Unmanned aerial vehicle dynamic path planning method based on spatial capacity and communication quality constraint

Abstract

The application discloses a dynamic path planning method for unmanned aerial vehicles based on airspace capacity and communication quality constraints. Firstly, the target low-altitude airspace is discretized into multiple three-dimensional grid cells in the spatial dimension, and the passable state of each grid cell and its spatial connectivity are determined based on the spatial distribution information of buildings. At the same time, the low-altitude airspace is discretized in the time dimension, and the corresponding airspace capacity constraints are set for each grid cell in combination with the minimum safety interval requirement of the unmanned aerial vehicle. A monitoring communication quality evaluation model is constructed to evaluate the corresponding communication quality indicators in real time according to the propagation conditions between the grid cell where the unmanned aerial vehicle is located and the ground monitoring station. A dynamic task arrival model is constructed to predict the occupation of each grid cell at different times in the future. The application generates a safe and efficient grid sequence flight path for the spatio-temporal joint path planning of the newly arrived tasks under the premise of meeting the grid passable constraint, the airspace capacity constraint and the communication quality constraint.

Description

Technical Field

[0001] This invention relates to the fields of low-altitude airspace management, UAV path planning and scheduling, and specifically to a dynamic path planning method for UAVs based on airspace capacity and communication quality constraints. Background Technology

[0002] With the rapid development of the low-altitude economy, drones are increasingly being used in urban logistics, inspection and monitoring, and emergency rescue. Urban low-altitude airspace is gradually evolving from the traditional single-drone, low-density operation mode to a multi-drone, high-density collaborative operation mode. Against this backdrop, how to achieve safe, orderly, and efficient drone operation management in complex urban environments has become a critical issue that urgently needs to be addressed.

[0003] Current UAV path planning methods often combine global path planning with local obstacle avoidance. Global planning typically focuses only on efficiency metrics such as flight distance, time, or energy consumption, while specific collision avoidance and safety control are handled by the individual UAV through local perception during execution. This approach is applicable to low-density, static environments, but it has significant limitations in complex urban low-altitude environments. Firstly, urban environments are characterized by dense buildings and complex spatial structures, with low-altitude airspace exhibiting significant discontinuities and occlusions, making it difficult for traditional two-dimensional or simplified three-dimensional models to accurately depict the structure of passable airspace. Secondly, with the increasing number of UAVs, relying solely on local obstacle avoidance mechanisms can easily lead to excessive local airspace load, making it difficult to effectively reduce potential conflict risks at the global level. Furthermore, UAVs typically rely on ground-based surveillance systems for status broadcasting and situational awareness during operation. Factors such as building obstruction and changes in propagation conditions significantly affect the quality of surveillance communication, and existing research often fails to systematically incorporate communication quality constraints into the path planning and airspace management processes. Meanwhile, existing methods generally do not adequately consider the dynamic nature of mission arrival and lack the ability to predict and model the temporal occupancy of allocated flight paths. This makes it difficult to perceive the availability of future airspace resources in a timely manner when planning new missions, thereby affecting the overall scheduling safety and airspace utilization efficiency.

[0004] Therefore, there is an urgent need for a dynamic path planning method for UAVs that can uniformly model across spatial, temporal, and communication dimensions and comprehensively consider accessibility, safety intervals, airspace capacity, and monitoring communication quality, in order to meet the needs of safe operation and efficient scheduling of multiple UAVs in urban low-altitude areas. Summary of the Invention

[0005] The purpose of this invention is to provide a dynamic path planning method for unmanned aerial vehicles (UAVs) based on airspace capacity and communication quality constraints. By performing spatiotemporal discrete modeling of urban low-altitude airspace and introducing airspace capacity constraints and monitoring communication quality constraints, a global, safe, and predictable traffic management method for multiple UAV operations can be achieved, thereby improving the safety, controllability, and utilization efficiency of urban low-altitude airspace operations.

[0006] To achieve the above-mentioned technical objectives, the technical solution adopted by the present invention is as follows:

[0007] Dynamic path planning methods for UAVs based on airspace capacity and communication quality constraints include:

[0008] Step 1: Discretize the target low-altitude airspace into multiple three-dimensional grid cells, identify the spatial connectivity between each grid cell, and determine the passability status of each grid cell based on the spatial distribution information of buildings.

