Spatial moving target collaborative monitoring resource scheduling method based on multiple agents

Abstract

The invention discloses a multi-agent-based space moving target cooperative monitoring resource scheduling method, which establishes a problem scheduling mathematical model aiming at a space moving target tracking-oriented satellite-ground resource integrated scheduling problem, designs a satellite-ground cooperative multi-agent framework based on a multi-agent system, and provides a multi-agent-based space moving target cooperative monitoring resource scheduling method. The invention provides a satellite-ground cooperative scheduling strategy for space moving target tracking.

Description

Technical Field

[0001] The present invention belongs to the field of satellite communication resource scheduling, and in particular relates to a multi-agent-based space mobile target collaborative monitoring resource scheduling method. Background Art

[0002] Space mobile targets refer to the main objects active in the space environment, including space debris, spacecraft in orbit, ballistic missiles, etc. Space mobile targets are a typical type of time-sensitive targets, which move at high speeds in space and are fleeting. At the same time, such targets often have extremely high tactical or strategic value.

[0003] Surveillance and tracking of moving targets in space is a fundamental capability for acquiring space situational awareness. However, because moving targets often move over long time spans and over large spatial distances, continuous tracking is difficult to achieve with a single device. Currently, tracking moving targets in space primarily relies on ground-based radar and space satellites. However, both methods have limitations and cannot independently meet the increasingly demanding needs of observing moving targets in space. Ground-based radars are not limited by capacity or power and can track multiple targets simultaneously, but their position is fixed and their tracking effectiveness is affected by weather. Space satellites primarily include high- and medium-orbit satellites. The former have a wider swath, observing against the Earth's background above geostationary orbit, requiring targets to be at a relatively high temperature for detection. Therefore, they can only detect active targets, such as approaching hypersonic vehicles and unfired ballistic missiles. When a target enters the coverage area, it can be automatically monitored, so there is no need for active tracking by high-orbit satellites; the latter has a smaller width of only 1° to 2° and needs to rely on its own rapid attitude maneuvering capabilities to observe the target from the side with deep space as the background. Therefore, it can observe passive targets such as space debris, spacecraft, and turned-off ballistic missiles. In order to achieve continuous tracking of the target, medium and low-orbit satellites need to continuously perform attitude maneuvers to place the target at the center of their field of view. Compared with high-orbit satellites, medium and low-orbit satellites have greater maneuverability and more complex coordinated scheduling. The development of a space-ground integrated mobile target tracking system can not only achieve complementary development between the sky and the earth on the basis of existing observation platforms, maximize the use of existing resources, and efficiently complete more complex tracking tasks, but also provide a reference direction for the research and development and coordination of various detection platforms in the future.

[0004] At present, there are few studies on multi-complex task scheduling in the context of large-scale heterogeneous satellite clusters, and the following problems are faced: (1) The study of multi-faceted cooperative goals does not take into account the dynamic uncertainty of space mobile targets. This is mainly reflected in: ① The uncertainty of the arrival of space mobile targets. That is, it is impossible to accurately obtain the area and time when the space mobile target first appears in advance. ② There are errors in the trajectory prediction of space mobile targets, and this error is cumulative. When the prediction error is large, the target is easily lost. (2) The collaborative scheduling of heterogeneous resources is a large-scale and difficult optimization problem, and the efficiency of traditional algorithms needs to be further improved. (3) Since space mobile targets maintain high-speed movement, higher requirements are placed on the real-time performance of collaborative scheduling strategies. Summary of the Invention

[0005] In order to solve the problems of the prior art, the present invention provides a multi-agent-based space mobile target collaborative monitoring resource scheduling method, which realizes satellite-ground collaborative scheduling for space mobile target tracking.

[0006] The technical solution adopted by the present invention to solve the above technical problems is:

[0007] A resource scheduling method for collaborative monitoring of space mobile targets based on multi-agents, the resource scheduling method specifically comprising:

[0008] Step 1: Read the target sequence information to be planned, calculate the dynamic priority of the target, and sort the target sequence according to the dynamic priority of the target;

[0009] Step 2: Traverse the target sequence information and send the target information to be planned to the radar and satellite. The radar and satellite perform parallel planning.

