Simulation method for constellation autonomous rapid cooperative observation at sea based on double-layer planning architecture

Abstract

The application provides a constellation autonomous rapid sea observation simulation method based on a double-layer planning framework, which comprises the following steps: initializing a satellite constellation, a target state and simulation parameters; in a simulation time, a planning cycle is circularly executed: generating an observation task request according to the target state; performing a first-layer constellation-level task allocation, and allocating appropriate satellites to the task based on coverage capacity, load balancing and orbit characteristics; performing a second-layer single-satellite task planning, and selecting an optimal scheduling window to generate a planning task by comprehensively considering time response, observation quality and window margin in the available window; performing observation and checking the effect; updating the target state; and until the simulation ends. The application adopts a double-layer planning framework, the first layer allocates appropriate satellites to each task request from a global perspective, and the second layer selects an optimal scheduling window to generate a specific planning task, which not only avoids the problem that local decision-making leads to global imbalance in single-layer planning, but also overcomes the defect that the calculation complexity of global optimization algorithms is too high.

Description

Technical Field

[0001] This invention relates to the field of autonomous real-time mission planning and scheduling technology for spacecraft, specifically to a constellation-based autonomous rapid maritime collaborative observation simulation method based on a two-layer planning architecture, applicable to the rapid allocation and scheduling of autonomous missions for multi-satellite collaborative observation of moving targets at sea. Background Technology

[0002] With the rapid development of aerospace technology, satellite constellations are playing an increasingly important role in fields such as ocean surveillance, maritime search and rescue, and fisheries monitoring. Traditional single-satellite observation modes are no longer sufficient to meet the demands for large-scale, high-frequency, and multi-target observation; multi-satellite collaborative observation has become an inevitable trend. However, rapid autonomous mission planning for satellite constellations faces numerous challenges: First, satellite observation of ground targets is limited by multiple factors, including orbital constraints, lighting conditions, and sensor field of view, resulting in uneven spatiotemporal distribution of visibility windows. Visibility windows for the same target may overlap or remain blank for different satellites, making the rational allocation of tasks a crucial issue.

[0003] Secondly, maritime targets are mobile, and their positions change constantly over time. Mission planning needs to consider the uncertainty of target motion prediction; traditional static planning methods are ill-suited to dynamically changing observation scenarios.

[0004] Secondly, satellite constellations typically consist of multiple satellites, each with different orbital parameters, sensor configurations, and current load status. Achieving load balancing at the global level, preventing some satellites from being overloaded while others remain idle, is key to improving the overall efficiency of the constellation.

[0005] In the existing technology, common satellite mission planning methods mainly include: single-satellite planning methods based on greedy algorithms, which are simple and efficient but lack global optimization capabilities; methods based on integer programming, which can obtain the global optimal solution but have high computational complexity and are difficult to meet real-time requirements; and methods based on heuristic algorithms, such as genetic algorithms and simulated annealing, which to some extent balance solution quality and computational efficiency, but are difficult to adjust parameters and are difficult to handle dynamic mission arrival situations.

[0006] Therefore, there is an urgent need for a constellation-based rapid mission planning method that can comprehensively consider multiple factors such as visibility constraints, target motion, and load balancing, in order to achieve efficient and automated scheduling of ocean observation missions. Summary of the Invention

[0007] This invention proposes a constellation-based autonomous and rapid collaborative maritime observation simulation method based on a two-layer planning architecture, which solves the problems of insufficient global optimization capability, poor load balancing effect, and difficulty in guaranteeing real-time performance in existing satellite constellation mission planning methods when dealing with multi-target and dynamic scenarios.

