An air-ground integrated unmanned aerial vehicle cluster cooperative control method

Abstract

The present application relates to the field of unmanned aerial vehicle cluster control, and particularly relates to an air-ground integrated unmanned aerial vehicle cluster cooperative control method. The method comprises the following steps: randomly obtaining a combination of any one unmanned aerial vehicle, a ground unit and a task to be executed in the cluster, and the current positions of the unmanned aerial vehicle and the ground unit in the combination and the task position of the task to be executed; calculating the cooperative gain of each combination; constructing a corresponding global optimization function based on the cooperative gain of all combinations, determining a task allocation scheme with the maximum global optimization function of the cluster; and issuing a cooperative execution instruction to the multiple unmanned aerial vehicles and the multiple ground units in response to the task allocation scheme. That is, the scheme of the present application can improve the cooperative efficiency of the air-ground integrated cluster.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) swarm control. More specifically, this invention relates to an integrated air-ground UAV swarm cooperative control method. Background Technology

[0002] In modern UAV swarm collaborative control technology, task allocation is a core element determining the overall system efficiency and mission success rate. Its main function is to determine the optimal matching scheme between units and tasks based on mission requirements and the current status (such as position and capabilities) of each unit in the swarm (including UAVs or ground units), using specific algorithms, thereby guiding the swarm to complete the task in the most efficient way.

[0003] In existing technologies, task allocation commonly employs models based on independent cost assessment. These models typically define a static cost function for each individual UAV or ground unit, which is mainly composed of linearly weighted physical quantities such as distance and energy consumption. The fundamental flaw of this model is that it treats each unit as an isolated decision-making entity, completely ignoring the synergistic gains that air and ground units can generate when performing tasks. For example, the monitoring cost of an air unit (UAV) can dynamically decrease because a ground unit can provide it with communication relay or close-range verification.

[0004] Therefore, because the existing model lacks a mathematical description of the dynamic and coupled collaborative relationship between air and ground units, its allocation decision is merely a superposition of local optima rather than a true global system optimum, thus limiting the overall performance of the integrated air and ground cluster.

[0005] Therefore, it is particularly important to address the issue of poor global performance in existing task allocation models due to the neglect of collaborative gains. Summary of the Invention

[0006] The purpose of this invention is to propose an integrated air-ground UAV swarm cooperative control method to solve the problem of poor global performance caused by neglecting cooperative gains in existing task allocation models; to this end, this invention provides a solution in one aspect.

[0007] This invention provides an integrated air-ground UAV swarm cooperative control method, comprising:

[0008] Randomly obtain a combination of a triple of any UAV, ground unit and task to be executed in the cluster, as well as the current position of the UAV, ground unit and task position of the task to be executed in the combination;

[0009] Calculate the collaborative gain of each combination; construct the corresponding global optimization function based on the collaborative gain of all combinations, and determine the task allocation scheme with the largest global optimization function of the cluster;

[0010] In response to the task allocation scheme, coordinated execution commands are issued to multiple UAVs and multiple ground units;

[0011] Wherein, the collaborative gain is the product of the pre-acquired task complementarity and collaborative reachability; the task complementarity is the dot product of the fusion capability vector and the task requirement vector; each element in the fusion capability vector is the maximum value of the corresponding dimension in the capability vector of the UAV and the capability vector of the ground unit in the combination; the capability vector and the task requirement vector have one-to-one correspondence and the same dimension; the task requirement vector represents the requirement of the corresponding task to be executed for the collaborative capability of the UAV and the ground unit.

[0012] The collaborative reachability is negatively correlated with the communication loss index of the combination and the absolute value of the difference between the current position of the UAV and the ground unit in the combination and the estimated time from the current position to the task position of the task to be executed.

[0013] The above scheme obtains the synergistic gain by comprehensively quantifying the task complementarity in capabilities and the collaborative reachability in spatiotemporal communication between air and ground units. Then, a global optimization function is constructed, which enables the task allocation scheme that maximizes global benefits from the perspective of the entire cluster. This overcomes the fundamental defect of existing technologies that ignore synergistic effects, where the superposition of local optima does not equal global optima. Thus, it achieves true global optimization of air and ground resources and significantly improves the overall collaborative efficiency and mission success rate of UAV clusters.