[0010] Step 3: All radars calculate the execution window information for the target and feed back the optimal execution window to the command and control center;

[0011] Step 4: According to the target position and satellite grouping, the corresponding satellite is determined for planning, and the satellite calculates the tracking window for the target;

[0012] Step 5: The satellite determines the optimal satellite tracking method based on the dual-satellite clipping and feeds it back to the command and control center;

[0013] Step 6: The command center selects the optimal planning method and sends the result to the execution satellite or radar to trigger the next target planning.

[0014] Furthermore, before executing the resource scheduling method, a scheduling model is constructed. The scheduling model is a satellite-ground resource integrated model for space mobile target tracking, specifically including: an objective function, ground-based radar tracking constraints, and space satellite tracking constraints;

[0015] 1) Objective function

[0016] The product of tracking time ratio and target importance is used as the optimization target, and the optimization objective function is expressed as:

[0017]

[0018] Where: N is the target number, I j is the importance of target j, is the tracking duration of target j, is the flight duration of target j;

[0019] 2) Ground-based radar tracking constraints

[0020] When the task is assigned to a ground-based radar, the tracking constraints for moving targets in space need to meet the radar's altitude constraint, distance constraint, azimuth constraint, and maximum tracking number constraint.

[0021]

[0022] Where: and is the target position and radar position; is the elevation angle between the target and the radar, which cannot exceed the maximum elevation angle Ele of the radar; The distance between the target and the radar, which cannot exceed the radar detection range; is the azimuth between the target and the radar [Clock min ,Clock max ], which cannot exceed the azimuth interval of the radar detection range; Num(t)≤Num max Indicates that the number of targets tracked by the radar at any time cannot be greater than its maximum tracking capacity;

[0023] 3) Space satellite tracking constraints

[0024] When a task is assigned to a satellite constellation, the task obtained by each satellite participating in the same target tracking is set as a subtask; the subtask status information is:

[0025]

[0026] Where: m is the subtask number, i is the executing satellite number, j is the target number, The subtask start time, The end time of the subtask, is the satellite attitude information at the beginning of the subtask, Satellite attitude information at the end of the subtask, Power consumption for subtasks;

[0027] For a set of subtasks with the same target, two satellites p and q are required to execute them simultaneously, which satisfies the following conditions:

[0028]

[0029] Formula (4) indicates that target tracking requires two different satellites p and q to start and end at the same time, and each mission execution consumes satellite power, and its power consumption increases linearly with the mission time, and the power consumption rate is α;

[0030] For each execution satellite, the following conditions are met:

[0031] (1) The time interval for each satellite to perform its mission must be within the visible time window of the target, [ws m ,we m ] represents the time window in which the satellite of mission m is visible to the target;

[0032]

[0033] (2) The time interval between adjacent missions of each satellite must meet the minimum satellite attitude transition time, Represents the attitude conversion time between adjacent satellite tasks;

[0034]

[0035] (3) The remaining power of the satellite must always be greater than the power threshold E;

[0036]

[0037] Furthermore, the method is implemented based on a multi-agent system, which includes a radar agent, a virtual master radar station agent, a satellite agent, a virtual master satellite agent, and a command and control center agent;

[0038] The radar agent is responsible for persistent monitoring of a fixed area. When a monitored target appears, the radar agent calculates and sends its own monitoring time window, performs window calculation and tracking tasks;

[0039] The virtual master radar station agent is responsible for receiving window information from all radar agents, selecting the optimal window and returning it to the command center agent, executing window calculation requests and selecting tracking modes between radars.

[0040] The satellite agent is responsible for moving according to its own orbit. When a surveillance target appears, the satellite agent calculates and sends its own surveillance time window, and performs window calculation and tracking tasks.

[0041] The virtual master satellite agent is responsible for receiving the window information of the satellite agent, selecting the optimal execution window information and returning it to the command and control center agent, executing the sending window calculation request, dual-satellite window clipping, and satellite tracking mode selection;

[0042] The intelligent agent in the command and control center is responsible for maintaining and updating target information, issuing monitoring requirements, receiving and coordinating window information from radars and satellites, selecting the optimal window as the target tracking window, performing target information maintenance, and sending mission information.