[0008] The technical solution of this invention is implemented as follows: The first aspect of this invention provides a constellation-based autonomous rapid collaborative ocean observation simulation method based on a two-layer planning architecture, comprising the following steps: Initialize the satellite constellation, target status data, and simulation parameters; Within the simulation timeframe, the following steps are executed cyclically according to a preset planning period: Generate an observation task request based on the current state of the target; Perform the first-level constellation-level mission allocation: based on the expected coverage of the target by the satellites, the overall load of the constellation, and the orbital characteristics, allocate appropriate satellites to each mission request; Perform second-level single-satellite mission planning: For mission requests allocated to each satellite, select the optimal scheduling window within the available time window and generate specific planned tasks; Perform collaborative marine observations based on the generated planning tasks, and verify the effectiveness of the task execution; Update the target status based on the verification results; Determine if the current simulation time has ended. If not, proceed to the next planning cycle; otherwise, output the simulation results.

[0009] Specifically, the method for generating observation task requests is as follows: Establish a target ship state machine that includes unknown states, discovered states, states tracked once, states tracked twice, states ready for identification, and states to be maintained. When the target vessel is in an unknown state, a search task request is generated; When the system is in the discovered state, an initial tracing task request is generated; When the system is in the state of tracking once or tracking twice, a subsequent tracking task request is generated. When in the ready-to-recognize state, a recognition task request is generated; When in maintenance mode, a maintenance tracking task request is generated; Different types of task requests are configured with different priorities and ranges.

[0010] Furthermore, for search mission requests, the search mode is selected as either a normal search mode or a wide-range search mode based on the degree of uncertainty regarding the target vessel's position. If the uncertainty diameter is less than or equal to the preset threshold, then the normal search mode is selected, and the width is the first preset value; otherwise, the wide search mode is selected, and the width is the second preset value; wherein, the second preset value is greater than the first preset value. The uncertainty diameter is calculated based on the product of the target ship's speed and the uncertainty diffusion time.

[0011] Specifically, the method for performing the first-level constellation-level task allocation includes the following steps: Get all pending task requests, sort them in descending order of priority, and prioritize the allocation of high-priority tasks; For each sorted task request, iterate through every available satellite in the constellation and calculate the satellite's overall score for the current task request. The comprehensive score is calculated as follows: calculate the satellite's time delay score and incident angle score, and add them together to obtain the coverage capability score; count the number of tasks assigned to each satellite, calculate the ratio of the current number of satellite tasks to the average number of tasks, and calculate the load balancing factor based on this ratio; calculate the difference between the satellite's orbital inclination and the target's latitude, and calculate the orbital coverage factor based on this difference; multiply the coverage capability score, load balancing factor, and orbital coverage factor together to obtain the comprehensive score. The satellite with the highest overall score will be selected as the allocation target for the current mission request.

[0012] Furthermore, methods for calculating the satellite's time delay score and incident angle score for the target include: Calculate the position vector from the satellite to the target; The cone angle is calculated based on the position vector, and the cone angle is the angle between the line connecting the satellite and the target and the line connecting the satellite and the Earth's center. Calculate the azimuth angle of the target in the satellite orbital plane; Select time periods that simultaneously satisfy the preset cone angle constraint and azimuth angle constraint, and use them as the visibility window; The earliest visible window is determined from the visibility window, and the average incident angle of that window is obtained. The time delay score is calculated based on the difference between the start time of the earliest visible window and the current time, and the incident angle score is calculated based on the average incident angle of the earliest visible window.

[0013] Specifically, the method for performing second-level single-satellite mission planning includes the following steps: For each satellite, obtain the list of mission requests assigned to that satellite and sort them in descending order according to the priority of each mission request; For each sorted task request, find all scheduling windows for that satellite that satisfy the sensor conic angle constraint and azimuth angle constraint for the task objective; For each scheduling window found, the corresponding window score is calculated. The window score is obtained by weighted summation of three parts: time delay score, incident angle score, and window margin score. The window margin score is calculated based on the difference between the window length and the task duration. The scheduling window with the highest window score is selected as the optimal scheduling window for the task request, and a planning task is generated based on the optimal scheduling window. The planning task includes at least one or more of the following: satellite identifier, target identifier, task type, start time, end time, duration, incident angle, and swath width. Update the satellite's idle time to the end time of the planned mission to avoid time conflicts between subsequent planned missions and the already planned mission.

