Multi-uav low-noise path planning method and system

Abstract

The present application relates to the technical field of path planning, in particular to a multi-unmanned aerial vehicle low-noise path planning method and system, comprising: modeling the flight environment of the unmanned aerial vehicle; establishing a sound source directivity model of the unmanned aerial vehicle, and evaluating the noise influence of the unmanned aerial vehicle flight on a noise-sensitive area according to a noise attenuation formula; utilizing an acoustic-driven artificial potential field algorithm to perform path planning, generating a flight path of the unmanned aerial vehicle by calculating an attractive potential field, a repulsive potential field and the interaction force between the unmanned aerial vehicles, and updating the path planning in real time. The present application converts noise characteristics into a repulsive force mechanism, combines noise characteristics with an artificial potential field, simplifies the noise evaluation process, not only realizes rapid and real-time path planning, but also dynamically avoids noise-sensitive areas when planning a path, greatly reducing the influence on noise-sensitive areas.

Description

Technical Field

[0001] This invention relates to the field of path planning technology, and in particular to a low-noise path planning method and system for multiple unmanned aerial vehicles (UAVs). Background Technology

[0002] In recent years, with the widespread application of drones and electric vertical takeoff and landing (eVTOL) aircraft in urban logistics, air transportation, and other fields, the noise pollution problem caused by their low-altitude flight has become increasingly prominent. Traditional path planning algorithms mostly focus on flight efficiency and obstacle avoidance, but neglect the impact of noise. Although existing research has carried out noise assessment and single-aircraft path planning, there are still shortcomings in multi-aircraft collaboration, dynamic environmental adaptability, and real-time low-noise path planning.

[0003] In the area of ​​low-noise path planning for unmanned aerial vehicles (UAVs), some existing studies have adopted linear acoustic propagation models combined with heuristic algorithms for path planning, failing to fully consider the noise characteristics of the aircraft and the environmental sound field features. Subsequent research has employed ray tracing methods to model the environmental noise of the aircraft, traversing all possible operating conditions and positions to form a noise database, and then using heuristic algorithms for optimization. While this method can effectively solve the low-noise optimization problem for a single UAV, the noise modeling cost becomes extremely high as the number and types of UAVs increase. The massive computational demands consume significant high-performance computing platform resources and time, and it lacks scalability, making it difficult to adapt to low-noise optimization scenarios involving multiple UAVs.

[0004] In summary, while existing path planning technologies have achieved some success in various scenarios, most do not adequately consider noise issues, making it difficult to meet the growing environmental requirements of low-altitude aircraft. Furthermore, existing low-noise path planning methods either fail to achieve real-time planning or are inadequate in multi-agent collaboration and dynamic environmental adaptability. This makes it difficult to effectively address the noise pollution problems caused by drones to residents and the environment in practical applications, especially in complex urban environments. Summary of the Invention

[0005] Therefore, the technical problem to be solved by the present invention is to overcome the problem of insufficient dynamic environment adaptability and difficulty in real-time low-noise path planning when multiple UAVs fly in cooperation in the prior art.

[0006] To address the aforementioned technical problems, this invention provides a low-noise path planning method for multiple unmanned aerial vehicles (UAVs), comprising:

[0007] Model the flight environment of the UAV and set noise-sensitive areas and their corresponding noise thresholds;

[0008] Obtain the angle between the observation point in the noise-sensitive area and the direction of the UAV's movement; use the UAV's sound source directivity model to obtain the sound pressure level of the observation point in the noise-sensitive area pointed by the UAV; calculate the received sound pressure level of the observation point in the noise-sensitive area to the UAV based on the sound pressure level of the observation point in the noise-sensitive area and the distance between the observation point in the noise-sensitive area and the UAV.

[0009] Establish an attractive potential field based on the destination position of the drone;

[0010] The difference between the received sound pressure level of the UAV and the noise threshold at the observation point in the noise-sensitive area is mapped to the first repulsive potential field coefficient; based on the first repulsive potential field coefficient and the distance between the observation point in the noise-sensitive area and the UAV, the repulsive potential field of the noise-sensitive area to the UAV is constructed.

[0011] Establish the repulsive potential field of static obstacles on the UAV and the repulsive potential field between adjacent UAVs;

[0012] The resultant force of the UAV's flight is calculated based on the attractive and repulsive potential fields, and the UAV's flight path is generated.

[0013] Preferably, the flight environment of the UAV is modeled, and noise-sensitive areas and their corresponding noise thresholds are set, including:

[0014] By using open-source databases to obtain urban environmental geometric information, buildings in the urban environment are grouped by category to obtain multi-category building areas;

[0015] Noise-sensitive areas and their corresponding noise thresholds are set according to the building area category.

[0016] Preferably, the UAV sound source directivity model is used to describe the sound pressure level of the UAV in different directions and is constructed based on simulators or acoustic experimental measurement methods.

[0017] Preferably, the observation point in the noise-sensitive area is located at the center of the noise-sensitive area.

