A bird-repelling method and device based on UAV array simulation

Abstract

This invention provides a bird deterrence method and apparatus based on drone array simulation, belonging to the field of drone bird deterrence technology. The method includes acquiring predation image information of birds of prey; performing motion modeling based on the predation image information to obtain a predation trajectory; mapping the predation trajectory to a drone array to generate a drone simulation strategy, with each drone equipped with an LED light; acquiring bird information and environmental parameters of the target area; generating a drone bird deterrence strategy based on the bird information, environmental parameters, and drone simulation strategy; and using the drone bird deterrence strategy to deter birds, obtaining the bird deterrence result. This invention solves the problem of limited effectiveness in existing drone bird deterrence technologies.

Description

Technical Field

[0001] This invention relates to the field of drone bird control technology, and more specifically, to a bird control method and apparatus based on drone array simulation. Background Technology

[0002] Traditional bird control methods, such as loud noises, laser beams, or manual deterrence, are often ineffective, and birds may gradually adapt. In recent years, drones, due to their high mobility and remote control capabilities, have become an emerging means of bird control, mainly used in areas susceptible to bird damage, such as airports, farmland, and power facilities, to reduce safety hazards and economic losses.

[0003] However, current drone-based bird control solutions have several shortcomings. First, traditional methods rely on propeller noise to scare away birds, but birds may gradually adapt, leading to a decrease in effectiveness. Second, drones have limited flight time, making it difficult to achieve long-term, efficient bird control. Furthermore, existing solutions mostly use fixed routes or manual operation, lacking intelligence and adaptive adjustments, making it difficult to respond effectively to different bird species. The coverage area of ​​a single drone is limited, requiring multiple drones to work together, but formation control and cost issues still need optimization. Additionally, due to complex environments, wind speed and weather changes can affect the stability of drones, reducing bird control efficiency.

[0004] Therefore, existing bird deterrence solutions have significant shortcomings in terms of noise adaptability, battery life, intelligence, coverage, and environmental adaptability. There is an urgent need for more efficient, lower power consumption, and more environmentally friendly methods to drive away birds, improve bird deterrence efficiency, and reduce interference with the target area and its surrounding environment, thus providing a new technological path for drone-based bird deterrence. Summary of the Invention

[0005] The purpose of this invention is to provide a bird-repelling method and device based on unmanned aerial vehicle (UAV) array simulation, in order to improve the aforementioned problems. To achieve the above objective, the technical solution adopted by this invention is as follows:

[0006] In a first aspect, this application provides a bird-repelling method based on unmanned aerial vehicle (UAV) array simulation, comprising:

[0007] Acquire predation image information of birds of prey, perform motion modeling based on the predation image information, and obtain the predation trajectory;

[0008] The drone array is mapped according to the predation trajectory to generate a drone simulation strategy, and each drone is equipped with an LED light;

[0009] Acquire bird information and environmental parameters for the target area;

[0010] A bird deterrence strategy is generated based on bird information, environmental parameters, and drone simulation strategies. The bird deterrence strategy is then used to deter birds and obtain bird deterrence results.

[0011] Secondly, this application also provides a bird-repelling device based on unmanned aerial vehicle (UAV) array simulation, comprising:

[0012] The first acquisition module is used to acquire predation image information of birds of prey, and to perform motion modeling based on the predation image information to obtain the predation trajectory;

[0013] The mapping module is used to map the UAV array according to the predation trajectory and generate a UAV simulation strategy. Each UAV is equipped with an LED light.

[0014] The second acquisition module is used to acquire bird information and environmental parameters of the target area;

[0015] The bird deterrence module is used to generate a drone bird deterrence strategy based on bird information, environmental parameters, and drone simulation strategy, and then use the drone bird deterrence strategy to deter birds and obtain bird deterrence results.