[0009] Step 2: Discretize the low-altitude airspace in the time dimension, and set corresponding airspace capacity constraints for each grid cell based on the minimum safe interval requirement of UAVs.

[0010] Step 3: Construct a surveillance communication quality assessment model and determine the corresponding communication quality indicators based on the propagation conditions between the grid cell where the UAV is located and the ground surveillance station;

[0011] Step 4: Construct a dynamic task arrival model, track the status of executed tasks, and build a predictive airspace occupancy model based on their assigned flight paths to predict the occupancy status of each grid cell at different future times.

[0012] Step 5: Under the premise of satisfying grid accessibility constraints, airspace capacity constraints, and communication quality constraints, perform spatiotemporal path planning for the newly arriving mission and generate the corresponding grid sequence flight path.

[0013] To optimize the above technical solution, the specific measures also include:

[0014] In step 1, the entire city's low-altitude area is... Regular discretization is performed to construct a three-dimensional raster model composed of cubic mesh elements. Represents the set of all grid cells, each grid cell From its center point coordinates and side length The corresponding three-dimensional volume region is defined as follows:

[0015]

[0016] To characterize the reachability relationships between meshes, a binary adjacency variable is introduced. To describe any two grids The connectivity, when When, it means that the two grids are directly adjacent in space and can be connected to each other. If there is no direct reachability between the two,

[0017] The urban environment includes several buildings, denoted by their index. Each building Its two-dimensional planar projection profile and height The determined region, which constitutes its occupied area in three-dimensional space, is represented as:

[0018]

[0019] If a certain grid If a grid intersects with the volume occupied by any building, it is considered non-navigable, formally defined as follows:

[0020] .

[0021] In step 2, a capacity constraint based on a minimum safety distance is set for each grid cell. It is assumed that all UAVs within the same grid are at the same vertical height level. For any grid... Its maximum capacity is defined as:

[0022]

[0023] in, This represents the horizontal projected area of ​​a cubic grid cell. This means that any two drones should maintain a minimum safe distance. Minimum required horizontal area This is the floor function.

[0024] The time dimension is discretized into equal-length time steps. ,set up Indicates time Occupy grid The number of drones, and their utilization rate is defined as:

[0025]

[0026] when When the value is close to 1, it indicates that the grid is under high load; if If the grid capacity is exceeded, it indicates a potential collision risk.

[0027] In step 3, a monitoring signal model is constructed to characterize the communication link quality between the UAV and the ground monitoring base station. The monitoring station is set as follows: Equipped with a receiving antenna, its location is In dense urban environments, due to the probability of partial obstruction by buildings, the signal transmitted by a drone within a certain grid cell may be in a line-of-sight (LoS) or non-line-of-sight (NLoS) state.

[0028] To characterize the spatial differences within the mesh, an area-weighted method is used to estimate the probabilities of Loss and Non-LoS states. and Representing grids respectively Internal surveillance station The probabilities of visible and invisible areas are as follows:

[0029]

[0030] in, Represents the area of ​​the corresponding region, based on the air-to-ground channel model, grid. To the monitoring station The path loss is expressed as:

[0031]

[0032] in, The Euclidean distance between the grid center and the monitoring station. The location of the grid center. At the speed of light, For carrier frequency, and Let represent the average additional loss of the LoS and NLoS links, respectively. The expected path loss of the mesh is:

[0033]

[0034] Assuming the drone's transmission power is The receiving antenna gain is Then at time Grid The signal-to-interference-plus-noise ratio (SIR) of an internal drone is expressed as:

[0035]

[0036] in, For noise power spectral density, To monitor signal bandwidth, As the effective interference factor, aggregated interference power Consider drone signals from the same grid and adjacent grids, i.e.:

[0037]

[0038] Indicates time Occupy adjacent grid The number of drones, when the signal-to-interference-plus-noise ratio of the drones is higher than a preset threshold. If the signal is received successfully, it is considered that the monitoring signal has been successfully received; otherwise, it is considered that a communication blind spot has occurred, affecting the reliability of the overall airspace monitoring.

