Construction method of unmanned system test task and electronic equipment

Abstract

The invention discloses an unmanned system test task construction method and electronic equipment, and relates to the technical field of unmanned systems. The method comprises the following steps: constructing a structured task library and a resource library, and constructing a joint constraint model based on the task library and the resource library; determining a task combination based on the test target and a task library, wherein the task combination comprises a plurality of subtasks; performing simulation scheduling based on the resource library, the task combination and the joint constraint model to determine a scheduling feasibility score; determining a to-be-adjusted sub-task according to the scheduling feasibility score; performing iterative adjustment on the to-be-adjusted sub-tasks based on a preset rule until the scheduling feasibility score of the task combination meets a requirement or the number of iterations reaches a preset number, so as to obtain a target task combination; and outputting a task structure and a resource allocation table based on the target task combination so as to import the task structure and the resource allocation table into the unmanned system. Therefore, the execution efficiency of the test process of the unmanned system and the improvement of the intelligent level of the system can be ensured.

Description

Technical Field

[0001] This application relates to the field of unmanned systems, and more particularly to a method for constructing an unmanned system test task and an electronic device. Background Technology

[0002] In the field of unmanned system testing (such as unmanned vehicles, drones, unmanned ships, and unmanned special-purpose robots), traditional static task design and linear resource scheduling methods are facing significant challenges as equipment intelligence increases and test task complexity grows. Most existing testing systems adopt a task-scheduling separation model, where a fixed task set is first manually or according to preset rules, and then resources are allocated and time is scheduled based on this task set. This approach fails to fully consider the dynamic state, availability, conflict potential, and equipment compatibility of scheduling resources during the task design phase, easily leading to resource conflicts, task blocking, or scheduling failures in the actual execution of the generated test plan, resulting in low testing efficiency and resource waste.

[0003] Some studies have attempted to improve scheduling performance through stress prediction or rule-based priority scheduling. However, these approaches often rely on static scoring models, making it difficult to dynamically assess the true schedulability of task combinations under complex resource conditions, and also failing to achieve coordinated optimization of task content design and resource allocation. Furthermore, for hard constraints commonly found in testing scenarios, such as "sequential dependencies between subtasks," "binding of multiple resource types," and "test window limitations," most existing scheduling models lack joint modeling and feedback correction mechanisms, making it difficult to support highly automated and intelligent testing processes. Summary of the Invention

[0004] This application aims to at least partially address one of the technical problems in related technologies. To this end, the first objective of this application is to propose a method for constructing test tasks for unmanned systems. The method includes: constructing a structured task library and a resource library, and constructing a joint constraint model based on the task library and resource library; determining task combinations based on test objectives and the task library, wherein each task combination includes multiple sub-tasks; performing simulated scheduling based on the resource library, task combinations, and joint constraint model to determine a scheduling feasibility score; determining sub-tasks to be adjusted based on the scheduling feasibility score; iteratively adjusting the sub-tasks to be adjusted based on preset rules until the scheduling feasibility score of the task combination meets the requirements or the number of iterations reaches a preset number, to obtain a target task combination; and outputting a task structure and resource allocation table based on the target task combination, so as to import the task structure and resource allocation table into the unmanned system. This application establishes a joint modeling mechanism for tasks and resources, introduces scheduling simulation and feasibility scoring feedback mechanisms, and combines optimization algorithms such as mixed integer programming and constraint programming to achieve deep integration of task content construction and scheduling resource allocation. This ensures that the generated test tasks are schedulable and have high execution efficiency under real resource constraints, thereby achieving deep integration and optimization of test tasks and resource allocation. This will guarantee the execution efficiency of the unmanned system testing process and improve the system's intelligence level.

[0005] The second objective of this application is to propose an electronic device.

[0006] To achieve the above objectives, the first aspect of this application proposes a method for constructing test tasks for unmanned systems. The method includes: constructing a structured task library and a resource library, and constructing a joint constraint model based on the task library and the resource library; determining a task combination based on the test objective and the task library, wherein the task combination includes multiple sub-tasks; performing simulated scheduling based on the resource library, the task combination, and the joint constraint model to determine a scheduling feasibility score; determining the sub-tasks to be adjusted based on the scheduling feasibility score; iteratively adjusting the sub-tasks to be adjusted based on preset rules until the scheduling feasibility score of the task combination meets the requirements or the number of iterations reaches a preset number, so as to obtain a target task combination; and outputting a task structure and a resource allocation table based on the target task combination, so as to import the task structure and resource allocation table into the unmanned system.

