A multi-load unmanned aerial vehicle system for simulated combat training
By dividing the training area into tactical semantic units and constructing the payload action function, a net tactical action field is generated, which solves the problem of occlusion modeling and simulation of multi-payload UAV systems in complex environments, and realizes high-precision tactical execution and cooperative control.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- ANHUI ZHONGKE XIANGYU INNOVATION TECH CO LTD
- Filing Date
- 2026-04-25
- Publication Date
- 2026-06-30
Smart Images
Figure CN122308462A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of unmanned aerial vehicle technology, and more specifically, to a multi-payload unmanned aerial vehicle system for simulating combat training. Background Technology
[0002] Existing multi-payload unmanned aerial vehicle (UAV) systems for simulated combat training mainly suffer from the following problems: With the widespread application of unmanned aerial vehicles (UAVs) in military training and combat missions, multi-payload UAV systems play a crucial role in simulated combat training, tactical assessment, and adversarial exercises. Existing multi-payload UAV simulation systems typically base mission planning and effectiveness evaluation on the UAV's flight parameters, payload characteristics, and target space information. However, these systems suffer from technical limitations in complex training environments, restricting the realism and decision-making effectiveness of the simulation systems.
[0003] Existing technologies struggle to provide unified, quantifiable, and reusable modeling of occlusion effects in complex 3D training environments. Traditional UAVs or target perception systems often employ coarse region judgments or single occlusion factors when considering environmental occlusion, neglecting the structural characteristics of the environment and the geometric distribution of spatial units. This makes it difficult to accurately describe the degree of occlusion when ray paths traverse different types of obstacles such as buildings, terrain, and vegetation. Path occlusion calculations lack a unified platform and quantification mapping method, making it impossible to accurately assess the comprehensive occlusion impact after a path traverses various spatial units, and also failing to consider the attenuation differences of different spatial medium properties.
[0004] When evaluating the effect of UAV payloads on a target area, existing systems typically only consider the payload's baseline strength, distance attenuation, or simple directional factors, failing to distinguish the environmental semantic categories of the target space, such as built-up areas, open areas, airspace, or electromagnetically sensitive areas. This leads to convergent calculations of the effect of the same payload in different types of spatial units, failing to truly reflect the differentiated responses of the payload under different environmental conditions.
[0005] Furthermore, in existing technologies, occlusion factors and target attribute factors are usually treated in isolation. Occlusion is only used as a simple distance correction or visibility judgment condition, which cannot accurately characterize the actual effective intensity of the payload under complex spatial path conditions. At the same time, most multi-payload UAV systems rely on fixed rules or manually set empirical parameters, lacking the ability to simulate and extrapolate adversarial scenarios and strategies under multiple types of payloads and tactical environments.
[0006] Existing technologies cannot closely approximate real battlefield conditions at the simulation level, which limits the application value of multi-payload unmanned aerial vehicle systems in simulated combat training, tactical situation analysis, and confrontation strategy evaluation. There is an urgent need for a simulation system that can comprehensively consider environmental obstruction, tactical semantic attributes, and the characteristics of multiple payloads.
[0007] In view of this, the present invention proposes a multi-payload unmanned aerial vehicle system for simulating combat training to solve the above problems. Summary of the Invention
[0008] To overcome the aforementioned deficiencies of the prior art and to achieve the above objectives, the present invention provides the following technical solution: a multi-payload unmanned aerial vehicle system for simulating combat training, comprising: The tactical environment construction module is used to divide the training area into spatial units with tactical semantic attributes, and calculate the environmental occlusion coefficient of each spatial unit based on spatial geometric relationships and occlusion relationships to form a set of tactical semantic units. The adversarial capability analysis module is used to construct a load action function based on the load of each drone in the different opposing factions, obtain the intensity of the load's effect on the tactical semantic unit, and superimpose all load effects to obtain the tactical action field of both sides. The rule coupling cancellation module is used to map the tactical action fields of both sides to the same rule confrontation framework; according to the preset rule coupling relationship, the actions of both sides are cancelled to generate a unified net tactical action field; The tactical situation evolution module is used to update the tactical state of each spatial unit within the spatial domain defined by the set of tactical semantic units through the net tactical action field, forming a tactical state field that changes continuously over time. The trimming and collaborative scheduling module is used to construct multi-dimensional rule functions based on the tactical state field, determine the rule consistency of UAV payload target combinations, determine the UAV collaborative scheduling scheme, and realize the collaborative control of multi-payload UAVs in simulated combat training.
[0009] Preferably, the method for obtaining the spatial unit includes: The spatial geometric information of the training area is acquired and then discretized according to a preset spatial resolution to form different spatial units. The spatial geometric information includes the terrain outline information, building distribution information, obstacle location information, and height information of the training area. Spatial units are constructed using regular voxels, with each spatial unit corresponding to a specific spatial range within the training area; each spatial unit is configured with corresponding spatial location parameters, spatial occupancy state parameters, and environmental medium attribute parameters. Based on spatial location parameters and spatial occupancy status parameters, basic tactical semantic annotations are performed on spatial units, enabling them to possess semantic attributes for tactical analysis, thereby forming spatial units with tactical semantic attributes.
[0010] Preferably, the method for forming the set of tactical semantic units includes: All spatial units are uniformly organized to form a spatial unit set; based on the interaction direction between the UAV and the target spatial unit, a ray path is constructed from the current position of the UAV to the target spatial unit; based on the geometric interaction relationship between the ray path and the spatial units, the spatial units traversed by the ray path are determined one by one, and all spatial units actually traversed by the ray path are counted to form a path voxel set. Based on the environmental medium properties of each spatial unit, the occlusion effect is cumulatively calculated according to the traversal length of the ray path within the corresponding spatial unit to obtain the cumulative path occlusion amount. The cumulative path occlusion amount is converted into an environmental occlusion coefficient according to a preset exponential decay law, and the spatial unit is tactically semantically divided according to the size range of the environmental occlusion coefficient, dividing the spatial unit into a visible area, a strong occlusion area, and a weak occlusion area. A first threshold and a second threshold for environmental occlusion coefficient are preset, with the first threshold being greater than the second threshold. When the environmental occlusion coefficient is greater than or equal to the first threshold, the corresponding spatial unit is determined to have a low degree of attenuation of the load effect, and the spatial unit is classified as a visible area. When the environmental occlusion coefficient is greater than or equal to the second threshold but less than the first threshold, the spatial unit is determined to have a partial attenuation of the load effect, and the spatial unit is classified as a weakly occluded area. When the environmental occlusion coefficient is less than the second threshold of the preset environmental occlusion coefficient, it is determined that the spatial unit has a high degree of attenuation of the load effect, and the spatial unit is divided into a strong occlusion zone; and spatial units with the same tactical semantic attributes are combined according to spatial adjacency to form a set of tactical semantic units. Preferably, the method for obtaining the intensity of the load's effect on the tactical semantic unit includes: After identifying the different factions involved in the confrontation and the drones included in each faction, the payload type, payload reference strength parameters, and payload main direction parameters of each drone are read; at the same time, the environmental occlusion coefficient is called, and the semantic category information of the tactical semantic unit is read. The semantic category response factor is obtained according to the correspondence between payload type and semantic category information. Based on the payload type, payload reference strength parameters, payload action direction parameters, and semantic category response factors carried by the UAV, a payload action function is constructed to calculate the effect strength of each type of payload of each UAV on each tactical semantic unit.