[0039] In step 4, the dynamically arriving flight mission flow is considered, denoted as... Each task Grid cells consisting of start and end points Arrival time and deadline As defined, upon arrival of a task, the control center needs to decide whether to approve the task; if the task is approved, the system allocates a feasible spatiotemporal path for it, denoted by a binary variable. Indicates task At any moment Whether it is approved, i.e.

[0040]

[0041] Each approved task Assign a grid sequence path ,in For the first on the path One grid cell, The path length is a fixed value, and the flight time of each drone within a single grid cell is a fixed value. This duration includes the nominal flight time as well as additional time reserved for safe maneuvers and operational adjustments. The total flight time is ,

[0042] To achieve forward-looking and safety-aware path planning, a predictive occupancy map needs to be constructed to estimate future airspace occupancy based on currently executed tasks. Tasks are divided into two categories: active tasks. The corresponding flight missions have already been executed along the assigned paths, and the missions have been reached. Then wait for path allocation.

[0043] For each task to be planned Predicting the occupancy map in its feasible planning time domain Internal build, in which For the current moment, For the task The deadline. For activated tasks. Its allocation path is Then for the grid At any moment The predicted occupancy contribution is:

[0044]

[0045] in, For the first on the path One grid, For mission takeoff time, For flight duration, mission The total predicted usage is calculated by summing the contributions of all activated tasks:

[0046]

[0047] The complete predicted occupancy chart is denoted as:

[0048] .

[0049] In step 5, spatiotemporal path planning is performed for the newly arriving task, generating a corresponding grid sequence flight path. The spatial state of each search node is represented as follows: This indicates that the drone is at any time. Occupy grid State transitions only allow the UAV to move from one grid to an adjacent grid in a fixed time increment. Specifically, from state... Once launched, the drone can move to any nearby location. ,in Indicates and For adjacent grid sets, the state expansion during the search process is based on the total estimated cost function:

[0050]

[0051] in, Indicates from the initial state to The cumulative cost, For heuristic estimation, reflecting from The remaining cost to reach the goal,

[0052] During the search process, from the state Transferred to The cost takes into account three objectives, namely

[0053]

[0054] in, , and These are non-negative weights used to control the relative importance of path length, congestion penalty, and signal quality penalty.

[0055] Basic cost Assigning a unit cost to each grid transfer incentivizes choosing shorter paths; the congestion penalty is defined as:

[0056]

[0057] in

[0058]

[0059] To predict occupancy rates and suppress drones from passing through congested areas, a penalty is imposed on states with low predicted signal-to-interference-plus-noise ratios to ensure communication quality. For state The predicted signal-to-interference-plus-noise ratio (SINNR) is then defined as follows:

[0060]

[0061] in, and These represent the signal-to-interference-plus-noise ratio (SIR) threshold and the ideal level, respectively. and By controlling the penalty magnitude, this design enables the planner to comprehensively consider path length, airspace congestion, and communication quality during the search process, thereby generating a safe, efficient flight path with stable signal.

[0062] Flight path planning is performed by the control center for each new arrival mission. Calculate feasible paths, specifically, search from the initial state. Add to open list Initially, the list stores candidate states to be explored. In each iteration, the method starts from... Extracting from the options has the minimum total cost. The state in which The cumulative cost to reach the current state. For the Manhattan distance heuristic estimation, for the selected state Check each adjacent grid And generate candidate next states. The following condition must be met throughout the entire transfer period: the arrival time must not exceed the task deadline. And the prediction occupancy of the next grid does not exceed the capacity limit. All feasible state transitions are evaluated using the defined multi-objective cost function; if a lower-cost path to s' is found, the state is updated. and All newly discovered or costly improvements will be re-inserted into the state. For future expansion, the search terminates upon reaching the target mesh, at which point the final path is generated using predecessor mapping. If no feasible path exists, the task is rejected.

