A target tracking trajectory planning method based on multi-unmanned aerial vehicle cooperation

Abstract

This invention specifically relates to a target tracking trajectory planning method based on multi-UAV cooperation. First, it utilizes a target dynamic allocation method based on consortium auctions, enabling the formation of a consortium for tracking tasks when the number of UAVs is greater than or equal to the number of targets. Furthermore, to address specific scenarios of cooperative tracking, a utility function is defined to consider constraints such as tracking distance and UAV dynamics. Then, the trajectory optimization problem is solved using the MINCO spatiotemporal decoupling framework based on differential flatness. This framework achieves independent spatiotemporal adjustment and designs multiple objective functions for cooperative tracking: maintaining tracking effectiveness through distance preservation, guiding the field of view using ray projection to avoid occlusion, and ensuring smooth maneuvering through velocity matching. By repeating these steps, UAV swarms can achieve stable and efficient cooperative tracking in complex scenarios.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) technology, and more specifically to a target tracking trajectory planning method based on multi-UAV collaboration. Background Technology

[0002] In related technologies, UAV cooperative autonomous target tracking refers to the process of using autonomous UAV systems to pursue and track non-cooperative targets in unknown or complex environments. This technology involves the perception, decision-making, and trajectory planning of autonomous UAVs to achieve stable tracking in complex scenarios. It has a wide range of applications, including photography, search and rescue, and patrol and surveillance.

[0003] Existing solutions, such as distributed task allocation algorithms, are mostly designed for the classic allocation problem of a single UAV performing a single task. They cannot handle complex allocation scenarios where multiple UAVs need to collaborate to complete the same task, which is precisely a key requirement in real-world applications such as cooperative tracking. Furthermore, most existing techniques are designed for static or quasi-static task allocation problems, making it difficult to effectively handle highly dynamic scenarios with real-time target movement and dynamic task changes (old targets disappearing, new targets appearing). During tracking, UAV cooperative trajectory planning must simultaneously satisfy multiple constraints, including obstacle avoidance, cooperative collision avoidance, tracking distance, and tracking time. However, existing methods often only focus on some constraints, lacking targeted optimization design for swarm tracking trajectories. Moreover, the high degree of coupling between various constraints makes the trajectory optimization problem extremely complex.

[0004] It should be noted that the information disclosed in the background section above is only used to enhance the understanding of the background of the present invention, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] This invention provides a target tracking trajectory planning method based on multi-UAV collaboration, which can realize collaborative target decision-making and collaborative tracking trajectory optimization, thereby effectively overcoming the defects of the prior art to a certain extent.

[0006] Other features and advantages of the invention will become apparent from the following detailed description, or may be learned in part by practice of the invention.

[0007] According to a first aspect of the present invention, a target tracking trajectory planning method based on multi-UAV cooperative operation is provided, the method comprising:

[0008] Acquire target information for each target to be tracked, and current status information for each drone in the drone swarm; the target information includes the current location of each target; the current status information of the drones includes their current location.

[0009] A distributed consortium auction method is invoked to perform collaborative target decision-making for each target, determining the tracking target corresponding to each UAV. The distributed consortium auction method estimates the profit of the UAV tracking target based on a utility function. The utility function includes: a tracking distance utility component for minimizing the travel distance and energy consumption of the UAV approaching the target, and a heading alignment utility component for evaluating the kinematic consistency between the current motion state of the UAV and the target orientation.

[0010] Based on the MINCO trajectory class, collaborative tracking trajectory planning is performed to determine the collaborative tracking trajectory corresponding to each UAV.

[0011] In some exemplary embodiments, the method further includes:

[0012] Based on the Euclidean distance between the UAV and the target, configure the tracking distance utility component in the utility function, including:

[0013]

[0014] in, , which represents the Euclidean distance between UAV i and target j; This represents the position vector of drone i; This represents the position vector of target j; This indicates the maximum value of the utility; This indicates a small amount to prevent singularity.

[0015] In some exemplary embodiments, the method further includes:

[0016] Based on the difference between the current velocity vector direction of the UAV and the azimuth angle of the target, the heading alignment utility component in the utility function is configured, including:

[0017]

[0018] in, This represents the difference between the current velocity vector direction of the UAV and the azimuth angle of the target.

[0019] In some exemplary embodiments, the utility function further includes a relative proximity rate utility component for assessing the feasibility of tracking a dynamic target;

[0020] The method further includes: configuring the proximity rate utility component of the utility function based on the relative velocity vector between the UAV and the corresponding target and the line-of-sight unit vector, including:

[0021]

[0022] in, This represents the relative velocity vector between the UAV and its corresponding target; This represents the line-of-sight unit vector, determined based on the UAV's position vector and the target's position vector.

[0023] In some exemplary embodiments, the utility function further includes a task persistence utility component for providing constraints for target switching;

[0024] The method further includes: utilizing The function configures additional utility for the target currently being tracked by the drone, including:

[0025]

[0026] in, This represents the target assigned to drone i at the previous moment.

[0027] In some exemplary implementations, a distributed consortium auction method is invoked to perform collaborative target decision-making for each objective, determining the tracking target corresponding to each drone, including:

[0028] For each drone Allocation status Initial values ​​of each target price in the target set Initial profits of drones drones Allocation targets, drones Initialize with partners;

[0029] Using an individual bidding strategy, each UAV bids for the target based on its utility function to determine the expected maximum profit and the corresponding optimal task; and to determine the expected second-highest profit and the corresponding second-best task.

[0030] When the number of drones exceeds the target number, a cooperative bidding strategy is used. Each drone determines its expected maximum cooperative profit and optimal cooperative task based on its utility function and the expected maximum profit of its partner drones from the target.

[0031] When the drone is currently idle, and the expected maximum profit based on the individual bidding strategy and / or the expected maximum profit based on the cooperative bidding strategy is greater than zero, the bidding type is configured according to the value of the maximum profit, and bidding is conducted; wherein, the bid for individual bidding is based on the utility function and the expected second-highest profit, combined with the preset bid increment configuration; the bid for cooperative bidding is based on the utility function and the expected maximum profit of the drone and the partner drone, combined with the preset bid increment configuration;

[0032] Construct separate sets of bids for individual bidding and sets of bids for cooperative bidding, and determine the winning tuple based on the bidding results; wherein the winning tuple includes: drone / drone alliance, and the corresponding target; and configure the corresponding drone as assigned;

[0033] Announcing information is generated based on the winning tuple, and the announcing information is then used to globally synchronize the status of each drone.

[0034] In some exemplary implementations, determining the winning tuple based on the bidding results includes:

[0035] When the profit from a single bid is no less than that from a replacement bid and no less than that from a cooperative bid, the winning tuple is configured based on the single bid strategy; or,

[0036] When the profit from individual bidding is lower than that from replacement bidding and lower than that from cooperative bidding, the winning tuple is configured based on the cooperative bidding strategy.

[0037] In some exemplary implementations, cooperative tracking trajectory planning is performed based on the MINCO trajectory class to determine the cooperative tracking trajectory corresponding to each UAV, including:

[0038] A trajectory planning problem for tracking is constructed based on target constraints; wherein, the target constraints include: a view perception constraint, wherein the view perception constraint is configured based on a penalty term of ray projection, used to control the UAV trajectory in a region with sparse obstacles to maintain observation of the target;

[0039] The trajectory planning problem is analyzed based on the current location information of the drone to obtain the corresponding collaborative tracking trajectory planning results.