[0043] To address the resource scheduling problem of collaborative monitoring of mobile space targets, this paper considers the uncertainty of mobile space targets and proposes a multi-agent-based collaborative monitoring resource scheduling method. To address the cumulative problem of target prediction errors, a dynamic target priority mechanism is designed to determine the planning order of targets, thereby prioritizing targets with high error. To address the uncertainty of target arrival, two optimal solution selection strategies are designed from the perspectives of global optimization and local optimization. Furthermore, a global satellite grouping strategy is designed to reduce the communication burden between satellites. BRIEF DESCRIPTION OF THE DRAWINGS

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

[0045] Figure 1 It is a flow chart of space moving target monitoring;

[0046] Figure 2 is the agent behavior graph;

[0047] Figure 3 This is a schematic diagram of global regional divisions;

[0048] Figure 4 This is a comparison chart of returns of different strategies;

[0049] Figure 5 It is a graph of system resource utilization under different strategies;

[0050] Figure 6 It is a task planning Gantt chart;

[0051] Figure 7 It is the cumulative number of messages handled by the master star and the message reduction ratio. DETAILED DESCRIPTION

[0052] The following description of exemplary embodiments of the present application is made in conjunction with the accompanying drawings, including various details of the embodiments of the present application to facilitate understanding, which should be considered as merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present application. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0053] like Figure 1 As shown in Figure 2, space mobile target surveillance and tracking is a dynamic decision-making process, which mainly goes through four steps:

[0054] ① Target information fusion: Update the target status and attributes based on prior information and tracking information of detection resources.

[0055] ② Target trajectory prediction: Based on the current state, the target’s trajectory for a period of time is predicted.

[0056] ③ Satellite-ground collaborative planning: Based on the target's predicted trajectory information, intelligent scheduling of constellation and ground-based radar station resources is performed to generate a collaborative tracking plan.

[0057] ④ Target tracking plan execution: Execute the tracking plan and continuously modify the target state attribute information.

[0058] Considering the uncertainty of moving targets in space, this patent proposes a multi-agent-based collaborative monitoring resource scheduling method for moving targets in space. Specifically, to address the problem of accumulated target prediction errors, a dynamic target priority mechanism is designed to determine the planning order of targets, thereby enabling preferential response to targets with high error rates.

[0059] The specific steps are:

[0060] Step 1: Read the target sequence information to be planned, calculate the dynamic priority of the target, and sort the target sequence according to the dynamic priority of the target;

[0061] Step 2: Traverse the target sequence information and send the target information to be planned to the main radar and main satellite. The radar and satellite perform parallel planning.

[0062] Step 3: All radars calculate the execution window information for the target and feed back the optimal execution window to the command and control center;

[0063] Step 4: According to the target position and satellite grouping, the corresponding satellite is determined for planning, and the satellite calculates the tracking window for the target;

[0064] Step 5: The main satellite determines the optimal satellite tracking solution based on the dual-satellite cropping and feeds it back to the command and control center.

[0065] Step 6: The command center selects the optimal plan and sends the result to the execution satellite or radar. After that, the next target plan is triggered.

[0066] Before executing the resource scheduling method, a scheduling model is constructed. The scheduling model is an integrated satellite-ground resource model for space mobile target tracking. Before constructing the multi-complex target integrated scheduling model, some reasonable assumptions are first made:

[0067] 1) Each satellite carries only one sensor, meaning that the same satellite can only track one target at a time;

[0068] 2) Ground-based radars are fixed in position, with fixed observation azimuth and elevation angles;

[0069] 3) Sufficient communication capabilities between satellites and between satellites and the ground to achieve real-time synchronization of data and information;

[0070] 4) The target trajectory prediction error increases linearly with time. When the target error is greater than a given threshold, the target is considered lost.

[0071] Based on the above assumptions, the problem will be modeled below.