[0018] Preferably, the received sound pressure level at the observation point is calculated based on the sound pressure level pointed at the observation point and the distance between the observation point and the drone, using the following formula:

[0019] ;

[0020] in, Let be the received sound pressure level of the j-th noise-sensitive area at time t for the i-th UAV. Let i be the sound pressure level of the observation point pointed by the i-th UAV to the j-th noise-sensitive area. For the directional model of the drone's sound source, Let be the angle between the observation point in the j-th noise-sensitive region and the forward direction of the i-th UAV. Let be the distance between the observation point in the j-th noise-sensitive area and the i-th UAV. This represents the distance between the sound source directional observation point and the center of the UAV in the UAV sound source directional model. and These are the influence factors of atmospheric parameters and ground parameters on sound propagation attenuation, respectively. Atmospheric attenuation coefficient, This is the ground attenuation coefficient.

[0021] Preferably, the received sound pressure level of the UAV at the observation point in the noise-sensitive area is mapped to the noise threshold as a first repulsive potential coefficient, as shown in the formula:

[0022] ;

[0023] in, Let be the first repulsive potential coefficient of the repulsive potential field of the j-th noise-sensitive region to the i-th UAV. Let be the difference between the received sound pressure level of the i-th UAV at the observation point in the j-th noise-sensitive area and the noise threshold of the j-th noise-sensitive area. As a penalty factor;

[0024] The repulsive potential field of the noise-sensitive region to the UAV is constructed based on the first repulsive potential field coefficient and the distance between the observation point in the noise-sensitive region and the UAV. The formula is as follows:

[0025] ;

[0026] in, Let be the repulsive force exerted by the j-th noise-sensitive region on the i-th UAV. Let be the distance between the observation point in the j-th noise-sensitive area and the i-th UAV. For minimum safe distance, Let be the direction vector from the i-th UAV to the j-th noise-sensitive area observation point. for Step function.

[0027] Preferably, an attractive potential field is established based on the destination position of the UAV, using the following formula:

[0028] ;

[0029] in, Let the destination position of the i-th drone be the attractiveness of that drone. The gravitational coefficient, Let be the distance between the i-th drone and its destination position. Let be the direction vector from the i-th drone to its endpoint position.

[0030] Preferably, a repulsive potential field of static obstacles on the UAV is established, with the following formula:

[0031] ;

[0032] in, Let p be the repulsive force exerted by the p-th static obstacle on the i-th drone. The coefficient of the second repulsive potential field. Let p be the distance between the p-th static obstacle and the i-th drone. For minimum safe distance, Let be the direction vector from the i-th drone to the p-th static obstacle. for Step function;

[0033] The repulsive potential field between adjacent UAVs is established by the following formula:

[0034] ;

[0035] in, For the first The repulsive force of one drone on the i-th drone, The coefficient of the third repulsive potential field. For the first The distance between the i-th drone and the i-th drone The radius of the drone. For the first The direction vector of the i-th drone is pointed to by the drone.

[0036] Preferably, when calculating the resultant force of the UAV's flight based on the attractive and repulsive potential fields, the method further includes:

[0037] Calculate the magnitude of the resultant force of the UAV's flight at each moment;

[0038] If the magnitude of the flight resultant force is less than the preset jamming threshold for a continuous preset time period, a disturbance vector with a random magnitude along the tangent direction of the nearest static obstacle is generated, and the disturbance vector is superimposed on the flight resultant force of the UAV at the current moment.

[0039] The present invention also provides a low-noise path planning system for multiple unmanned aerial vehicles (UAVs), comprising:

[0040] The environment modeling module is used to model the flight environment of the UAV and set noise-sensitive areas and their corresponding noise thresholds.

[0041] The attractive potential field construction module is used to establish the attractive potential field based on the destination position of the UAV.

[0042] The received sound pressure level calculation module is used to obtain the angle between the observation point in the noise-sensitive area and the direction of the UAV's movement, and to obtain the sound pressure level of the observation point in the noise-sensitive area pointed by the UAV using the UAV sound source directivity model; based on the sound pressure level of the observation point in the noise-sensitive area pointed by the UAV and the distance between the observation point in the noise-sensitive area and the UAV, the received sound pressure level of the UAV at the observation point in the noise-sensitive area is calculated.

[0043] The first repulsive potential field construction module is used to map the difference between the received sound pressure level of the UAV and the noise threshold at the observation point in the noise-sensitive area to the first repulsive potential field coefficient; and to construct the repulsive potential field of the noise-sensitive area to the UAV based on the first repulsive potential field coefficient and the distance between the observation point in the noise-sensitive area and the UAV.

[0044] The second repulsive potential field construction module is used to establish the repulsive potential field of static obstacles on the UAV and the repulsive potential field between adjacent UAVs.

[0045] The path generation module is used to calculate the resultant force of the UAV's flight based on the attractive and repulsive potential fields, and to generate the UAV's flight path.