[0014] Optionally, the construction of the task requirement vector includes:

[0015] Construct an ontology library, which includes standard task requirement vectors for different historically executed tasks;

[0016] Based on the ontology library, the standard task requirement vector that matches the task to be executed in the combination is used as the task requirement vector of the task to be executed.

[0017] The ontology library in the above scheme provides a unified and quantified benchmark for the task requirement vector of the task to be executed.

[0018] Optionally, the capability vector of the UAV includes the UAV's wide-area aerial reconnaissance capability, ground close-range observation capability, and communication relay capability; the capability vector of the ground unit includes the ground unit's wide-area aerial reconnaissance capability, ground close-range observation capability, and communication relay capability.

[0019] The capability vectors of the UAV and the ground unit obtained above ensure that the task complementarity calculated in the subsequent calculation can truly reflect the operational potential of the UAV and the ground unit.

[0020] Optionally, collaborative accessibility for:

[0021] ;

[0022] in, , drones Ground Unit The current position is respectively the position of the task to be executed. The first estimated time and the second estimated time for the task location; The set time tolerance; For drones Ground Unit and tasks to be performed Communication loss metrics between them This is the maximum value of the communication loss metric among all combinations, and The expression is not zero, exp() is an exponential function, and i, j, and k are the numbers of the UAV, ground unit, and task to be executed, respectively.

[0023] By comprehensively considering the estimated time difference between the current location of the air-ground unit and the location of the task to be performed, as well as the communication loss indicators among the three, the feasibility and stability of the combined collaborative action can be dynamically and precisely evaluated.

[0024] Optionally, the method for obtaining the first estimated time includes:

[0025] Obtain the three-dimensional Euclidean distance between the current position of the drone and the position of the task to be performed;

[0026] Divide the three-dimensional Euclidean distance by the standard cruise speed of the corresponding UAV to obtain the first estimated time; the standard cruise speed is the average flight speed in the standard cruise mode.

[0027] The above scheme provides a relatively accurate estimated arrival time for the drone.

[0028] Optionally, the method for obtaining the second estimated time includes:

[0029] Constructing an environmental cost map based on a digital elevation model;

[0030] use The search algorithm plans the optimal path from the current location of the ground unit to the task location on the environmental cost map;

[0031] The sum of the travel times of the ground unit traversing all grids in the optimal path is used as the second estimated time; the travel time is positively correlated with the corresponding grid slope and the preset vehicle performance curve.

[0032] The above scheme can accurately calculate the actual travel time of ground units in complex terrain, which greatly improves the accuracy of subsequent collaborative accessibility assessment.

[0033] Optionally, the method for obtaining the communication loss index includes:

[0034] Using radio frequency simulation software, a communication coverage map is generated based on the terrain data of the work area and the communication equipment parameters of each unit;

[0035] The communication loss index is negatively correlated with the signal strength of the corresponding combination in the communication coverage map.

[0036] Optionally, the global optimization function for:

[0037] ;

[0038] in, This is a binary decision variable; a value of 1 indicates that the drone will be... and ground unit Assigned to tasks to be executed ; For drones Ground Unit and tasks to be performed The combined synergistic gain; and drones Ground Unit Arrival at the task to be executed The first execution cost and the second execution cost of the task location; These are the weighting coefficients; I, J, and K represent the number of UAVs, ground units, and the total number of tasks to be executed, respectively.

[0039] The global optimization model also includes the following constraints:

[0040] Each task can be executed by a combination of one drone and a ground unit in an air-to-ground manner.

[0041] Each drone belongs to a maximum of one group;

[0042] Each ground unit belongs to at most one group.

[0043] The global optimization function described above aims to maximize the net benefit of the cluster in order to balance the relationship between high synergy and low resource consumption.

[0044] Optionally, the first execution cost is the sum of time cost and energy cost;

[0045] Wherein, energy consumption cost is the ratio of the estimated energy consumption from the current position of the drone to the mission position to the total battery capacity of the drone; the estimated energy consumption is the product of the first estimated time and the average cruise power of the drone in standard cruise mode; time cost is the ratio of the first estimated time to the preset total mission time limit.

