefore, the total observation task priority is a known quantity, which can be obtained according to formula (3):

[0122]

[0123] In the formula, α1, α2, α3, and α4 represent priority coefficients respectively, and α1+α2+α3+α4=1, βj represents the user-defined importance, and β j ∈[1,10].

[0124] 2.2 Decision variable definition

[0125] Define the allocation matrix A = [x i ] 1×K , the decision variable x i Expressed as a Boolean value, that is, x i ∈{0,1}, 1 means that platform i observes target j in the kth observation window, and 0 means no observation.

[0126]

[0127] 2.3 Objective Function

[0128] Based on the above definition, a model is established. This optimization scheduling problem is a multi-objective optimization problem, that is, an initial scheduling plan is formed, which comprehensively considers the total revenue of the observation task, the total time to complete the task, the total maneuvering distance of the ground-based mobile observation platform, and the average load value of the platform. The uniform design concept is used, and the comprehensive weighted multi-objective optimization algorithm is adopted. Table 2 is the objective function table used by the comprehensive weighted multi-objective optimization algorithm of this application.

[0129] Table 2 Objective function table

[0130]

[0131] 2.4 Constraints

[0132] The constraint function used in the comprehensive weighted multi-objective optimization algorithm of this application is specifically:

[0133] When calculating the observation window, factors such as observation platform performance, azimuth angle, and observation feasibility are taken into account, and some invalid observation windows have been eliminated. Therefore, only constraints such as observation task allocation, observation platform preparation, and rationality of rescheduling time are considered here.

[0134] (1) Observation task uniqueness constraint

[0135] In order to ensure the success rate of a single task, this paper assumes that each platform can only observe one target at the same time, as shown in formula (6).

[0136]

[0137] (2) Logical constraints of the observation platform

[0138] Each type of observation platform has a complete set of processes within an observation cycle, such as Figure 5As shown. It is usually represented by a time constraint network diagram. The focus of this paper is on the scheduling of collaborative observation tasks on heterogeneous platforms. Therefore, the observation process is simplified, ignoring constraints such as target positioning, space-based platform task upload and data download, ground-based observation equipment heating and cooling time, and effect evaluation time. The logical constraints of the observation platform are specifically shown in formula (7)

[0139]

[0140] (3) Observation platform capability constraints

[0141] Due to the constraints of space target observation technology, the observation platform cannot work continuously and consumes a certain amount of energy during the observation process. The maximum observation time of a single circle of the space-based observation platform is limited. In order to simplify the model, the capacity constraint is expressed by the percentage of energy consumption per observation and the maximum number of power on and off times per day, as shown in formula (8).

[0142]

[0143] 3 Pre-scheduling model solution

[0144] From the above model, we can see that this scheduling is a multi-objective optimization problem, similar to the backpack salesman problem (TSP problem), which is an NP-Hard problem. Traditional linear programming is difficult to deal with nonlinear benefits and nonlinear constraints. It can only solve the Pareto effective (optimal) solution non-inferior solution set by constructing a heuristic (intelligent optimization) algorithm. At present, intelligent optimization algorithms have been maturely applied to multi-objective optimization problems, including evolutionary algorithms, particle swarm algorithms, NSGA-II, etc. This paper realizes the pre-allocation of the above problem by improving the NSGA-II method.

[0145] 3.1 Coding Design

[0146] Since there is a one-to-one mapping relationship between the observation window and the observation target, it is only necessary to assign the observation platform to the observation window. Therefore, the chromosome adopts a double-layer encoding, which is formed according to the time sequence of the observation window. To define a single chromosome, each gene bit on the first layer is encoded in binary to represent the decision variable, 1 means execution, and 0 means no execution. The corresponding gene bit on the second layer is encoded in decimal to represent the observation platform number, such as Figure 6 The total length of the code corresponds to the total number of observation windows. Assume there are 3 observation platforms and 1 observation target. At a given time t total There are 7 observation windows. After arranging by time, windows numbered 1, 4, 5, 6, and 7 are executed by platforms 2, 3, 1, 2, and 1 respectively, and other windows are not executed.

[0147] In the decoding design, the observation window numbers with genetic code "0" are eliminated, and the "platform-target" allocation scheme arranged in time sequence is solved according to the one-to-one mapping relationship between each observation window and the observation platform (station) and observation target.

[0148] 3.2 Fitness function construction

[0149] In order to avoid the explosion of solution space and ensure the global distribution balance of initial solution, the uniform design idea is adopted to improve the fitness function. The number of factors selects the number of objective functions a, the number of levels selects the search direction, that is, the number of fitness functions b, and the uniform design table U is queried. b (b a ), and then assign the weights of each objective function according to formula (10).