[0016] The beneficial effects of this invention are as follows: This invention combines raptor predation behavior modeling with UAV array collaborative control, improving bird deterrence efficiency and intelligence. By simulating predation trajectories, LED visual simulation effectively stimulates birds' escape responses, avoiding the adaptation problems of traditional noise deterrence. Simultaneously, the collaborative operation of multiple UAVs expands the coverage area, making it suitable for large areas. By acquiring real-time bird flock and environmental information and dynamically selecting the optimal bird deterrence action using reinforcement learning, the adaptability to different bird species and environmental conditions is enhanced. Furthermore, by optimizing the reward function to balance bird deterrence effectiveness, energy consumption, and safety, the operating time of the UAVs is extended, while reducing energy consumption and risk. At the same time, the dynamic adjustment of LED brightness ensures all-weather bird deterrence effectiveness, reducing interference with the surrounding environment, providing an efficient, intelligent, and environmentally friendly innovative solution for UAV bird deterrence.

[0017] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing embodiments of the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description, claims, and drawings. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram of the bird-repelling method based on UAV array simulation described in an embodiment of the present invention;

[0020] Figure 2 This is a schematic diagram of key skeletal points in an embodiment of the present invention;

[0021] Figure 3 This is a schematic diagram showing the position of the UAV array at a certain moment in an embodiment of the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0023] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this invention, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0024] Example 1:

[0025] This embodiment provides a bird-repelling method based on unmanned aerial vehicle (UAV) array simulation.

[0026] See Figure 1 The figure shows that the method includes steps S1, S2, S3 and S4.

[0027] Step S1: Obtain predation image information of birds of prey, perform motion modeling based on the predation image information, and obtain the predation trajectory;

[0028] It is understandable that birds' fear of birds of prey stems from innate instincts, and their reactions to single signals such as noise and light are easily adapted to traditional bird deterrence methods (e.g., propeller noise, fixed lasers) because they do not address the core trigger points of their predators. Therefore, by breaking down the actual predation stages of birds of prey and extracting skeletal points for modeling, the dynamic characteristics of raptor predation (such as speed changes during dives and claw posture during strikes) can be reconstructed, allowing drone simulations to more closely resemble natural scenarios and preventing birds from ignoring threats due to simulation distortion.

[0029] Therefore, this step involves motion modeling of the raptor's predation process to extract its predation trajectory. By using realistic predation trajectories, the realism of the drone-based bird deterrence simulation is improved, enhancing the birds' stress response to deterrence signals.

[0030] Step S1 includes:

[0031] Step S11: Obtain predation image sequences for different predation stages based on predation image information, wherein the predation stages include circling stage, diving stage, and pouncing stage;

[0032] In this embodiment, predation image information of birds of prey such as eagles, falcons, and hawks is acquired. At the same time, multiple keyframes are extracted according to the predation stage to obtain multiple sets of predation image sequences. Each set of predation image sequences represents a predation sample of a bird of prey in a predation stage.

[0033] Step S12: Perform pose estimation based on the predation image sequence and extract key skeleton points from different time frames. The key skeleton points include skeleton point categories and skeleton point coordinates.

[0034] In this embodiment, an animal pose estimation algorithm is used to analyze the raptor in each keyframe, extracting key skeletal points such as the head, wings, claws, and tail feathers, and recording the 3D coordinates of each skeletal point, such as... Figure 2 The image shows the key skeletal points of a raptor when facing it. The expression for the key skeletal points is as follows:

[0035] ;

[0036] ;

[0037] In the formula, Indicates time The set of key skeletal points Indicates the first Key skeletal points in time The coordinates of the skeletal points, , and They represent the first The x, y, and y coordinates of each key skeletal point in 3D space. Indicates the first Bone point categories for key skeletal points Indicates the number of key skeleton points. .

[0038] Step S13: Perform time-series processing on key skeleton points in different time frames to obtain predation trajectories in different predation stages.

[0039] In this embodiment, multiple representative predation trajectories are obtained through different image sequences, different birds of prey, and different scenarios. That is, multiple predation trajectories are generated for each predation stage, and each predation trajectory can subsequently generate a drone simulation strategy.

[0040] Step S2: Map the UAV array according to the predation trajectory and generate a UAV simulation strategy. Each UAV is equipped with an LED light.