[0011] Preferably, the method for obtaining the tactical action field of both sides includes: At the same time, using tactical semantic units as spatial computing primitives, the intensity of the effect of various payloads on each tactical semantic unit is calculated for all drones belonging to the same faction, and the process is traversed according to the drone index and payload type index. The intensity of the same type of payload is accumulated in the functional category dimension. The intensity of the same type of payload generated by different UAVs in the same camp at the same tactical semantic unit is aggregated to obtain the payload component of the camp at the corresponding tactical semantic unit. The payload components of our own faction at the corresponding tactical semantic units are organized according to the tactical semantic unit index to obtain our own tactical action field in each tactical semantic unit; at the same time, the payload components of the enemy faction at the corresponding tactical semantic units are also organized according to the tactical semantic unit index to obtain the enemy's tactical action field in each tactical semantic unit.
[0012] Preferably, the method for mapping the tactical action fields of both sides to the same rule-based adversarial framework includes: The payload action components corresponding to each tactical semantic unit in the friendly tactical action field and the enemy tactical action field are mapped to the rule domain. Based on the preset correspondence table between payload function categories and rule function domains, the payload action components of different payload function categories are converted into different rule function domains. The load components of both sides in the corresponding rule functional domains are normalized; using the tactical semantic unit index as a unified spatial coordinate, a one-to-one correspondence is established between the rule functional domains of both sides and the rule functional domains of the enemy at the same tactical semantic unit index, thereby completing the mapping of the tactical action fields of both sides to a unified rule confrontation framework.
[0013] Preferably, the method for generating a uniform net tactical action field includes: For each tactical semantic unit, a friendly rule functional domain and an enemy rule functional domain are constructed respectively. A pre-established rule coupling relationship matrix is introduced, and the adversarial influence intensity between different rule functional domains is characterized by the rule coupling relationship matrix. The coupling coefficients preset in the rule coupling relationship matrix describe the weight of the weakening, suppression or enhancement effect of one rule functional domain on another rule functional domain; at the same tactical semantic unit, the rule action components of the friendly and the enemy in the same rule functional domain are canceled out in the same domain, and the net rule action component of the rule functional domain is determined according to the difference in rule strength between the two sides. Based on the rule coupling relationship matrix, the net rule action components of each rule functional domain are mapped to other associated rule functional domains according to the corresponding coupling coefficients, and cross-rule functional domain coupling operations are performed to obtain the modified rule action components that include cross-domain effects. The rule functional domain components after in-domain cancellation and cross-domain coupling are recombined to form the net rule functional domain vector of the tactical semantic unit. Based on the preset reverse mapping relationship between the rule functional domain and the tactical action dimension, the net rule functional domain vector is converted into the corresponding net tactical action component. The process is repeated for all tactical semantic units to obtain the net tactical action distribution covering all tactical semantic units, thereby constructing a unified net tactical action field.
[0014] Preferably, the method for forming a tactical state field that changes continuously over time includes: Within the spatial domain defined by the set of tactical semantic units, a tactical state vector is established for each tactical semantic unit. At each calculation moment, the net tactical action component of the corresponding tactical semantic unit is read, and the net tactical action component is used as the state-driven input. The net tactical action component is converted into a tactical state increment through a preset state response function. The state response function includes a response sensitivity parameter, a saturation limit parameter, and a recovery decay parameter; the tactical state increment is recursively superimposed with the tactical state vector of the previous moment to obtain the tactical state vector of the current moment. Simultaneously, based on the preset spatial adjacency relationship, a weighted diffusion calculation is performed on the tactical state vector differences between adjacent tactical semantic units, and all tactical semantic units are processed repeatedly in time order to form a tactical state field that evolves continuously over time.
[0015] Preferably, the method for determining the UAV collaborative scheduling scheme includes: At the current moment, obtain the tactical state vector of each tactical semantic unit and use the tactical state vector as the input for rule calculation; construct the UAV-payload-target combination for any UAV, available payload type and candidate tactical semantic unit; A multidimensional rule evaluation function is constructed. By weighting the components of each dimension of the tactical state vector of the corresponding tactical semantic unit with the rule weights corresponding to the payload, the rule score of the UAV-payload-target combination is obtained. The rule score is compared with a preset rule score threshold. If the rule score meets the preset rule score threshold, the combination is determined to be a rule-consistent combination; otherwise, it is determined to be a rule-conflicting combination and is removed. Within a set of combinations that meet the preset rule scoring thresholds, the allocation relationship between UAVs, payloads, and tactical semantic units is optimized by using rule scoring as the optimization objective; each UAV is limited to executing only one payload target task within the same scheduling cycle; and the optimization result is determined as the UAV collaborative scheduling scheme for the current moment.
[0016] Preferably, the method for achieving coordinated control of multi-payload UAVs in simulated combat training includes: After obtaining the UAV collaborative scheduling scheme, the payload target allocation information of each UAV is sent to the UAV control terminal, instructing the UAVs to perform operations according to the planned payload task type, direction of action, and target tactical semantic unit; at the same time, UAV status information, payload output status information, and environmental feedback information are collected in real time, input into the tactical situation evolution module, and the tactical state field is updated. Based on the updated tactical state field, the trimming and collaborative scheduling module performs rule consistency judgment and collaborative optimization on the combination of UAV payload targets according to a preset cycle, dynamically adjusts the UAV flight path, payload usage order and task allocation, and realizes collaborative control of multi-payload UAVs in simulated combat training.