[0150]

[0151] Taking the pre-scheduling stage as an example, the objective function is taken from (5) to (9), the number of factors (the number of objective functions a) is 4, and the number of experimental schemes (the number of fitness functions b) is 7, that is, U7(7 4 ), and the weights of each factor are shown in Table 3.

[0152] Table 3 U7(7 4 ) and the weights of each factor

[0153] Column number <![CDATA[U7(7 4 )]]> <![CDATA[μ1]]> <![CDATA[μ2]]> <![CDATA[μ3]]> <![CDATA[μ4]]> 1 (1,3,5,7) 0.063 0.188 0.313 0.438 2 (2,6,2,6) 0.125 0.375 0.125 0.375 3 (3,1,7,5) 0.188 0.063 0.438 0.313 4 (4,4,4,4) 0.250 0.250 0.250 0.250 5 (5,7,1,3) 0.313 0.438 0.063 0.188 6 (6,2,6,2) 0.375 0.125 0.375 0.125 7 (7,5,3,1) 0.438 0.313 0.188 0.063

[0154] Therefore, the fitness function can be obtained as shown in formula (10):

[0155]

[0156] Secondly, in order to unify the standard, the objective function value is dimensionless processed by the maximum value method of formula (11) and converted into the interval [0,1].

[0157]

[0158] In the formula, i represents the number of objective functions, j represents the number of optional scheduling schemes, They represent the maximum and minimum values ​​of the objective function i in solution j respectively.

[0159] Next, the present application is described in detail with more specific embodiments.

[0160] In order to verify the effectiveness of the proposed pre-scheduling method for emergency collaborative observation of space targets, a scenario with 4 observation targets, 10 observation platforms (3 space-based observation platforms, 4 ground-based fixed observation platforms, 4 ground-based mobile stations, and 2 mobile observation platforms) and a scheduling interval of 12 hours was designed (simulation time 22Jun 2020 08:00:00UTGG~23Jun2020 08:00:00UTGG). The relevant parameters were set as α1=α2=α3=α4=0.25, λ1=λ2=0.5, platform usage priorities pr1=0.9, pr2=0.8, pr3=0.7, and operating cost C S =1, C GMP =0.5, C G =0.2, and the initial settings of other parameters are shown in Tables 4 and 5. The visible observation windows are calculated by STK software. Under the initial conditions, there are 304 observation windows, and some of the windows are shown in Table 6. Based on the improved NSGA-II scheduling algorithm, the scheduling algorithm is simulated in Matlab 2016b software on a laptop computer with Intel(R) Core I5-4200U CPU@2.30GHz, 8G memory, and Windows 10 Professional 64-bit operating system.

[0161] Table 4 Initial parameters of space-based observation platform and observation targets

[0162]

[0163] Table 5 Parameter settings for ground-based fixed observation platform / station position and ground-based mobile platform initial position

[0164]

[0165] Table 6 O1 observation window table (partial)

[0166]

[0167] O1 observation window STK simulation see Figure 7 .

[0168] The initial conditions are set as follows: the population size is 200 randomly generated, the selection probability is 0.5, the crossover probability is 0.7, the single point mutation probability is 0.01, and the iterations are 100 times and repeated 50 times. The experimental results are shown in Figure 2. Figures 8 to 10 And as shown in Table 7.

[0169] Table 7 Final scheduling plan

[0170]

[0171] from Figures 8 to 10It can be seen that the improved NSGA-II algorithm can converge quickly after an average of 35 iterations, the time is 120.45 seconds, the objective function value is 8.408, and the average value of the fitness function basically converges to about 278.5. From Table 7, it can be seen that the number of times the space-based platform is used is relatively large. This is because there are more windows corresponding to the space-based observation platform, accounting for 40.3% (134) of the total number of windows. However, considering the scarcity of space-based platform resources, the platform usage priority is set low, and only 5 are selected; the ground-based mobile platform has a large uncertainty factor, and the corresponding restriction parameters are set in the experiment. The utilization rate is also relatively low, and GM1 and GM2 each move once. From Table 8, it can be seen that the interval between the use of each platform is long, and it has the ability to cope with uncertain risks and the ability to respond to rescheduling, which shows that the scheme has good robustness and balanced benefits. In general, these situations are basically in line with reality.

[0172] Table 8 Concentration of usage time, interval time and total time of each platform

[0173]

[0174] Aiming at the practical problems of platform diversity, single observation uncertainty, task hierarchy priority and so on in space target observation, the present invention uniformly considers space-based, ground-based fixed and ground-based mobile observation platforms from the perspective of multi-platform task collaborative allocation, and establishes a heterogeneous platform pre-scheduling model for space target emergency observation. It is verified based on STK and improved NSGA-Ⅱ algorithm, which provides a useful reference for building a space-ground integrated space target observation system.