[0014] Specifically, methods for verifying the effectiveness of task execution include: For each planning task, the aiming point is calculated using dead reckoning based on the target ship's last known position, initial speed, and initial heading. Obtain the true position of the target ship at the end of the mission; Calculate the distance deviation between the aiming point and the true position; Determine if the distance deviation is less than or equal to half the width of the planned task. If so, the task is considered successful; otherwise, the task is considered a failure.

[0015] Specifically, the methods for updating the target state based on the verification results include: If the task verification is successful, then advance the target ship's state machine according to the task type: If the current state is unknown and the search task is successful, then the process will transition to the discovered state. If the current state is "discovered" and the tracking task is successful, then transition to the "tracked once" state. If the current state is one-time tracking and the tracking task is successful, then transition to two-time tracking. If the current state is tracking twice and the tracking task is successful, then transition to the preparation for recognition state; If the current state is ready for identification and the identification task is successful, then the process will transition to the maintenance state. If the current state is maintenance and the tracking task is successful, then maintain the maintenance state; If the task verification fails, the execution status will be rolled back according to the task type.

[0016] A second aspect of the present invention provides an electronic device, including a memory and a processor, wherein the memory stores a computer program executable on the processor, and the processor executes the computer program to implement the steps of the simulation method.

[0017] A third aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the simulation method.

[0018] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) This invention adopts a two-layer planning architecture. The first layer comprehensively evaluates the coverage capability of each satellite to the target, the overall load balance of the constellation, and the matching characteristics between the satellite orbit and the target position from a global perspective, and allocates a suitable satellite for each task request. The second layer, at the single-satellite level, selects the optimal scheduling window to generate a specific planning task within the available time window for the assigned task, taking into account the time response, observation quality, and window sufficiency. The two planning layers form a close progressive relationship, which not only avoids the problem of global imbalance caused by local decisions in single-layer planning, but also overcomes the defect of excessive computational complexity of global optimization algorithms. (2) The present invention adaptively selects the normal search mode or the wide-span search mode according to the size of the target position uncertainty diameter, so that the allocation of observation resources can closely match the observation stage and position uncertainty of the target. When the target first appears, the wide-span search is used to quickly cover a large area. After the target status is gradually clarified, the narrow-span tracking is switched to improve the observation accuracy, effectively reducing invalid observations and improving the utilization efficiency of observation resources. (3) In the first-level constellation-level task allocation, the present invention calculates the comprehensive score of all available satellites for each task request. The comprehensive score is obtained by multiplying the coverage capability score, load balancing factor and orbital coverage factor, and the satellite with the highest score is selected for allocation. This comprehensively takes into account the observation timeliness, observation geometric quality, task load balancing within the constellation and orbital kinematic matching characteristics, avoiding the problem of satellite overload or poor observation conditions caused by single-factor decision-making, and making the task allocation result more reasonable in the global scope. (4) In the second-layer single-satellite mission planning, this invention finds all scheduling windows that satisfy the sensor cone angle and azimuth angle constraints for each mission, and calculates the window score for each window by weighted summation of time delay score, incident angle score and window margin score. The window with the highest score is selected to generate the planned mission. Finally, the idle time of the satellite is updated to avoid time conflicts of subsequent missions. This not only takes into account the timeliness of mission execution and the impact of the observation incident angle on imaging quality, but also ensures that the selected window has enough margin to complete the full observation through window margin score. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating the constellation-based autonomous rapid collaborative ocean observation simulation method based on a two-layer planning architecture, as described in this invention.

[0021] Figure 2 This is a schematic diagram of the two-layer planning architecture of the present invention.

[0022] Figure 3 This is a flowchart illustrating the method for performing the first-level constellation-level task allocation in an embodiment of the present invention.