[0046] Compared with the prior art, the above-described technical solution of the present invention has the following advantages:

[0047] This invention discloses a low-noise path planning method for multiple unmanned aerial vehicles (UAVs). Based on an artificial potential field algorithm, it considers the noise impact of UAVs on noise-sensitive areas. First, it calculates the sound pressure level of the UAV pointing towards the noise-sensitive area using a sound source directivity model. Then, it assesses the received sound pressure level of the UAV in the noise-sensitive area based on noise attenuation. The coefficients in the repulsive potential field are dynamically adjusted according to the received noise and a noise threshold. This increases the repulsive force when the received noise level in the noise-sensitive area exceeds the threshold, causing the UAV to move away from the noise-sensitive area. This invention innovatively integrates noise factors into the artificial potential field algorithm. By transforming noise characteristics into a repulsive force mechanism and combining noise characteristics with an artificial potential field, it simplifies the noise assessment process. This not only achieves fast, real-time path planning but also dynamically avoids noise-sensitive areas in real time during path planning, significantly reducing the impact on noise-sensitive areas and significantly improving the environmental friendliness of aircraft operation. This provides key technical support for sustainable development in the low-altitude economy era. Attached Figure Description

[0048] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings, wherein:

[0049] Figure 1This is a flowchart of a low-noise path planning method for multiple unmanned aerial vehicles (UAVs) according to the present invention;

[0050] Figure 2 This is a structural diagram of a low-noise path planning method for multiple unmanned aerial vehicles (UAVs) according to the present invention.

[0051] Figure 3 This is a schematic diagram of the directional expression of the sound source of a drone;

[0052] Figure 4 This is a diagram illustrating the change in repulsive force during the flight of a drone.

[0053] Figure 5 This is a schematic diagram of the environment setup in the second embodiment, wherein... Figure 5 (a) in the diagram is a schematic diagram of the directional model of the drone's sound source. Figure 5 (b) is a visualization of the observation points and the start and end points of the UAV in the conceptual urban environment;

[0054] Figure 6 This is a test result diagram of the second embodiment, in which... Figure 6 In the figure, (a) and (b) are the flight trajectory calculated by the AD-APF method proposed in this invention and its noise at the observation point, respectively. Figure 6 In the figure, (c) and (d) show the flight trajectory calculated by the traditional artificial potential field method and its noise at the observation point, respectively.

[0055] Figure 7 This is a schematic diagram of the environment setup in the third embodiment, wherein... Figure 7 Figures (a) and (b) are visualizations of actual urban environment models. Figure 7 (c) is a visualization of the observation points and the starting and ending points of the UAV in the actual urban environment;

[0056] Figure 8 This is a schematic diagram of the drone sound source directivity model according to the third embodiment, wherein... Figure 8 (a) in the diagram shows the setup of spherically distributed sound source directivity observation points. Figure 8 (b) is a schematic diagram of the noise directivity measured at the spherical observation point in the XY plane;

[0057] Figure 9 This is a test result diagram of the third embodiment, in which... Figure 9 In the figure, (a) and (b) are the flight trajectory calculated by the AD-APF method proposed in this invention and its noise at the observation point, respectively. Figure 9 (c) and (d) in the figure represent the flight trajectory calculated by the traditional artificial potential field method and the noise situation at the observation point, respectively. Detailed Implementation

[0058] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.

[0059] Reference Figure 1 and Figure 2 As shown, this invention provides a low-noise path planning method for multiple unmanned aerial vehicles (UAVs), comprising:

[0060] S1: Model the flight environment of the UAV and set noise-sensitive areas and their corresponding noise thresholds;

[0061] S2: Obtain the angle between the observation point in the noise-sensitive area and the direction of the UAV's flight, and use the UAV sound source directivity model to obtain the sound pressure level of the observation point in the noise-sensitive area pointed by the UAV; based on the sound pressure level of the observation point in the noise-sensitive area pointed by the UAV and the distance between the observation point in the noise-sensitive area and the UAV, calculate the received sound pressure level of the UAV at the observation point in the noise-sensitive area to assess the noise impact of the UAV's flight on the noise-sensitive area.

[0062] S3: Path planning using the Acoustically Driven Artificial Potential Field (AD-APF) algorithm, including:

[0063] Establish an attractive potential field based on the destination position of the drone;

[0064] The difference between the received sound pressure level of the UAV and the noise threshold at the observation point in the noise-sensitive area is mapped to the first repulsive potential field coefficient; based on the first repulsive potential field coefficient and the distance between the observation point in the noise-sensitive area and the UAV, the repulsive potential field of the noise-sensitive area to the UAV is constructed.

[0065] Establish the repulsive potential field of static obstacles on the UAV and the repulsive potential field between adjacent UAVs;

[0066] The resultant force of the UAV's flight is calculated based on the attractive and repulsive potential fields, and the UAV's flight path is generated.

[0067] Specifically, S1 models the UAV's flight environment, sets noise-sensitive areas and their corresponding noise thresholds, including:

[0068] S11: Use open-source databases to obtain urban environmental geometry information, group buildings in the urban environment by category, and obtain multi-category building areas.

[0069] Urban environmental geometry information, including building distribution and ground undulation, is obtained using open-source databases (such as OpenStreetMap). This data is typically presented as a 3D mesh, accurately describing the physical structure of the urban environment. The raw 3D mesh data is then preprocessed, including uniformly structured mesh generation and density clustering of building clusters. Buildings in the urban environment are grouped by category, resulting in multi-category building regions to improve the accuracy and efficiency of noise calculation. For example, the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm can be used to group buildings in the urban environment by category, facilitating subsequent environmental perception and noise calculation.

[0070] S12: Set noise-sensitive areas and their corresponding noise thresholds according to the building area category.