[0007] According to one embodiment of this application, simulated scheduling is performed based on a resource pool, task combination, and joint constraint model to determine a scheduling feasibility score. This includes: during the simulated scheduling process, calculating a score for the percentage of successfully scheduled subtasks, a score for the number of concurrent conflicts, a score for the probability of high-priority tasks being blocked, a score for the maximum resource load rate and idle rate, and a score for the total scheduling time and maximum latency; and determining the scheduling feasibility score based on the score for the percentage of successfully scheduled subtasks and its corresponding first weight coefficient, the score for the number of concurrent conflicts and its corresponding second weight coefficient, the score for the probability of high-priority tasks being blocked and its corresponding third weight coefficient, the score for the maximum resource load rate and idle rate and its corresponding fourth weight coefficient, and the score for the total scheduling time and maximum latency and its corresponding fifth weight coefficient.

[0008] According to one embodiment of this application, the percentage score of successfully scheduled subtasks is obtained based on the following formula:

[0009] Where S represents the percentage score of successfully scheduled subtasks, Ns represents the number of successfully scheduled subtasks, and N represents the number of subtasks.

[0010] According to one embodiment of this application, the concurrent conflict count score is obtained based on the following formula:

[0011] in, The maximum conflict threshold, For smoothing terms, Rate the conflict.

[0012] According to one embodiment of this application, a high-priority task obstruction probability score is obtained based on the following formula:

[0013] Where Fp represents the probability score of high-priority tasks being blocked, Nb represents the number of blocked tasks, and Nh represents the number of high-priority subtasks. This is a smoothing term.

[0014] According to one embodiment of this application, the maximum resource load rate and idle rate scores are obtained based on the following formula:

[0015] Where Fu represents the maximum resource load rate and idle rate score. The target utilization rate is defined by Ui, which represents the utilization rate of each device within the scheduling cycle, and m, which represents the number of critical devices.

[0016] According to one embodiment of this application, the total scheduling duration and the maximum latency score are obtained based on the following formula:

[0017] Among them, F T This represents the total scheduling time and the maximum delay score. Indicates the total scheduling duration. Indicates the maximum task completion delay. Indicates the reference scheduling time threshold. Represents the coefficient. This is a smoothing term.

[0018] According to one embodiment of this application, the preset rules include: determining the corresponding replacement subtask based on the target function of the subtask to be adjusted; or splitting the subtask to be adjusted into multiple first subtasks; and deleting the subtask to be adjusted.

[0019] According to one embodiment of this application, the joint constraint model includes: subtask resource matching constraints, resource exclusivity and concurrency restriction constraints, subtask execution order constraints, subtask time window constraints, test terminal queuing restrictions, and conflict penalty constraints. The subtask resource matching constraints characterize that the resources allocated to each selected subtask belong to the resource set of that subtask type; the resource exclusivity and concurrency restriction constraints characterize that at the same time, the same device is allocated one subtask; the subtask execution order constraints characterize the priority order of subtask execution; the subtask time window constraints characterize that subtasks are completed within a set time window; the test terminal queuing restrictions characterize that each test terminal executes one subtask at any time; and the conflict penalty constraints characterize that when two subtasks conflict, the penalty is minimized using an optimization objective function.

[0020] To achieve the above objectives, a second aspect of this application provides an electronic device, including a memory, a processor, and a program for constructing unmanned system test tasks stored in the memory and executable on the processor. When the processor executes the program for constructing unmanned system test tasks, it implements the aforementioned method for constructing unmanned system test tasks.

[0021] According to the unmanned system test task construction method and electronic device of this application, a structured task library and resource library are constructed, and a joint constraint model is constructed based on the task library and resource library; a task combination is determined based on the test target and the task library, and the task combination includes multiple sub-tasks; simulation scheduling is performed based on the resource library, task combination, and joint constraint model to determine the scheduling feasibility score; the sub-tasks to be adjusted are determined based on the scheduling feasibility score; the sub-tasks to be adjusted are iteratively adjusted based on preset rules until the scheduling feasibility score of the task combination meets the requirements or the number of iterations reaches the preset number, so as to obtain the target task combination; the task structure and resource allocation table are output based on the target task combination, and the task structure and resource allocation table are imported into the unmanned system. This application establishes a joint modeling mechanism for tasks and resources, introduces a scheduling simulation and feasibility score feedback mechanism, and combines optimization algorithms such as mixed integer programming and constraint programming to achieve deep integration of task content construction and scheduling resource configuration, ensuring that the generated test tasks have schedulability and high execution efficiency under real resource constraints, and achieving deep integration and optimization of test tasks and resource configuration, so as to ensure the execution efficiency of the unmanned system test process and the improvement of the system intelligence level. Attached Figure Description

[0022] Figure 1 This is a flowchart illustrating a method for constructing unmanned system test tasks according to some embodiments of this application; Figure 2 This is a block diagram of an electronic device according to some embodiments of this application. Detailed Implementation

[0023] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0024] The following describes in detail, with reference to the accompanying drawings, the method for constructing unmanned system test tasks and the electronic equipment according to embodiments of this application.