[0046] The second execution cost is the sum of the time component and the energy component; the time component is the normalized value of the second estimated time; the energy component is the normalized value of the sum of the instantaneous power of the ground units of all path segments in the optimal path; the instantaneous power is calculated based on the terrain slope and traffic speed of the corresponding path segment and by calling the pre-established ground unit power consumption model.

[0047] Optionally, determining the task allocation scheme that maximizes the global optimization function of the cluster includes optimizing the global optimization function using a heuristic algorithm.

[0048] The beneficial effects of this invention are as follows:

[0049] The present invention quantifies the collaborative potential of air-ground units from two dimensions: task complementarity and collaborative reachability. It constructs a global optimization function with the goal of maximizing the total cluster gain, which can find the truly globally optimal task allocation scheme. This solves the problem of limited overall system performance caused by ignoring collaborative effects and significantly improves the collaborative efficiency of air-ground integrated clusters. Attached Figure Description

[0050] Figure 1 The flowchart illustrating the steps of an air-ground integrated UAV swarm cooperative control method in this embodiment is shown in the diagram. Detailed Implementation

[0051] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0052] This invention provides an integrated air-ground UAV swarm cooperative control method, aiming to solve the technical problem of poor global performance caused by neglecting the cooperative gains between air and ground units in existing task allocation models. Specifically, it quantifies and evaluates the cooperative gains that each potential combination of "UAV-ground unit-task to be executed" can generate, and constructs a global optimization function to find a task allocation scheme that maximizes the overall swarm efficiency.

[0053] Specifically, such as Figure 1 As shown in this embodiment, an air-ground integrated UAV swarm cooperative control method includes the following steps:

[0054] Step S1: Obtain the cluster status information and the task information of the task to be executed, and randomly obtain a combination of any UAV, ground unit and the task to be executed in the cluster.

[0055] The cluster in this embodiment includes multiple drones and multiple ground units; one group of drones and ground units constitutes one air-to-ground unit. Specifically, the cluster acquires real-time data on all drones within the drone swarm. and ground unit The status information is as follows: I represents the total number of UAVs, and J represents the total number of ground units.

[0056] For any drone Its status information includes current location, standard cruise speed, etc. For any ground unit... Its status information includes its current location.

[0057] The current position is a three-dimensional position coordinate; the standard cruise speed is the average flight speed of the UAV in standard cruise mode.

[0058] At the same time, obtain all tasks to be executed. For any task to be executed Its task information includes the task location. This represents the total number of tasks to be executed.

[0059] The above combination is any combination of a UAV, a ground unit, and a task to be performed. Multiple UAVs, ground units, and tasks to be performed can form multiple combinations.

[0060] Step S2: Calculate the synergistic gain of each combination.

[0061] An efficient collaborative task allocation requires not only that the combined capabilities of the participating air and ground units meet the task requirements, but also that they can form effective coordination in time, space and communication. Therefore, in this embodiment, it is necessary to calculate the collaborative gain of each combination.

[0062] For example, for any drone Ground Unit and tasks to be performed Combinations of triplets Calculate the number of drones in any combination With ground unit For tasks to be executed Synergistic gain .

[0063] Specifically, the process of synergistic gain includes:

[0064] Step S21: Calculate the task complementarity.

[0065] The process of obtaining task complementarity is as follows:

[0066] First, obtain the capability vectors of the UAVs and ground units in the combination to construct a fused capability vector.

[0067] Accordingly, for each drone in the cluster Ground Unit Construction and Task Requirement Vector Same ability vector and .

[0068] For example, drones The capability vector is Where i is the serial number of the drone. For drones Normalized wide-area aerial reconnaissance capability; Representing drones Normalized ground-based close-range observation capabilities; Representing drones The normalized communication relay capability, where T represents the transpose of the matrix.

[0069] Specifically, taking L=3 as an example, the wide-area aerial reconnaissance capability reflects the reconnaissance coverage area per unit time; specifically, the coverage rate is obtained using a single-aircraft coverage probability model.

[0070] Ground close-range observation capability refers to the resolution and accuracy of UAV observations of ground targets at close range, which is mainly quantified by ground sampling distance (GSD).