[0040] In some exemplary embodiments, the target constraints further include: velocity matching constraints and / or tracking distance constraints; wherein,

[0041] The velocity matching constraint is configured based on the velocity matching cost function to control the difference between the velocity vectors of the UAV and the target to remain as expected;

[0042] The tracking distance constraint is configured based on the spatiotemporal cooperative target tracking penalty function to control the UAV and the target to maintain the expected tracking distance.

[0043] In some exemplary embodiments, the target constraints further include: end constraints, intermediate point constraints, and time constraints; smoothing constraints, obstacle avoidance constraints, dynamic constraints, and cooperative inter-machine obstacle avoidance constraints.

[0044] Trajectory planning problems oriented towards tracking based on objective constraints include:

[0045]

[0046] in, , , , These represent smoothing constraints, obstacle avoidance constraints, dynamic constraints, and cooperative inter-machine obstacle avoidance constraints, respectively. Indicates tracking distance constraints; Indicates visual perception constraints; Indicates speed matching constraints; This represents the corresponding weighting coefficient.

[0047] According to a second aspect of the present invention, a computer program product is provided, on which a computer program is stored, and when the computer program is executed by a processor, the above-described target tracking trajectory planning method based on multi-UAV cooperation is implemented.

[0048] According to a third aspect of the present invention, an electronic device is provided, comprising:

[0049] Processor; and

[0050] Memory for storing the executable instructions of the processor;

[0051] The processor is configured to implement the above-described target tracking trajectory planning method based on multi-UAV cooperation when executing the executable instructions.

[0052] According to a fourth aspect of the present invention, a storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the above-described target tracking trajectory planning method based on multi-UAV cooperative operation.

[0053] The target tracking trajectory planning method based on multi-UAV collaboration provided by the embodiments of the present invention proposes a distributed mechanism based on consortium auction when making target decisions. This mechanism achieves autonomous negotiation through an iterative process of bidding, clearing, and notification. Simultaneously, by constructing a utility function using tracking distance utility and heading alignment utility, and combining this utility function with profit estimation during the bidding phase, the method effectively integrates the comprehensive utility of multiple dimensions such as flight, speed, and distance. This provides precise quantitative input for target decision-making, thereby effectively improving the quality and robustness of the decision-making process. Ultimately, this ensures the generation of high-quality tracking trajectories that meet constraints.

[0054] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Attached Figure Description

[0055] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention. It is obvious that the drawings described below are merely some embodiments of the invention, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0056] Figure 1 The illustration shows a schematic diagram of a target tracking trajectory planning method based on multi-UAV cooperation, an exemplary embodiment of the present invention.

[0057] Figure 2 The illustration shows a schematic diagram of a collaborative target decision-making method flow according to an exemplary embodiment of the present invention;

[0058] Figure 3 This diagram illustrates an exemplary embodiment of the present invention of a scenario in which parallel tracking leads to tracking failure.

[0059] Figure 4 The diagram illustrates an exemplary embodiment of the present invention, a view perception penalty.

[0060] Figure 5 The illustration shows a schematic diagram of a sparse obstacle scene according to an exemplary embodiment of the present invention;

[0061] Figure 6 This schematic diagram illustrates a scene with dense obstacles, as exemplified by an embodiment of the present invention.

[0062] Figure 7 This schematic diagram illustrates a tracking start state in a sparse obstacle scene according to an exemplary embodiment of the present invention;

[0063] Figure 8 This schematic diagram illustrates the state during tracking in a sparse obstacle scene according to an exemplary embodiment of the present invention.

[0064] Figure 9 This schematic diagram illustrates an exemplary embodiment of the present invention of tracking termination state in a sparse obstacle scene;

[0065] Figure 10 This schematic diagram illustrates a tracking start state in a dense obstacle scene, as an exemplary embodiment of the present invention.

[0066] Figure 11 This schematic diagram illustrates the state during tracking in a dense obstacle scene, as per an exemplary embodiment of the present invention.

[0067] Figure 12 This schematic diagram illustrates the termination of tracking in a dense obstacle scene, an exemplary embodiment of the present invention.

[0068] Figure 13 This schematic diagram illustrates an exemplary embodiment of the present invention of a drone exiting a tracking state;

[0069] Figure 14 This schematic diagram illustrates an exemplary embodiment of the present invention where a drone continues to track a target even after it has exited the system.

[0070] Figure 15 The diagram illustrates the composition of an electronic device according to an exemplary embodiment of the present invention. Detailed Implementation

[0071] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the invention will be more comprehensive and complete, and will fully convey the concept of the exemplary embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0072] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0073] In related technologies, existing collaborative decision-making algorithms are difficult to adapt to dynamically changing task sets and alliance collaboration requirements; and under multiple constraints, existing methods have the problem of struggling to plan the optimal trajectory that takes into account obstacle avoidance, visibility and dynamics.

[0074] To address the shortcomings and deficiencies of existing technologies, this example implementation provides a target tracking trajectory planning method based on multi-UAV cooperation. (Reference) Figure 1 As shown, the method may specifically include:

[0075] Step S11: Obtain the target information corresponding to each target to be tracked, and the current status information of each drone in the drone cluster; wherein, the target information includes the current position of each target; the current status information of the drone includes the current position.

[0076] Step S12: Invoke the distributed consortium auction method to make collaborative target decisions for each target and determine the tracking target corresponding to each UAV; wherein, the distributed consortium auction method estimates the profit of the target tracked by the UAV based on a utility function; the utility function includes: a tracking distance utility component for minimizing the travel distance and energy consumption of the UAV approaching the target, and a heading alignment utility component for evaluating the kinematic consistency between the current motion state of the UAV and the target orientation;

[0077] Step S13: Perform collaborative tracking trajectory planning based on the MINCO trajectory class to determine the collaborative tracking trajectory corresponding to each UAV.

[0078] The method disclosed herein addresses the problem that existing collaborative decision-making algorithms are unable to adapt to dynamically changing task sets and alliance collaboration requirements; and that existing methods struggle to plan optimal trajectories that balance obstacle avoidance, visibility, and dynamics under multiple constraints. It proposes a UAV collaborative autonomous tracking trajectory planning method for complex scenarios, enabling UAV swarms to achieve stable and efficient collaborative tracking in complex scenarios.

[0079] The following will describe in more detail each step of the target tracking trajectory planning method based on multi-UAV cooperation in this exemplary embodiment, with reference to the accompanying drawings and embodiments.

[0080] In step S11, target information corresponding to each target to be tracked and current status information of each drone in the drone cluster are obtained; wherein, the target information includes the current position of each target; and the current status information of the drone includes the current position.

[0081] For example, the aforementioned drone can be a micro-rotor drone. For instance, the trajectory planning method described above can be executed at the control terminal of the drone swarm to generate planned paths for each drone.

[0082] Specifically, in the system tracing task, you can configure... A drone and One target. The set of drones is represented as... Furthermore, in a drone swarm, all drones can communicate with each other and synchronize their clocks.