[0017] Compared with the prior art, the present invention has the following beneficial effects: This invention achieves refined spatial modeling of ray paths by dividing the environmental space into unified spatial units and determining each spatial unit traversed by the UAV ray path to form a path voxel set. Simultaneously, by introducing path occlusion accumulation and environmental occlusion coefficient, the occlusion impact under different paths, directions, and spatial units is transformed into a unified numerical index. This enables comprehensive analysis of multi-obstacle occlusion in complex 3D environments, achieving quantitative comparison and ranking of occlusion levels. It avoids the problems of vague and incomparable occlusion descriptions caused by lighting, background interference, or terrain complexity in existing technologies. Tactical semantic spatial partitioning provides clear operational guidance for UAV mission planning, enhancing the UAV's tactical execution capabilities in complex environments. Through the unified organization and attribute representation of spatial units, path occlusion assessment and tactical semantic partitioning results can be automatically generated without the need for manual experience rules, making it suitable for different mission scenarios and environmental complexities. The quantified path occlusion information and tactical semantic unit set can assist UAVs in path planning, obstacle avoidance strategy selection, and observation position deployment, improving mission success rate and reducing the risk of being detected or interfered with.
[0018] By introducing semantic category response factors and based on a preset correspondence rule table, the effects of different loads in different semantic environments are adjusted, enabling the same load to exhibit differentiated equivalent intensity in different tactical semantic units such as building, terrain, airspace, and personnel activity areas, significantly improving the realism and precision of load effect assessment. An environmental occlusion coefficient is introduced into the load effect function, and through an exponential decay term, it participates in the intensity calculation along with the semantic category response factor, achieving a coupled expression of occlusion attenuation effects and spatial semantic response characteristics. This allows the load effect intensity to simultaneously reflect the degree of spatial path occlusion and target environmental sensitivity, overcoming the problem of separate treatment of these two aspects in existing technologies. By establishing a preset correspondence rule table between load types and tactical semantic categories, and parameterizing it in the form of response weights, this invention can flexibly adapt to different load functional attributes and different tactical environment requirements, supporting rule expansion and parameter adjustment, and is suitable for simulation applications in multi-load, multi-faction, and multi-tactical scenarios. The load action function comprehensively considers the load reference strength, action direction, spatial distance attenuation, occlusion index attenuation, and semantic category response factor, making the load's effect on the tactical semantic unit closer to the real physical action process, thereby significantly improving the simulation accuracy and reliability of multi-load UAV systems in actual combat training simulation, confrontation assessment, and tactical simulation. Attached Figure Description
[0019] Figure 1 This is a schematic diagram of a multi-payload unmanned aerial vehicle system for simulating combat training according to the present invention; Figure 2 This is a schematic diagram of a multi-payload UAV method for simulating combat training according to the present invention. Detailed Implementation
[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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 skilled in the art without creative effort are within the scope of protection of the present invention. Example
[0021] Please see Figure 1 As shown, this embodiment provides a multi-payload unmanned aerial vehicle system for simulating combat training, specifically including the following steps: The tactical environment construction module is used to divide the training area into spatial units with tactical semantic attributes, and calculate the environmental occlusion coefficient of each spatial unit based on spatial geometric relationships and occlusion relationships to form a set of tactical semantic units. The adversarial capability analysis module is used to construct a load action function based on the load of each drone in the different opposing factions, obtain the intensity of the load's effect on the tactical semantic unit, and superimpose all load effects to obtain the tactical action field of both sides. The rule coupling cancellation module is used to map the tactical action fields of both sides to the same rule confrontation framework; according to the preset rule coupling relationship, the actions of both sides are cancelled to generate a unified net tactical action field; The tactical situation evolution module is used to update the tactical state of each spatial unit within the spatial domain defined by the set of tactical semantic units through the net tactical action field, forming a tactical state field that changes continuously over time. The trimming and collaborative scheduling module is used to construct multi-dimensional rule functions based on the tactical state field, determine the rule consistency of UAV payload target combinations, determine the UAV collaborative scheduling scheme, and realize the collaborative control of multi-payload UAVs in simulated combat training.
[0022] Methods for obtaining spatial units include: The spatial geometric information of the training area is acquired and then discretized according to a preset spatial resolution to form different spatial units. The spatial geometric information includes the terrain outline information, building distribution information, obstacle location information, and height information of the training area. It should be noted that, based on the three-dimensional spatial coordinate system of the training area, spatial resolution parameters are set, and a regular grid is constructed in three-dimensional space. The training area is divided into equally spaced segments along the three coordinate axes. Each grid cube is defined as a spatial unit, and a unique spatial index number is assigned to each spatial unit. Spatial units are constructed using regular voxels, with each spatial unit corresponding to a specific spatial range within the training area; each spatial unit is configured with corresponding spatial location parameters, spatial occupancy state parameters, and environmental medium attribute parameters. Among them, the spatial position parameter is used to describe the relative positional relationship of the corresponding spatial unit in the training area; the spatial occupancy status parameter is used to indicate whether the corresponding spatial unit is occupied by buildings, terrain or other physical structures; the environmental medium attribute parameter is used to characterize the ability of the corresponding spatial unit to influence UAV reconnaissance, communication or action path, and can be set according to the physical structure type or spatial characteristics corresponding to the spatial unit. For example, for spatial units inside buildings, the environmental medium attribute parameter can be set to strong shading ability; for spatial units in open spaces, the environmental medium attribute parameter can be set to weak shading ability; and for spatial units in woodlands or complex terrains, the shading ability can be set to medium. Based on spatial location parameters and spatial occupancy status parameters, basic tactical semantic annotations are performed on spatial units, enabling them to possess semantic attributes for tactical analysis, thereby forming spatial units with tactical semantic attributes.
[0023] It should be noted that basic tactical semantic annotation is used to perform preliminary classification of spatial units from a tactical perspective, enabling spatial units to possess semantic attributes that can be directly used in tactical analysis; basic tactical semantic annotation includes passable areas, obscured areas, potential threat areas, and impassable areas; For example: Space units located above open roads and not occupied by physical structures can be marked as passable areas for UAV path planning, formation maneuvers, or mission switching; space units located behind buildings and occupied by physical structures can be marked as shielded areas because the physical structures block external views or communication paths, for training UAVs in covert flight and evasive behavior under hostile reconnaissance conditions; space units located on the top of high-rise buildings or near high points in mountains have strong line-of-sight and coverage capabilities, and can be marked as potential threat areas to simulate the deployment areas of enemy radar, surveillance equipment, or jamming devices; and narrow space units located inside physical structures of buildings, inside mountains, or with limited height can be marked as impassable areas to prohibit UAVs from entering.