[0025] Among related technologies, unmanned system test task scheduling technology has the following key defects, which seriously restrict its application effect and intelligence level in actual complex task system test scenarios.

[0026] First, the task design and scheduling processes are disconnected. Traditional systems typically generate a static task list first, and then independently execute scheduling and resource allocation. This process fails to incorporate scheduling constraints such as resource availability, device capability matching, and time window limitations into the task combination phase. This results in generated test task combinations that may not be actually scheduled for execution, leading to issues such as task delays, conflicts, or interruptions, thus affecting the integrity and efficiency of the test plan.

[0027] Secondly, there is a lack of dynamic prediction and feedback mechanisms for scheduling feasibility. Existing solutions often rely on fixed priorities, static scoring, or empirical rules for scheduling and sorting, which makes it difficult to cope with dynamic environmental conditions such as frequent changes in resource status and complex dependencies between task submodules. They also cannot predict whether the combination of test tasks has high scheduling feasibility in the early stages of task design.

[0028] Secondly, there is a lack of integrated optimization capabilities for task and resource allocation. In complex testing scenarios, there are often strong constraints such as the order of tasks, regional requirements, and device binding between task sub-items. Meanwhile, resource types are heterogeneous and capacity is limited. Traditional scheduling methods lack joint modeling capabilities, making it difficult to achieve global optimization of task selection, time arrangement, and resource allocation, resulting in low resource utilization and insufficient task throughput.

[0029] Based on this, this application proposes a method for constructing unmanned system test tasks. By establishing a joint modeling mechanism for tasks and resources, introducing a scheduling simulation and feasibility scoring feedback mechanism, and combining optimization algorithms such as mixed integer programming and constraint programming, the method achieves a deep integration of task content construction and scheduling resource configuration. This ensures that the generated test tasks have schedulability and high execution efficiency under real resource constraints, thereby achieving a deep integration and optimization of test tasks and resource configuration. This ensures the efficiency of the unmanned system test process and the improvement of the system's intelligence level.

[0030] Figure 1 This is a flowchart illustrating a method for constructing an unmanned system test task according to some embodiments of this application. (Refer to...) Figure 1 The method for constructing an unmanned system test task according to an embodiment of this application may include the following steps: S110, construct a structured task library and resource library, and build a joint constraint model based on the task library and resource library.

[0031] Specifically, building a structured task library involves organizing and storing the various components of a test task according to certain rules and formats to enable efficient management and use of these tasks. Specifically, it is first necessary to define the basic building blocks of each test task. These basic building blocks include the task ID, subtask structure (e.g., multiple subtasks included in each task), task functional objectives, task execution time estimation, task type classification, and task resource requirements (e.g., test area, test equipment, test terminal). By clarifying this key information, each test task can be described in detail and stored in the task library according to a certain structure, thus constructing a structured task library.

[0032] Building a resource repository involves detailed modeling and categorized storage of various resources involved in the test scenario, enabling rapid retrieval and rational allocation of resources during task scheduling and execution. Specifically, the resource repository includes test area models, test device models, time window models, and dependency models. The test area model includes the area ID, maximum concurrency, and supported task types; the test device model covers device ID, device type, whether it is exclusive, and supported sub-task types; the test terminal model, such as a tablet, records the number of terminals, command issuance capabilities, and queuing capacity; the time window model defines the available resource time period and task start / end time requirements; and the dependency model describes the sequential execution relationships between tasks and the mutual exclusion of device occupancy. Through the establishment of these models, the resource repository can comprehensively and accurately reflect the status and constraints of various resources, providing strong resource support and scheduling basis for unmanned system testing tasks.

[0033] Furthermore, information from the task library and resource library is combined, and a joint constraint model is constructed by defining variables and constraints.

[0034] First, define the variables as follows: The task set T = {t1, t2, ..., tn} in the task library contains multiple subtasks sij, where ti represents the i-th task in the task set, sij represents the j-th subtask of the i-th task, dij represents the estimated execution time of the j-th subtask sij of the i-th task, type_ij represents the task type of the j-th subtask sij of the i-th task, A = {a1, a2, ..., ap} represents the test region set, Ck represents the maximum concurrency capacity of region ak, and E = {e1, ..., eq} P represents the set of test devices; P={p1,...,pr} represents the set of test terminals; x_ij∈{0,1} indicates whether the j-th subtask sij of the i-th task is selected, for example, x_ij=1 indicates that the j-th subtask sij of the i-th task is selected, and x_ij=0 indicates that the j-th subtask sij of the i-th task is not selected; st_ij indicates the start time of the j-th subtask sij of the i-th task; res_ij indicates the resources (area, device, terminal) allocated to the j-th subtask sij of the i-th task.