[0023] Figure 4 This is a flowchart illustrating the method for performing second-layer single-satellite mission planning in an embodiment of the present invention. Detailed Implementation

[0024] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0025] Reference Figure 1 , 2 The first aspect of this invention provides a constellation-based autonomous rapid collaborative ocean observation simulation method based on a two-layer planning architecture, comprising the following steps: Step S1: Initialize satellite constellation, target status data, and simulation parameters; The satellite constellation includes multiple satellites, each equipped with various sensors, each with conic angle constraints and azimuth angle constraints; the target vessel data includes the vessel's initial position, velocity, and heading information; based on the satellite's TLE orbital parameters, the satellite position and velocity vectors at each time step within the simulation time range are calculated; Within the simulation timeframe, the following steps are executed cyclically according to a preset planning period: Step S2: Generate an observation task request based on the current state of the target; Establish a target ship state machine that includes unknown states, discovered states, states tracked once, states tracked twice, states ready for identification, and states to be maintained. When the target vessel is in an unknown state, a search task request is generated; When the system is in the discovered state, an initial tracing task request is generated; When the system is in the state of tracking once or tracking twice, a subsequent tracking task request is generated. When in the ready-to-recognize state, a recognition task request is generated; When in maintenance mode, a maintenance tracking task request is generated; Different types of task requests are configured with different priorities and ranges.

[0026] Furthermore, for search mission requests, the search mode is selected as either a normal search mode or a wide-range search mode based on the degree of uncertainty regarding the target vessel's position. If the uncertainty diameter is less than or equal to the preset threshold, then the normal search mode is selected, and the width is the first preset value; otherwise, the wide search mode is selected, and the width is the second preset value; wherein, the second preset value is greater than the first preset value. The uncertainty diameter is calculated based on the product of the target vessel's speed and the uncertainty diffusion time, where the uncertainty time is the time difference between the last time the target was confirmed and the current time.

[0027] Step S3: Perform the first-level constellation-level task allocation: Based on the expected coverage of the target by the satellites, the overall load status of the constellation, and the orbital characteristics, allocate appropriate satellites to each task request; Specifically, such as Figure 3 As shown, the method for performing the first-level constellation-level task allocation includes the following steps: S31, obtain all task requests to be assigned, sort the task requests in descending order according to priority, and assign high priority tasks first; S32, For each sorted task request, iterate through each available satellite in the constellation and calculate the satellite's overall score for the current task request; The calculation method for the comprehensive score is as follows: Calculate the satellite's time delay score and angle of incidence score for the target, and add them together to obtain the coverage capability score: ; in, For comprehensive scoring, Scoring based on time delay, Rate the angle of incidence; The number of tasks assigned to each satellite is counted, the ratio of the current number of tasks to the average number of tasks is calculated, and the load balancing factor is calculated based on this ratio (the load balancing factor is inversely proportional to this ratio). The calculation formula is as follows: ; in, As a load balancing factor, This is the ratio of the current number of satellite missions to the average number of missions. Calculate the difference between the satellite's orbital inclination and the target's latitude, and then calculate the orbital coverage factor based on this difference (the orbital coverage factor is inversely proportional to this difference). The calculation formula is as follows: ; in, For orbital coverage factor, This is the difference between the satellite's orbital inclination and the absolute value of the target's latitude. Multiply the coverage score, load balancing factor, and track coverage factor together to obtain the comprehensive score. : ; S33: Select the satellite with the highest comprehensive score as the allocation target for the current mission request. If all satellites fail to meet the visibility constraint, mark the mission as unassigned.

[0028] Furthermore, methods for calculating the satellite's time delay score and incident angle score for the target include: Calculate the position vector from the satellite to the target; The cone angle is calculated based on the position vector, and the cone angle is the angle between the line connecting the satellite and the target and the line connecting the satellite and the Earth's center. Calculate the azimuth angle of the target in the satellite orbital plane; Select time periods that simultaneously satisfy the preset cone angle constraint and azimuth angle constraint, and use them as the visibility window; The earliest visible window is determined from the visibility window, and the average incident angle of that window is obtained. The time delay score is calculated based on the difference between the start time of the earliest visible window and the current time, and the incident angle score is calculated based on the average incident angle of the earliest visible window. The calculation formula is as follows: ; ; in, This is the difference between the start time of the earliest visible window and the current time. The average angle of incidence of the earliest visible window.