[0024] Methods for forming a set of tactical semantic units include: All spatial units are uniformly organized to form a spatial unit set; for the interaction direction between the UAV and the target spatial unit, a ray path is constructed from the current position of the UAV to the target spatial unit; based on the geometric interaction relationship between the ray path and the spatial unit, the spatial units traversed by the ray path are determined one by one, and all spatial units actually traversed by the ray path are counted to form a path voxel set, which is used to limit the spatial range of occlusion calculation. Based on the environmental medium properties of each spatial unit, the occlusion effect is cumulatively calculated according to the traversal length of the ray path within the corresponding spatial unit to obtain the cumulative path occlusion amount, which is used to characterize the comprehensive occlusion degree of the ray path during its traversal of the spatial unit. Specifically, the cumulative path occlusion is calculated as follows: ;in, Indicates ray path The cumulative amount of path occlusion; Indicates from drones Pointing to tactical semantic units ray path; Indicates ray path The set of path voxels consisting of all spatial units actually traversed; Indicates the index number of the drone; Indicates the index number of the tactical semantic unit; Indicates the index number of the spatial unit in the path voxel set; Indicates the first The unit length attenuation coefficient of each spatial unit is used to distinguish the differences in shading intensity between different spatial units, such as buildings, terrain, and airspace. Indicates the ray path at the 1st... The length of travel within a spatial unit; The cumulative path occlusion is converted into an environmental occlusion coefficient according to a preset exponential decay law. Based on the size range of the environmental occlusion coefficient, the spatial unit is divided into a visible area, a strong occlusion area, and a weak occlusion area. A first threshold and a second threshold for the environmental occlusion coefficient are preset, and the first threshold is greater than the second threshold. When the environmental occlusion coefficient is greater than or equal to the first preset environmental occlusion coefficient threshold, it is determined that the corresponding spatial unit has a low degree of attenuation of the load effect, and the spatial unit is divided into a visible area; when the environmental occlusion coefficient is greater than or equal to the second preset environmental occlusion coefficient threshold and less than the first preset environmental occlusion coefficient threshold, it is determined that the spatial unit has a partial attenuation of the load effect, and the spatial unit is divided into a weak occlusion area. When the environmental occlusion coefficient is less than the second threshold of the preset environmental occlusion coefficient, it is determined that the spatial unit has a high degree of attenuation of the load effect, and the spatial unit is divided into a strong occlusion zone; and spatial units with the same tactical semantic attributes are combined according to spatial adjacency to form a set of tactical semantic units. Specifically: traverse all spatial units that have completed tactical semantic partitioning, and check the adjacent units of each spatial unit; merge spatial units and adjacent units with the same tactical semantic attributes into one tactical semantic unit; repeat the above operation until all spatially continuous and tactically semantically identical spatial units are merged one by one; each merged tactical semantic unit consists of several spatial units, forming a connected aggregation region; summarize all tactical semantic units to form a tactical semantic unit set, which is used for subsequent tactical situation analysis, path planning, or visibility assessment.
[0025] It should be noted that the environmental shading factor is calculated as follows: ;in, Indicates the first The environmental occlusion coefficient corresponding to each tactical semantic unit is used to quantify the actionable intensity of the target spatial unit under the corresponding ray path, and satisfies the law that the actionable intensity decreases exponentially when the cumulative amount of path occlusion increases. The base of the natural logarithm is used to construct the exponential decay function so that the occlusion effect conforms to the preset exponential decay law. Methods for obtaining the strength of the effect of the payload on the tactical semantic unit include: After identifying the different factions involved in the confrontation and the drones included in each faction, the payload type, payload reference strength parameters, and payload main direction parameters of each drone are read. At the same time, the environmental occlusion coefficient is called to characterize the degree of occlusion attenuation of the spatial path between the drone and the tactical semantic unit. The semantic category information of the tactical semantic unit is read, and the semantic category response factor is obtained according to the correspondence between payload type and semantic category information to characterize the response differences of different categories of spatial units to different payloads. It should be noted that the semantic category response factor is obtained by establishing a preset correspondence rule table between payload type and tactical semantic category. The correspondence rule table is set according to the physical action mechanism of different payloads on different environmental targets, and is used to characterize the difference in sensitivity of spatial units of different semantic categories to payload action. Specifically, the tactical semantic categories are divided into semantic types such as terrain, buildings, airspace, personnel activity area, and equipment target, and the payload types are divided into functional categories such as reconnaissance, suppression, jamming, and damage. Based on the matching relationship between payload functional attributes and tactical semantic categories, a response weight value is preset for each pair of "payload type-semantic category" combinations. The response weight value is used to adjust the equivalent proportion of the payload baseline intensity in the semantic category spatial unit. After reading the semantic category information of the tactical semantic unit and the UAV payload type, the corresponding semantic category response factor is determined by looking up a table, so that the payload intensity produces differentiated responses in different tactical semantic environments.
[0026] Based on the payload type, payload reference strength parameters, payload action direction parameters, and semantic category response factors carried by the UAV, a payload action function is constructed to calculate the effect strength of each type of payload of each UAV on each tactical semantic unit.
[0027] The load application function is: ;in, Indicates drone load The intensity of the effect on each tactical semantic unit; This represents the load reference strength parameter, characterizing the maximum output capability of the load without attenuation; This indicates the load type index, such as reconnaissance load, jamming load, strike load, etc. Indicates the index of the time when the action occurred; Indicates drone To tactical semantic units The spatial distance is a three-dimensional Euclidean distance; This represents the angle between the principal direction of the load action and the line connecting the UAV to the tactical semantic unit. This represents the space attenuation coefficient, used to describe the energy attenuation rate of different payload types during space propagation; This represents the shading index attenuation term. The stronger the shading, the weaker the transmittance. This represents the semantic category response factor, used to describe the coupling response characteristics between load type and spatial semantic category; This represents the semantic category information of tactical semantic units (e.g., built-up areas, open areas, woodlands, electromagnetically sensitive areas, etc.). Methods for obtaining the tactical field of effect for both sides include: At the same time, using tactical semantic units as spatial computing primitives, the intensity of the effect of various payloads on each tactical semantic unit is calculated for all drones belonging to the same faction, and the process is traversed according to the drone index and payload type index. The intensity of the same type of payload is accumulated in the functional category dimension. The intensity of the same type of payload generated by different UAVs in the same camp at the same tactical semantic unit is aggregated to obtain the payload component of the camp at the corresponding tactical semantic unit. It should be noted that the accumulation is performed separately along the functional dimension, and the intensity of different functional categories of payloads, such as reconnaissance payloads, jamming payloads, suppression payloads, and damage payloads, is accumulated independently to form payload action components. Under a unified set of tactical semantic units and a unified spatial indexing system, each tactical semantic unit is associated with the payload action components of its own tactical action field and the payload action components of the enemy's tactical action field at the corresponding moment, thus forming a distributed tactical action field of both sides in the tactical space.