[0035] Then, constraints are defined, such as source matching constraints, resource exclusivity and concurrency limits, subtask execution order constraints, subtask time window constraints, test terminal queuing limits, and conflict penalty constraints for each subtask. Specifically, the source matching constraint for a subtask constrains the resources allocated to it, such as region support type, device support type, and terminal support type; the resource exclusivity and concurrency limits constrain the number of tasks that can be supported by the same device at any given time, and the number of concurrent regional resources that can be supported by the same region at any given time; the subtask execution order constraint constrains the execution order of subtasks; the subtask time window constraint constrains the execution time of subtasks; the test terminal queuing limit constrains the number of tasks executed by each test terminal at any given time; and the conflict penalty constraint is used to optimize the objective function.

[0036] S120 determines task combinations based on test objectives and task libraries. The task combinations include multiple sub-tasks.

[0037] Specifically, the task combinations to be executed are determined based on the test objectives. For example, the subtasks corresponding to the test objectives are determined by querying a two-dimensional mapping table between test objectives and subtasks. This two-dimensional mapping table includes multiple test objectives and the subtasks corresponding to each test objective. Then, the tasks corresponding to each subtask are determined based on the task library, and the corresponding task combinations are determined based on the tasks corresponding to each subtask.

[0038] S130 uses a simulation scheduling model based on a resource pool, task combination, and joint constraint model to determine the scheduling feasibility score.

[0039] Specifically, the system determines the current resource status based on the resource library and invokes the scheduling engine accordingly to allocate resources, time, and execution order for the subtasks in the task combination. If scheduling failure occurs (e.g., scheduling failure, resource conflict, or task delay), the task is marked as unschedulable, and the cause of the conflict is recorded, such as the number of concurrent conflicts, the probability of high-priority tasks being blocked, and the maximum resource load rate and idle rate (for critical equipment or regions). The number of concurrent conflicts refers to the number of times two or more subtasks simultaneously request the same resource (such as equipment or region), and the probability of high-priority tasks being blocked refers to the frequency at which high-priority tasks cannot obtain the required resources in a timely manner.

[0040] Furthermore, by calculating the percentage of successfully scheduled subtasks, the score for the number of concurrent conflicts, the score for the probability of high-priority tasks being blocked, the score for the maximum resource load rate and idle rate, and the score for the total scheduling time and maximum delay, and inputting the percentage of successfully scheduled subtasks, the score for the number of concurrent conflicts, the score for the probability of high-priority tasks being blocked, the score for the maximum resource load rate and idle rate, and the score for the total scheduling time and maximum delay into a preset formula, the scheduling feasibility score of each task in the task combination is output.

[0041] S140, determine the sub-tasks to be adjusted based on the scheduling feasibility score.

[0042] Specifically, after determining the scheduling feasibility score of each task in the task combination, the scheduling feasibility score is compared with a preset score threshold to determine whether the task is a low-feasibility task, and the subtasks corresponding to the low-feasibility task are identified as subtasks to be adjusted. For example, if the scheduling feasibility score of a task is less than the preset score threshold, it means that the task is a low-feasibility task, and the subtasks of the low-feasibility task need to be adjusted; if the scheduling feasibility score of a task is greater than or equal to the preset score threshold, it means that the task is a feasible task, and the subtasks of the feasible task do not need to be adjusted. The preset score threshold can be set according to the actual situation, and no specific restrictions are imposed here.

[0043] S150: Based on preset rules, iteratively adjust the subtasks to be adjusted until the scheduling feasibility score of the task combination meets the requirements or the number of iterations reaches the preset number, so as to obtain the target task combination.

[0044] Specifically, after identifying the subtasks to be adjusted, the subtasks are iteratively adjusted based on preset rules, such as replacing, splitting, and deleting subtasks. After each adjustment, a new task is obtained. The scheduling simulation of the new task is performed, and the feasibility score of the new task is recalculated. If the feasibility score of the new task meets the requirements, such as the feasibility score of the new task being greater than or equal to a preset score threshold, or the number of iterations reaching a preset number, or the tolerance for subtask changes reaching a preset tolerance threshold, the new task is updated to the task combination to be identified as the target task combination, thus completing the initial scheduling feasibility screening of the task combination.