[0071] Specifically, the ground-based close-range observation capability is the final GSD obtained by fusing the horizontal and vertical GSDs. Since the calculation methods for the horizontal and vertical GSDs are existing technologies, they will not be elaborated upon here.

[0072] Communication relay capability refers to the effectiveness of a UAV as a relay node in extending communication coverage; in this embodiment, communication rate is used as the communication relay capability.

[0073] Since the communication rate is based on existing technology, it will not be discussed further here.

[0074] After acquiring wide-area aerial reconnaissance capabilities, ground close-in observation capabilities, and communication relay capabilities, these capabilities are normalized to obtain the UAV's capability vector.

[0075] The normalization process employs a maximum-minimum normalization method. Taking ground-to-ground observation capability as an example, the maximum value among all combinations of ground-to-ground observation capabilities is obtained, and this maximum value is used to normalize the ground-to-ground observation capability of the UAV in the current combination.

[0076] Among them, the capability vector of the ground unit The meaning of each dimension in the vector corresponds one-to-one with the dimension of the UAV's capability vector. For each ground unit, an L-dimensional capability vector for the ground unit is constructed. , Where j is the serial number of the ground unit, For the wide-area aerial reconnaissance capability of ground unit j, For ground-based close-range observation capabilities of ground unit j, This refers to the communication relay capability of ground unit j.

[0077] It should be noted that the values ​​of each component in the capability vector of a ground unit are determined based on the specific capabilities of the ground unit; for capabilities that the ground unit cannot perform, such as "wide-area air reconnaissance", the corresponding component values ​​are usually set to 0.

[0078] As for the capabilities that ground units excel at, such as "close-range ground observation," their component values ​​are calculated based on the specific performance of their ground sensors using existing technical models such as the Johnson criterion.

[0079] Finally, the component values ​​in the capability vector of the ground unit also need to be normalized in the same way as the capability vector of the UAV to ensure that they are compared with the mission requirement vector under the same standard.

[0080] In this embodiment, each element in the fusion capability vector is composed of the maximum value of the corresponding dimension elements in the capability vectors of the two units.

[0081] The aforementioned fusion capability vector represents the capabilities of unmanned aerial vehicles (UAVs). and ground unit The combined capabilities. When drones... and ground unit When working collaboratively, the combined capability level of these components depends on the stronger unit. For example, in the wide-area aerial monitoring dimension, the capability value of the UAV is much greater than that of the ground unit (which has a value of 0), so the fused capability value is the UAV's capability value. In the ground sample collection dimension, the ground unit's close-range ground observation capability is much greater than that of the UAV, so the ground unit's capability value is used. In this embodiment, taking the maximum value is used to simulate the collaborative effect of complementing each other's strengths.

[0082] Secondly, obtain the task requirement vector of the tasks to be executed in the combination.

[0083] Specifically, the process of obtaining the task requirement vector for the task to be executed is as follows:

[0084] An ontology library is constructed, which includes standard task requirement vectors corresponding to different historical execution tasks. Based on the ontology library, the standard task requirement vector that matches the task to be executed in the combination is used as the task requirement vector of the task to be executed.

[0085] The standard task requirement vector is a standardized scoring vector based on its operational logic and task understanding; it is an L-dimensional vector. Each task requirement vector is mapped one-to-one with the task type of a historically executed task.

[0086] In this embodiment, L can be 3, where the dimensions are wide-area aerial surveillance capability, ground-based close-range observation capability, and communication relay capability, respectively. In this case, the mission requirement vector includes scores for wide-area aerial surveillance capability, ground-based close-range observation capability, and communication relay capability.

[0087] Specifically, the scoring rule is a 4-level scale: [1=Not needed, 2=Helpful, 3=Important, 4=Crucial]. When the task requirement corresponding to an execution task is [3, 1, 2], max-min normalization is used to map these integer levels to a standardized floating-point vector, so the task requirement vector is [1.0, 0.0, 0.5]. The mapping relationship from different historical tasks to task requirement vectors is stored to form an ontology library.