[0028] The payload components of our own faction at the corresponding tactical semantic units are organized according to the tactical semantic unit index to obtain our own tactical action field in each tactical semantic unit; at the same time, the payload components of the enemy faction at the corresponding tactical semantic units are also organized according to the tactical semantic unit index to obtain the enemy's tactical action field in each tactical semantic unit.
[0029] Methods for mapping the tactical action fields of both sides to the same rule-based adversarial framework include: The payload action components corresponding to each tactical semantic unit in the friendly tactical action field and the enemy tactical action field are mapped to the rule domain. Based on the preset correspondence table between payload function categories and rule function domains, the payload action components of different payload function categories are converted into different rule function domains. It should be noted that, based on a pre-established correspondence table between payload function categories and rule function domains, the system associates payload function categories such as reconnaissance, jamming, suppression, and damage to rule function domains such as information acquisition domain, information suppression domain, action restriction domain, and entity weakening domain in the rule layer. When the payload action component at a certain tactical semantic unit in the tactical action field is obtained, the corresponding rule function domain is determined by looking up the table according to the payload function category to which the component belongs. The payload action component is then re-identified and assigned to the corresponding rule function domain component through the function domain mapping function, thereby realizing the conversion from the payload physical mechanism dimension to the rule function dimension. The load components of both sides in the corresponding rule functional domains are normalized; using the tactical semantic unit index as a unified spatial coordinate, a one-to-one correspondence is established between the rule functional domains of both sides and the rule functional domains of the enemy at the same tactical semantic unit index, thereby completing the mapping of the tactical action fields of both sides to a unified rule confrontation framework.
[0030] Methods for generating a uniform net tactical action field include: For each tactical semantic unit, a friendly rule functional domain and an enemy rule functional domain are constructed respectively. A pre-established rule coupling relationship matrix is introduced, and the adversarial influence intensity between different rule functional domains is characterized by the rule coupling relationship matrix. It should be noted that the rule coupling matrix is the fundamental structural parameter for rule-layer adversarial computation, used to characterize the mutual influence relationships between different rule functional domains. First, the rule functional domains participating in adversarial modeling in the system are standardized and defined. Different load function categories are abstracted into several rule functional domains at the rule layer. Each rule functional domain represents a rule capability dimension with unified rule semantics and action mechanisms. Based on this set of rule functional domains, a rule coupling matrix is constructed. The rule coupling matrix is a square matrix with a dimension equal to the number of rule functional domains. Any element in the matrix represents the influence strength and direction of the rule functional domain corresponding to a row on the rule functional domain corresponding to a column. The coupling coefficients are determined through at least one of the following methods: a pre-built rule mechanism library, statistical analysis results of historical adversarial samples, and simulation calibration results. These coefficients are written into the rule parameter library before system deployment, thus making the rule coupling matrix a fixed structural parameter model that can be directly called at runtime.
[0031] The coupling coefficients preset in the rule coupling relationship matrix describe the weight of the weakening, suppression or enhancement effect of one rule functional domain on another rule functional domain; at the same tactical semantic unit, the rule action components of the friendly and the enemy in the same rule functional domain are canceled out in the same domain, and the net rule action component of the rule functional domain is determined according to the difference in rule strength between the two sides. It should be noted that when performing co-domain cancellation calculations at the same tactical semantic unit, the system first obtains the friendly rule function domain vector and the enemy rule function domain vector respectively. Each component of the vector represents the rule action component in the corresponding rule function domain, and all have undergone dimension unification and standardization. For any rule function domain, under the same tactical semantic unit index, the friendly rule function domain component and the corresponding enemy rule function domain component are subtracted to obtain the net rule action component of that rule function domain. The difference result retains sign information; its positive and negative values respectively represent the direction of rule advantage attribution, and the absolute value represents the strength of the advantage. Thus, a net rule result after direct confrontation between the two sides is formed at each rule function domain. The net rule action components of each rule function domain are combined to form the net rule function domain vector of that tactical semantic unit.
[0032] Based on the rule coupling relationship matrix, the net rule action components of each rule functional domain are mapped to other associated rule functional domains according to the corresponding coupling coefficients, and cross-rule functional domain coupling operations are performed to obtain the modified rule action components that include cross-domain effects. After obtaining the net rule functional domain vector, cross-rule functional domain coupling operations are performed. The system uses the net rule functional domain vector as an input vector and performs matrix operations with a pre-established rule coupling relationship matrix. Through the coupling coefficients in the matrix, the net effect strength of a certain rule functional domain is mapped to other associated rule functional domains, forming additional correction components for other rule functional domains. The correction components received by each rule functional domain from other rule functional domains are accumulated and then superimposed with the original net rule effect component of the current domain to obtain the corrected rule effect component that includes cross-domain influence. This process expands the rule layer calculation from direct adversarial interaction within a single domain to a collaborative propagation calculation between multiple rule domains, resulting in a corrected rule functional domain vector that reflects the overall coupling characteristics of the rule system.
[0033] The rule functional domain components after in-domain cancellation and cross-domain coupling are recombined to form the net rule functional domain vector of the tactical semantic unit. Based on the preset reverse mapping relationship between the rule functional domain and the tactical action dimension, the net rule functional domain vector is converted into the corresponding net tactical action component. The process is repeated for all tactical semantic units to obtain the net tactical action distribution covering all tactical semantic units, thereby constructing a unified net tactical action field.
[0034] It should be noted that after completing the rule-layer coupling operation, the system converts the modified rule-functional domain vector into net tactical action components of the tactical action layer based on the preset reverse mapping relationship from rule functional domains to tactical action dimensions. The reverse mapping relationship describes the contribution weight of each rule functional domain to each tactical action dimension, and can be predetermined through simulation calibration and stored in the mapping parameter table. The system performs a weighted combination operation on the modified rule-functional domain vector and the mapping relationship, projecting each component of the rule layer onto the tactical action dimension space to obtain the net tactical action components in the corresponding tactical dimensions.