[0029] Step S4: Perform the second-level single-satellite mission planning: Select the optimal scheduling window within the available time window for the mission requests allocated to each satellite, and generate specific planned tasks. Specifically, such as Figure 4 As shown, the method for performing second-level single-satellite mission planning includes the following steps: S41, For each satellite, obtain the list of mission requests assigned to that satellite and sort them in descending order according to the priority of each mission request; S42, For each sorted task request, find all scheduling windows for that satellite that satisfy the sensor conic angle constraint and azimuth angle constraint for the task objective; S43, for each found scheduling window, calculate the corresponding window score. The window score is obtained by weighted summation of three parts: time delay score, incident angle score, and window margin score. The window margin score is calculated based on the difference between the window length and the task duration, and the calculation formula is as follows: ; ; in, Score the window margin. This is the difference between the window length and the task duration. The window is scored; the weight parameters of the three scores can be set in different scenarios, and the default weight is 1 in this embodiment. S44, select the scheduling window with the highest window score as the optimal scheduling window for the task request, and generate a planning task based on the optimal scheduling window. The planning task includes satellite identifier, target identifier, task type, start time, end time, duration, incident angle and swath width. S45, update the satellite's idle time to the end time of the planned mission to avoid time conflicts between subsequent planned missions and the already planned mission.

[0030] Step S5: Perform collaborative ocean observation according to the generated planning task, and verify the task execution effect; Specifically, methods for verifying the effectiveness of task execution include: For each planning task, the aiming point is calculated using dead reckoning based on the target ship's last known position, initial speed, and initial heading. Obtain the true position of the target ship at the end of the mission; Calculate the distance deviation between the aiming point and the true position; Determine if the distance deviation is less than or equal to half the width of the planned task. If so, the task is considered successful; otherwise, the task is considered a failure.

[0031] Furthermore, the specific calculation formula for dead reckoning is as follows: ; ; in, and These are the latitude and longitude of the aiming point, respectively. and Given the latitude and longitude of the target's last known location. For the target ship speed, The target vessel's heading angle, For time intervals, 111.0 represents the average latitude, and 111.0 represents the number of kilometers corresponding to the average unit of latitude and longitude.

[0032] Step S6: Update the target status based on the verification result; Specifically, the methods for updating the target state based on the verification results include: If the task verification is successful, update the ship's last known position to the true position, update the last observed time to the task end time, and advance the target ship's state machine according to the task type: If the current state is unknown and the search task is successful, then the process will transition to the discovered state. If the current state is "discovered" and the tracking task is successful, then transition to the "tracked once" state. If the current state is one-time tracking and the tracking task is successful, then transition to two-time tracking. If the current state is tracking twice and the tracking task is successful, then transition to the preparation for recognition state; If the current state is ready for identification and the identification task is successful, then the process will transition to the maintenance state. If the current state is maintenance and the tracking task is successful, then maintain the maintenance state; If the task verification fails, the state rollback or retry strategy is executed according to the task type; if the tracking task fails, the ship's state rolls back to the discovered state; if the identification task fails, the ship's state remains in the ready-to-identify state.

[0033] Step S7: Determine whether the current simulation time has ended. If it has not ended, proceed to the next planning cycle; otherwise, output the simulation results, including statistical information such as the number of successes / failures of each type of task and the average position deviation.

[0034] A second aspect of the present invention provides an electronic device, including a memory and a processor, wherein the memory stores a computer program executable on the processor, and the processor executes the computer program to implement the steps of the simulation method.

[0035] A third aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the simulation method.

[0036] The following example illustrates the technical solution of this invention by using a Walker constellation consisting of 12 satellites to schedule observation missions for 60 ships at sea. Effectiveness verification data is also provided to further demonstrate the feasibility and beneficial effects of this invention.