[0072] Before executing the resource scheduling method, a scheduling model is constructed. The scheduling model is a satellite-ground resource integrated model for space mobile target tracking, specifically including: an objective function, ground-based radar tracking constraints, and space satellite tracking constraints;

[0073] 1) Objective function

[0074] The product of tracking time ratio and target importance is used as the optimization target, and the optimization objective function is expressed as:

[0075]

[0076] Where: N is the target number, I j is the importance of target j, is the tracking duration of target j, is the flight duration of target j;

[0077] 2) Ground-based radar tracking constraints

[0078] When the task is assigned to a ground-based radar, the tracking constraints for moving targets in space need to meet the radar's altitude constraint, distance constraint, azimuth constraint, and maximum tracking number constraint.

[0079]

[0080] Where: and is the target position and radar position; is the elevation angle between the target and the radar, which cannot exceed the maximum elevation angle Ele of the radar; The distance between the target and the radar, which cannot exceed the radar detection range; is the azimuth between the target and the radar [Clock min ,Clock max ], which cannot exceed the azimuth interval of the radar detection range; Num(t)≤Num max Indicates that the number of targets tracked by the radar at any time cannot be greater than its maximum tracking capacity;

[0081] 3) Space satellite tracking constraints

[0082] When a task is assigned to a satellite constellation, the task obtained by each satellite participating in the same target tracking is set as a subtask; the subtask status information is:

[0083]

[0084] Where: m is the subtask number, i is the executing satellite number, j is the target number, The subtask start time, The end time of the subtask, is the satellite attitude information at the beginning of the subtask, Satellite attitude information at the end of the subtask, Power consumption for subtasks;

[0085] For a set of subtasks with the same target, two satellites p and q are required to execute them simultaneously, which satisfies the following conditions:

[0086]

[0087] Formula (4) indicates that target tracking requires two different satellites p and q to start and end at the same time, and each mission execution consumes satellite power, and its power consumption increases linearly with the mission time, and the power consumption rate is α;

[0088] For each execution satellite, the following conditions are met:

[0089] (1) The time interval for each satellite to perform its mission must be within the visible time window of the target, [ws m ,we m ] represents the time window in which the satellite of mission m is visible to the target;

[0090]

[0091] (2) The time interval between adjacent missions of each satellite must meet the minimum satellite attitude transition time, Represents the attitude conversion time between adjacent satellite tasks;

[0092]

[0093] (3) The remaining power of the satellite must always be greater than the power threshold E;

[0094]

[0095] In order to achieve the rapid response requirements of the space mobile target tracking system, a heuristic solution framework based on multi-agent is designed. It contains 5 types of agents, and the behavior diagram of each type of agent is as follows Figure 2 As shown in Figure 2, the functional behaviors of various intelligent agents are described as follows:

[0096] Radar Agent: Responsible for persistent monitoring of a fixed area. When a target appears, the Radar Agent calculates and sends its own monitoring time window. This includes window calculation and tracking task execution.

[0097] The virtual master radar station agent is responsible for receiving window information from all radar agents and selecting the optimal window to return to the command center agent. This includes sending window calculation requests and selecting inter-radar tracking solutions.

[0098] Satellite Agent: Moves according to its own orbit. When a target appears, the Satellite Agent calculates and sends its own monitoring time window. This includes behaviors such as window calculation and tracking task execution.

[0099] Virtual Master Satellite Agent: Responsible for receiving window information from satellite agents, selecting the optimal execution window, and returning it to the C2C agent. This includes sending window calculation requests, performing dual-satellite window cropping, and selecting a satellite tracking solution.

[0100] The Command and Control Center Agent maintains and updates target information, issues surveillance requests, receives and coordinates radar and satellite window information, and selects the optimal window as the target tracking window. This includes actions such as maintaining target information and sending mission information.

[0101] The entire MAS employs layered negotiation based on the traditional Contract Net Protocol (CNP). Interaction between agents is achieved through the Agent Communication Language (ACL). Based on the ACL, agents can interact by sharing information, reporting status, issuing requests, and responding to requests.

[0102] The main process of satellite-ground layered negotiation includes:

[0103] ① Overall ranking of target priorities to determine the planning order of targets.

[0104] ② Target trajectory prediction data acquisition: Based on the current state, the target trajectory prediction data for a period of time in the future is obtained.