[0035] Methods for forming a tactical state field that changes continuously over time include: Within the spatial domain defined by the set of tactical semantic units, a tactical state vector is established for each tactical semantic unit. The tactical state vector represents the comprehensive tactical attributes of the spatial unit. At each calculation moment, the net tactical action component of the corresponding tactical semantic unit is read. The net tactical action component is used as the state-driven input and converted into a tactical state increment through a preset state response function. The state response function includes a response sensitivity parameter, a saturation limit parameter, and a recovery decay parameter; the tactical state increment is recursively superimposed with the tactical state vector of the previous moment to obtain the tactical state vector of the current moment. State response function: ;in, Represents tactical semantic units At any moment Belongs to the Tactical state increments in the tactical state dimension; This represents the response sensitivity parameter, which controls how sensitive this state dimension is to tactical actions; Indicates at time Acting on tactical semantic units Above, belonging to the first Net tactical effect component of the tactical state dimension; This represents the saturation limit parameter, which is adjusted. The rate at which the function reaches saturation; This represents the recovery decay parameter, indicating the natural decay rate of this tactical attribute when there is no continuous input. This represents the hyperbolic tangent function, with an output range of (-1, 1), used to limit the range of variation and prevent divergence. Indicates the first Each tactical semantic unit, at time... The Tactical state components of the tactical state dimension; Indexes representing tactical state dimension types; A time inertia factor is introduced in the state recursion process to limit the magnitude of state change at a single moment, and a natural decay factor is introduced to characterize the trend of tactical state regressing to the initial equilibrium state when there is no continuous input. Simultaneously, based on the preset spatial adjacency relationship, a weighted diffusion calculation is performed on the tactical state vector differences between adjacent tactical semantic units, and all tactical semantic units are processed repeatedly in time order to form a tactical state field that evolves continuously over time.
[0036] Methods for determining a drone collaborative scheduling scheme include: At the current moment, obtain the tactical state vector of each tactical semantic unit and use the tactical state vector as the input for rule calculation; construct the UAV-payload-target combination for any UAV, available payload type and candidate tactical semantic unit; A multidimensional rule evaluation function is constructed. By weighting the components of each dimension of the tactical state vector of the corresponding tactical semantic unit with the rule weights corresponding to the payload, the rule score of the UAV-payload-target combination is obtained. The multidimensional rule evaluation function is: ;in, The rule-based score represents the combination of drone, payload, and target. Indicates the total number of tactical state dimension types; Represents the rule weight coefficient, representing the load. In the Functional adaptation weights in the tactical state dimension; The rule score is compared with a preset rule score threshold. If the rule score meets the preset rule score threshold, the combination is determined to be a rule-consistent combination; otherwise, it is determined to be a rule-conflicting combination and is removed. Within a set of combinations that meet the preset rule scoring thresholds, the allocation relationship between UAVs, payloads, and tactical semantic units is optimized by using rule scoring as the optimization objective; each UAV is limited to executing only one payload target task within the same scheduling cycle; and the optimization result is determined as the UAV collaborative scheduling scheme for the current moment.
[0037] Methods for achieving coordinated control of multi-payload UAVs in simulated combat training include: After obtaining the UAV collaborative scheduling scheme, the payload target allocation information of each UAV is sent to the UAV control terminal, instructing the UAVs to perform operations according to the planned payload task type, direction of action, and target tactical semantic unit; at the same time, UAV status information, payload output status information, and environmental feedback information are collected in real time, input into the tactical situation evolution module, and the tactical state field is updated. It should be noted that the UAV status information, payload output status information, and environmental feedback information specifically include: UAV status information mainly includes the UAV's real-time position coordinates, flight speed, heading angle, attitude angle, flight altitude, remaining endurance, and power system status, which reflect the UAV's motion and maneuverability; payload output status information includes the working mode, output power, direction of action, duration of action, payload resource consumption, and whether the payload is currently active or in standby state for each payload, which is used to assess the payload's actual ability to act on the target; environmental feedback information includes the terrain, building obstruction, weather and lighting conditions, communication link status, enemy and friendly unit activities, and the current tactical state vector of the tactical semantic unit in the space unit where the UAV is located, which is used to correct UAV actions, payload usage strategies, and collaborative scheduling schemes in real time, thereby ensuring that the UAV swarm can achieve accurate, continuous, and coordinated mission execution in complex tactical environments.
[0038] Based on the updated tactical state field, the trimming and collaborative scheduling module performs rule consistency judgment and collaborative optimization on the combination of UAV payload targets according to a preset cycle, dynamically adjusts the UAV flight path, payload usage order and task allocation, and realizes collaborative control of multi-payload UAVs in simulated combat training.
[0039] The preset semantic centrality score threshold is set by staff. By collecting different semantic centrality scores, the average of multiple semantic centrality scores is taken as the preset semantic centrality score threshold. Similarly, preset switching threshold and preset cosine similarity threshold are set.
[0040] In this embodiment, by dividing the environmental space into unified spatial units and determining each spatial unit traversed by the UAV ray path to form a path voxel set, a refined spatial modeling of the ray path is achieved. At the same time, by introducing the cumulative path occlusion and the environmental occlusion coefficient, the occlusion effects under different paths, directions, and spatial units are transformed into a unified numerical index, thereby comprehensively analyzing the occlusion situation of multiple obstacles in complex 3D environments, realizing quantitative comparison and ranking of occlusion degree, and avoiding the problems of vague and incomparable occlusion descriptions caused by lighting, background interference, or terrain complexity in the prior art. Tactical semantic spatial partitioning provides clear operational guidance for UAV mission planning, enhancing the UAV's tactical execution capabilities in complex environments. Through the unified organization and attribute representation of spatial units, path occlusion assessment and tactical semantic partitioning results can be automatically generated without the need for manual experience rules, making it suitable for different mission scenarios and environmental complexities. The quantified path occlusion information and tactical semantic unit set can assist UAVs in path planning, obstacle avoidance strategy selection, and observation position deployment, improving mission success rate and reducing the risk of being detected or interfered with.