[0037] Step 1, System Initialization: In this embodiment, the satellite constellation adopts a Walker 12 / 3 / 1 configuration, including 3 orbital planes, 4 satellites in each orbital plane, and each satellite is equipped with the following two types of sensors: Sensor 1: Supports search and tracking tasks, with a cone angle range of 10 to 30 degrees and an azimuth angle range of 10 to 170 degrees and 190 to 350 degrees.

[0038] Sensor 2: Supports recognition tasks, with a cone angle range of 5 to 25 degrees and an azimuth angle range of 20 to 160 degrees and 200 to 340 degrees.

[0039] The initial positions of the target vessels are distributed within a pre-defined sea lane area, the initial speed is randomly selected between 12 and 20 knots, and the initial course is based on the route direction with random disturbances added.

[0040] The simulation parameters are set as follows: the simulation start time is 00:00:00 on March 1, 2026 UTC, the total simulation duration is 24 hours, the planning cycle is 180 seconds, and the task cooldown time is 5 minutes.

[0041] Based on the two-line orbital element file of the satellite, the SGP4 orbital propagation model is used to calculate the satellite position and velocity vectors at each time step within the simulation time range; in this embodiment, the time step size is consistent with the planning period, which is 180 seconds.

[0042] For each satellite, calculate its position vector (in kilometers) and velocity vector (in kilometers per second) in the geocentric inertial coordinate system at each time step, and store them in memory for use in subsequent visibility window calculations.

[0043] Step 2, task request generation: At the start of each planning cycle, iterate through all target ships and generate task requests based on their current states: 1) If the vessel is in an unknown state, generate a search task request; calculate the position uncertainty diameter based on the time difference between the vessel's last observation time and the current time. If the uncertainty diameter is less than or equal to 140 km, select the normal search mode with a swath width of 140 km for 2 seconds; otherwise, select the wide-swath search mode with a swath width of 280 km for 4 seconds; the priority of the search task is 10. 2) If the vessel is already detected, generate an initial tracking task request with a priority of 7 (base priority 5 plus 2), a swath width of 30 kilometers, and a duration of 1 second; 3) If the vessel is in the state of being tracked once or twice, a subsequent tracking task request will be generated with a priority of 5, a swath width of 30 kilometers, and a duration of 1 second; 4) If the vessel is in the ready-to-identify state, generate an identification task request with a priority of 8, a width of 50 kilometers, and a duration of 1 second; 5) If the vessel is under maintenance, generate a maintenance tracking task request with a priority of 4 (basic priority 5 minus 1), a width of 30 kilometers, and a duration of 1 second.

[0044] Step 3, First-tier constellation-level task allocation: First, sort all task requests in descending order of priority to ensure that high-priority tasks are assigned first.

[0045] For each task request, perform the following steps: Step 1: Obtain the current position (latitude and longitude) of the target vessel.

[0046] Step 2: Traverse all satellites and calculate the overall score for the mission for each satellite.

[0047] Step 3: For each satellite, first determine whether the satellite is available, i.e., whether the current time has exceeded the satellite's idle time; if the satellite is unavailable, skip the satellite.

[0048] Step 4: Calculate the satellite visibility window to the target. Call the visibility window calculation function of the physics engine, input the satellite index, sensor configuration, target position, earliest start time and end time, and output a list of visibility windows. Each window contains the start time, end time and average incident angle.

[0049] Step 5: If visibility windows exist, calculate the overall score for each visibility window.

[0050] Step 6: Select the satellite with the highest overall score and assign the task to that satellite; if no satellite has a visibility window, mark the task as unassigned.

[0051] Step 4, Second-tier single-star mission planning: For each satellite, perform the following steps: Step 1: Obtain the list of mission requests assigned to the satellite and sort them in descending order of priority.

[0052] Step 2: For each task request, find a scheduling window that satisfies the sensor constraints.

[0053] Step 3: Iterate through the sensors that support this task type. For each sensor, call the physics engine's visibility window calculation function to obtain a list of visibility windows.

[0054] Step 4: For each visible window, calculate the window score.

[0055] Step 5: Select the window with the highest rating to generate the planning task.