[0088] In this embodiment, when a task to be executed needs to be configured, the task type (e.g., "region search") is used as the key to retrieve the corresponding standard task requirement vector ([1.0,0.0, 0.5]) in the ontology library to complete the matching, and the obtained standard task requirement vector is used as the task requirement vector of the task to be executed.

[0089] Then, the dot product of the fusion capability vector and the task requirement vector is used as the task complementarity.

[0090] In this embodiment, the task complementarity represents the degree of matching between the capabilities of the air-ground unit and the task requirements; that is, if the capabilities most needed by the task happen to be the ones that this air-ground combination excels at (i.e., the values ​​of the corresponding dimensions are all large), then the dot product result is... A high score indicates a highly complementary and efficient combination; conversely, a low score indicates a mismatch between capabilities and needs.

[0091] Step S22: Calculate cooperative reachability.

[0092] The process of obtaining collaborative reachability is as follows:

[0093] First, acquire drones respectively Arrival at the task to be executed First estimated time for mission location and ground units Arrival at the task to be executed The second estimated time for the mission location.

[0094] The first estimated timeframe is: acquiring the drone. Current location and tasks to be performed The three-dimensional Euclidean distance between the mission location and the target location is then used as the first estimated time, and the ratio of this three-dimensional Euclidean distance to the standard cruise speed (i.e., average flight speed) of the UAV in its standard cruise mode is used as the first estimated time.

[0095] It should be noted that the above three-dimensional Euclidean distance is based on the assumption that the UAV's flight path in three-dimensional space is an unobstructed straight line, and other obstacles are negligible.

[0096] The second estimated timeframe is: Based on digital elevation model (DEM) data, an environmental cost map is constructed using... The search algorithm maps the location of ground units to tasks on the environmental cost map. The optimal path to the location is determined by summing the cost values ​​of all grid cells along that optimal path to obtain the second estimated time.

[0097] In this map, the cost of each grid cell is the travel time required for a ground unit to traverse that grid cell. This travel time is positively correlated with the grid's slope and a preset vehicle performance curve. For example, the steeper the slope, the slower the travel speed, and the longer the travel time. For grid cells that exceed their climbing capacity or have obstacles, their cost is set to infinity.

[0098] The vehicle performance curves mentioned above are speed-gradient curves; gradient and speed performance curves are usually provided by equipment manufacturers after empirical testing on a controlled gradient test platform.

[0099] Secondly, obtain the communication loss indicators between the current position of the UAV, the current position of the ground unit, and the position of the task to be performed.

[0100] In this embodiment, the process of obtaining the communication loss index is as follows:

[0101] First, radio frequency simulation software is used to calculate and generate a communication coverage map covering the entire work area based on the terrain data of the work area and the communication equipment parameters of each unit. Then, by querying the signal strength of the corresponding location combination in the communication coverage map, the communication loss index is obtained. The communication loss index is negatively correlated with the signal strength, that is, the stronger the signal, the smaller the value of the communication loss index.

[0102] The communication loss metric characterizes the cost of an effective communication link between the UAV and ground units near the mission location.

[0103] In one embodiment, the communication loss index is the value of a negative exponential function with the natural constant e as the base and the signal strength as the parameter.

[0104] The communication coverage map stores the signal strength or path loss between any combinations in space. The aforementioned RF simulation software is the open-source Radio Mobile or the commercial software ATDI HTZ, and the communication equipment parameters include antenna gain, transmit power, and operating frequency.

[0105] The above method uses radio frequency simulation software to pre-calculate and generate communication coverage maps, eliminating the need for complex real-time link calculations. Communication loss indicators between any combination of locations can be obtained simply by quick querying, ensuring both the professionalism and accuracy of communication assessment and meeting the real-time requirements of task allocation decisions.

[0106] Then, based on the first estimated time, the second estimated time, and the communication loss index, the cooperative reachability is obtained.

[0107] The specific formula for calculating collaborative reachability is as follows:

[0108] ;

[0109] in, For collaborative reachability; For drones Arrival at the task to be executed The first estimated time for the mission location; For ground unit Arrival at the task to be executed The second estimated time for the mission location; The set time tolerance is a positive real number used to adjust the sensitivity to the time difference of arrival; For drones Ground Unit and tasks to be performed Communication loss metrics between them Let be the maximum value of the communication loss index among all combinations, i, j, and k be the serial numbers of the UAV, ground unit, and task to be executed, respectively, and exp() be the exponential function.