[0045] S160 outputs a task structure and resource allocation table based on the target task combination, so as to import the task structure and resource allocation table into the unmanned system.

[0046] Specifically, after determining the target task combination, the process enters a joint solution phase based on optimization algorithms. This involves constructing an integrated joint optimization model for task selection and resource allocation, comprehensively considering multiple decision variables such as task selection, subtask start time, and resource allocation. Simultaneously, scheduling constraints are set for each subtask, including source matching constraints, resource exclusivity and concurrency limits, subtask execution order constraints, subtask time window constraints, test terminal queuing limits, and conflict penalty constraints. The model aims to maximize the number of schedulable tasks, minimize resource conflicts, maximize resource utilization, and minimize overall scheduling time, constructing a scheduling performance function under multi-objective trade-offs. For different task structures and resource constraints, appropriate optimization methods are selected, such as mixed integer programming for linearly expressible models, constraint programming for handling discrete and complex constraint problems, and reinforcement learning for policy generation in highly dynamic environments. After optimization, the task structure and resource allocation table are output, i.e., a set of optimal task schemes and their detailed scheduling plans, including the execution time period, resource number, and terminal path for each subtask. Finally, the task structure and resource allocation table are imported into the unmanned system for testing to ensure that the target task combination is highly feasible, efficient, and robust under existing resource conditions.

[0047] This application establishes a joint modeling mechanism for tasks and resources, introduces scheduling simulation and feasibility scoring feedback mechanisms, and combines optimization algorithms such as mixed integer programming and constraint programming to achieve deep integration of task content construction and scheduling resource allocation. This ensures that the generated test tasks are schedulable and have high execution efficiency under real resource constraints, thereby achieving deep integration and optimization of test tasks and resource allocation. This will guarantee the execution efficiency of the unmanned system testing process and improve the system's intelligence level.

[0048] In some embodiments, simulated scheduling is performed based on a resource pool, task combination, and joint constraint model to determine a scheduling feasibility score. This includes: during the simulated scheduling process, calculating a score for the percentage of successfully scheduled subtasks, a score for the number of concurrent conflicts, a score for the probability of high-priority tasks being blocked, a score for the maximum resource load rate and idle rate, and a score for the total scheduling time and maximum latency; and determining the scheduling feasibility score based on the score for the percentage of successfully scheduled subtasks and its corresponding first weight coefficient, the score for the number of concurrent conflicts and its corresponding second weight coefficient, the score for the probability of high-priority tasks being blocked and its corresponding third weight coefficient, the score for the maximum resource load rate and idle rate and its corresponding fourth weight coefficient, and the score for the total scheduling time and maximum latency and its corresponding fifth weight coefficient.

[0049] Specifically, during the simulated scheduling process, the percentage of successfully scheduled subtasks is scored based on the following formula:

[0050] Where S represents the percentage score of successfully scheduled subtasks, Ns represents the number of successfully scheduled subtasks, and N represents the number of subtasks.

[0051] The score for the number of concurrent conflicts is obtained based on the following formula:

[0052] in, The maximum conflict threshold, For smoothing terms, Conflict scoring is performed. Among these, the maximum conflict threshold is... It can be dynamically adjusted based on the scheduling and allocation situation each time. The closer to 1, the fewer the conflicts.

[0053] The probability score of high-priority tasks being blocked is obtained based on the following formula:

[0054] Where Fp represents the probability score of high-priority tasks being blocked, Nb represents the number of blocked tasks, and Nh represents the number of high-priority subtasks. The closer Fp is to 1, the more smoothly the high-priority tasks are completed. This is a smoothing term.

[0055] The maximum resource load rate and idle rate scores are obtained based on the following formula:

[0056] Where Fu represents the maximum resource load rate and idle rate score. The target utilization rate is defined by Ui, which represents the utilization rate of each device within the scheduling cycle, and m, which represents the number of critical devices. It can be 90%.

[0057] The total scheduling time and maximum latency score are obtained based on the following formulas:

[0058] Among them, F T This represents the total scheduling time and the maximum delay score. Indicates the total scheduling duration. Indicates the maximum task completion delay. Indicates the reference scheduling time threshold. For smoothing terms, Represents the coefficient. To determine the relative importance of total scheduling time and maximum latency, It can be between 0.5 and 2.0. F T A higher score indicates a more compact scheduling and lower latency.