[0083] The set of targets is represented as A subset of drones consisting of no more than two drones is called a coalition. A set of drone coalitions is represented as... The assignment pair is a pair of alliances and tasks. (By alliance) and tasks The assigned pair is represented as If and only if the task It can be done by the alliance Upon completion, the allocation pair It is feasible. The set of all feasible assignment pairs is represented as: To maximize the advantages of multi-drone collaborative tracking, two assignment pairs with shared drones or tasks cannot be selected simultaneously. That is, for two assignment pairs... and ,have and For a single drone The alliance formed, we use its index It indicates that it consists of two different drones. and The alliance formed is represented as ,Will and The assigned pair is represented as ,Depend on and The composition of the pairing is represented as . and The returns are expressed as follows: and .

[0084] Each feasible allocation pair is associated with a utility. This is determined by the utility function. Determined, making for .

[0085] For example, when multiple targets to be tracked are identified through signal acquisition equipment such as radar, the current position of each target, as well as the current position of each drone in the drone swarm, can be obtained. Simultaneously, tracking tasks can be created for each target separately, triggering the allocation of collaborative tracking tasks across the drone swarm.

[0086] In step S12, a distributed consortium auction method is invoked to perform collaborative target decision-making for each target and determine the tracking target corresponding to each UAV. The distributed consortium auction method estimates the profit of the target tracked by the UAV based on a utility function. The utility function includes a tracking distance utility component for minimizing the travel distance and energy consumption of the UAV approaching the target, and a heading alignment utility component for evaluating the kinematic consistency between the current motion state of the UAV and the target orientation.

[0087] For example, the method further includes configuring the tracking distance utility portion of the utility function based on the Euclidean distance between the UAV and the target.

[0088] For example, the method further includes configuring the heading alignment utility component in the utility function based on the difference between the current velocity vector direction of the UAV and the azimuth angle of the target.

[0089] Specifically, in existing technologies, most multi-UAV task allocation algorithms focus on maximizing the number of tasks completed or improving the overall utilization of the cluster. However, in real-world cooperative tracking trajectory planning scenarios, the simple metric of the number of tasks completed often struggles to cope with complex physical constraints. The system faces complex constraints and multiple challenges related to the high coupling between spatial distance, UAV motion dynamics, and temporal stability in UAV clusters. Therefore, it is necessary to design a utility function for cooperative tracking trajectory planning, so that the allocation problem is no longer a point-to-point matching, but a comprehensive dynamic process that integrates multiple constraints.

[0090] Considering that actual tracking task allocation is not a single-dimensional point-to-point allocation, but a multi-dimensional dynamic process, this method can pre-configure utility functions. Specifically,

[0091] First, there is the need to reconcile tracking distance with the feasibility of movement. Tracking distance is the foundation for task allocation in cooperative tracking. Theoretically, tracking the closest drone should yield the highest efficiency. However, drones are limited by their dynamics models and cannot perform maneuvers with excessively large angles. If distance is used as the sole metric, it is easy for a drone to be assigned to a target in the opposite direction of its movement, thus forcing it to perform maneuvers that violate dynamic constraints. Therefore, the algorithm can pursue the closest distance while also considering heading alignment to ensure the smoothness of the generated trajectory.

[0092] Secondly, real-time assessment of dynamic target motion can be considered. The relative motion state between the target and the drone determines the ease or difficulty of tracking the target. In high-speed dynamic scenarios, the feasibility and difficulty of tracking a target in an approaching state and a target in an escape state are drastically different, even if the current instantaneous distance is the same. By introducing the consideration of relative velocity vectors, the tracking priority of targets can be effectively distinguished, avoiding the system wasting drone resources on targets that are difficult to track.

[0093] Finally, the stability of the allocation algorithm should also be considered. During the execution frequency of the allocation algorithm, noise at the perception layer or uncertainty in the appearance of the target can induce jumps in the allocation results; frequent target switching can lead to oscillations in subsequent trajectory planning and control. Therefore, the allocation strategy can possess a certain degree of task persistence, and a lag mechanism can be introduced to maintain the robustness of the system.

[0094] Based on the above considerations, a comprehensive utility function integrating target motion situational awareness and stability is proposed. .

[0095] First, the tracking distance utility can be configured. This is designed to minimize the drone's travel distance and energy consumption when approaching a target. drones With the goal The position vector, and the tracking distance utility are expressed as:

[0096]

[0097] in, , which represents the Euclidean distance between UAV i and target j; This represents the position vector of drone i; This represents the position vector of target j; This indicates the maximum value of the utility; This indicates a small amount to prevent singularity.

[0098] The closer the distance between the drone and the target, the higher the utility. This represents the maximum value of the utility. 。

[0099] Secondly, heading alignment utility can be configured. This evaluates the kinematic consistency between the UAV's current motion state and the target's orientation.

[0100] Define the line-of-sight unit vector as The heading alignment utility is expressed as:

[0101]

[0102] in, Indicates the drone's current speed; This represents the difference between the current velocity vector direction of the UAV and the azimuth angle of the target.

[0103] In a physical sense, this term represents the normalized projection of the current velocity vector onto the line-of-sight direction. When the drone is flying directly toward the target, it can track without drastic maneuvers, and the utility function value is at its maximum.

[0104] Correspondingly, utility functions can include:

[0105]

[0106] in, These are the weighting coefficients for each sub-item.

[0107] Furthermore, in some exemplary embodiments, the utility function further includes a relative approach rate utility portion for evaluating the feasibility of tracking a dynamic target. The method also includes configuring the relative approach rate utility portion of the utility function based on the relative velocity vector between the UAV and the corresponding target, and the line-of-sight unit vector.

[0108] Specifically, for relative proximity rate utility, this term incorporates target velocity information to assess the feasibility of tracking dynamic targets. Let Let the velocity estimate vector of the target be denoted by , then the relative velocity vector is . Focusing on the projection of relative velocity onto the line of sight, the relative approach velocity utility is expressed as:

[0109]

[0110] in, This represents the relative velocity vector between the UAV and its corresponding target; This represents the line-of-sight unit vector, determined based on the UAV's position vector and the target's position vector.

[0111] If this value is greater than 0, it indicates that the distance between the two is decreasing, the task is easy to perform, and a reward should be given. If this value is less than 0, it indicates that the target is fleeing or moving away, tracking requires more energy, and the effectiveness is reduced.

[0112] Correspondingly, utility functions can include:

[0113]

[0114] in, , , These are the weighting coefficients for each sub-item.

[0115] Furthermore, in some exemplary embodiments, the utility function further includes a task persistence utility component for providing constraints on target switching. The method also includes: utilizing... The function configures additional utility for the target currently being tracked by the drone.

[0116] Specifically, to suppress the frequent task switching phenomenon common in task decision-making algorithms, a lagged reward term, namely task persistence utility, is introduced.

[0117] set up For the drone at the previous moment The assigned objective and the task's sustained utility are represented as follows:

[0118]

[0119] in, This represents the target assigned to drone i at the previous moment.

[0120] use The function gives the current tracking target additional utility, ensuring that the system will only trigger a target switch if the utility of the new target is significantly higher than that of the old target.