[0041] Step S2 includes:

[0042] Step S21: Normalize the predation trajectory to obtain a standardized trajectory;

[0043] It is understandable that different birds of prey vary greatly in size, and directly using the original trajectory would lead to chaotic drone array scale (for example, the trajectory of a small falcon cannot be adapted to a large drone). Normalization and scaling allow the trajectory to match the real space of the target area (such as an airport or farmland).

[0044] In this embodiment, the predation trajectory is normalized to adapt to the spatial scale of the UAV array. The normalization formula is as follows:

[0045] ;

[0046] In the formula, Indicates the first Key skeletal points in time Normalized skeletal point coordinates, Indicates the first Key skeletal points in time The coordinates of the skeletal points, Indicates the reference center point. This represents the normalized scale, where the reference center point is selected from key bone points in the head, and the normalized scale is the maximum bone span.

[0047] Step S22: Assign the corresponding UAV to each key skeleton point in the standardized trajectory to obtain the standard trajectory of each UAV;

[0048] In this embodiment, since raptors rely on the coordination of multiple parts (such as wing flapping in coordination with tail feather adjustment) to hunt, a single drone cannot simulate this. By matching skeletal points and interpolating to complete the data, the array is made to form a raptor-like body structure for coordinated flight, thus restoring the overall hunting posture.

[0049] Specifically, each key skeleton point is assigned to a corresponding UAV to obtain a standard trajectory. Then, based on structural priority or motion complexity settings, some UAVs that do not correspond one-to-one with key skeleton points are assigned to specific structural regions. For example, multiple UAVs can be mapped to a wing. For UAVs that do not correspond one-to-one with key skeleton points, a standard trajectory for this type of UAV can be generated based on the standard trajectory of the assigned key skeleton points using the Bézier curve interpolation method. This enables cooperative flight and dynamic simulation of multiple UAVs on the same motion trajectory.

[0050] For drones that do not correspond one-to-one with key skeleton points, Bézier interpolation is performed using the standardized trajectory corresponding to the key skeleton points to obtain the standard trajectory of the drone without a corresponding relationship. For example... Figure 3 The image shows the position of the drone array at a certain moment after the interpolation is performed.

[0051] Step S23: Map the standard trajectory to the real three-dimensional space according to the scaling factor to obtain the actual flight trajectory of each UAV;

[0052] In this embodiment, the formula for mapping the standard trajectory to the real three-dimensional space is:

[0053] ；

[0054] In the formula, Indicates the first A drone in time The true three-dimensional coordinates No. A drone in time Normalized coordinates, Indicates the scaling factor. This represents the position offset center, which can be set as the center of the target area. It is used to translate the standard trajectory to the target spatial position, thereby achieving a true spatial mapping of the motion trajectory.

[0055] Step S24: Calculate the velocity sequence and acceleration sequence of each UAV based on the actual flight trajectory, and generate the velocity control command and acceleration control command for each UAV;

[0056] In this embodiment, the velocity sequence and acceleration sequence of the UAV can be obtained by differentiating the actual flight trajectory.

[0057] Step S25: Generate the UAV simulation strategy corresponding to each predation trajectory based on the speed control command and acceleration control command.

[0058] In this embodiment, the generated control commands can clearly define when the UAV accelerates and when it changes direction (such as needing to continuously accelerate during the dive phase), ensuring that the array flies strictly according to the predation trajectory and avoiding deviation that could cause the simulation to fail.

[0059] Step S3: Obtain bird information and environmental parameters for the target area;

[0060] In this embodiment, cameras can be installed on the ground or in a control tower, and deep learning models (such as YOLO and DeepSORT) can be used for image recognition and multi-target tracking. Alternatively, an Avian Radar bird radar system can be deployed to detect and track the long-range three-dimensional trajectories of birds within the target area, thereby obtaining bird information for the target area. Environmental parameters can be obtained through environmental sensors or remote meteorological data interfaces.

[0061] Step S4: Generate a drone bird deterrence strategy based on bird information, environmental parameters, and drone simulation strategy, and then use the drone bird deterrence strategy to deter birds and obtain bird deterrence results.