[0059] After determining the scores for the percentage of successfully scheduled subtasks, the number of concurrent conflicts, the probability of high-priority tasks being blocked, the maximum resource load and idle rate, and the total scheduling time and maximum latency, input these scores into the following formula to determine the scheduling feasibility score of the task combination:

[0060] in, Indicates the scheduling feasibility score; Indicates the first weighting coefficient; The score indicates the number of concurrent conflicts. This represents the second weighting coefficient; This indicates a score representing the probability of high-priority tasks being blocked. This represents the third weighting coefficient; This indicates the maximum resource load rate and idle rate score; This represents the fourth weighting coefficient; This represents the total scheduling time and the maximum delay score. This represents the fifth weighting coefficient.

[0061] In some embodiments, a corresponding replacement subtask is determined based on the target function of the subtask to be adjusted; or the subtask to be adjusted is split into multiple first subtasks; or the subtask to be adjusted is deleted.

[0062] Specifically, the degree of conflict for each subtask to be adjusted can be evaluated using the following formula:

[0063] Among them, C j This refers to task t j The degree of conflict, R j It is task t j Dependent on the set of resources This is based on the current or historical conflict rate. The subtask with the highest conflict level is then replaced. For example, based on the target function of the subtask with the highest conflict level, a replacement subtask is determined, and this replacement subtask replaces the subtask with the highest conflict level. For instance, a two-dimensional mapping between the target function and subtasks can be used to identify the subtask with the same target function as the subtask to be adjusted as the replacement subtask. This two-dimensional mapping includes multiple subtasks and the corresponding function for each subtask.

[0064] If the execution time of the subtask to be adjusted exceeds a preset time threshold, the subtask is determined to be too long. If the resource utilization rate of the subtask to be adjusted exceeds a preset resource utilization rate threshold, the resource utilization of the subtask to be adjusted is determined to be concentrated. If the subtask to be adjusted is determined to be long and has concentrated resource utilization, it can be split into multiple first subtasks, for example, subtask t... j Split into multiple first subtasks {t j1 , t j2 The first subtask is shorter than the subtask to be adjusted, which reduces the concentrated load on resources.

[0065] Furthermore, if the scheduling feasibility score of the task combination still does not meet the requirements after the subtask to be adjusted is adjusted, or if the adjustment cost is too high, the subtask to be adjusted can be deleted, which to a certain extent ensures that the task finally used for testing has high schedulability and execution stability.

[0066] In some embodiments, the joint constraint model includes: subtask resource matching constraints, resource exclusivity and concurrency limit constraints, subtask execution order constraints, subtask time window constraints, test terminal queuing limits, and conflict penalty constraints. The subtask resource matching constraints characterize that the resources allocated to each selected subtask belong to the resource set of that subtask type; the resource exclusivity and concurrency limit constraints characterize that at the same time, the same device is allocated one subtask; the subtask execution order constraints characterize the priority order of subtask execution; the subtask time window constraints characterize that subtasks are completed within a set time window; the test terminal queuing limits characterize that each test terminal executes one subtask at any time; and the conflict penalty constraints characterize that when two subtasks conflict, the penalty is minimized using the optimization objective function.

[0067] Specifically, the subtask resource matching constraint is used to characterize the set of resources allocated to each selected subtask that belongs to that subtask type. The expression for the subtask resource matching constraint is as follows:

[0068] Where xij=1 represents the j-th subtask in the task combination; Aij represents the set of test regions for the j-th subtask in the task combination, ak represents the test region of xij; Eij represents the set of test devices for the j-th subtask in the task combination; el represents the test device of xij; Pij represents the set of test terminals for the j-th subtask in the task combination, pm represents the test terminal of xij.

[0069] Resource exclusivity and concurrency constraints are used to characterize the situation where the same device is assigned a subtask at the same time. Specifically, resource exclusivity constraint means that for the same time t, the same device... The limit on the number of tasks that can be assigned, at the same time t, on the same device When assigned to a single task, the resource exclusivity and concurrency constraints are expressed as follows:

[0070] in, =1 indicates a subtask In time Occupying equipment .

[0071] Concurrency constraints refer to the number of tasks that can run concurrently in the same region ak at the same time t. The expression for concurrency constraints is as follows:

[0072] in, Indicates time The number of subtasks occupying the AK area. This indicates the maximum concurrency capacity of region ak.

[0073] Subtask execution order constraints are used to characterize the priority order of subtask execution in task composition. subtasks Must precede In the case of execution, the subtask execution order constraint expression is as follows:

[0074] in, Indicates task combination A subtask; Indicates task combination Another subtask; Subtasks The estimated execution time.

[0075] Subtask time window constraints are used to characterize the completion of a subtask within a specified time window. Need to be in the window If the task is completed within a certain timeframe, the expression for the subtask time window constraint is as follows:

[0076] in, Indicates a subtask; Indicates the lower limit of the window; Indicates the upper limit of the window; Subtasks The estimated execution time.