[0121] Correspondingly, utility functions can be configured. Represented as:

[0122]

[0123] in, These are the weighting coefficients for each sub-item.

[0124] Specifically, This represents a static tracking distance minimization utility function, used to minimize the travel distance and energy consumption of the UAV when approaching the target; This represents the utility function for evaluating heading alignment, used to assess the kinematic consistency between the UAV's current motion state and the target's orientation; This represents the relative proximity rate utility function, used to evaluate the feasibility of tracking dynamic targets; This represents the task persistence utility function, used to provide constraints for target switching.

[0125] By pre-constructing the utility function and designing its components for collaborative tracking trajectory planning, the allocation problem is transformed from point-to-point matching into a comprehensive dynamic process that integrates multiple constraints.

[0126] For example, in step S12 above, the distributed consortium auction method is invoked to perform collaborative target decision-making for each target, determining the tracking target corresponding to each drone, including:

[0127] Step S21, for each UAV Allocation status Initial values ​​of each target price in the target set Initial profits of drones drones Allocation targets, drones Initialize with partners;

[0128] Step S22: Using an individual bidding strategy, each UAV bids for the target individually based on its utility function to determine the expected maximum profit and the corresponding optimal task; and to determine the expected second-highest profit and the corresponding second-best task.

[0129] Step S23: When the number of drones exceeds the target number, a cooperative bidding strategy is used to determine the expected maximum cooperative profit and the optimal cooperative task for each drone based on its utility function and the expected maximum profit of the partner drones to the target.

[0130] Step S24: When the drone is currently idle and the expected maximum profit based on the individual bidding strategy and / or the expected maximum profit based on the cooperative bidding strategy is greater than zero, configure the bidding type according to the value of the maximum profit and conduct the bidding; wherein, the bid for individual bidding is based on the utility function and the expected second-highest profit, combined with the preset bid increment configuration; the bid for cooperative bidding is based on the utility function and the expected maximum profit of the drone and the partner drone, combined with the preset bid increment configuration;

[0131] Step S25: Construct the bid sets for individual bidding and cooperative bidding respectively, and determine the winning tuple based on the bidding results; wherein, the winning tuple includes: drone / drone alliance, and the corresponding target; and configure the corresponding drone as assigned;

[0132] Step S26: Generate notification information based on the winning tuple, and synchronize the notification information with the global status of each drone.

[0133] Specifically, the distributed consortium auction algorithm addresses the following problem: given a tuple Find the allocation that yields the highest return. . refer to Figure 2 As shown, the method may include the following steps: Bidding initialization. First, the algorithm performs profit estimation. The drone first calculates two bidding strategies: individual bidding. and cooperative bidding The algorithm calculates the expected maximum profit. Then, when the drone is currently in an "idle" state and the maximum potential profit among feasible strategies is strictly greater than zero, the drone begins bidding. The algorithm first evaluates the payoff of "individual bidding." If the profit of individual bidding is no less than that of replacement bidding and no less than that of cooperative bidding, the agent will prioritize this strategy. If these conditions are not met, the algorithm further evaluates "cooperative bidding." Finally, the algorithm broadcasts the information. Once the optimal bidding triplet is determined, it is encapsulated in a bidding message and broadcast to other drones. If the drone is not in an idle state, it inherits the utility value from the previous round.

[0134] Specifically, during the bidding phase, initialization can be performed first for each drone. The following variables need to be initialized. The subscript of the variable represents the drone's index, and the superscript represents the auction iteration round. Initialization may include:

[0135] 1) Drones Allocation status Each agent can be in one of three different allocation states: idle, bidding, or allocated. The corresponding labels are "idle," "bid," and "asg," respectively. Initially, all drones are in the idle state.

[0136] 2) The price of each target price in the target set For all The price is initially set to 0.

[0137] 3) Profits of drones Drones The profit, which is the maximum value of the return function among all objectives, is expressed as: .

[0138] 4) Drones Allocation target The initial value is -1 when no target is assigned.

[0139] 5) Drones Partners The initial value is -1 when no alliance is assigned, and it is also -1 if there are no partners.

[0140] During the profit calculation phase, individual bidding and cooperative bidding can be implemented.

[0141] Individual bidding is conducted for tasks performed independently. (Drone) The maximum profit that can be obtained through individual bidding is expressed as:

[0142]

[0143] in, This indicates the maximum profit for drones in a single auction, in the bid... Indicates separate bidding. This represents the set of tasks currently being assigned. Let i represent the utility function of drone i with respect to target j.

[0144] Optimal Task Choose according to the following formula:

[0145]

[0146] In addition, to accelerate the convergence of the auction algorithm and reduce the number of iterations, the second-highest profit in the individual bidding process can also be calculated:

[0147]

[0148] For example, in an iteration cycle of the bidding phase, for each drone in the drone swarm, the current state of each drone can be determined first. This can be done by calculating the corresponding profit for each target in the target set for the drones in the idle state, thereby filtering out the highest profit and the corresponding target, as well as the second highest profit and the corresponding target.

[0149] Alternatively, in some exemplary embodiments, during the bidding stage, the profit of targets for both idle and assigned drones can be calculated together, which facilitates target replacement for assigned drones.

[0150] Specifically, when the number of drones exceeds the number of targets, such as when there is redundancy in the number of drones or when a target is removed from the target set, two drones need to form an alliance to track the same target. In this case, the alliance needs to jointly bid for the target.

[0151] Regarding the alliance's cooperative bidding strategy, drones Partner drones and its tasks The allocation pairs satisfy:

[0152]

[0153] Among them, the subscript Indicates cooperative bidding. This indicates the offer made by the cooperating drone for the target.

[0154] Correspondingly, the optimal profit from their cooperative bidding is expressed as:

[0155]

[0156] in, Valuation of the drone partner in the bidding process.

[0157] Since the partners' bids are unknown at this stage, the bids can be made using the partners' expected maximum profit from the target.

[0158] For example, during the initialization phase, the number of drones and targets can be compared. If the number of targets is less than the number of drones, then during the bidding phase, an individual bidding strategy can be executed for each drone. The utility function is used to calculate the expected maximum profit and corresponding optimal task (i.e., target) based on the individual bidding, as well as the expected second-highest profit and corresponding task. Afterwards, a cooperative bidding strategy can be executed. Based on the utility function and the expected maximum profit of cooperative drones for the target, partner drones are selected, and the expected maximum profit of cooperative bidding is calculated. Specifically, for the current drone, other drones in the drone set can be traversed. Based on the profits of other drones for the target, combined with the utility function and the current drone's price for the target in the previous round, the cooperative bidding profit is calculated, and the drone with the highest profit is selected as the partner. The utility function for cooperative bidding is calculated based on the current drone, cooperative drones, and the target.

[0159] These other drones can be either simply idle drones, or they can be both idle and assigned drones.

[0160] For example, the above-mentioned determination of the winning tuple based on the bidding result includes: configuring the winning tuple based on the individual bidding strategy when the profit of the individual bidding is not lower than the replacement bidding and not lower than the cooperative bidding; or, configuring the winning tuple based on the cooperative bidding strategy when the profit of the individual bidding is not lower than the replacement bidding and not lower than the cooperative bidding.