[0063] The present invention has the following beneficial effects:

[0064] This invention discretizes urban low-altitude airspace into three-dimensional grid units, fully leveraging the precision and flexibility of grid modeling in depicting the structure of navigable airspace in complex urban architectural environments. This improves the accuracy of UAVs' control over airspace availability and safety during the global path planning phase. By introducing airspace capacity constraints in the time dimension and combining them with minimum safe interval requirements for UAVs, this invention effectively avoids excessive aggregation of multiple UAVs in local airspace, achieving higher airspace load balancing compared to traditional global path planning methods. This invention fully utilizes the real-time assessment capability of the surveillance communication quality evaluation model for propagation conditions between ground surveillance stations and UAVs in dynamic environments, incorporating communication quality as a unified path planning constraint into the spatiotemporal planning framework, which is beneficial for improving UAV flight safety and situational awareness reliability. By constructing a predictive airspace occupancy model, the temporal occupancy of allocated flight paths is modeled in advance, allowing new arrival tasks to fully consider the availability of future airspace resources during the planning phase, thereby achieving safer and more efficient spatiotemporal path scheduling management for multiple UAVs. Attached Figure Description

[0065] Figure 1 This invention relates to a scenario diagram of a low-altitude unmanned aerial vehicle (UAV) traffic management system.

[0066] Figure 2 This is a schematic diagram of the average signal-to-interference-plus-noise ratio distribution of the communication monitoring signal involved in this invention;

[0067] Figure 3 This is a graph showing the experimental simulation results of the relationship between the average signal-to-interference-plus-noise ratio and the average grid occupancy ratio and the task arrival rate of this invention;

[0068] Figure 4 This is a simulation result of the relationship between the task completion rate and average path efficiency and the task arrival rate of the present invention.

[0069] Figure 5 This is a graph showing the experimental simulation results of the relationship between the average task planning time and the average maximum grid load violation count and the task arrival rate according to the present invention.

[0070] Figure 6 This is a comparison chart of the average grid occupancy ratio between the path planning method of this invention and the traditional method. Detailed Implementation

[0071] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.

[0072] Figure 1This paper illustrates an application scenario for a low-altitude unmanned aerial vehicle (UAV) traffic management system. In this scenario, the airspace is divided into multiple three-dimensional grid cells, each capable of accommodating multiple UAVs simultaneously while maintaining a minimum safe separation distance. Ground-based monitoring base stations are deployed, and the UAVs in the air transmit monitoring signals to these stations via their equipped monitoring equipment. Each UAV's dynamic mission involves planning a flight path from a designated starting point to a destination; this flight path is centrally allocated and managed by the UAV traffic management system.

[0073] Figure 2 The simulation test scenario and the average signal-to-interference-plus-noise ratio distribution of the monitoring signal used in this invention visually demonstrate the impact of building obstruction on the quality of the monitoring signal.

[0074] Figure 3 The figures show the simulation results of the average signal-to-interference-plus-noise ratio (SIR) and average grid occupancy ratio under different mission arrival rates using the UAV path planning method proposed in this invention. As the mission arrival rate increases, the average SIR decreases, while the grid occupancy ratio increases. At a mission arrival rate of 3, the proposed method achieves the highest average SIR among all methods, improving by approximately 28% compared to traditional planning methods, while also improving airspace utilization uniformity by approximately 35% compared to traditional methods. Because the method emphasizes maintaining surveillance communication quality, its average grid occupancy is slightly higher than the variant that does not consider SIR, while the traditional planning method performs the worst overall.

[0075] Figure 4 The figures show the experimental simulation results of task completion rate and average path efficiency under different task arrival rates using the UAV path planning method proposed in this invention. The task completion rate and average path efficiency (defined as the ratio of the shortest path length to the actual path length assigned by the planner) are compared. Both indicators show a slight decrease with increasing task arrival rate, reflecting the trade-offs brought about by introducing airspace capacity and communication quality constraints in path planning. Notably, when the task arrival rate is 3, a total of 4,694 tasks were processed during the 500-second simulation, and the task completion rate of the proposed method was approximately 90%. Unfinished tasks were mainly due to exceeding the expected deadline, highlighting the sacrifices made to ensure airspace safety.

[0076] Figure 5 The figures show the experimental simulation results of the average task planning time and average maximum grid load violation number under different task arrival rates using the UAV path planning method proposed in this invention. The average planning time of the proposed method is slightly higher than that of the traditional method, reaching 0.26 seconds at a task arrival rate of 3, but still meets the requirements for real-time operation. Both the proposed method and its congestion-considered variant maintain zero violations of maximum grid capacity, while the traditional method, which does not consider airspace capacity and communication quality constraints, has an average of 33.7 grids exceeding the capacity limit per task during flight.