[0077] The test terminal queuing limit is used to characterize the execution of a subtask by each test terminal at any given time. The expression for the test terminal queuing limit is as follows:

[0078] in, =1 indicates a subtask In time The test terminal was occupied. .

[0079] Conflict penalty constraints are used to characterize the minimization of penalties when two subtasks conflict, using the optimization objective function. Conflict penalty constraints are used to optimize the objective function and define variables. This represents a conflict caused by overlapping resources between two subtasks. The expression for minimizing the penalty in the objective function is as follows:

[0080] Where ij represents the j-th subtask in the i-th task, and kl represents the l-th subtask in the k-th task.

[0081] As a concrete example, the system for constructing unmanned system test tasks includes a task content combination module, a resource scheduling simulation module, a scheduling feasibility feedback module, and an optimization output module. By deeply integrating task content design with resource scheduling, it achieves collaborative optimization and coordinated scheduling of test tasks and resource solutions, thereby improving the task execution efficiency and resource utilization efficiency of the unmanned system test field.

[0082] Based on the task knowledge base, the task content construction module dynamically generates multiple task combinations from the candidate subtask set according to the preset functional objectives, task capability models, and testing requirements of the unmanned system test tasks. The task combinations not only consider the completeness of the capability dimensions covered, but also use parameters such as preliminary estimates of execution time and resource requirements as inputs for subsequent scheduling and evaluation.

[0083] The resource scheduling simulation module performs rapid scheduling simulations based on the sub-task requirements within a task combination, utilizing test field resource modeling information (such as the number of concurrent connections in the test area, available time periods for devices, and task types supported by the test terminal). This module employs a scheduling constraint modeling language to ensure that the execution of each sub-task does not conflict with other task resources, while also satisfying rules such as device exclusivity and area restrictions. The scheduling simulation uses an event-driven discrete-time scheduling method, combined with an optimization engine, to output a schedulability score for each task combination.

[0084] The scheduling feasibility feedback module is used to provide real-time feedback of scheduling simulation results to the task selection module. The system introduces scheduling feasibility scoring during the task selection phase, employing a dynamic scoring weighting mechanism to eliminate or replace task combinations with severe scheduling conflicts or excessive resource consumption, thereby significantly reducing the subsequent scheduling failure rate.

[0085] Finally, the output module is optimized with the objective function of maximizing scheduling feasibility. While ensuring task coverage and diversity requirements, a joint solution algorithm is used to output the task test paper and resource allocation scheme in an integrated manner. The output results include: the test task structure (including task combinations and sub-task sequences) and the resource allocation plan (including test area, equipment, and terminal resource allocation tables), which can be directly imported into the test execution system or the main control platform.

[0086] In summary, this application provides a dynamic task content generation mechanism for unmanned systems under test, based on their capability models and evaluation requirements. This mechanism supports the dynamic combination of functional characteristics, task capability models, and specific test objectives (such as navigation stability, path planning capability, and environmental perception accuracy) of different types of unmanned systems (e.g., unmanned vehicles, drones, and unmanned special platforms). This avoids the problem of traditional static task libraries failing to cover actual evaluation needs, achieving a "capability-driven, goal-oriented" task assembly process and significantly improving the relevance, adaptability, and scalability of task content. Furthermore, regarding the scheduling feasibility prediction mechanism for real test sites and resource conditions, this application uses a simulation scheduling module to evaluate the execution feasibility of tasks under current test resource constraints (e.g., test area, equipment, data terminal) in real time. It quantifies indicators such as conflict probability, equipment resource obstruction probability, and critical path time delay, generating a "scheduling feasibility score." This allows for the early identification of potential execution conflicts in the test plan, ensuring high scheduling success rate and execution efficiency of the task scheme in real unmanned system test scenarios. To address the closed-loop linkage mechanism of task generation and scheduling evaluation, this application introduces a mechanism that uses scheduling evaluation results to guide the selection and adjustment of task content. This mechanism can automatically eliminate severely conflicting or unschedulable task combinations, reconstruct or replace key sub-tasks, and form multiple rounds of iterative optimization. This ensures that the final output task has strong resource adaptability and execution robustness, overcomes the risk of implementation failure caused by the disconnect between traditional task generation and scheduling, and better meets the testing needs of complex and ever-changing unmanned systems.

[0087] Corresponding to the above embodiments, this application also proposes an electronic device.

[0088] See Figure 2 As shown, the electronic device 300 of this application includes a memory 310, a processor 320, and a construction program for an unmanned system test task stored in the memory 310 and capable of running on the processor 320. When the processor executes the construction program for the unmanned system test task, it implements the aforementioned method for constructing the unmanned system test task.