[0105] ③ Ground-based radar / satellite calculates the visible time window of the target.

[0106] ④ Crop the visible time window to generate candidate tracking tasks.

[0107] ⑤Selection of the optimal solution at the main radar / main satellite level.

[0108] ⑥Select the optimal solution between the satellite and the ground.

[0109] ⑦ Execute the optimal solution and update radar, satellite status and target information.

[0110] In order to deal with the problem of target prediction error accumulation, a target dynamic priority mechanism is designed, which specifically involves target dynamic priority coordination, optimal solution selection strategy and global satellite grouping strategy.

[0111] (1) Dynamic priority coordination of goals

[0112] The priority of target j at time t Mainly consider the importance of its target and prediction error Related. Here it can be defined as:

[0113]

[0114] Where: The importance of each target I j ∈[I min ,I max ], Er max is the maximum prediction error that the system can tolerate, Er ref is the minimum error. c is a constant that makes I j tgt Lower goals can also be allocated certain tracking resources.

[0115] When planning is triggered, the system calculates the priority of the target to be planned according to formula (8) and determines the order of target planning. The dynamic priority of the target reflects the urgency of the target tracking demand and also reflects the tracking benefit of the target. In each planning process, targets with higher dynamic priorities should be tracked first. When a target is not tracked for a long time, its prediction error will accumulate and become larger, and its priority will increase accordingly. After the target is tracked, the target prediction error decreases and the priority is reduced. At this time, the device is allowed to track other targets with higher priorities. By changing the dynamic priority of the target, the tracking resource can poll the target, thereby ensuring that each target can be arranged for tracking when the prediction error accumulates too large.

[0116] (2) Optimal solution selection strategy

[0117] In terms of selecting the optimal solution, this patent sets two selection strategies.

[0118] (1) From a global perspective, the optimal selection strategy based on the temporal perspective (TPB) is adopted. The starting time of the satellite and ground solutions is compared, and the solution that starts the earliest is selected as the optimal solution. According to formula (1), the longer the target tracking time is, the greater the benefit. From a global perspective, the earlier the target tracking starts, the longer it is likely to be continuously tracked, and the lower the probability of target loss. At the same time, this strategy can arrange tracking for newly appeared targets as early as possible. Therefore, the optimal selection strategy based on the temporal perspective is adopted, that is, the earliest executable tracking subtask is adopted. Its strategy is shown in formula (9), and the tracking solution that starts the earliest is selected from all feasible solutions Ns.

[0119]

[0120] (2) From a local perspective, the best tracking solution is selected based on the best tracking effect (TEB). The tracking solution with the best tracking effect is selected each time, which makes the target prediction error decrease rapidly. The satellite tracking effect is determined by the average satellite angle (K) and the average distance between the satellite and the target Dis (K):

[0121]

[0122] The tracking effect of ground radar is determined by radar parameters (resolution Res and error Err):

[0123]

[0124] Each time the optimal solution is selected, it will be carried out at the main radar station, main satellite and command center. The command center only receives the unique optimal solution fed back by the radar coordination layer and the unique optimal solution fed back by the satellite coordination layer, and compares and selects the optimal solution to obtain the global optimal solution.

[0125] (3) Global satellite grouping strategy

[0126] In order to meet actual needs, the communication burden between satellites needs to be reduced as much as possible. This patent reference

[18] designs a global satellite grouping strategy based on the geographic satellite grouping idea. The satellite grouping strategy of this application mainly considers the visibility relationship between satellites and targets. When the target and the satellite are distributed in different hemispheres, the observation line of sight is easily blocked by the earth. Therefore, this strategy effectively reduces the number of satellites participating in information reception while ensuring benefits, thereby significantly reducing the communication burden of the entire satellite link and improving the operating efficiency of the system.

[0127] like Figure 3 As shown in the figure, the world is divided into eight normal zones (G1 to G8) and four transition zones (C1 to C4) based on longitude and longitude. Each normal zone spans 90° in longitude and latitude. For example, the range of the G1 zone is: longitude [0°, 90°E], latitude [0°, 90°N]. Each transition zone spans 30° in longitude and latitude. For example, the range of the C1 zone is: longitude [30°W, 30°E], latitude [30°S, 30°N].