[0161] Specifically, drones are only incentivized to bid when there is a type of bidding that generates a positive profit. That is, based on the expected maximum profit from individual bidding. The expected maximum profit corresponds to cooperative bidding. If at least one of the following is positive, then bidding will proceed. Furthermore, based on the profit value, the strategy corresponding to the maximum profit can be selected to determine whether to bid individually or in a cooperative manner.

[0162] In determining the drone After the bidding method is selected, its allocation status will be set to "bid", and a bid price will be provided for the task with the highest profit.

[0163] Specifically, in order to maximize the chances of winning the auction, the bidding price is set to the highest possible value.

[0164] drones The individual bid for the target can be configured based on the utility function and the second-highest profit, as expressed by the formula:

[0165]

[0166] in, This represents the target value corresponding to the highest expected profit based on drone i and a single bidding strategy. A defined utility function.

[0167] Furthermore, as a preferred option, a bid increment can be configured for each individual bid. This applies to drones. The formula for a separate bid on a target is expressed as:

[0168]

[0169] in, This indicates the increase in bid volume.

[0170] This price ensures that the robot's profit is no less than the second-highest profit from a standalone bid.

[0171] Simply put, for a standalone bidding strategy, the bid price equals the task reward. Profit from the second-best option + This means that if the bid price is greater than 0, then the current task is the one with the best profit.

[0172] Specifically, for cooperative bidding within an alliance, the bidding price can be calculated based on the expected maximum profit of the drone and partner drones corresponding to the target, combined with a utility function. The formula is as follows:

[0173]

[0174] in, Partner The profits can be viewed as drone profits. A "cooperation fee" that must be paid to teammates in order to achieve cooperation.

[0175] Furthermore, as a preferred option, a bid increment can be configured for cooperative bidding. The bid price for cooperative bidding can then be expressed as:

[0176]

[0177] If the total revenue of the alliance is less than the sum of the current revenues of the two drones, this item will result in a negative calculated profit, thus preventing cooperative bidding from occurring.

[0178] In the above embodiments, by configuring the bidding increment, the price is ensured to rise continuously as the bidding process proceeds, thereby ensuring that the algorithm terminates within a finite time.

[0179] In some embodiments, if the number of drones equals the number of targets, then only a single bid is made, and the bidding process executes a classic auction algorithm.

[0180] Specifically, during the market clearing and announcement phase, it's possible to determine whether a drone in the bidding process is the auction winner. This can be done by analyzing the collected bids. Extract the winning tuple. (Bid set) ;in, This indicates a bid submitted in a standalone auction. This indicates the bid submitted by the alliance in the auction. Among them,

[0181]

[0182]

[0183] For a set of bids, the drone with the highest bid can be selected as the winner. Furthermore, the drone's status can be configured to be assigned. Simultaneously, the current market clearing price can be determined, and an announcement message containing updated status, target, and price information can be broadcast to other drones. Conversely, if a drone fails to win the auction, it is considered a loser, reset to an idle state, and its target mission index, partner identifier, and profit value are reset to zero, awaiting the next round of auction.

[0184] For example, for a drone that is already assigned a task in the current round, we can first detect the tasks it already has. Does a new valid bid exist? If the bid set is not empty, it indicates that the task faces competition from other drones, and the algorithm needs to recalculate the winner. In this scenario, if the drone still exists in the winning alliance... This indicates that the node has successfully defended ownership of the mission, at which point the node will update its partners based on the composition of the winning alliance. Among them, partners It may be empty and inherits the utility value from the previous round. However, if a node is not in the winning coalition, it signifies that the drone has been replaced by a higher-value bid, thus forcing it to leave the task, resetting to an idle state, resetting its utility to zero, and broadcasting an announcement message containing empty target information to the network to announce the release of the resource. Finally, if the task does not receive any new bids, the algorithm will execute a state-preserving strategy, and the node will directly inherit the utility value from the previous round.

[0185] During the notification phase, the algorithm performs global state synchronization. Regardless of the agent's current state, it first updates its locally maintained task price vector based on all notification messages received in the previous phase (market clearing phase). Profit status of neighboring nodes Subsequently, for agents in an idle state, their utility value in the next round must be calculated. It also broadcasts to other drones.

[0186] Specifically, the distributed auction algorithm went through three stages: bidding, market clearing, and announcement, completing one round of iterations. The algorithm terminated when all drones were assigned to the target.

[0187] Specifically, in the collaborative goal decision-making process, a "distributed consortium auction" strategy is adopted. Through an iterative closed loop between "bidding (profit calculation)" and "settlement and notification (information synchronization)," task allocation and conflict resolution among multiple agents are achieved, meeting the needs of tasks with an imbalance between the number of tasks and the number of drones. Meanwhile, to support the efficient operation of this method, the system constructs a decision-making basis based on a utility function. This function not only fully considers the constraints and dynamic characteristics of the task but also deeply integrates multi-dimensional key indicators, including distance, heading, speed, and persistence, providing precise quantitative input for the auction process.

[0188] In step S13, collaborative tracking trajectory planning is performed based on the MINCO trajectory class to determine the collaborative tracking trajectory corresponding to each UAV.

[0189] For example, step S13 described above may include:

[0190] Step S31: Construct a trajectory planning problem for tracking based on target constraints; wherein the target constraints include: line-of-sight perception constraints, wherein the line-of-sight perception constraints are configured based on the penalty term of ray projection, used to control the UAV trajectory in areas with sparse obstacles to maintain observation of the target;

[0191] Step S32: Analyze the trajectory planning problem based on the current location information of the UAV to obtain the corresponding cooperative tracking trajectory planning result.

[0192] Specifically, when using a rotary-wing UAV, considering the differential flatness characteristic of rotary-wing UAVs, trajectory planning can be directly performed on a time-differentiable curve. To facilitate spatiotemporal joint optimization of the trajectory and fully utilize the maneuverability of the UAV, all trajectory planning in this invention uses the MINCO trajectory class to represent the UAV's motion trajectory.

[0193] The MINCO trajectory class is essentially a minimal control polynomial trajectory class, defined as:

[0194]

[0195] in, It indicates a Vie, A segment of polynomial locus, the order of which is ; Let be the order of the relevant integral chain. represent Segmented polynomial coefficient matrix, Represents intermediate path points. This represents the duration of each polynomial trajectory segment.

[0196] Specifically, trajectory Defined as:

[0197]

[0198] No. Segment trajectory Defined as:

[0199]

[0200] in, Represents a polynomial natural basis. Representing the The polynomial coefficient matrix of the segment trajectory, Representing the The duration of the trajectory segment, represent The total duration of the polynomial trajectory.

[0201] The core lies in parameter mapping This means that any MINCO trajectory can be utilized. and This is used for parameterization. Simultaneously, a defined MINCO trajectory represents the trajectory under specified initial and terminal conditions. dimension The minimum control trajectory of the integrator chain satisfies the following control minimization:

[0202]

[0203] in, It is a positive definite diagonal matrix. and These are the initial and terminal conditions, respectively. Given intermediate path points, the trajectory is in It must be passed at any time .