[0062] In this embodiment, a hierarchical reinforcement learning approach is adopted, employing a fitness evaluation and reward function mechanism. The predation phase is determined first, followed by the determination of the drone's bird-scaring strategy. By comprehensively considering bird-scaring effectiveness, flight efficiency, power consumption, safe distance, and environmental adaptability, the optimal strategy is ultimately selected for execution. This strategy evaluation mechanism balances efficiency and safety, improving energy utilization efficiency and flight mission safety while ensuring bird-scaring effectiveness. It is particularly suitable for application scenarios where drone resources are limited or under complex weather conditions.

[0063] Step S4 includes:

[0064] Step S41: Generate the current state based on bird information and environmental parameters;

[0065] Step S41:

[0066] Step S411: Obtain the drone status, which includes the location and remaining battery power of each drone;

[0067] It is understandable that drones are the carriers for bird deterrence, and their status directly determines whether the strategy can be completed. That is, the position information ensures that the array can fly in coordination along the trajectory (avoiding collisions between drones), and the remaining power determines the duration of a single bird deterrence (to prevent power failure and crashes midway, power needs to be reserved for return).

[0068] Step S412: Generate bird flock status based on bird information and generate environmental status based on environmental parameters. The bird information includes bird species, bird number, bird location and bird density. The environmental parameters include wind speed and visibility.

[0069] In this embodiment, the flock status is strategically targeted. For example, a single large raptor (such as an eagle) requires a circling phase for deterrence; a dense flock of small birds requires a swooping phase for rapid dispersal. The number, location, and density of the flock determine the coverage area and approach distance of the drone array. Environmental parameters are the basis for the feasibility of the strategy. For example, when the wind speed exceeds the drone's maximum wind resistance capability, the flight speed needs to be adjusted (to avoid trajectory deviation); when visibility is low, the LED brightness needs to be increased (to ensure that birds can recognize the simulated threat).

[0070] Step S413: Generate the current state based on the bird flock state, environmental state, and drone state.

[0071] In this embodiment, the expression for the current state is:

[0072] ;

[0073] In the formula, Indicates time The current state at that time, Indicates time The state of the bird flock at that time Indicates time The environmental conditions at that time, Indicates time The status of the drone at that time.

[0074] Step S42: Select the drone simulation strategy based on the current state to obtain the optimal drone simulation strategy;

[0075] Step S42 includes:

[0076] Step S421: Select the predation stage based on the state of the bird flock to obtain the target predation stage;

[0077] In this embodiment, the real-time status of the bird flock is used to determine its flight pattern. By analyzing information such as the flock's position and speed, it can be inferred whether the flock is in an escape state (fleeing towards the ground) or a cruising state (flying at a higher altitude). Based on this information, the corresponding predation phase is selected for the UAV array.

[0078] Specifically, when flocks of birds are in higher altitudes, especially at the upper level of their flight (preset altitude), the circling phase can be chosen. This is because circling is a common hunting behavior for birds of prey, particularly when prey or flocks are at high altitudes, allowing them to find hunting opportunities. When the flock begins to descend, especially as they approach the ground, the dive phase should be chosen. In this situation, the raptor's hunting behavior shifts from circling to diving, approaching the prey with a faster and more threatening posture. When the flock is close to the drone array (within a preset distance), the swooping phase can be chosen for close-range engagement. Therefore, the choice of the target hunting phase is crucial for effective bird deterrence.

[0079] Step S421 includes:

[0080] Step A1: Obtain the average height of the flock, the average speed of the flock, the magnitude of the change in the flock's direction, and the current distance of the flock to the drone array based on the flock's location;

[0081] Step A2: Calculate the circling fitness based on the average height of the flock, the average speed of the flock, and the circling phase fitness function;

[0082] In this embodiment, when calculating the fitness of each predation stage, normalization is performed to unify it to the same scoring scale before fitness comparison is conducted. Specifically, the formula for the fitness function in the circling stage is as follows:

[0083] ;

[0084] In the formula, Indicates hovering fitness. and Both represent the spiral weight parameters. Indicates the average height of the flock of birds. Indicates the maximum observation altitude. Indicates the average speed of the flock of birds. This indicates the maximum speed of the flock of birds.