[0175] The above are only a few embodiments of the present application and do not constitute any form of limitation to the present application. Although the present application is disclosed as above with preferred embodiments, it is not intended to limit the present application. Any technician familiar with the profession, without departing from the scope of the technical solution of the present application, using the technical content disclosed above to make slight changes or modifications are equivalent to equivalent implementation cases and fall within the scope of the technical solution.

Claims

1. A multi-platform space target observation collaborative scheduling method, characterized in that: include: Determining the observation target and the observation platform to be used, wherein the observation platform includes at least two of a ground-based fixed observation platform, a ground-based mobile observation platform and a space-based observation platform; Acquire a set of available observation windows of the observation platform for the observation target; Using a weighted multi-objective optimization algorithm, select observation windows that meet the conditions from the set of available observation windows, and schedule them in sequence; The weighted multi-objective optimization algorithm includes an objective function, which includes: a task completion benefit function, a task completion concentration function, a ground-based mobile total consumption function, and a platform average load function; The task completion benefit function is determined according to a first formula, which is: In the formula, is the number of observation tasks successfully scheduled, is the observation task priority, For observation target The success rate of scheduling tasks, is the decision variable, For the An observation window, For the The cost consumed by an observation window; among them, , and They are: ； ； ； In the formula, For observation platform The priority of For observation target The priority of It is a ground-based fixed observation platform. It is a ground-based mobile observation platform. It is a space-based observation platform. , , , represent the priority coefficients respectively, and , represents the user-defined importance, and , Indicates the observation platform For observation target No. Observation duration, is the duration of the entire scheduling interval, For observation target The number of all observation windows of is the number of observation platforms, is the number of observed targets, For observation platform For observation target The number of observation windows, , Indicates Observation platform corresponding to observation window For the observation target The success rate of observation.

2. The multi-platform space target observation coordinated scheduling method according to claim 1 is characterized in that: The task completion concentration function is determined according to a second formula, which is: In the formula, For the The start time of an observation window, For the The end time of the observation window, For the The start time of the window to be scheduled.

3. The multi-platform space target observation coordinated scheduling method according to claim 2 is characterized in that: The total consumption function of the foundation movement is determined according to the third formula, which is: In the formula, , , They are the number of space-based observation platforms, ground-based fixed observation platforms, and ground-based mobile observation platforms. Ground-based mobile platform The corresponding An observation window, Ground-based mobile platform The corresponding An observation window, Ground-based mobile platform The corresponding and The distance between the geographical coordinates corresponding to the observation windows, is the decision variable, Specifically: 。 4. The multi-platform space target observation coordinated scheduling method according to claim 3 is characterized in that: The platform average load function is determined according to a fourth formula, which is: In the formula, is the number of the observation platform, is the total number of observation platforms, For the observation platform The number of observation tasks, each of which contains one observation target, For the observation platform The total number of observation windows.

5. The multi-platform space target observation coordinated scheduling method according to any one of claims 1 to 4, characterized in that: The weighted multi-objective optimization algorithm includes a constraint function, and the constraint function includes an observation task uniqueness constraint function, and the observation task uniqueness constraint function is determined according to the fifth formula, and the fifth formula is: In the formula, For the observation platform For the observation target No. The decision variables for executing the observation in the observation window.

6. The multi-platform space target observation coordinated scheduling method according to claim 5 is characterized in that: The constraint function also includes an observation platform logic constraint function, which is determined according to a sixth formula: In the formula, For the observation platform For the observation target No. The end time of the observation window. For the observation platform Preparation time, For the observation platform For the observation target No. The start time of the observation window, For the ground-based mobile observation platform The distance between the corresponding observation station and the next station corresponding to the platform, is the maneuvering speed of the ground-based mobile observation platform, is the number of space-based observation platforms, is the number of ground-based fixed observation platforms.

7. The multi-platform space target observation coordinated scheduling method according to claim 5 is characterized in that: The constraint function also includes an observation platform capability constraint function, and the observation platform capability constraint function is determined according to the seventh formula, and the seventh formula is: In the formula, Space-based observation platform Percentage of energy consumption per single observation, Space-based observation platform Maximum available energy for a single track turn, The maximum number of power on and off times per day.

8. The multi-platform space target observation coordinated scheduling method according to claim 5 is characterized in that: The weighted multi-objective optimization algorithm also includes a fitness function, which is determined according to the eighth formula: In the formula, are the weights obtained from the uniform design table, is the task completion reward function, is the task completion concentration function, is the total consumption function of foundation movement, is the platform average load function.

9. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

Images