[0110] In practical applications, communication loss is an objective reality, therefore It is a real number greater than 0, that is Not zero.

[0111] Time tolerance The value can be dynamically adjusted according to the urgency of the task, with the preferred range being [value missing]. Seconds. For time-sensitive tasks, A smaller value can be chosen to ensure close collaboration; for routine tasks, the value can be appropriately relaxed. This represents the absolute value of the difference between the first and second estimated times from the current position of the UAV and ground unit to the task location to be executed. A larger absolute value indicates greater difficulty in synchronizing their actions. Time tolerance. Normalizing the absolute value of the difference allows the sensitivity of this term to time differences to be adjusted. The larger the value, the worse the time synchronization.

[0112] The cost of maintaining an effective communication link between air-to-ground units near the mission location was evaluated. The lower the value, the smoother the communication and the higher the feasibility of space collaboration.

[0113] As can be seen from the above formula, collaborative reachability It is between and The value between [the estimated time difference between the drone and the ground unit]. If the absolute value of the estimated time difference between the drone and the ground unit is smaller, that is... The smaller the value, and the smaller the communication loss index, the closer the negative value of the exponential part is to... The closer the value of collaborative reachability is to Conversely, the larger the absolute value of the estimated time difference or the larger the communication loss index, the more exponentially the cooperative reachability will decrease.

[0114] Efficient collaboration requires units to arrive at the mission area at approximately the same time and maintain reliable communication during collaboration. Therefore, collaborative accessibility is used to evaluate the feasibility and stability of collaborative operations between air and ground units at the temporal, spatial, and communication levels.

[0115] Step S23: Calculate the collaborative gain of the corresponding combination based on the task complementarity and collaborative reachability.

[0116] Specifically, cooperative gain The calculation formula is as follows:

[0117] ;

[0118] in, For drones Ground Unit and tasks The task complementarity of the combined tasks For drones Ground Unit and tasks The collaborative accessibility of the combined components.

[0119] Step S3: Construct the corresponding global optimization function based on the collaborative gain of all combinations, and determine the task allocation scheme with the largest global optimization function of the cluster.

[0120] The process of obtaining the task allocation scheme is as follows:

[0121] First, construct a global optimization function.

[0122] After calculating the synergistic gain for all possible combinations, a global optimization function is constructed and solved to find the task allocation scheme that maximizes the total benefit of the entire cluster.

[0123] Specifically, the global optimization function Defined as the maximum sum of the net profits of all tasks to be executed, specifically expressed as:

[0124] ;

[0125] in, This is a binary decision variable; when it takes a value of 1, it indicates that the drone will be... and ground unit The groups were assigned to tasks to be performed. A value of 0 indicates no allocation; For drones Ground Unit and tasks to be performed The combined synergistic gain; and drones Ground Unit Arrival at the task to be executed The first execution cost and the second execution cost; I, J, and K represent the total number of UAVs, ground units, and tasks to be executed, respectively. The weighting coefficient is set to balance the importance of synergistic gains and execution costs in decision-making; the empirical value is 1.5.

[0126] The first execution cost of the drone is the sum of its time cost and energy cost; where the energy cost is the ratio of the estimated energy consumption from the drone's current location to the mission location to the drone's total battery capacity. The estimated energy consumption is the product of a first estimated time and the drone's average cruise power in its known standard cruise mode. The time cost is the ratio of the first estimated time to the preset total mission time limit.

[0127] The preset total task time limit is the maximum time required to complete the corresponding task to be executed.

[0128] The second execution cost of a ground unit is the sum of its time and energy components. The time component is the normalized value of the second estimated time of the ground unit in any combination; the energy component is the normalized value of the sum of the instantaneous power of all path segments in the optimal path. The instantaneous power of each path segment is calculated based on the terrain slope and traffic speed of that path segment, and by calling a pre-established ground unit power consumption model (such as a vehicle power consumption model).