[0071] Based on the actual application scenario, noise-sensitive areas, such as schools, hospitals, and residential areas, are identified in the preprocessed grid data and marked as areas requiring key protection. A noise threshold is then set for each noise-sensitive area; for example, the noise threshold for the school area could be set to 50 dB.

[0072] Specifically, refer to Figure 3 As shown in Figure S2, the UAV sound source directivity model is used to describe the sound pressure level of the UAV in different directions and is constructed based on simulator or acoustic experimental measurement methods. Based on the UAV's flight attitude and the sound source directivity model, the sound pressure level of the UAV in different directions can be calculated. Sound source directivity can be expressed as a function... ,in The angle between the direction of the drone's movement and the direction the drone is pointing towards the observation point in the noise-sensitive area.

[0073] During the flight of the UAV, flight parameters such as position, attitude, and sound source characteristics are acquired. Based on the angle between the observation point in the noise-sensitive area and the direction of the UAV's flight, the sound pressure level of the observation point pointed by the UAV to the noise-sensitive area is obtained using the UAV sound source directivity model.

[0074] Preferably, the observation point in the noise-sensitive area is located at the center of the noise-sensitive area, because the center location can represent the general noise level within the noise-sensitive area. Alternatively, the observation point can be set as the observation point closest to the UAV in the noise-sensitive area.

[0075] Existing environmental noise modeling methods include: physical model-based noise prediction methods, such as boundary element method-based noise prediction, which can accurately simulate noise but has high requirements for input parameters and high computational complexity. Furthermore, the complexity of these numerical simulation methods increases sharply in high-frequency sound wave calculations. Ray tracing-based computational methods can greatly improve computational efficiency, but still cannot meet the needs of real-time computing. Data-driven noise prediction methods, such as machine learning and deep learning methods, can handle nonlinear relationships and have fast prediction speeds, but rely on large amounts of high-quality training data and have complex model structures. In contrast, the ISO 9613-2 standard proposed by the International Organization for Standardization expresses the results of acoustic experiments and simulations through scientific empirical functions, incorporating sound frequency, propagation distance, sound absorption coefficient, and atmospheric medium information into noise calculations, and can be used for fast and accurate noise prediction.

[0076] Therefore, this invention utilizes the empirical model in the ISO 9613-2 standard to calculate the attenuation of noise propagating from the sound source to the observation point, in order to calculate the noise impact of the UAV on noise-sensitive areas under different positions and flight conditions, and characterize the received sound pressure level of the UAV at the observation point.

[0077] Based on the sound pressure level at the observation point in the noise-sensitive area pointed by the drone and the distance between the observation point in the noise-sensitive area and the drone, the received sound pressure level at the observation point in the noise-sensitive area is calculated using the following formula:

[0078] ;

[0079] in, Let be the received sound pressure level of the j-th noise-sensitive area at time t for the i-th UAV. Let i be the sound pressure level of the observation point pointed by the i-th UAV to the j-th noise-sensitive area. For the directional model of the drone's sound source, Let be the angle between the observation point in the j-th noise-sensitive region and the forward direction of the i-th UAV. Let be the distance between the observation point in the j-th noise-sensitive area and the i-th UAV. This represents the distance between the sound source directional observation point and the center of the UAV in the UAV sound source directional model. and These are the influence factors of atmospheric parameters and ground parameters on sound propagation attenuation, respectively. Atmospheric attenuation coefficient, This is the ground attenuation factor, as defined in ISO 9613-2.

[0080] This invention is based on the noise attenuation modeling of the ISO 9613-2 standard, which ensures that the attenuation of noise during propagation can be accurately calculated under different environmental conditions (such as different ground materials, meteorological parameters, etc.) and integrated into the subsequent calculation of artificial potential field, thereby achieving accurate simulation of noise environment and precise guidance of path planning.

[0081] Several algorithms are widely used in the field of agent path planning. For example, the A* algorithm quickly finds the shortest path by combining actual and estimated distances through heuristic search, but it struggles to handle dynamic environmental changes. Reinforcement learning methods, such as deep Q-networks, learn optimal strategies through agent-environment interaction, offering strong adaptability but requiring long training times and being highly dependent on the environment. In multi-agent path planning algorithms, the conflict problem of UAVs is central. Centralized algorithms can plan globally optimal paths but have high computational complexity, while distributed algorithms are computationally efficient but prone to path conflicts. Hybrid algorithms combine the advantages of both but are complex to implement. Therefore, this invention uses an artificial potential field method combined with rapid noise modeling to achieve acoustically driven path planning.

[0082] Specifically, in S3, an attractive potential field is established based on the destination position of the UAV, including:

[0083] The attractive potential field points from the drone's current position to the destination position. The attractive potential energy is calculated using a quadratic function. Based on the definition of the objective function, the force exerted by the attractive potential field on the drone is calculated. In the algorithm, the attractive force of the destination position on the i-th drone is expressed as follows:

[0084] ;

[0085] in, Let the destination position of the i-th drone be the attractiveness of that drone. The gravitational coefficient, Let be the distance between the i-th drone and its destination position. Let be the direction vector from the i-th drone to its endpoint position.

[0086] The repulsive potential field points from the drone's current position towards a static obstacle or noise-sensitive area. This repulsive potential field is active when the distance between the drone and the obstacle or area is less than a safe distance. The strength of the repulsive potential field is related to the distance between the drone and the obstacle or area; the closer the distance, the stronger the repulsive force.