[0085] in, The larger the value, the higher the altitude of the flock, which better matches the scenario of birds of prey circling and searching for prey at high altitudes. The larger the value, the slower the flock flies (e.g., in cruising mode), making it easier for raptors to lock onto their targets by circling. By quantifying the degree of adaptation of the circling phase to the current flock, subjective judgment is avoided. For example, for flocks cruising at low speeds at high altitudes, a high circling adaptation score indicates that this phase can effectively simulate a raptor hunting scenario.

[0086] Step A3: Calculate the dive fitness based on the average height of the flock, the magnitude of the change in flock direction, and the fitness function during the dive phase;

[0087] In this embodiment, the formula for the fitness function during the dive phase is as follows:

[0088] ;

[0089] In the formula, Indicates dive adaptability. and Both represent the dive weighting parameters. Indicates the average height of the flock of birds. Indicates the maximum observation altitude. Indicates the magnitude of the change in the direction of the flock of birds. This represents the maximum directional change, where, and Both represent changes per unit of time.

[0090] in, The larger the value, the lower the altitude of the flock, which better matches the scenario of a raptor swooping down to approach its prey. The higher the score, the more frequently the flock changes its flight direction (e.g., panicked flight upon sensing a threat), making it easier for birds of prey to close the distance by diving. Therefore, by quantifying the adaptability of the dive phase, for example, a high dive adaptability score indicates that for flocks of birds with chaotic low-altitude flight, the ability to rapidly approach and increase the sense of threat can force the flock to move away.

[0091] Step A4: Calculate the attack fitness based on the average speed of the flock, the current distance of the flock to the drone array, and the attack phase fitness function;

[0092] In this embodiment, the formula for the fitness function during the attack phase is as follows:

[0093] ;

[0094] In the formula, Indicates pouncing fitness. and All represent the attack weight parameters. Indicates the average speed of the flock of birds. Indicates the maximum speed of the flock of birds. This indicates the current distance from the flock of birds to the drone array. This indicates the maximum distance from the flock of birds to the drone array.

[0095] in, The larger the value, the faster the flock flies (e.g., in an emergency escape), requiring raptors to use their swooping phase (maximum speed) to catch up. The larger the value, the closer the flock of birds is to the drone, requiring a swooping phase (close-range strong deterrence) to quickly drive them away and prevent them from entering the target area (such as an airport runway). By quantifying the adaptability of the swooping phase, for example, a high swooping adaptability score indicates that a flock of birds approaching the target area at high speed at close range can be forcibly blocked through a short-range, rapid swooping maneuver.

[0096] Step A5: Select the target predation phase based on the maximum value among hovering fitness, diving fitness, and pouncing fitness.

[0097] Step S422: Calculate the reward value of each UAV simulation strategy during the target predation phase based on the reward function;

[0098] In step S422, the steps for constructing the reward function are as follows:

[0099] Step B1: Construct a bird-repelling effect reward based on the current distance of the bird flock to the target area and the change in the distance of the bird flock;

[0100] In this embodiment, the effectiveness of the drone array in driving away flocks of birds is measured by a bird-repelling effect reward. The reward is calculated based on the change in distance between the flock of birds and the target area. The formula for calculating the bird-repelling effect reward is as follows:

[0101] ;

[0102] In the formula, This indicates a reward for effective bird-repelling. and All of these represent weighted parameters for bird-repelling effectiveness. Indicates the distance traveled by the flock of birds. This indicates the current distance of the flock of birds to the target area, where the current distance of the flock to the target area represents the distance from the center point of the flock to the center point of the target area.