[0056] Step 6: Update the satellite's idle time to the mission end time to ensure that subsequent missions do not conflict with planned missions.

[0057] Step 5, Task Execution and Verification: For each planning task, perform the following steps: Step 1: Obtain the target ship object.

[0058] Step 2: Calculate the aiming point. If the ship has a last known position, calculate the aiming point using dead reckoning based on the last known position, initial speed, and heading; otherwise, perform dead reckoning based on the ship's initial position.

[0059] Step 3: Obtain the true position of the target at the end of the mission (find the state point closest to the end of the mission from the ship trajectory data).

[0060] Step 4: Calculate the distance deviation between the aiming point and the true position (use the semi-versus formula to calculate the spherical distance).

[0061] Step 5: Determine if the distance deviation is less than or equal to half the task width. If so, the task is successful; otherwise, the task fails.

[0062] Step 6: If the mission is successful, update the ship's last known position to the true position, update the last observed time to the mission end time, and advance the ship's state machine according to the mission type; at the same time, record the success log.

[0063] Step 7: If the task fails, roll back the execution status according to the task type; at the same time, record the failure log.

[0064] Step 6, Simulation Loop: The simulation main loop proceeds according to the planned cycle. Each cycle executes steps such as task request generation, two-layer planning, task execution and verification, and state update. After the simulation ends, the simulation results are output.

[0065] Simulation results in this embodiment show that, using the method of the present invention, the success rate of the search task is over 90%, the success rate of the tracking task is over 85%, and the success rate of the identification task is over 80% within a 24-hour simulation period, with the average position deviation controlled within 15 kilometers, thus verifying the effectiveness of the method of the present invention.

[0066] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A simulation method for autonomous and rapid collaborative ocean observation based on a two-layer planning architecture, characterized in that, Includes the following steps: Initialize the satellite constellation, target status data, and simulation parameters; Within the simulation timeframe, the following steps are executed cyclically according to a preset planning period: Generate an observation task request based on the current state of the target; Perform the first-level constellation-level mission allocation: based on the expected coverage of the target by the satellites, the overall load of the constellation, and the orbital characteristics, allocate appropriate satellites to each mission request; Perform second-level single-satellite mission planning: For mission requests allocated to each satellite, select the optimal scheduling window within the available time window and generate specific planned tasks; Perform collaborative marine observations based on the generated planning tasks, and verify the effectiveness of the task execution; Update the target status based on the verification results; Determine if the current simulation time has ended. If not, proceed to the next planning cycle; otherwise, output the simulation results.

2. The constellation-based autonomous rapid collaborative ocean observation simulation method based on a two-layer planning architecture as described in claim 1, characterized in that, The method for generating observation task requests is as follows: Establish a target ship state machine that includes unknown states, discovered states, states tracked once, states tracked twice, states ready for identification, and states to be maintained. When the target vessel is in an unknown state, a search task request is generated; When the system is in the discovered state, an initial tracing task request is generated; When the system is in the state of tracking once or tracking twice, a subsequent tracking task request is generated. When in the ready-to-recognize state, a recognition task request is generated; When in maintenance mode, a maintenance tracking task request is generated; Different types of task requests are configured with different priorities and ranges.

3. The constellation-based autonomous rapid collaborative ocean observation simulation method based on a two-layer planning architecture as described in claim 2, characterized in that, For search mission requests, select either the normal search mode or the wide-span search mode based on the degree of uncertainty regarding the target vessel's location: If the uncertainty diameter is less than or equal to the preset threshold, then the normal search mode is selected, and the width is the first preset value; otherwise, the wide search mode is selected, and the width is the second preset value; wherein, the second preset value is greater than the first preset value. The uncertainty diameter is calculated based on the product of the target ship's speed and the uncertainty diffusion time.