[0087] If the noise level received by the noise-sensitive area exceeds its corresponding noise threshold, the magnitude of the repulsive force of the noise-sensitive area on the drone is increased, causing the drone to move away from the noise-sensitive area. Specifically, in the noise-driven artificial potential field framework, this invention maps the "physical-acoustic coupling field" into a composite mapping operator of "dynamic repulsive gain" in real time, as the dimensionless first repulsive potential field coefficient.

[0088] Specifically, the difference between the received sound pressure level and the noise threshold at the observation point in the noise-sensitive area is mapped to the first repulsive potential coefficient, as shown in the formula:

[0089] ;

[0090] in, Let be the first repulsive potential coefficient of the repulsive potential field of the j-th noise-sensitive region to the i-th UAV. The received sound pressure level of the i-th UAV at the observation point in the j-th noise-sensitive area. Noise threshold of the j-th noise sensitive region The difference, This is a penalty factor.

[0091] when When the first repulsive potential coefficient remains zero, no additional potential field is generated; when At dB, the first repulsive potential coefficient is linearly interpolated between 0 and 1 to reflect the noise avoidance behavior as it approaches the upper threshold; while when At dB, the first repulsive potential coefficient is amplified by introducing a penalty factor β=20 to characterize the sharp amplification effect of excessive noise on behavioral decision-making.

[0092] The repulsive potential field of the noise-sensitive region to the UAV is constructed based on the first repulsive potential field coefficient and the distance between the observation point in the noise-sensitive region and the UAV. The formula is as follows:

[0093] ;

[0094] in, Let be the repulsive force exerted by the j-th noise-sensitive region on the i-th UAV. Let be the distance between the observation point in the j-th noise-sensitive area and the i-th UAV. For minimum safe distance, Let be the direction vector from the i-th UAV to the j-th noise-sensitive area observation point. for Step function.

[0095] This invention effectively transforms complex noise characteristics (including the dynamic directivity of the sound source, frequency characteristics, atmospheric absorption, and ground reflection) into repulsive force factors in an artificial potential field, enabling aircraft to sense and respond to changes in the noise environment in real time.

[0096] The repulsive potential field of the static obstacle on the UAV is established by the following formula:

[0097] ;

[0098] in, Let p be the repulsive force exerted by the p-th static obstacle on the i-th drone. The coefficient of the second repulsive potential field. Let p be the distance between the p-th static obstacle and the i-th drone. For minimum safe distance, Let be the direction vector from the i-th drone to the p-th static obstacle. for The step function is defined as follows: That is, when hour Conversely, it is 0.

[0099] The repulsive potential field between adjacent UAVs is established by the following formula:

[0100] ;

[0101] in, For the first The repulsive force of one drone on the i-th drone, The coefficient of the third repulsive potential field. For the first The distance between the i-th drone and the i-th drone The radius of the drone. For the first The direction vector of the i-th drone is pointed to by the drone.

[0102] In multi-agent scenarios, this invention introduces a repulsive force mechanism between drones to ensure a safe distance between them, thereby avoiding the risk of collision. This method not only improves the operational safety of drones in dense airspace but also enables collaborative operations among drones, enhancing the overall system's operational efficiency. Compared to existing technologies, the AD-APF algorithm has significant advantages in multi-agent collaborative planning, better balancing the mutual influence between aircraft, enabling them to complete tasks efficiently and safely in complex environments, and effectively handling path planning problems when multiple aircraft are operating simultaneously.

[0103] Figure 4This is an example diagram illustrating the change in repulsive force during the flight of a drone.

[0104] To address the local minima problem inherent in traditional artificial potential field methods, this invention, when calculating the resultant flight force of a UAV based on attractive and repulsive potential fields, also includes:

[0105] Calculate the magnitude of the resultant force of the UAV's flight at each moment;

[0106] If the magnitude of the flight resultant force is less than the preset jamming threshold for a continuous preset time period, a disturbance vector with a random magnitude along the tangent direction of the nearest static obstacle is generated, and the disturbance vector is superimposed on the flight resultant force of the UAV at the current moment.

[0107] Specifically, when the magnitude of the flight resultant force is less than the preset jamming threshold, the number of consecutive "low resultant force" steps can be recorded by accumulating the counter. Once the accumulator exceeds the preset number of steps, such as 10 steps, it is determined that the UAV has entered a local minimum, and a perturbation vector with a random magnitude is generated along the tangent direction of the nearest static obstacle. And add it to the flight force of the drone at the current moment. This breaks the local equilibrium, drives the agent out of the stagnant area, and then resets the counter to restart the stuck detection.

[0108] This invention discloses a low-noise path planning method for multiple unmanned aerial vehicles (UAVs). Based on an artificial potential field algorithm, it considers the noise impact of UAVs on noise-sensitive areas. First, it calculates the sound pressure level of the UAV pointing towards the noise-sensitive area using a sound source directivity model. Then, it assesses the received sound pressure level of the UAV in the noise-sensitive area based on noise attenuation. The coefficients in the repulsive potential field are dynamically adjusted according to the received noise and a noise threshold. This increases the repulsive force when the received noise level in the noise-sensitive area exceeds the threshold, causing the UAV to move away from the noise-sensitive area. This invention innovatively integrates noise factors into the artificial potential field algorithm. By transforming noise characteristics into a repulsive force mechanism and combining noise characteristics with an artificial potential field, it simplifies the noise assessment process. This not only achieves fast, real-time path planning but also dynamically avoids noise-sensitive areas in real time during path planning, significantly reducing the impact on noise-sensitive areas and significantly improving the environmental friendliness of aircraft operation. This provides key technical support for sustainable development in the low-altitude economy era.