[0041] By introducing semantic category response factors and based on a preset correspondence rule table, the effects of different loads in different semantic environments are adjusted, enabling the same load to exhibit differentiated equivalent intensity in different tactical semantic units such as building, terrain, airspace, and personnel activity areas, significantly improving the realism and precision of load effect assessment. An environmental occlusion coefficient is introduced into the load effect function, and through an exponential decay term, it participates in the intensity calculation along with the semantic category response factor, achieving a coupled expression of occlusion attenuation effects and spatial semantic response characteristics. This allows the load effect intensity to simultaneously reflect the degree of spatial path occlusion and target environmental sensitivity, overcoming the problem of separate treatment of these two aspects in existing technologies. By establishing a preset correspondence rule table between load types and tactical semantic categories, and parameterizing it in the form of response weights, this invention can flexibly adapt to different load functional attributes and different tactical environment requirements, supporting rule expansion and parameter adjustment, and is suitable for simulation applications in multi-load, multi-faction, and multi-tactical scenarios. The load action function comprehensively considers the load reference strength, action direction, spatial distance attenuation, occlusion index attenuation, and semantic category response factor, making the load's effect on the tactical semantic unit closer to the real physical action process, thereby significantly improving the simulation accuracy and reliability of multi-load UAV systems in actual combat training simulation, confrontation assessment, and tactical simulation. Example
[0042] Please see Figure 2 As shown, for parts not described in detail in this embodiment, please refer to the description in Embodiment 1. A method for multi-payload unmanned aerial vehicles (UAVs) used for simulating combat training is provided, including: S1. Divide the training area into spatial units with tactical semantic attributes, and calculate the environmental occlusion coefficient of each spatial unit based on spatial geometric relationships and occlusion relationships to form a set of tactical semantic units. S2. Based on the payload of each drone in the different opposing factions, construct the payload action function to obtain the intensity of the payload's effect on the tactical semantic unit; and superimpose all payload effects to obtain the tactical action field of both sides. S3. Map the tactical action fields of both sides to the same rule-based confrontation framework; perform cancellation calculations on the actions of both sides according to the preset rule coupling relationship to generate a unified net tactical action field; S4. Within the spatial domain defined by the set of tactical semantic units, the tactical state of each spatial unit is updated by driving the net tactical action field, forming a tactical state field that changes continuously with time. S5. Construct a multi-dimensional rule function based on the tactical state field, determine the rule consistency of the UAV payload target combination, determine the UAV collaborative scheduling scheme, and realize the collaborative control of multi-payload UAVs in simulated combat training.
[0043] Since the electronic device described in this embodiment is the electronic device used in implementing a multi-payload unmanned aerial vehicle (UAV) system for simulated combat training as described in this application, those skilled in the art can understand the specific implementation methods and various variations of the electronic device in this embodiment based on the multi-payload UAV system for simulated combat training described in this application. Therefore, how the electronic device implements the method in this application will not be described in detail here. Any electronic device used by those skilled in the art in implementing a multi-payload UAV system for simulated combat training as described in this application falls within the scope of protection of this application.
[0044] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.
[0045] The above description is merely a preferred embodiment of the present invention, and the scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for users of ordinary technical skills, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.
Claims
1. A multi-payload unmanned aerial vehicle system for simulating combat training, characterized in that, include: The tactical environment construction module is used to divide the training area into spatial units with tactical semantic attributes, and calculate the environmental occlusion coefficient of each spatial unit based on spatial geometric relationships and occlusion relationships to form a set of tactical semantic units. The adversarial capability analysis module is used to construct a load action function based on the load of each drone in the different opposing factions, obtain the intensity of the load's effect on the tactical semantic unit, and superimpose all load effects to obtain the tactical action field of both sides. The rule coupling cancellation module is used to map the tactical action fields of both sides to the same rule-based adversarial framework; Based on the preset rules of coupling, the actions of both sides are canceled out to generate a unified net tactical action field; The tactical situation evolution module is used to update the tactical state of each spatial unit within the spatial domain defined by the set of tactical semantic units through the net tactical action field, forming a tactical state field that changes continuously over time. The trimming and collaborative scheduling module is used to construct multi-dimensional rule functions based on the tactical state field, determine the consistency of rules for the combination of UAV payload targets, determine the UAV collaborative scheduling scheme, and realize the collaborative control of multi-payload UAVs in simulated combat training.
2. A multi-payload unmanned aerial vehicle system for simulating combat training according to claim 1, characterized in that, The method for obtaining the spatial unit includes: The spatial geometric information of the training area is acquired and then discretized according to a preset spatial resolution to form different spatial units. The spatial geometric information includes the terrain outline information, building distribution information, obstacle location information, and height information of the training area. Spatial units are constructed using regular voxels, with each spatial unit corresponding to a specific spatial range within the training area; each spatial unit is configured with corresponding spatial location parameters, spatial occupancy state parameters, and environmental medium attribute parameters. Based on spatial location parameters and spatial occupancy status parameters, basic tactical semantic annotations are performed on spatial units, enabling them to possess semantic attributes for tactical analysis, thereby forming spatial units with tactical semantic attributes.
3. A multi-payload unmanned aerial vehicle system for simulating combat training according to claim 2, characterized in that, The method for forming a set of tactical semantic units includes: All spatial units are uniformly organized to form a spatial unit set; based on the interaction direction between the UAV and the target spatial unit, a ray path is constructed from the current position of the UAV to the target spatial unit; based on the geometric interaction relationship between the ray path and the spatial units, the spatial units traversed by the ray path are determined one by one, and all spatial units actually traversed by the ray path are counted to form a path voxel set. Based on the environmental medium properties of each spatial unit, the occlusion effect is cumulatively calculated according to the traversal length of the ray path within the corresponding spatial unit to obtain the cumulative path occlusion amount. The cumulative path occlusion amount is converted into an environmental occlusion coefficient according to a preset exponential decay law, and the spatial unit is tactically semantically divided according to the size range of the environmental occlusion coefficient, dividing the spatial unit into a visible area, a strong occlusion area, and a weak occlusion area. A first threshold and a second threshold for environmental occlusion coefficient are preset, with the first threshold being greater than the second threshold. When the environmental occlusion coefficient is greater than or equal to the first threshold, the corresponding spatial unit is determined to have a low degree of attenuation of the load effect, and the spatial unit is classified as a visible area. When the environmental occlusion coefficient is greater than or equal to the second threshold but less than the first threshold, the spatial unit is determined to have a partial attenuation of the load effect, and the spatial unit is classified as a weakly occluded area. When the environmental occlusion coefficient is less than the second threshold of the preset environmental occlusion coefficient, it is determined that the spatial unit has a high degree of attenuation of the load effect, and the spatial unit is divided into a strong occlusion zone; and spatial units with the same tactical semantic attributes are combined according to spatial adjacency to form a set of tactical semantic units.