[0128] The group number of the satellite and the space mobile target is determined by the latitude and longitude coordinates of the projection points on the ground, and the candidate satellite set of the target is determined accordingly.

[0129] The master satellite maintains a dynamic table of constellation satellite groupings, which changes periodically. Upon receiving target planning information from the command and control center, the master satellite assigns satellites based on the group number of the target's current projection. Given that targets falling within the handover zone often experience frequent changes in group number, dynamic triggering of four zones is required for planning.

[0130] ① When the target falls in the normal zone, the current zone and its left, right, up, and down zones are assigned for tracking. For example, if the target is in the G2 zone, the G1, G2, G3, and G6 zones are assigned for tracking.

[0131] ② When the target falls in the handover zone, the satellites in the boundary area are assigned to track it. For example, if the target falls in area C2, the satellites in areas G1, G2, G5, and G6 are assigned to track it.

[0132] This application's experiments are based on the JADE (Java Agent Development Framework) platform to develop a satellite-ground resource scheduling simulation system for space mobile target tracking. JADE is a Java-based agent development framework that provides a series of tools for monitoring and managing agents throughout their lifecycle.

[0133] The simulation experiment uses a 4x9 low-orbit Walker satellite constellation and four ground radar stations to collaboratively track ballistic missiles as a case study. Since the prediction error of a ballistic missile is cumulative, it can be set to increase linearly with the prediction time, using the formula (12). and are the start and end time of task m, is the error of target j at time t. When the target is not tracked, its prediction error accumulates linearly, with a growth rate of l, whose size is set to: 0.06514398. When the target is tracked, its prediction error decreases linearly with time, and the reduction rate is the tracking effect of the current task, which is calculated using formulas (10) and (11). At the same time, it is set that when the target prediction error is greater than half the satellite width, the target is considered lost, and subsequent planning and scheduling for the lost target will no longer be performed.

[0134]

[0135] The satellite constellation parameters, ground-based radar parameters, and tracking ballistic missile parameters used in the simulation experiment are shown in Tables 1, 2, and 3, respectively. The satellite's swath width is set to 80 km, and the detection range is 1000 km to 4000 km.

[0136] Table 1 Satellite constellation parameters of simulation experiment

[0137]

[0138] Table 2 Parameters of ground-based radar in simulation experiment

[0139]

[0140] Table 3 Ballistic missile parameters in simulation experiments

[0141]

[0142] The present invention proposes two different optimal tracking task selection strategies, namely, temporal perspective based (TPB) and track effect based (TEB). Here, the present invention will compare the effects of the two strategies through simulation experiments. At the same time, this patent will use the task selection strategy based on comprehensive time (CTB) as a comparison method. CTB sets weights and comprehensively considers the earliest start time and tracking duration of the solution. Its strategy is shown in formula (13), where w p and w stare the weights of tracking duration and tracking start time, respectively. They are set to 0.85 and 0.15 according to the literature. During the experiment, c is set to 1.5.

[0143]

[0144] During the experiment, the strategy was evaluated by using system tracking benefits and resource utilization. The benefits of each target are the importance of each tracking target. j Tracking duration Flight duration The product of the ratios. is the execution time of all tasks related to target j, which is obtained by calculating the sum of the duration of all tracking tasks of target j. The strategy benefit is a ratio, calculated as shown in formula (14). is the benefit of target j.

[0145]

[0146] Resource utilization is defined as the ratio of the time that resources spend on executing tracking tasks, and its calculation is shown in formula (15). is the resource utilization of device i, T is the total simulation scenario time, and the average resource utilization is the average resource utilization of all devices.

[0147]

[0148] Figure 4 The following table shows the tracking returns and average returns for each target under different strategies. Since the returns are calculated according to formula (14), they do not have units. Analysis shows that the TPB strategy outperforms the TEB and CTB strategies. Figure 5 The utilization and average resource utilization of various tracking resources under different strategies are shown. Because radars outperform satellites in tracking performance and have a wider coverage area, radar resource utilization under the TEB and CTB strategies is higher than that under the TPB strategy. Furthermore, because the TPB strategy selects the task planning result based on the earliest start time, its resource utilization is slightly higher than that of the other two strategies. Figure 6 Plan a Gantt chart for tasks under the TPB strategy.