[0103] Step B2: Construct an efficiency reward based on the drone's total flight distance, total flight time, and power consumption;

[0104] Understandably, to balance bird-scaring effectiveness with resource consumption, and to avoid overloading the drone with bird-scaring (such as taking long detours) and preventing it from performing other tasks, this step uses an efficiency reward-based strategy that prioritizes short paths, short times, and low energy consumption. Specifically, efficiency rewards measure how effectively the drone uses resources during flight, and the formula for calculating these rewards is as follows:

[0105] ;

[0106] In the formula, Indicates an efficiency reward. This indicates the total flight distance of the drone. This indicates the total flight time of the drone. This indicates the drone's power consumption. and Both represent efficiency weight parameters.

[0107] Step B3: Construct environmental adaptation rewards based on wind speed and visibility;

[0108] Understandably, in order to ensure that the strategy is executable in complex environments and to avoid using difficult strategies (such as rapid dives) in strong winds or low visibility, which could lead to trajectory deviations or simulation failures, this step constructs an environment adaptation reward.

[0109] Specifically, the adaptability of drones in specific environments is assessed based on wind speed and visibility. Drones that perform well while performing missions in harsh environments receive a higher environmental adaptation bonus. The formula for calculating the environmental adaptation bonus is as follows:

[0110] ;

[0111] In the formula, This indicates an environmental adaptation reward. and All represent environmental weight parameters. Indicates the maximum acceptable wind speed. Indicates the current wind speed. Indicates the maximum acceptable visibility. Indicates the current visibility.

[0112] Step B4: Construct a safety reward system by setting a minimum distance between the bird flock and the drone and a preset safe distance threshold;

[0113] In this embodiment, the safety reward is mainly used to ensure that the drone maintains a safe distance from the flock of birds, avoiding unnecessary risks caused by getting too close (such as the drone approaching the flock of birds too closely and causing a collision). The formula for calculating the safety reward is as follows:

[0114] ;

[0115] In the formula, Indicates a safety reward. Represents the safety weight parameter. This indicates the minimum distance between the flock of birds and the drone. This indicates the preset safe distance threshold.

[0116] Step B5: Construct a reward function using bird deterrence effect rewards, efficiency rewards, environmental adaptation rewards, and safety rewards.

[0117] In this embodiment, a reward function is designed to avoid biases from a single objective (such as focusing only on performance while ignoring power consumption, or focusing only on safety while ignoring vehicle removal) through multi-dimensional balancing. The formula for the reward function is as follows:

[0118] ;

[0119] In the formula, Represents the reward function, , , and All represent weight parameters. This indicates a reward for effective bird-repelling. Indicates an efficiency reward. This indicates an environmental adaptation reward. This indicates a safety reward.

[0120] Step S423: Select the optimal drone simulation strategy based on the reward value.

[0121] In this embodiment, the optimal drone simulation strategy is selected by updating the reward value and Q-learning. Simultaneously, when updating using the reward value and Q-learning, the invention simulates the collective behavior of a flock of birds under different drone simulation strategies using the Boids Algorithm, thereby calculating the corresponding reward value and obtaining the optimal drone simulation strategy.

[0122] Step S43: Generate a drone bird-repelling strategy based on the optimal drone simulation strategy. The drone bird-repelling strategy includes speed control commands, acceleration control commands, and LED brightness control commands for each drone.

[0123] In this embodiment, each drone not only receives speed and acceleration control commands based on its own flight trajectory, but also adjusts the brightness of its LEDs according to the external visibility. Through dynamically adjusted LED control, the visual simulation intensity is matched with the external environment, thereby maintaining a high bird-repelling effect under different weather conditions such as daytime, dusk, and fog.

[0124] Step S43 includes:

[0125] Step S431: Obtain the speed control command and acceleration control command for each UAV according to the optimal UAV simulation strategy;

[0126] Step S432: Calculate the LED brightness at each time step according to the visibility and LED brightness calculation formula, wherein the LED brightness calculation formula is constructed using visibility, maximum visibility and minimum effective visibility;

[0127] In this embodiment, the worse the visibility, the brighter the LED; conversely, the better the visibility, the dimmer (or off) the LED. This saves power, avoids visual interference, and ensures that the drone's visual signal is transmitted to birds more clearly and effectively in low visibility conditions. The formula for calculating the LED brightness is as follows:

[0128] ;

[0129] In the formula, Indicates LED brightness. Indicates the maximum brightness of the LED. Indicates the current visibility. Indicates maximum visibility. This represents the minimum effective visibility, where the current visibility is calculated from weather visibility and ambient light intensity.