[0204] In addition, parameter mapping It has linear complexity, and its more specific correspondence can be expressed as:

[0205]

[0206] Among them, when hour, It is a non-singular banded matrix. This represents two ways of representing the MINCO trajectory class. and The transformation between them can be achieved through striped PLU matrix decomposition. The time and space complexity.

[0207] For a defined objective function that allows the use of gradients Both MINCO trajectory classes can be used. To the objective function Parameterization can be specifically represented as:

[0208]

[0209] That is, to change the original... The objective function is represented Transform into The objective function is represented Subsequently, mapping Allows for the original gradient to be computed in a linear computational complexity. and Calculate gradient and This greatly facilitates subsequent optimization.

[0210] Specifically, based on existing constraints, this invention adds custom target constraints to complete the modeling of the spatiotemporal trajectory optimization problem for cooperative target tracking. The custom target constraints include at least one: a view perception constraint. This view perception constraint is configured based on a penalty term of ray projection to control the UAV trajectory in obstacle-sparse regions to maintain target observation.

[0211] Preferably, in some exemplary embodiments, the target constraint further includes: a velocity matching constraint and / or a tracking distance constraint; wherein the velocity matching constraint is configured based on a velocity matching cost function to control the difference between the velocity vectors of the UAV and the target to remain as expected; and the tracking distance constraint is configured based on a spatiotemporal cooperative target tracking penalty function to control the UAV and the target to maintain the expected tracking distance.

[0212] Specifically, the custom target constraints can include: spatiotemporal cooperative target tracking penalties, line-of-sight perception exclusion penalties, and velocity matching penalties, etc., specifically designed for cooperative target tracking problems, thereby completing the modeling of the spatiotemporal trajectory optimization problem for cooperative target tracking. Regarding the yaw angle setting, this invention always points it towards the target to ensure that the UAV always faces the target during tracking. The position trajectory planning problem for tracking can be specifically expressed in the following form:

[0213]

[0214] in, This represents end constraints, intermediate point constraints, and time constraints.

[0215] This represents a safety constraint; the trajectory must be planned within a safe space free of obstacles.

[0216] ; indicates the dynamic constraints of the drone, the maximum speed and maximum acceleration on the trajectory cannot exceed the dynamic limits of the drone.

[0217] This indicates that the drone's trajectory should be planned in areas with sparse obstacles in order to maintain observation of the target.

[0218] This indicates that during the tracking process, the drone should maintain an appropriate tracking distance from the target.

[0219] This means that the velocity vectors of the drone and the target should be as consistent as possible during the tracking process to ensure a smooth tracking process.

[0220] By constructing soft constraints, the constraints at the endpoints, intermediate points, and outside of time are transformed into the following unconstrained trajectory optimization problems:

[0221]

[0222] in, , , , These represent smoothing constraints, obstacle avoidance constraints, dynamic constraints, and cooperative inter-machine obstacle avoidance constraints, respectively. Indicates tracking distance constraints; Indicates visual perception constraints; Indicates speed matching constraints; This represents the corresponding weighting coefficient.

[0223] Specifically, regarding the constraints mentioned above:

[0224] 1) Smoothing constraint To ensure that the generated trajectory is smooth and easy to control, this method uses minimizing the jerk (the third derivative of the position trajectory) as a regularization term.

[0225] For those defined in the time interval trajectory The smoothness cost function is expressed as:

[0226]

[0227] 2) Obstacle avoidance constraints Environmental safety is a fundamental constraint for UAV flight. This method utilizes an Euclidean Symbolic Distance Field (ESDF) map to obtain environmental gradient information. For the trajectory... For any point on the gradient, if its distance from an obstacle is less than a safety threshold, a penalty is applied. To ensure the continuity of the gradient, a cubic penalty function is used, expressed as:

[0228]

[0229] 3) Dynamic constraints To ensure the generated trajectory is physically executable, the maximum speed of the trajectory must be limited. and maximum acceleration The penalties for exceeding the limits of velocity and acceleration are defined as follows:

[0230]

[0231] The total dynamic feasibility cost is .

[0232] 4) Inter-machine collision avoidance constraints In multi-drone collaboration, collisions between drones must be prevented. Considering the influence of airflow on drones, this method models the safe zone between drones as an ellipsoid rather than a sphere. For drones... and drones Their positions are respectively Define ellipsoidal distance metric :

[0233]

[0234] in, a,b These are the semi-axis lengths of the ellipsoid in the vertical and horizontal directions, respectively.

[0235] The cost of mutual obstacle avoidance is defined as:

[0236]

[0237] 5) For the tracking distance constraint, a spatiotemporal cooperative target tracking penalty function is constructed to optimize the tracker's trajectory during tracking. This ensures that the trajectory always closely follows the target's path, thereby reducing the risk of target loss. The penalty term is defined as follows:

[0238]

[0239]

[0240] in, and These are the tracking distance thresholds.

[0241] To facilitate solving the problem using an optimizer, the gradient of the cost function with respect to the UAV's position is given:

[0242]

[0243] By configuring this penalty, it can be ensured that when the drone lags behind the target, it generates a pull towards the target, and when the distance is too close, it generates a thrust, thus maintaining an appropriate tracking distance with the target. This solves the problem of... Figure 3 The scenario shown illustrates how the target can be easily obscured by obstacles during "parallel" tracking.

[0244] 6) For velocity matching constraints, a velocity matching cost function can be configured to control the difference between the velocity vectors of the UAV and the target to remain as expected.

[0245] Specifically, in dynamic target cooperative tracking tasks, relying solely on position and distance constraints often leads to insufficient dynamic response of the tracking system. In essence, analogous to controller design, a simple position error penalty is equivalent to introducing a P-controller between the tracker and the target. This inevitably produces "phase lag" and "overshoot oscillation" when the target performs rapid maneuvers or changes in speed. To compensate for this deficiency and improve the agility and robustness of the tracking system, this method introduces a velocity matching cost into the optimization objective function. This term aims to minimize the difference between the planned trajectory and the predicted target trajectory in the velocity vector domain, and is defined as follows:

[0246]

[0247] Specifically, by jointly optimizing location costs With speed cost The drone can not only maintain the desired relative distance, but also keep synchronized with the target in both tangential and normal velocity components, achieving high-quality tracking.

[0248] Specifically, a trajectory optimization framework is used in the collaborative tracking trajectory planning process, employing a spatiotemporal decoupling design. Based on the theory of differential flatness, the original high-dimensional state planning problem is transformed into an optimization problem in a low-dimensional flat output space, and parameterized expression is achieved using the MINCO trajectory class. The MINCO trajectory, by separating spatial path points from temporal variables, enables independent adjustment of the spatial form and temporal characteristics of the trajectory. To adapt to the specific needs of collaborative tracking scenarios, a systematic objective function is designed. The spatiotemporal collaborative tracking penalty term, using distance penalty, ensures that the UAV maintains an appropriate tracking distance from the target within a specific time period. The vision perception penalty term evaluates the local obstacle distribution density through ray projection, guiding the UAV to actively fly towards areas with open vision, significantly reducing the probability of target occlusion. Furthermore, the velocity matching penalty term aims to minimize the velocity vector difference between the UAV and the target, making tracking maneuvers smoother.