[0149] In the experiment, the cumulative communication throughput of the main satellite during the simulation process (unit: piece) was counted to explore the impact of the global satellite grouping strategy on inter-satellite communication and the impact on mission planning results.

[0150] Figure 7The authors compared the cumulative number of primary satellite communications before and after the global grouping strategy was implemented. They also analyzed the system tracking benefits before and after the implementation of the global grouping strategy. This analysis shows that the global satellite grouping strategy significantly reduces satellite communication pressure by between 20% and 40%, while maintaining approximately the same tracking benefits.

[0151] This experiment primarily studies the effectiveness of tracking strategies at varying scales in dense target scenarios. To construct this scenario, the radar and satellite data used in the simulation remain unchanged. Based on the target data from Experiment 3.3.1, simulation scenarios with varying resource constraints are generated at scales of 10, 20, 30, and 40 targets. Each additional 10 targets maintains the same trajectory as the initial 10, with the launch time pushed back 30 seconds.

[0152] In all simulation scenarios, both the TPB and CTB strategies outperformed the TEB strategy. In small-scale scenarios, the TPB strategy significantly outperformed the CTB strategy. In large-scale scenarios, the TPB and CTB strategies achieved similar returns. This may be due to the fact that when the number of targets in a scenario increases, the TPB strategy frequently switches targets due to the real-time changes in target dynamic priorities. While the TPB strategy ensures that each target is tracked as quickly as possible, it also wastes considerable time in the posture transitions required to switch targets. The CTB strategy, on the other hand, prefers to track a single target for a longer period of time.

[0153] Table 4 Average tracking gains under different c values in dense target scenarios

[0154]

[0155] Table 4 shows how the average system benefit changes with target size and the value of parameter c when using the TPB strategy in a dense target scenario. c is the error weight factor in formula (8), which represents the weight of the target error in the target dynamic priority. Analysis shows that when there are fewer tracking targets, choosing a smaller c can obtain greater benefits; when there are more tracking targets, choosing a larger c can obtain greater benefits. The larger the c value, the more the target dynamic priority is affected by the target error. When tracking resources are sufficient, a smaller c value can ensure that targets with high importance obtain more resources, thereby obtaining greater tracking benefits. When tracking resources are limited, a larger c value can ensure that targets with large errors obtain more resources, thereby tracking as many targets as possible to obtain greater tracking benefits.

[0156] In a randomized scenario, the effectiveness of tracking strategies with varying numbers of targets was investigated. To construct the randomized scenario, the radar and satellite data used in the simulation remained unchanged. 50 target trajectories were randomly generated. The target flight durations ranged from 20 minutes 52 seconds to 32 minutes 1 second, and the target trajectory distribution was determined. The target importance distribution was consistent with Table 3. The target appearance times were randomly distributed at the simulation start time, 30 seconds after the simulation start time, 60 seconds after the simulation start time, and 90 seconds after the simulation start time.

[0157] Scaled test scenarios were randomly selected with 10, 20, 30, and 40 targets. Each algorithm was run five times at each target scale, and the average return was taken. The average return of the system targets at different target scales in the randomized scenarios was studied. In all simulation scenarios, the TPB strategy outperformed both the TEB and CTB strategies. As the number of targets increased, the gap between the returns of the TPB and CTB strategies gradually narrowed.

[0158] Table 5 Average tracking returns under different c values in random scenarios

[0159]

[0160] Table 5 shows how the average system return varies with target size and parameter c under the TPB strategy in a random scenario. Under the random scenario, the return exhibits high stability when c = 2.5. The corresponding return fluctuates relatively little under different target size scenarios.

[0161] The above specific embodiments do not constitute a limitation on the scope of protection of this application. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application shall be included within the scope of protection of this application.