[0077] Figure 6 To use the UAV path planning method proposed in this invention ( Figure 6 (b) and traditional methods ( Figure 6 (a) Comparison of average grid occupancy ratios. Similar to the spatial distribution of the average signal-to-interference-plus-noise ratio (SIR) of the surveillance signal shown in Figure 2, it can be seen that the proposed algorithm can guide the UAV to areas with high SIR while effectively avoiding areas obstructed by tall buildings. Furthermore, the distribution of UAVs in the airspace is more uniform, and no local congestion was observed, fully demonstrating the algorithm's synergistic ability to optimize flight safety and surveillance signal quality.

[0078] The UAV dynamic path planning method based on airspace capacity and communication quality constraints described in this invention includes the following steps:

[0079] In step 1, the entire city's low-altitude area is... Regular discretization is performed to construct a three-dimensional raster model composed of cubic mesh elements. Let... Represents the set of all grid cells, each grid cell From its center point coordinates and side length Define it. The corresponding three-dimensional volume region is represented as:

[0080]

[0081] This discretization method effectively characterizes key spatial features in urban environments, such as narrow airways, building obstructions, and height hierarchy, providing a unified spatial reference framework for subsequent path planning and signal calculation.

[0082] To characterize the reachability relationships between meshes, a binary adjacency variable is introduced. To describe any two grids Connectivity. When When , it means that the two grids are directly adjacent in space and can be connected to each other. If If the two are not directly reachable, then there is no direct reachability between them.

[0083] The urban environment includes several buildings, denoted by their index. Each building Its two-dimensional planar projection profile and height The determined region, which constitutes its occupied area in three-dimensional space, is represented as:

[0084]

[0085] If a certain grid A grid is considered non-navigable if it intersects with the volume occupied by any building. The formal definition is as follows:

[0086]

[0087] In step 2, a capacity constraint based on a minimum safety distance is set for each grid cell, assuming that all UAVs within the same grid are at the same vertical height level. For any grid... Its maximum capacity is defined as:

[0088]

[0089] in, This represents the horizontal projected area of ​​a cubic grid cell. This means that any two drones should maintain a minimum safe distance. Minimum required horizontal area This is a floor function. This capacity constraint characterizes the upper limit of the number of drones that can safely remain in each grid cell simultaneously, providing a basis for subsequent airspace congestion assessment and path planning.

[0090] The time dimension is discretized into equal-length time steps. .set up Indicates time Occupy grid The number of drones, and their utilization rate is defined as:

[0091]

[0092] when When the value is close to 1, it indicates that the grid is under high load; if This indicates that the grid capacity has been exceeded, posing a potential collision risk and reflecting a state of airspace congestion and safety constraints.

[0093] In step 3, a monitoring signal model is constructed to characterize the communication link quality between the UAV and the ground monitoring base station. Let the monitoring station... Equipped with a receiving antenna, its location is In dense urban environments, due to the potential partial obstruction by buildings, the signal transmitted by a drone within a grid cell may be in either line-of-sight (LoS) or non-line-of-sight (NLoS) mode.

[0094] To characterize the spatial differences within the mesh, an area-weighted method is used to estimate the probabilities of Loss-of-Stake (LoS) and Non-LoS (NLoS) states. Let... and Representing grids respectively Internal surveillance station The probabilities of visible and invisible areas are as follows:

[0095]

[0096] in, This represents the area of ​​the corresponding region. Based on the air-to-ground channel model, the grid... To the monitoring station The path loss is expressed as:

[0097]

[0098] in, The Euclidean distance between the grid center and the monitoring station. At the speed of light, For carrier frequency, and These represent the average additional loss for LoS and NLoS links, respectively. The expected path loss of the mesh is:

[0099]

[0100] Assuming the drone's transmission power is The receiving antenna gain is Then at time Grid The signal-to-interference-plus-noise ratio (SIR) of an internal drone is expressed as:

[0101]

[0102] in, For noise power spectral density, To monitor signal bandwidth, Effective interference factor. Aggregated interference power. Consider drone signals from the same grid and adjacent grids, i.e.:

[0103]

[0104] When the signal-to-interference-plus-noise ratio of the drone is higher than the preset threshold When the signal is received, it is considered that the monitoring signal has been successfully received; otherwise, a communication blind spot may occur, which may affect the reliability of the overall airspace monitoring.