[0089] It should be noted that the above-described embodiments and explanations of the beneficial effects of the method for constructing unmanned system test tasks also apply to the electronic devices in the embodiments of this application. To avoid redundancy, they will not be elaborated in detail here.

[0090] It should be noted that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0091] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0092] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0093] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0094] In this application, unless otherwise expressly specified and limited, the terms "installation," "connection," "joining," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components, unless otherwise expressly limited. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0095] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A method for constructing a test task of an unmanned system, characterized in that, The method comprises: constructing a structured task library and a resource library, and constructing a joint constraint model based on the task library and the resource library; determining a task combination based on a test target and the task library, the task combination comprising a plurality of subtasks; performing simulated scheduling based on the resource library, the task combination, and the joint constraint model to determine a scheduling feasibility score; determining a to-be-adjusted subtask according to the scheduling feasibility score; iteratively adjusting the to-be-adjusted subtask based on a preset rule until the scheduling feasibility score of the task combination meets a requirement or the number of iterations reaches a preset number, to obtain a target task combination; outputting a task structure and a resource allocation table based on the target task combination, to import the task structure and the resource allocation table into the unmanned system.

2. The method of claim 1, wherein, The simulated scheduling based on the resource library, the task combination, and the joint constraint model to determine a scheduling feasibility score comprises: during the simulated scheduling, calculating a scheduling success subtask proportion score, a concurrent conflict number score, a high-priority task blocking probability score, a maximum resource load rate and idle rate score, and a scheduling total time length and maximum delay score; determining the scheduling feasibility score based on the scheduling success subtask proportion score and a corresponding first weight coefficient, the concurrent conflict number score and a corresponding second weight coefficient, the high-priority task blocking probability score and a corresponding third weight coefficient, the maximum resource load rate and idle rate score and a corresponding fourth weight coefficient, and the scheduling total time length and maximum delay score and a corresponding fifth weight coefficient.

3. The method of claim 2, wherein, The scheduling success subtask proportion score is obtained based on the following formula: wherein S represents the scheduling success subtask proportion score, Ns represents the number of successfully scheduled subtasks, and N represents the number of subtasks.

4. The method of claim 2, wherein, The concurrent conflict number score is obtained based on the following formula: wherein, is a maximum conflict threshold, is a smoothing term, is a conflict score.

5. The method of claim 2, wherein, The high-priority task blocking probability score is obtained based on the following formula: wherein Fp represents the high-priority task blocking probability score, Nb represents the number of blocked tasks, Nh represents the number of high-priority sub-tasks, is a smoothing term.

6. The method of claim 2, wherein, The maximum resource load rate and idle rate score is obtained based on the following formula: wherein Fu denotes the maximum resource load rate and idle rate score, wherein Uti denotes the utilization of each device in the scheduling period, and m denotes the number of critical devices.

7. The method of claim 2, wherein, The scheduling total time length and maximum delay score is obtained based on the following formula: wherein F T represents the total scheduling duration and the maximum delay score, represents the total scheduling duration, represents the maximum task completion delay, represents the reference scheduling time threshold, represents the coefficient, is a smoothing term.

8. The method of claim 1, wherein, The preset rule comprises: determining a corresponding replacement subtask based on a target function of the to-be-adjusted subtask; or splitting the to-be-adjusted subtask into a plurality of first subtasks; deleting the to-be-adjusted subtask.

9. The method of claim 1, wherein, The joint constraint model comprises a subtask resource matching constraint, a resource exclusive and concurrent restriction constraint, a subtask execution order constraint, a subtask time window constraint, a test terminal queuing restriction, and a conflict penalty constraint, wherein the subtask resource matching constraint is used to represent that the resource allocated to each selected subtask belongs to the resource set of the subtask type; the resource exclusive and concurrent restriction constraint is used to represent that at the same time, the same device is allocated a subtask; the subtask execution order constraint is used to represent the priority order of subtask execution; the subtask time window constraint is used to represent that a subtask is completed within a set time window; the test terminal queuing restriction is used to represent that each test terminal executes one subtask at any time; and the conflict penalty constraint is used to represent that a conflict between subtasks is penalized. The conflict penalty constraint is used to represent that when two sub-tasks conflict, a penalty minimization is performed using an optimization objective function.

10. An electronic device, comprising: The application further provides a computer readable storage medium storing the construction program of the unmanned system test task, and the construction program of the unmanned system test task is stored in the memory and can be run on the processor. When the processor executes the construction program of the unmanned system test task, the construction method of the unmanned system test task according to any one of claims 1-9 is implemented.

Images