[0129] Specifically, the normalized value of the second estimated time is the ratio of the second estimated time of the ground unit in any combination to the maximum value of the second estimated time of all ground units; the normalized value of the sum of instantaneous power is the ratio of the sum of instantaneous power of all path segments in the optimal path of the ground unit in any combination to the maximum value of the sum of instantaneous power of multiple ground units.

[0130] The power consumption model of the ground unit mentioned above can be constructed using existing vehicle dynamics simulation software (such as Simulink Powertrain Blockset). By inputting the vehicle's physical parameters (mass, size, motor characteristics, etc.), power consumption data under different operating conditions can be simulated and generated.

[0131] When obtaining the maximum value of the global optimization function, the following constraints must also be satisfied:

[0132] Each task can be executed by a combination of one drone and a ground unit in an air-to-ground manner.

[0133] Each drone belongs to a maximum of one group;

[0134] Each ground unit belongs to at most one group.

[0135] Secondly, an optimization algorithm is used to optimize the global optimization function in order to obtain the optimal task allocation scheme.

[0136] In this embodiment, after obtaining the global optimization function, it is necessary to solve the global optimization function to obtain the optimal task allocation scheme.

[0137] The above solution is an optimization problem, which is an integer linear programming problem. It can be solved in a specified time using heuristic algorithms (such as genetic algorithms and simulated annealing algorithms).

[0138] Since heuristic algorithms are existing technology, they will not be discussed further here.

[0139] In this embodiment, an optimal set of... The values ​​clearly define the drone. With ground unit The pairing situation and the corresponding tasks to be performed. This constitutes the final task allocation scheme.

[0140] Furthermore, in response to the final task allocation scheme, the control system sends instructions to the selected drone. and ground unit Issue collaborative execution tasks The specific instructions, including target waypoints and coordination strategies, guide the cluster to complete the task in a globally optimal manner.

[0141] For example, if The system then sends a signal to the drone. and ground unit Issue collaborative execution tasks The instructions include target waypoints and coordination strategies.

[0142] This invention constructs a dynamic collaborative gain for each potential combination of UAVs, ground units, and tasks by utilizing collaborative accessibility and task complementarity. This dynamic collaborative gain quantifies the potential value of air-ground collaboration. Collaborative accessibility assesses the feasibility and stability of air-ground unit cooperation from temporal, spatial, and communication perspectives. Task complementarity, through vectorized analysis, accurately calculates the matching degree between the capability combinations of air-ground units and specific task requirements. Finally, by solving a global optimization problem aimed at maximizing the total system gain, optimal task allocation is achieved. In other words, this invention overcomes the limitations of traditional methods that only consider independent costs, enabling truly global optimization of air-ground integrated UAV swarm resources. This significantly improves the overall combat effectiveness and mission success rate of the swarm, thereby enhancing the collaborative operational efficiency of the entire swarm.

[0143] In the description of this specification, "multiple" means at least two, such as two, three or more, etc., unless otherwise expressly and specifically defined.

[0144] While various embodiments of the invention have been shown and described in this specification, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and essence of the invention.

Claims

1. A method for integrated air-ground UAV swarm cooperative control, characterized in that, include: Randomly obtain a combination of a triple of any UAV, ground unit and task to be executed in the cluster, as well as the current position of the UAV, ground unit and task position of the task to be executed in the combination; Calculate the collaborative gain of each combination; construct the corresponding global optimization function based on the collaborative gain of all combinations, and determine the task allocation scheme with the largest global optimization function of the cluster; In response to the task allocation scheme, coordinated execution commands are issued to multiple UAVs and multiple ground units; Wherein, the collaborative gain is the product of the pre-acquired task complementarity and collaborative reachability; the task complementarity is the dot product of the fusion capability vector and the task requirement vector; each element in the fusion capability vector is the maximum value of the corresponding dimension in the capability vector of the UAV and the capability vector of the ground unit in the combination; the capability vector and the task requirement vector have one-to-one correspondence and the same dimension; the task requirement vector represents the requirement of the corresponding task to be executed for the collaborative capability of the UAV and the ground unit. The collaborative reachability is negatively correlated with the communication loss index of the combination and the absolute value of the difference between the current position of the UAV and the ground unit in the combination and the estimated time from the current position to the task position of the task to be executed. Global optimization function for: ;in, This is a binary decision variable; when it takes a value of 1, it indicates that the drone will be... and ground unit The groups were assigned to tasks to be performed. A value of 0 indicates no allocation; For drones Ground Unit and tasks to be performed The combined synergistic gain; and drones Ground Unit Arrival at the task to be executed The first execution cost and the second execution cost; I, J, and K represent the total number of UAVs, ground units, and tasks to be executed, respectively. The weighting coefficients are set to balance the importance of synergistic gains and execution costs in decision-making; Collaborative reachability for: ;in, , drones and ground unit The current position is respectively the position of the task to be executed. The first estimated time and the second estimated time for the task location; The set time tolerance; For drones Ground Unit and tasks to be performed Communication loss metrics between them This is the maximum value of the communication loss metric among all combinations, and The expression is not zero, exp() is an exponential function, and i, j, and k are the numbers of the UAV, ground unit, and task to be executed, respectively.