[0249] For example, to fully verify the effectiveness and superiority of the proposed UAV trajectory planning method for cooperative tracking, simulation experiments were conducted to illustrate the technical effects. The experiments constructed a complete system including target state prediction, dynamic task allocation, and real-time trajectory planning, comprehensively considering and verifying the trajectory planning method for cooperative target tracking in a complex dynamic environment.

[0250] The specific parameter settings of the algorithm used in the experiment are shown in Table 1.

[0251] Table 1

[0252]

[0253] To test the effectiveness of the proposed method for UAV cooperative tracking trajectory planning, this embodiment selected two complex experimental scenarios with randomly generated obstacles. The first experimental scenario is one where obstacles are relatively sparse, such as... Figure 5 As shown, the map size is The first scenario features 40 obstacles, and the target will perform a certain degree of maneuvering in this simulation environment to evade tracking. The second experimental scenario has denser obstacles, and the map size is... There are 40 obstacles, such as Figure 6 As shown, the target's trajectory is relatively concentrated, and the drone is subject to significant environmental interference. When a target is captured by the sensor field of view of one drone, its position is broadcast to all drones. When the target is obscured or escapes the drone's sensor field of view, the drone will no longer receive the target's position information. To quantify the performance evaluation, we record and analyze the position error, target visibility time percentage, and trajectory planning success rate during the tracking process as standards for measuring the algorithm's robustness. Each experiment was conducted 10 times in each scenario.

[0254] Experimental results and related data in sparse obstacle environment simulation are as follows: Figure 7 , Figure 8 , Figure 9 As shown, the red drone is the target, the black drone is the tracker, and the three targets are traversing the obstacle area.

[0255] Position error, target visibility time percentage, and trajectory planning success rate were used as indicators to evaluate the results of the cooperative tracking experiment. A smaller position error indicates higher tracking accuracy of the UAVs and their ability to closely follow the movement of non-cooperative targets. A higher target visibility time percentage indicates that the UAV swarm can maintain continuous observation of the target through effective cooperation in complex obstacle environments, demonstrating the system's tracking robustness. A higher trajectory planning success rate proves the algorithm's ability to stably generate feasible solutions under multiple constraints such as obstacle avoidance, dynamic constraints, and formation maintenance, verifying the adaptability and reliability of the planning algorithm in dynamic environments. (Reference) Figure 4 As shown in the experiment, the position error is mitigated by the introduction of the velocity matching term, enabling the UAV to respond more agilely to sudden changes in target velocity and sharp turns. Combined with the line-of-sight perception penalty and re-acquisition strategy, the UAV is guided to actively avoid line-of-sight obstructions from scattered obstacles, minimizing the occurrence of target loss due to obstruction during tracking. Furthermore, the spatiotemporal cooperative tracking penalty allows the UAV to maintain an appropriate tracking distance from the target, reducing the probability of tracking failure.

[0256] Table 2 shows the tracking quality evaluation indicators in sparse obstacle simulation scenarios.

[0257] Table 2

[0258]

[0259] Experimental results in a dense obstacle environment are as follows: Figure 10 , Figure 11 , Figure 12 As shown, the red drone is the target and the black drone is the tracker. During the tracking process, we caused one drone to exit the tracking mission to verify the robustness of this method against interference.

[0260] In dense obstacle environment simulations, the view perception penalty addresses frequent obstacle occlusion. The gradient generated by this optimization term drives the UAV to proactively plan observation paths that bypass obstacle occlusion, reducing the risk of target loss. The spatiotemporal cooperative tracking penalty and velocity matching term provide predictive compensation in this scenario. When a target briefly enters the blind spot, this term drives the UAV to maintain proximity to the target based on the target's motion trend, ensuring the velocity does not diverge. Thanks to the coalition allocation mechanism, the system can maintain continuous tracking of all targets even after target loss; the entire allocation algorithm does not fail. Figure 13 and Figure 14 As shown in Table 3, these are the tracking quality evaluation indicators in a dense obstacle simulation scenario.

[0261] Table 3

[0262]

[0263] The method provided in this invention addresses the challenges of existing collaborative decision-making algorithms in adapting to dynamically changing task sets and alliance collaboration requirements. Furthermore, under multiple constraints, existing methods struggle to plan optimal trajectories that balance obstacle avoidance, visibility, and dynamics. This invention proposes a UAV collaborative autonomous tracking trajectory planning method for complex scenarios. First, a target dynamic allocation method based on alliance auctions is used to form an alliance for tracking tasks when the number of UAVs exceeds the number of targets. To address specific collaborative tracking scenarios, a dedicated utility function is designed to consider constraints such as tracking distance and UAV dynamics. Then, the trajectory optimization problem is solved using the MINCO spatiotemporal decoupling framework based on differential flatness. This framework enables independent spatiotemporal adjustment and designs multiple objective functions for collaborative tracking: distance preservation ensures tracking effectiveness, ray projection guides the field of view to avoid occlusion, and velocity matching ensures smooth maneuvering. By repeating these steps, UAV swarms can achieve stable and efficient collaborative tracking in complex scenarios.

[0264] It should be noted that the above figures are merely illustrative of the processes included in the method according to exemplary embodiments of the present invention, and are not intended to be limiting. It is readily understood that the processes shown in the above figures do not indicate or limit the temporal order of these processes. Furthermore, it is readily understood that these processes may, for example, be executed synchronously or asynchronously in multiple modules.

[0265] It should be noted that although several modules or units of the device for performing actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to embodiments of the present invention, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0266] Figure 15 A schematic diagram of an electronic device suitable for implementing embodiments of the present invention is shown.

[0267] It should be noted that, Figure 15 The electronic device 1000 shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.

[0268] like Figure 15 As shown, the electronic device 1000 includes a Central Processing Unit (CPU) 1001, which can perform various appropriate actions and processes based on programs stored in Read-Only Memory (ROM) 1002 or programs loaded from storage section 1008 into Random Access Memory (RAM) 1003. The RAM 1003 also stores various programs and data required for system operation. The CPU 1001, ROM 1002, and RAM 1003 are interconnected via a bus 1004. An Input / Output (I / O) interface 1005 is also connected to the bus 1004. Furthermore, the electronic device 1000 also includes an FPGA device and a System-on-a-Chip (SoC) device.

[0269] The following components are connected to I / O interface 1005: an input section 1006 including a keyboard, mouse, etc.; an output section 1007 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 1008 including a hard disk, etc.; and a communication section 1009 including a network interface card such as a LAN (Local Area Network) card, modem, etc. The communication section 1009 performs communication processing via a network such as the Internet. A drive 1010 is also connected to I / O interface 1005 as needed. Removable media 1011, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 1010 as needed so that computer programs read from them can be installed into storage section 1008 as needed.

[0270] In particular, according to embodiments of the present invention, the processes described below with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a storage medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 1009, and / or installed from removable medium 1011. When the computer program is executed by central processing unit (CPU) 1001, it performs various functions defined in the system of this application.

[0271] Specifically, the aforementioned electronic devices can be intelligent electronic devices, such as computers, tablets, etc. These electronic devices can connect to and communicate with the sensor platform, execute the methods described above based on the received data, and output an estimate of the target's position.