4. The constellation-based autonomous rapid collaborative ocean observation simulation method based on a two-layer planning architecture as described in claim 1, characterized in that, The method for performing the first-level constellation-level task assignment includes the following steps: Get all pending task requests, sort them in descending order of priority, and prioritize the allocation of high-priority tasks; For each sorted task request, iterate through every available satellite in the constellation and calculate the satellite's overall score for the current task request. The comprehensive score is calculated as follows: calculate the satellite's time delay score and incident angle score, and add them together to obtain the coverage capability score; count the number of tasks assigned to each satellite, calculate the ratio of the current number of satellite tasks to the average number of tasks, and calculate the load balancing factor based on this ratio; calculate the difference between the satellite's orbital inclination and the target's latitude, and calculate the orbital coverage factor based on this difference; multiply the coverage capability score, load balancing factor, and orbital coverage factor together to obtain the comprehensive score. The satellite with the highest overall score will be selected as the allocation target for the current mission request.

5. The constellation-based autonomous rapid collaborative ocean observation simulation method based on a two-layer planning architecture as described in claim 4, characterized in that, Methods for calculating the satellite's time delay score and angle of incidence score for a target include: Calculate the position vector from the satellite to the target; The cone angle is calculated based on the position vector, and the cone angle is the angle between the line connecting the satellite and the target and the line connecting the satellite and the Earth's center. Calculate the azimuth angle of the target in the satellite orbital plane; Select time periods that simultaneously satisfy the preset cone angle constraint and azimuth angle constraint, and use them as the visibility window; The earliest visible window is determined from the visibility window, and the average incident angle of that window is obtained. The time delay score is calculated based on the difference between the start time of the earliest visible window and the current time, and the incident angle score is calculated based on the average incident angle of the earliest visible window.

6. The constellation-based autonomous rapid collaborative ocean observation simulation method based on a two-layer planning architecture as described in claim 1, characterized in that, The method for performing second-level single-satellite mission planning includes the following steps: For each satellite, obtain the list of mission requests assigned to that satellite and sort them in descending order according to the priority of each mission request; For each sorted task request, find all scheduling windows for that satellite that satisfy the sensor conic angle constraint and azimuth angle constraint for the task objective; For each scheduling window found, the corresponding window score is calculated. The window score is obtained by weighted summation of three parts: time delay score, incident angle score, and window margin score. The window margin score is calculated based on the difference between the window length and the task duration. The scheduling window with the highest window score is selected as the optimal scheduling window for the task request, and a planning task is generated based on the optimal scheduling window. The planning task includes at least one or more of the following: satellite identifier, target identifier, task type, start time, end time, duration, incident angle, and swath width. Update the satellite's idle time to the end time of the planned mission to avoid time conflicts between subsequent planned missions and the already planned mission.

7. The constellation-based autonomous rapid collaborative ocean observation simulation method based on a two-layer planning architecture as described in claim 1, characterized in that, Methods for verifying the effectiveness of task execution include: For each planning task, the aiming point is calculated using dead reckoning based on the target ship's last known position, initial speed, and initial heading. Obtain the true position of the target ship at the end of the mission; Calculate the distance deviation between the aiming point and the true position; Determine if the distance deviation is less than or equal to half the width of the planned task. If so, the task is considered successful; otherwise, the task is considered a failure.

8. The constellation-based autonomous rapid collaborative ocean observation simulation method based on a two-layer planning architecture as described in claim 1, characterized in that, Methods for updating the target state based on the verification results include: If the task verification is successful, then advance the target ship's state machine according to the task type: If the current state is unknown and the search task is successful, then the process will transition to the discovered state. If the current state is "discovered" and the tracking task is successful, then transition to the "tracked once" state. If the current state is one-time tracking and the tracking task is successful, then transition to two-time tracking. If the current state is tracking twice and the tracking task is successful, then transition to the preparation for recognition state; If the current state is ready for identification and the identification task is successful, then transition to the maintenance state; If the current state is maintenance and the tracking task is successful, then maintain the maintenance state; If the task verification fails, the execution status will be rolled back according to the task type.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the simulation method as described in any one of claims 1 to 8.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the simulation method as described in any one of claims 1 to 8.

Images