[0109] Specifically, compared with the traditional artificial potential field method, the AD-APF method significantly improves noise control performance. For example, in simulation experiments, paths planned using the AD-APF method effectively reduce noise levels, lowering the maximum sound pressure level (Lmax) and sound exposure level (SEL) by 10.4 dB and 4.4 dB, respectively, significantly outperforming traditional methods. Secondly, the invention optimizes noise modeling and calculation, effectively solving the problems of slow computation speed and difficulty in real-time evaluation in existing low-noise path planning methods. Traditional numerical simulation-based low-noise path planning methods require enormous computational resources and time costs, while AD-APF, by combining noise characteristics with an artificial potential field, keeps computation time within a reasonable range under different scenarios. Furthermore, the computation time increases relatively gradually with the number of drones, meeting the needs of real-time path planning. This allows drones to quickly adjust their flight paths according to real-time environmental changes, adapting to complex and ever-changing urban environments and greatly improving the operational efficiency and flexibility of the aircraft.

[0110] Furthermore, this invention possesses excellent scalability and adaptability, making it suitable for various aircraft types and complex environments. Whether it's a small UAV or a large electric vertical takeoff and landing (EVTOL) aircraft, the AD-APF can be adapted to its specific noise characteristics and flight requirements by adjusting relevant parameters. This broad applicability gives the invention wide application prospects in various scenarios, providing strong technical support for the widespread application of low-altitude aircraft.

[0111] In summary, this invention has achieved significant technological advancements in noise control, real-time planning, multi-agent collaboration, adaptability, and scalability, providing an efficient and practical solution for low-noise path planning of low-altitude aircraft, and possessing significant application value and promising prospects for widespread adoption.

[0112] To verify the effectiveness of the method of the present invention, the second embodiment of the present invention selects a conceptual urban environment for flight environment modeling. This environment includes 46 buildings, covering an area of ​​1.5 × 1.5 km², of which 13 are defined as noise-sensitive areas (such as schools, hospitals, residential areas, etc.). Figure 5 As shown, the noise-sensitive area is represented by a green cylinder, in which... Figure 5 (a) in the diagram is a schematic diagram of the directional model of the drone's sound source. Figure 5 (b) in the diagram is a visualization of the observation points and the start and end points of the UAV in the conceptual urban environment. The noise threshold for all observation points is set to 50 dB, meaning that the flight noise of the UAV near the protected area cannot exceed 50 dB.

[0113] In this embodiment, four UAVs were selected for path planning tests. The starting and ending points of each UAV are shown in Table 1. The maximum flight altitude of the UAVs was set to 120 meters above the ground, and the flight speed was maintained at 5 m / s. The noise source model of the UAVs adopted a dipole-shaped sound source model, with a sound pressure level between 80 and 85 dB.

[0114] Table 1. Starting and ending coordinates of the four drones in the second embodiment.

[0115]

[0116] In this embodiment, AD-APF is used for path planning, and the specific steps include:

[0117] Initialize the flight environment model, including building distribution, ground undulations, and the location of noise-sensitive areas;

[0118] Initialize the attractive and repulsive potential field parameters in the AD-APF algorithm based on the UAV's starting position and target position;

[0119] During path planning, noise propagation between the drone and noise-sensitive areas is calculated in real time. The magnitude and direction of the repulsive force are adjusted according to the noise level to guide the drone to avoid noise-sensitive areas. At the same time, other drones are considered as dynamic obstacles, and their repulsive forces on the current drone are calculated to avoid collisions.

[0120] The test results of this embodiment are as follows: Figure 6 As shown, where Figure 6 In the figure, (a) and (b) are the flight trajectory calculated by the AD-APF method proposed in this invention and its noise at the observation point, respectively. Figure 6 (c) and (d) in the figure show the flight trajectory calculated by the traditional artificial potential field method and the noise level at the observation point, respectively. It can be seen that the low-noise path (LNP) planned by the AD-APF method significantly reduces the maximum sound pressure level (Lmax) and sound exposure level (SEL) within the protected area compared to the reference path (RP) planned by the traditional artificial potential field method. Specifically, using LNP, Lmax is reduced by 10.4 dB, SEL is reduced by 4.4 dB, while the path length only increases by about 205 meters (11.5%). This indicates that the AD-APF method can effectively reduce the impact on noise-sensitive areas while ensuring flight efficiency.

[0121] The third embodiment of the present invention was simulated in a real urban environment in Kowloon City, Hong Kong. This area also covered 1.5 × 1.5 km² and contained many high-rise and low-rise buildings. Sixteen noise observation points were set up within the area to monitor the noise level of the aircraft, referring to... Figure 7As shown, the noise-sensitive area is represented by a green cylinder, in which... Figure 7 Figures (a) and (b) are visualizations of actual urban environment models. Figure 7 (c) is a visualization of the observation points and the start and end points of the UAV in the actual urban environment.