[0272] It should be noted that the storage medium shown in the embodiments of the present invention can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In the present invention, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, wherein computer-readable program code is carried. Such transmitted data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The computer-readable signal medium can also be any storage medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the storage medium can be transmitted using any suitable medium, including but not limited to wireless, wired, etc., or any suitable combination thereof.

[0273] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0274] The units described in the embodiments of the present invention can be implemented in software or hardware, and the described units can also be located in a processor. The names of these units do not necessarily limit the specific unit itself.

[0275] It should be noted that, as another aspect, this application also provides a storage medium, which may be included in an electronic device or may exist independently without being assembled into the electronic device. The aforementioned storage medium carries one or more programs, which, when executed by an electronic device, cause the electronic device to perform the methods described in the following embodiments. For example, the electronic device may perform... Figure 1 The steps of the method shown.

[0276] In one embodiment, this application provides a computer program product including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0277] Furthermore, the above figures are merely illustrative of the processes included in the method according to exemplary embodiments of the present invention, and are not intended to be limiting. It is readily understood that the processes shown in the above figures do not indicate or limit the temporal order of these processes. Additionally, it is readily understood that these processes may be executed synchronously or asynchronously, for example, in multiple modules.

[0278] Other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention herein. This application is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of the invention are indicated by the claims.

[0279] It should be understood that the present invention is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A target tracking trajectory planning method based on multi-UAV collaboration, characterized in that The method includes: Acquire target information for each target to be tracked, and current status information for each drone in the drone swarm; the target information includes the current location of each target; the current status information of the drones includes their current location. A distributed consortium auction method is invoked to perform collaborative target decision-making for each target, determining the tracking target corresponding to each UAV. The distributed consortium auction method estimates the profit of the UAV tracking target based on a utility function. The utility function includes: a tracking distance utility component for minimizing the travel distance and energy consumption of the UAV approaching the target, and a heading alignment utility component for evaluating the kinematic consistency between the current motion state of the UAV and the target orientation. Based on the MINCO trajectory class, cooperative tracking trajectory planning is performed to determine the cooperative tracking trajectory corresponding to each UAV. The tracking distance utility component of the utility function is configured based on the Euclidean distance between the UAV and the target, and includes: Among them, , which represents the Euclidean distance between the drone i and the target j; represents the position vector of the drone i; represents the position vector of the target j; represents the maximum value of this utility; represents a small quantity to prevent singularity; The heading alignment utility component in the utility function is configured based on the difference between the UAV's current velocity vector direction and the target's azimuth angle, and includes: Among them, represents the difference between the current speed vector direction of the UAV and the azimuth angle of the target; Utility functions include: Among them, are the weight coefficients of each utility part respectively.

2. The method according to claim 1, wherein The utility function also includes a relative proximity rate utility component for assessing the feasibility of tracking a dynamic target; The method further includes: configuring the relative proximity rate utility component of the utility function based on the relative velocity vector between the UAV and the corresponding target and the line-of-sight unit vector, including: wherein, represents the relative velocity vector between the UAV and the corresponding target; represents the line-of-sight unit vector, which is determined based on the position vector of the UAV and the position vector of the target.

3. The method according to claim 1, wherein The utility function further includes a task persistence utility component that provides constraints for target switching; The method further includes: using a function to configure additional utility for the target currently tracked by the UAV, including: Among them, represents the target assigned to UAV i at the previous moment.

4. The method according to any one of claims 1 to 3, characterized in that, The distributed consortium auction method is invoked to perform collaborative target decision-making for each objective, determining the tracking target for each drone, including: For each unmanned aerial vehicle Allocation status Initial values of the prices of each target in the target set Initial profit of the unmanned aerial vehicle Unmanned aerial vehicle Allocation target, unmanned aerial vehicle Partners are initialized; Using an individual bidding strategy, each UAV bids for the target based on its utility function to determine the expected maximum profit and the corresponding optimal task; and to determine the expected second-highest profit and the corresponding second-best task. When the number of drones exceeds the target number, a cooperative bidding strategy is used. Each drone determines its expected maximum cooperative profit and optimal cooperative task based on its utility function and the expected maximum profit of its partner drones from the target. When the drone is currently idle, and the expected maximum profit based on the individual bidding strategy and / or the expected maximum profit based on the cooperative bidding strategy is greater than zero, the bidding type is configured according to the value of the maximum profit, and bidding is conducted; wherein, the bid for individual bidding is based on the utility function and the expected second-highest profit, combined with the preset bid increment configuration; the bid for cooperative bidding is based on the utility function and the expected maximum profit of the drone and the partner drone, combined with the preset bid increment configuration; Construct separate sets of bids for individual bidding and sets of bids for cooperative bidding, and determine the winning tuple based on the bidding results; the winning tuple includes: drone / drone alliance, and the corresponding target; and configure the corresponding drone as assigned. Announcing information is generated based on the winning tuple, and the announcing information is then used to globally synchronize the status of each drone.

5. The method according to claim 4, characterized in that, The winning tuple is determined based on the bidding results, including: When the profit from a single bid is no less than that from a replacement bid and no less than that from a cooperative bid, the winning tuple is configured based on the single bid strategy; or, When the profit from individual bidding is lower than that from replacement bidding and lower than that from cooperative bidding, the winning tuple is configured based on the cooperative bidding strategy.

6. The method according to claim 1, characterized in that, Based on the MINCO trajectory class, cooperative tracking trajectory planning is performed to determine the cooperative tracking trajectory for each UAV, including: A trajectory planning problem for tracking is constructed based on target constraints; wherein, the target constraints include: line-of-sight perception constraints; wherein, the line-of-sight perception constraints are configured based on the penalty term of ray projection, which is used to control the UAV trajectory in the obstacle-sparse region to maintain observation of the target; The trajectory planning problem is analyzed based on the current location information of the drone to obtain the corresponding collaborative tracking trajectory planning results.

7. The method according to claim 6, characterized in that, The target constraints also include: velocity matching constraints and / or tracking distance constraints; wherein... The velocity matching constraint is configured based on the velocity matching cost function to control the difference between the velocity vectors of the UAV and the target to remain as expected; The tracking distance constraint is configured based on the spatiotemporal cooperative target tracking penalty function to control the UAV and the target to maintain the expected tracking distance.

8. The method according to claim 7, characterized in that, The target constraints also include: end constraints, intermediate point constraints, and time constraints; smoothing constraints, obstacle avoidance constraints, dynamic constraints, and cooperative inter-machine obstacle avoidance constraints. Trajectory planning problems oriented towards tracking based on objective constraints include: in, , , , These represent smoothing constraints, obstacle avoidance constraints, dynamic constraints, and cooperative inter-machine obstacle avoidance constraints, respectively. Indicates tracking distance constraints; Indicates visual perception constraints; Indicates speed matching constraints; , , , , , , These represent the corresponding weighting coefficients.

9. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the target tracking trajectory planning method based on multi-UAV cooperation as described in any one of claims 1 to 8.

10. An electronic device, characterized in that, include: processor; as well as Memory for storing the executable instructions of the processor; The processor is configured to implement the target tracking trajectory planning method based on multi-UAV cooperation as described in any one of claims 1 to 8 by executing the executable instructions.

Images