[0130] Step S433: Generate LED brightness control instructions based on the LED brightness at each time step;

[0131] Step S434: Obtain the drone bird-repelling strategy through the speed control command, acceleration control command and LED brightness control command of each drone.

[0132] Step S44: Use a drone-based bird deterrence strategy to deter birds and obtain bird deterrence results;

[0133] Step S45: Update the current state based on the bird deterrence results, generate the drone bird deterrence strategy for the next time step, and carry out bird deterrence.

[0134] In summary, this invention effectively stimulates birds' instinctive escape response by simulating real-world raptor hunting behavior, reconstructing hunting trajectories using an array of drones, and combining this with LED visual simulation. This solves the problems of traditional noise-based bird deterrence being easily adapted to and having diminishing effectiveness. The collaborative bird deterrence using multiple drones covers a wide area, suitable for large-scale areas such as airports and power facilities, overcoming the low efficiency of single-drone operations. Furthermore, by integrating bird flock behavior and environmental perception information, reinforcement learning is used to dynamically optimize bird deterrence actions, improving the ability to cope with different bird species and weather conditions.

[0135] In addition, this invention balances bird deterrence effect, energy consumption and safety through a reward function, thereby improving the operational efficiency and endurance of drones. The adaptive adjustment of LED brightness also enhances the all-weather bird deterrence effect and reduces interference with the surrounding environment.

[0136] Example 2:

[0137] This embodiment provides a bird-repelling device based on unmanned aerial vehicle (UAV) array simulation, the device comprising:

[0138] The first acquisition module is used to acquire predation image information of birds of prey, and to perform motion modeling based on the predation image information to obtain the predation trajectory;

[0139] The mapping module is used to map the UAV array according to the predation trajectory and generate a UAV simulation strategy. Each UAV is equipped with an LED light.

[0140] The second acquisition module is used to acquire bird information and environmental parameters of the target area;

[0141] The bird deterrence module is used to generate a drone bird deterrence strategy based on bird information, environmental parameters, and drone simulation strategy, and then use the drone bird deterrence strategy to deter birds and obtain bird deterrence results.

[0142] It should be noted that the specific manner in which each module performs its operation in the apparatus described in the above embodiments has been described in detail in the embodiments of the method, and will not be elaborated here.

[0143] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0144] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A bird-repelling method based on unmanned aerial vehicle (UAV) array simulation, characterized in that, include: Acquire predation image information of birds of prey, perform motion modeling based on the predation image information, and obtain the predation trajectory; The drone array is mapped based on the predation trajectory to generate a drone simulation strategy, and each drone is equipped with an LED light; Acquire bird information and environmental parameters for the target area; A bird deterrence strategy is generated based on bird information, environmental parameters, and drone simulation strategies. The bird deterrence strategy is then used to deter birds and obtain bird deterrence results.

2. The bird-repelling method based on UAV array simulation according to claim 1, characterized in that... The obtained predator trajectory includes: Based on predation image information, a predation image sequence of different predation stages is obtained, wherein the predation stages include the circling stage, the diving stage, and the pouncing stage; Pose estimation is performed based on the predation image sequence, and key skeleton points are extracted from different time frames. The key skeleton points include skeleton point categories and skeleton point coordinates. By performing time-series processing on key skeleton points in different time frames, predation trajectories at different predation stages are obtained.

3. The bird-repelling method based on UAV array simulation according to claim 1, characterized in that... The generated drone simulation strategy includes: The predation trajectory is normalized to obtain a standardized trajectory; Assign a corresponding UAV to each key skeleton point in the standardized trajectory to obtain the standard trajectory of each UAV; The standard trajectory is mapped to the real three-dimensional space based on the scaling factor to obtain the actual flight trajectory of each UAV. Calculate the velocity and acceleration sequences of each UAV based on the actual flight trajectory, and generate speed control and acceleration control commands for each UAV. The drone simulation strategy for each predation trajectory is generated based on the speed control command and the acceleration control command.