[0105] In step 4, the dynamically arriving flight mission flow is considered, denoted as... Each task Grid cells consisting of start and end points Arrival time and deadline Defined. Upon arrival of a task, the control center needs to decide whether to approve it; if approved, the system assigns a feasible spatiotemporal path to it. Let there be a binary variable. Indicates task At any moment Whether it is approved, i.e.

[0106]

[0107] Each approved task Assign a grid sequence path ,in For the first on the path One grid cell, This represents the path length. The flight time of each drone within a single grid cell is a fixed value. This duration includes the nominal flight time as well as additional time reserved for safe maneuvers and operational adjustments. Mission The total flight time is .

[0108] To achieve proactive and safety-aware path planning, a predictive occupancy map needs to be constructed to estimate future airspace occupancy based on currently executed tasks. Tasks are categorized into two types: active tasks. The corresponding flight missions have already been executed along the assigned paths, and the missions have been reached. Then wait for path allocation.

[0109] For each task to be planned Predicting the occupancy map in its feasible planning time domain Internal build, in which For the current moment, For the task The deadline. For activated tasks. Its allocation path is Then for the grid At any moment The predicted occupancy contribution is:

[0110]

[0111] in, For the first on the path One grid, For mission takeoff time, For flight duration. Mission The total predicted usage is calculated by summing the contributions of all activated tasks:

[0112]

[0113] The complete predicted occupancy chart is denoted as:

[0114]

[0115] In step 5, spatiotemporal path planning is performed for the newly arriving task, generating the corresponding grid sequence flight path. The spatial state of each search node is represented as follows: This indicates that the drone is at any time. Occupy grid State transitions only allow the UAV to move from one grid to an adjacent grid in a fixed time increment. Specifically, from state... Once launched, the drone can move to any nearby location. ,in Indicates and The set of adjacent grids. The state expansion during the search process is based on the total estimated cost function:

[0116]

[0117] in, Indicates from the initial state to The cumulative cost, For heuristic estimation, reflecting from The remaining cost to reach the goal.

[0118] During the search process, from the state Transferred to The cost takes into account three objectives, namely

[0119]

[0120] in, , and These are non-negative weights used to control the relative importance of path length, congestion penalty, and signal quality penalty.

[0121] Basic cost Assigning a unit cost to each grid transfer incentivizes choosing shorter paths. The congestion penalty is defined as:

[0122]

[0123] in

[0124]

[0125] To predict occupancy rates, this is used to suppress drones from passing through congested areas. To ensure communication quality, a penalty is imposed on states with low predicted signal-to-interference-plus-noise ratios. Let... For state The predicted signal-to-interference-plus-noise ratio (SINNR) is then defined as follows:

[0126]

[0127] in, and These represent the signal-to-interference-plus-noise ratio (SIR) threshold and the ideal level, respectively. and Controlling the penalty magnitude. This design enables the planner to comprehensively consider path length, airspace congestion, and communication quality during the search process, thereby generating safe, efficient flight paths with stable signals.

[0128] Flight path planning is performed by the UTM control center for each new arrival mission. Calculate feasible paths. Specifically, the search starts from the initial state. Add to open list Initially, this list stores candidate states to be explored. In each iteration, the method starts from... Extracting from the options has the minimum total cost. The state in which The cumulative cost to reach the current state. For the Manhattan distance heuristic estimation. For the selected state. The method checks each adjacent grid. And generate candidate next states. The following condition must be met throughout the entire transfer period: the arrival time must not exceed the task deadline. And the prediction occupancy of the next grid does not exceed the capacity limit. All feasible state transitions are evaluated using the defined multi-objective cost function; if a lower-cost path to s' is found, the state is updated. and All new discoveries or cost improvements will be re-inserted into the state. For future expansion. The search terminates upon reaching the target mesh, at which point the final path is generated using predecessor mapping. If no feasible path exists, the task is rejected.

[0129] In another embodiment, the present invention provides a computer device including a processor and a memory; wherein, when the processor executes a computer program stored in the memory, it implements the steps of the above-described UAV dynamic path planning method based on airspace capacity and communication quality constraints.

[0130] For more detailed information on the above methods, please refer to the relevant content disclosed in the foregoing embodiments, which will not be repeated here.