2. The air-ground integrated UAV swarm cooperative control method according to claim 1, characterized in that, The construction of the task requirement vector includes: Construct an ontology library, which includes standard task requirement vectors for different historically executed tasks; Based on the ontology library, the standard task requirement vector that matches the task to be executed in the combination is used as the task requirement vector of the task to be executed.

3. The air-ground integrated UAV swarm cooperative control method according to claim 1, characterized in that, The capability vector of the UAV includes its wide-area aerial reconnaissance capability, ground close-in observation capability, and communication relay capability; the capability vector of the ground unit includes its wide-area aerial reconnaissance capability, ground close-in observation capability, and communication relay capability.

4. The air-ground integrated UAV swarm cooperative control method according to claim 1, characterized in that, The method for obtaining the first estimated time includes: Obtain the three-dimensional Euclidean distance between the current position of the drone and the position of the task to be performed; Divide the three-dimensional Euclidean distance by the standard cruise speed of the corresponding UAV to obtain the first estimated time; the standard cruise speed is the average flight speed in the standard cruise mode.

5. The air-ground integrated UAV swarm cooperative control method according to claim 1, characterized in that, The second method for obtaining the estimated time includes: Constructing an environmental cost map based on a digital elevation model; use The search algorithm plans the optimal path from the current location of the ground unit to the task location on the environmental cost map; The sum of the travel times of the ground unit traversing all grids in the optimal path is used as the second estimated time; the travel time is positively correlated with the corresponding grid slope and the preset vehicle performance curve.

6. The air-ground integrated UAV swarm cooperative control method according to claim 1, characterized in that, The method for obtaining the communication loss index includes: Using radio frequency simulation software, a communication coverage map is generated based on the terrain data of the work area and the communication equipment parameters of each unit; The communication loss index is negatively correlated with the signal strength of the corresponding combination in the communication coverage map.

7. The air-ground integrated UAV swarm cooperative control method according to claim 5, characterized in that, The global optimization model also includes the following constraints: Each task can be executed by a combination of one drone and a ground unit in an air-to-ground manner. Each drone belongs to a maximum of one group; Each ground unit belongs to at most one group.

8. The air-ground integrated UAV swarm cooperative control method according to claim 7, characterized in that, The first execution cost is the sum of time cost and energy cost; Wherein, energy consumption cost is the ratio of the estimated energy consumption from the current position of the drone to the task position to the total battery capacity of the drone; the estimated energy consumption is the product of the first estimated time and the average cruise power of the drone in standard cruise mode; time cost is the ratio of the first estimated time to the preset total task time limit for completing the corresponding task to be executed. The second execution cost is the sum of the time component and the energy component; the time component is the normalized value of the second estimated time; the energy component is the normalized value of the sum of the instantaneous power of the ground units of all path segments in the optimal path; the instantaneous power is calculated based on the terrain slope and traffic speed of the corresponding path segment and by calling the pre-established ground unit power consumption model.

9. The air-ground integrated UAV swarm cooperative control method according to claim 1, characterized in that, The task allocation scheme that maximizes the global optimization function of the cluster is determined. This includes using heuristic algorithms to optimize the global optimization function.

Images