[0122] This embodiment uses four 7kg DJI Matrices-350 drones for testing. The starting and ending points of the drones are shown in Table 2. The maximum flight altitude is also set to 120 meters above the ground, and the flight speed is maintained at 5 m / s. Their directional characteristics are as follows: Figure 8 As shown, where Figure 8 (a) in the diagram shows the setup of spherically distributed sound source directivity observation points. Figure 8 (b) is a schematic diagram of the noise directivity measured at the spherical observation point in the XY plane.

[0123] Table 2. Starting and ending coordinates of the four UAVs in the third embodiment.

[0124]

[0125] In this embodiment, AD-APF is used for path planning, and the specific steps include:

[0126] The geometric environment model of the area was obtained using an open-source database and preprocessed, including uniform structured mesh generation and density clustering of building clusters.

[0127] The processed environment model is input into the AD-APF algorithm to initialize the relevant parameters;

[0128] During the path planning process, the noise propagation between the UAV and the noise observation point is calculated in real time. The influence of factors such as atmospheric absorption and ground reflection on noise is considered according to the ISO9613-2 standard, and the repulsion force is dynamically adjusted.

[0129] At the same time, the flight path is adjusted in real time, taking into account the impact of other drones and static obstacles on the current drone.

[0130] The test results of this embodiment are as follows: Figure 9 As shown, where Figure 9 In the figure, (a) and (b) are the flight trajectory calculated by the AD-APF method proposed in this invention and its noise at the observation point, respectively. Figure 9(c) and (d) in the figure show the flight trajectory calculated by the traditional artificial potential field method and its noise level at the observation point, respectively. Simulation results in a real urban environment demonstrate that the AD-APF method can also effectively reduce the impact of the aircraft on noise-sensitive areas. Compared with the RP method, LNP reduces Lmax by 2.1 dBA and SEL by 1.3 dBA within the protected area, while the average path length increases by only about 55 meters (3.0%). This further validates the effectiveness and practicality of the AD-APF method in complex urban environments.

[0131] Based on the above-mentioned low-noise path planning method for multiple unmanned aerial vehicles (UAVs), the present invention also provides a low-noise path planning system for multiple UAVs, comprising:

[0132] The environment modeling module is used to model the flight environment of the UAV and set noise-sensitive areas and their corresponding noise thresholds.

[0133] The attractive potential field construction module is used to establish the attractive potential field based on the destination position of the UAV.

[0134] The received sound pressure level calculation module is used to obtain the angle between the observation point in the noise-sensitive area and the direction of the UAV's movement, and to obtain the sound pressure level of the observation point in the noise-sensitive area pointed by the UAV using the UAV sound source directivity model; based on the sound pressure level of the observation point in the noise-sensitive area pointed by the UAV and the distance between the observation point in the noise-sensitive area and the UAV, the received sound pressure level of the UAV at the observation point in the noise-sensitive area is calculated.

[0135] The first repulsive potential field construction module is used to map the difference between the received sound pressure level of the UAV and the noise threshold at the observation point in the noise-sensitive area to the first repulsive potential field coefficient; and to construct the repulsive potential field of the noise-sensitive area to the UAV based on the first repulsive potential field coefficient and the distance between the observation point in the noise-sensitive area and the UAV.

[0136] The second repulsive potential field construction module is used to establish the repulsive potential field of static obstacles on the UAV and the repulsive potential field between adjacent UAVs.

[0137] The path generation module is used to calculate the resultant force of the UAV's flight based on the attractive and repulsive potential fields, and to generate the UAV's flight path.

[0138] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0139] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0140] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0141] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0142] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A low-noise path planning method for multiple unmanned aerial vehicles (UAVs), characterized in that, include: Model the flight environment of the UAV and set noise-sensitive areas and their corresponding noise thresholds; Obtain the angle between the observation point in the noise-sensitive area and the direction of the UAV's movement; use the UAV's sound source directivity model to obtain the sound pressure level of the observation point in the noise-sensitive area pointed by the UAV; calculate the received sound pressure level of the observation point in the noise-sensitive area to the UAV based on the sound pressure level of the observation point in the noise-sensitive area and the distance between the observation point in the noise-sensitive area and the UAV. Establish an attractive potential field based on the destination position of the drone; The difference between the received sound pressure level of the UAV and the noise threshold at the observation point in the noise-sensitive area is mapped to the first repulsive potential coefficient. Based on the first repulsive potential coefficient and the distance between the observation point in the noise-sensitive region and the UAV, a repulsive potential field of the noise-sensitive region on the UAV is constructed. Establish the repulsive potential field of static obstacles on the UAV and the repulsive potential field between adjacent UAVs; The resultant force of the UAV's flight is calculated based on the attractive potential field and all repulsive potential fields, and the UAV's flight path is generated.

2. The low-noise path planning method for multiple unmanned aerial vehicles according to claim 1, characterized in that, Modeling the flight environment of the UAV, setting noise-sensitive areas and their corresponding noise thresholds, including: By using open-source databases to obtain urban environmental geometric information, buildings in the urban environment are grouped by category to obtain multi-category building areas; Noise-sensitive areas and their corresponding noise thresholds are set according to the building area category.