4. The bird-repelling method based on UAV array simulation according to claim 1, characterized in that... The process of generating a drone bird-repelling strategy based on bird information, environmental parameters, and drone simulation strategies includes: The current status is generated based on bird information and environmental parameters; Select the drone simulation strategy based on the current state to obtain the optimal drone simulation strategy; A drone bird-repelling strategy is generated based on the optimal drone simulation strategy. The drone bird-repelling strategy includes speed control commands, acceleration control commands, and LED brightness control commands for each drone. Bird control was achieved by using drones as a bird deterrent strategy, and the bird deterrent results were obtained. Update the current state based on the bird control results, generate the drone bird control strategy for the next time step, and carry out bird control.

5. The bird-repelling method based on UAV array simulation according to claim 4, characterized in that... The process of generating the current state based on bird information and environmental parameters includes: Acquire the drone status, which includes the location and remaining battery power of each drone; Bird flock status is generated based on bird information, and environmental status is generated based on environmental parameters. The bird information includes bird species, bird number, bird location, and bird density. The environmental parameters include wind speed and visibility. The current state is generated based on the flock status, environmental status, and drone status.

6. The bird-repelling method based on UAV array simulation according to claim 5, characterized in that... The process of obtaining the optimal UAV simulation strategy includes: Select the predation phase based on the state of the flock to obtain the target predation phase; The reward value for each drone simulation strategy during the target predation phase is calculated based on the reward function; The optimal drone simulation strategy is selected based on the reward value.

7. The bird-repelling method based on UAV array simulation according to claim 6, characterized in that... The target predation phase includes: The average height, average speed, change in direction, and current distance from the flock to the drone array are obtained based on the flock's location. Calculate the hovering fitness based on the average height of the flock, the average speed of the flock, and the hovering phase fitness function. Dive fitness is calculated based on the average height of the flock, the magnitude of the change in flock direction, and the fitness function during the dive phase. The attack fitness is calculated based on the average speed of the flock, the current distance of the flock to the drone array, and the attack phase fitness function. The target predation phase is selected based on the maximum value among hovering fitness, diving fitness, and pouncing fitness.

8. The bird-repelling method based on UAV array simulation according to claim 6, characterized in that... The steps for constructing the reward function are as follows: A bird-repelling effect reward is constructed based on the current distance of the flock to the target area and the change in the distance of the flock. Efficiency rewards are built based on the total flight distance, total flight time, and power consumption of the drone. Develop environmental adaptation rewards based on wind speed and visibility; Establish a safety reward system by defining the minimum distance between bird flocks and drones and a preset safe distance threshold; A reward function is constructed by rewarding bird deterrence effectiveness, efficiency, environmental adaptation, and safety.

9. The bird-repelling method based on UAV array simulation according to claim 4, characterized in that... The step of generating a drone bird-repelling strategy based on the optimal drone simulation strategy includes: Obtain the speed control and acceleration control commands for each UAV based on the optimal UAV simulation strategy; The LED brightness at each time step is calculated based on the visibility and LED brightness calculation formula, which is constructed using visibility, maximum visibility, and minimum effective visibility. LED brightness control instructions are generated based on the LED brightness at each time step; The bird-repelling strategy for drones is obtained by using speed control commands, acceleration control commands, and LED brightness control commands for each drone.

10. A bird-repelling device based on unmanned aerial vehicle (UAV) array simulation, characterized in that, include: The first acquisition module is used to acquire predation image information of birds of prey, and to perform motion modeling based on the predation image information to obtain the predation trajectory; The mapping module is used to map the UAV array according to the predation trajectory and generate a UAV simulation strategy. Each UAV is equipped with an LED light. The second acquisition module is used to acquire bird information and environmental parameters of the target area; The bird deterrence module is used to generate a drone bird deterrence strategy based on bird information, environmental parameters, and drone simulation strategy, and then use the drone bird deterrence strategy to deter birds and obtain bird deterrence results.

Images