4. A multi-payload unmanned aerial vehicle system for simulating combat training according to claim 3, characterized in that, The method for obtaining the intensity of the load's effect on the tactical semantic unit includes: After identifying the different factions involved in the confrontation and the drones included in each faction, the payload type, payload reference strength parameters, and payload main direction parameters of each drone are read; at the same time, the environmental occlusion coefficient is called, and the semantic category information of the tactical semantic unit is read. The semantic category response factor is obtained according to the correspondence between payload type and semantic category information. Based on the payload type, payload reference strength parameters, payload action direction parameters, and semantic category response factors carried by the UAV, a payload action function is constructed to calculate the effect strength of each type of payload of each UAV on each tactical semantic unit.
5. A multi-payload unmanned aerial vehicle system for simulating combat training according to claim 4, characterized in that, The methods for obtaining the tactical action field of both sides include: At the same time, using tactical semantic units as spatial computing primitives, the intensity of the effect of various payloads on each tactical semantic unit is calculated for all drones belonging to the same faction, and the process is traversed according to the drone index and payload type index. The intensity of the same type of payload is accumulated in the functional category dimension. The intensity of the same type of payload generated by different UAVs in the same camp at the same tactical semantic unit is aggregated to obtain the payload component of the camp at the corresponding tactical semantic unit. The payload components of our own faction at the corresponding tactical semantic units are organized according to the tactical semantic unit index to obtain our own tactical action field in each tactical semantic unit; at the same time, the payload components of the enemy faction at the corresponding tactical semantic units are also organized according to the tactical semantic unit index to obtain the enemy's tactical action field in each tactical semantic unit.
6. A multi-payload unmanned aerial vehicle system for simulating combat training according to claim 5, characterized in that, The method for mapping the tactical action fields of both sides to the same rule-based adversarial framework includes: The payload action components corresponding to each tactical semantic unit in the friendly tactical action field and the enemy tactical action field are mapped to the rule domain. Based on the preset correspondence table between payload function categories and rule function domains, the payload action components of different payload function categories are converted into different rule function domains. The load components of both sides in the corresponding rule functional domains are normalized; using the tactical semantic unit index as a unified spatial coordinate, a one-to-one correspondence is established between the rule functional domains of both sides and the rule functional domains of the enemy at the same tactical semantic unit index, thereby completing the mapping of the tactical action fields of both sides to a unified rule confrontation framework.
7. A multi-payload unmanned aerial vehicle system for simulating combat training according to claim 6, characterized in that, The method for generating a uniform net tactical action field includes: For each tactical semantic unit, a friendly rule functional domain and an enemy rule functional domain are constructed respectively. A pre-established rule coupling relationship matrix is introduced, and the adversarial influence intensity between different rule functional domains is characterized by the rule coupling relationship matrix. The coupling coefficients preset in the rule coupling relationship matrix describe the weight of the weakening, suppression or enhancement effect of one rule functional domain on another rule functional domain; at the same tactical semantic unit, the rule action components of the friendly and the enemy in the same rule functional domain are canceled out in the same domain, and the net rule action component of the rule functional domain is determined according to the difference in rule strength between the two sides. Based on the rule coupling relationship matrix, the net rule action components of each rule functional domain are mapped to other associated rule functional domains according to the corresponding coupling coefficients, and cross-rule functional domain coupling operations are performed to obtain the modified rule action components that include cross-domain effects. The rule functional domain components after in-domain cancellation and cross-domain coupling are recombined to form the net rule functional domain vector of the tactical semantic unit. According to the preset reverse mapping relationship from the rule functional domain to the tactical action dimension, the net rule functional domain vector is converted into the corresponding net tactical action component. The process is repeated for all tactical semantic units to obtain the net tactical action distribution covering all tactical semantic units, thereby constructing a unified net tactical action field.
8. A multi-payload unmanned aerial vehicle system for simulating combat training according to claim 7, characterized in that, The method for forming a tactical state field that changes continuously over time includes: Within the spatial domain defined by the set of tactical semantic units, a tactical state vector is established for each tactical semantic unit. At each calculation moment, the net tactical action component of the corresponding tactical semantic unit is read, and the net tactical action component is used as the state-driven input. The net tactical action component is converted into a tactical state increment through a preset state response function. The state response function includes a response sensitivity parameter, a saturation limit parameter, and a recovery decay parameter; the tactical state increment is recursively superimposed with the tactical state vector of the previous moment to obtain the tactical state vector of the current moment. Simultaneously, based on the preset spatial adjacency relationship, a weighted diffusion calculation is performed on the tactical state vector differences between adjacent tactical semantic units, and all tactical semantic units are processed repeatedly in time order to form a tactical state field that evolves continuously over time.
9. A multi-payload unmanned aerial vehicle system for simulating combat training according to claim 8, characterized in that, The method for determining the UAV collaborative scheduling scheme includes: At the current moment, obtain the tactical state vector of each tactical semantic unit and use the tactical state vector as the input for rule calculation; construct the UAV-payload-target combination for any UAV, available payload type and candidate tactical semantic unit; A multidimensional rule evaluation function is constructed. By weighting the components of each dimension of the tactical state vector of the corresponding tactical semantic unit with the rule weights corresponding to the payload, the rule score of the UAV-payload-target combination is obtained. The rule score is compared with a preset rule score threshold. If the rule score meets the preset rule score threshold, the combination is determined to be a rule-consistent combination; otherwise, it is determined to be a rule-conflicting combination and is removed. Within a set of combinations that meet the preset rule scoring thresholds, the allocation relationship between UAVs, payloads, and tactical semantic units is optimized using rule scoring as the optimization objective; each UAV is limited to executing only one payload target task within the same scheduling cycle; and the optimization result is determined as the UAV collaborative scheduling scheme for the current moment.
10. A multi-payload unmanned aerial vehicle system for simulating combat training according to claim 9, characterized in that, The method for achieving coordinated control of multi-payload UAVs in simulated combat training includes: After obtaining the UAV collaborative scheduling scheme, the payload target allocation information of each UAV is sent to the UAV control terminal, instructing the UAVs to perform operations according to the planned payload task type, direction of action, and target tactical semantic unit; at the same time, UAV status information, payload output status information, and environmental feedback information are collected in real time, input into the tactical situation evolution module, and the tactical state field is updated. Based on the updated tactical state field, the trimming and collaborative scheduling module performs rule consistency judgment and collaborative optimization on the combination of UAV payload targets according to a preset cycle, dynamically adjusts the UAV flight path, payload usage order and task allocation, and realizes collaborative control of multi-payload UAVs in simulated combat training.