3. The low-noise path planning method for multiple unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, The UAV sound source directivity model is used to describe the sound pressure level of the UAV in different directions and is constructed based on simulator or acoustic experimental measurement methods.

4. The low-noise path planning method for multiple unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, The observation point in the noise-sensitive area is set at the center of the noise-sensitive area.

5. The low-noise path planning method for multiple unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, The sound pressure level received by the drone at the observation point is calculated based on the sound pressure level pointed at the observation point and the distance between the observation point and the drone. The formula is as follows: ； in, Let be the received sound pressure level of the j-th noise-sensitive area at time t for the i-th UAV. Let i be the sound pressure level of the observation point pointed by the i-th UAV to the j-th noise-sensitive area. For the directional model of the drone's sound source, Let be the angle between the observation point in the j-th noise-sensitive region and the forward direction of the i-th UAV. Let be the distance between the observation point in the j-th noise-sensitive area and the i-th UAV. This represents the distance between the sound source directional observation point and the center of the UAV in the UAV sound source directional model. and These are the influence factors of atmospheric parameters and ground parameters on sound propagation attenuation, respectively. Atmospheric attenuation coefficient, This is the ground attenuation coefficient.

6. The low-noise path planning method for multiple unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, The received sound pressure level of the UAV at the observation point in the noise-sensitive area is mapped to the noise threshold as the first repulsive potential coefficient, and the formula is: ； in, Let be the first repulsive potential coefficient of the repulsive potential field of the j-th noise-sensitive region to the i-th UAV. Let be the difference between the received sound pressure level of the i-th UAV at the observation point in the j-th noise-sensitive area and the noise threshold of the j-th noise-sensitive area. As a penalty factor; The repulsive potential field of the noise-sensitive region to the UAV is constructed based on the first repulsive potential field coefficient and the distance between the observation point in the noise-sensitive region and the UAV. The formula is as follows: ； in, Let be the repulsive force exerted by the j-th noise-sensitive region on the i-th UAV. Let be the distance between the observation point in the j-th noise-sensitive area and the i-th UAV. For minimum safe distance, Let be the direction vector from the i-th UAV to the j-th noise-sensitive area observation point. for Step function.

7. The low-noise path planning method for multiple unmanned aerial vehicles according to claim 1, characterized in that, The attractive potential field is established based on the destination position of the drone, and the formula is: ； in, Let the destination position of the i-th drone be the attractiveness of that drone. The gravitational coefficient, Let be the distance between the i-th drone and its destination position. Let be the direction vector from the i-th drone to its endpoint position.

8. The low-noise path planning method for multiple unmanned aerial vehicles according to claim 1, characterized in that, The repulsive potential field of the static obstacle on the UAV is established by the following formula: ； in, Let p be the repulsive force exerted by the p-th static obstacle on the i-th drone. The coefficient of the second repulsive potential field. Let p be the distance between the p-th static obstacle and the i-th drone. For minimum safe distance, Let be the direction vector from the i-th drone to the p-th static obstacle. for Step function; The repulsive potential field between adjacent UAVs is established by the following formula: ； in, For the first The repulsive force of one drone on the i-th drone, The coefficient of the third repulsive potential field. For the first The distance between the i-th drone and the i-th drone The radius of the drone. For the first The direction vector of the i-th drone is pointed to by the drone.

9. A low-noise path planning method for multiple unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, When calculating the resultant force of a UAV's flight based on the attractive and repulsive potential fields, it also includes: Calculate the magnitude of the resultant force of the UAV's flight at each moment; If the magnitude of the flight resultant force is less than the preset jamming threshold for a continuous preset time period, a disturbance vector with a random magnitude along the tangent direction of the nearest static obstacle is generated, and the disturbance vector is superimposed on the flight resultant force of the UAV at the current moment.

10. A low-noise path planning system for multiple unmanned aerial vehicles (UAVs), characterized in that, include: The environment modeling module is used to model the flight environment of the UAV and set noise-sensitive areas and their corresponding noise thresholds. The attractive potential field construction module is used to establish the attractive potential field based on the destination position of the UAV. The received sound pressure level calculation module is used to obtain the angle between the observation point in the noise-sensitive area and the direction of the UAV's movement, and to obtain the sound pressure level of the observation point in the noise-sensitive area pointed by the UAV using the UAV sound source directivity model; based on the sound pressure level of the observation point in the noise-sensitive area pointed by the UAV and the distance between the observation point in the noise-sensitive area and the UAV, the received sound pressure level of the UAV at the observation point in the noise-sensitive area is calculated. The first repulsive potential field construction module is used to map the difference between the received sound pressure level of the UAV and the noise threshold at the observation point in the noise-sensitive area to the first repulsive potential field coefficient. Based on the first repulsive potential coefficient and the distance between the observation point in the noise-sensitive region and the UAV, a repulsive potential field of the noise-sensitive region on the UAV is constructed. The second repulsive potential field construction module is used to establish the repulsive potential field of static obstacles on the UAV and the repulsive potential field between adjacent UAVs. The path generation module is used to calculate the resultant force of the UAV's flight based on the attractive potential field and all repulsive potential fields, and to generate the UAV's flight path.

Images