Claims

1. A resource scheduling method for collaborative monitoring of space mobile targets based on multi-agents, characterized in that: The resource scheduling method specifically includes: Step 1: Read the target sequence information to be planned, calculate the dynamic priority of the target, and sort the target sequence according to the dynamic priority of the target; Step 2: Traverse the target sequence information and send the target information to be planned to the radar and satellite. The radar and satellite perform parallel planning. Step 3: All radars calculate the execution window information for the target and feed back the optimal execution window to the command and control center; Step 4: According to the target position and satellite grouping, the corresponding satellite is determined for planning, and the satellite calculates the tracking window for the target; Step 5: The satellite determines the optimal satellite tracking method based on the dual-satellite clipping and feeds it back to the command and control center; Step 6: The command center selects the optimal planning method and sends the result to the execution satellite or radar to trigger the next target planning.

2. The method according to claim 1, characterized in that Before executing the resource scheduling method, a scheduling model is constructed. The scheduling model is a satellite-ground resource integrated model for space mobile target tracking, specifically including: an objective function, ground-based radar tracking constraints, and space satellite tracking constraints; 1) Objective function The product of tracking time ratio and target importance is used as the optimization target, and the optimization objective function is expressed as: Where: N is the target number, I j is the importance of target j, is the tracking duration of target j, is the flight duration of target j; 2) Ground-based radar tracking constraints When the task is assigned to a ground-based radar, the tracking constraints for moving targets in space need to meet the radar's altitude constraint, distance constraint, azimuth constraint, and maximum tracking number constraint. Where: and is the target position and radar position; is the elevation angle between the target and the radar, which cannot exceed the maximum elevation angle Ele of the radar; The distance between the target and the radar, which cannot exceed the radar detection range; is the azimuth between the target and the radar [Clock min ,Clock max ], which cannot exceed the azimuth interval of the radar detection range; Num(t)≤Num max Indicates that the number of targets tracked by the radar at any time cannot be greater than its maximum tracking capacity; 3) Space satellite tracking constraints When a task is assigned to a satellite constellation, the task obtained by each satellite participating in the same target tracking is set as a subtask; the subtask status information is: Where: m is the subtask number, i is the executing satellite number, j is the target number, The subtask start time, The end time of the subtask, is the satellite attitude information at the beginning of the subtask, Satellite attitude information at the end of the subtask, Power consumption for subtasks; For a set of subtasks with the same target, two satellites p and q are required to execute them simultaneously, which satisfies the following conditions: Formula (4) indicates that target tracking requires two different satellites p and q to start and end at the same time, and each mission execution consumes satellite power, and its power consumption increases linearly with the mission time, and the power consumption rate is α; For each execution satellite, the following conditions are met: (1) The time interval for each satellite to perform its mission must be within the visible time window of the target, [ws m ,we m ] represents the time window in which the satellite of mission m is visible to the target; (2) The time interval between adjacent missions of each satellite must meet the minimum satellite attitude transition time, Represents the attitude conversion time between adjacent satellite tasks; (3) The remaining power of the satellite must always be greater than the power threshold E; 3. The method according to claim 2, characterized in that The method is implemented based on a multi-agent system, which includes a radar agent, a virtual master radar station agent, a satellite agent, a virtual master satellite agent, and a command and control center agent. The radar agent is responsible for persistent monitoring of a fixed area. When a monitored target appears, the radar agent calculates and sends its own monitoring time window, performs window calculation and tracking tasks; The virtual master radar station agent is responsible for receiving window information from all radar agents, selecting the optimal window and returning it to the command center agent, executing window calculation requests and selecting tracking modes between radars. The satellite agent is responsible for moving according to its own orbit. When a surveillance target appears, the satellite agent calculates and sends its own surveillance time window, and performs window calculation and tracking tasks. The virtual master satellite agent is responsible for receiving the window information of the satellite agent, selecting the optimal execution window information and returning it to the command and control center agent, executing the sending window calculation request, dual-satellite window clipping, and satellite tracking mode selection; The intelligent agent in the command and control center is responsible for maintaining and updating target information, issuing monitoring requirements, receiving and coordinating window information from radars and satellites, selecting the optimal window as the target tracking window, performing target information maintenance, and sending mission information.

Images