[0131] In another embodiment, the present invention provides a computer-readable storage medium for storing a computer program; when the computer program is executed by a processor, it implements the steps of the above-described UAV dynamic path planning method based on airspace capacity and communication quality constraints.

[0132] For more detailed information on the above methods, please refer to the relevant content disclosed in the foregoing embodiments, which will not be repeated here.

[0133] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. The systems, devices, and storage media disclosed in the embodiments are described simply because they correspond to the methods disclosed in the embodiments; relevant details can be found in the method section.

[0134] Those skilled in the art will clearly understand that the techniques in the embodiments of the present invention can be implemented using software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solutions in the embodiments of the present invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in various embodiments or certain parts of the embodiments of the present invention.

[0135] 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 functions specified in one or more boxes. These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable apparatus 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.

[0136] The above are merely preferred embodiments of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should be considered within the scope of protection of the present invention.

Claims

1. A dynamic path planning method for unmanned aerial vehicles (UAVs) based on airspace capacity and communication quality constraints, characterized in that, include: Step 1: Discretize the target low-altitude airspace into multiple three-dimensional grid cells, identify the spatial connectivity between each grid cell, and determine the passability status of each grid cell based on the spatial distribution information of buildings. Step 2: Discretize the low-altitude airspace in the time dimension, and set corresponding airspace capacity constraints for each grid cell based on the minimum safe interval requirement of UAVs. Step 3: Construct a surveillance communication quality assessment model and determine the corresponding communication quality indicators based on the propagation conditions between the grid cell where the UAV is located and the ground surveillance station; Step 4: Construct a dynamic task arrival model, track the status of executed tasks, and build a predictive airspace occupancy model based on their assigned flight paths to predict the occupancy status of each grid cell at different future times. Step 5: Under the premise of satisfying grid accessibility constraints, airspace capacity constraints, and communication quality constraints, perform spatiotemporal path planning for the newly arriving mission and generate the corresponding grid sequence flight path.

2. The UAV dynamic path planning method based on airspace capacity and communication quality constraints according to claim 1, characterized in that, In step 1, the entire city's low-altitude area is... Regular discretization is performed to construct a three-dimensional raster model composed of cubic mesh elements. Represents the set of all grid cells, each grid cell From its center point coordinates and side length The corresponding three-dimensional volume region is defined as follows: ； To characterize the reachability relationships between meshes, a binary adjacency variable is introduced. To describe any two grids The connectivity, when When, it means that the two grids are directly adjacent in space and can be connected to each other. If there is no direct reachability between the two, The urban environment includes several buildings, denoted by their index. Each building Its two-dimensional planar projection profile and height The determined region, which constitutes its occupied area in three-dimensional space, is represented as: ； If a certain grid If a grid intersects with the volume occupied by any building, it is considered non-navigable, formally defined as follows: 。 3. The UAV dynamic path planning method based on airspace capacity and communication quality constraints according to claim 2, characterized in that, In step 2, a capacity constraint based on a minimum safety distance is set for each grid cell. It is assumed that all UAVs within the same grid are at the same vertical height level. For any grid... Its maximum capacity is defined as: ； in, This represents the horizontal projected area of ​​a cubic grid cell. This means that any two drones should maintain a minimum safe distance. Minimum required horizontal area This is the floor function. The time dimension is discretized into equal-length time steps. ,set up Indicates time Occupy grid The number of drones, and their utilization rate is defined as: ； when When the value is close to 1, it indicates that the grid is under high load; if If the grid capacity is exceeded, it indicates a potential collision risk.

4. The UAV dynamic path planning method based on airspace capacity and communication quality constraints according to claim 3, characterized in that, In step 3, a monitoring signal model is constructed to characterize the communication link quality between the UAV and the ground monitoring base station. The monitoring station is set as follows: Equipped with a receiving antenna, its location is In dense urban environments, due to the probability of partial obstruction by buildings, the signal transmitted by a drone within a certain grid cell may be in a line-of-sight (LoS) or non-line-of-sight (NLoS) state. To characterize the spatial differences within the mesh, an area-weighted method is used to estimate the probabilities of Loss and Non-LoS states. and Representing grids respectively Internal surveillance station The probabilities of visible and invisible areas are as follows: ； in, Represents the area of ​​the corresponding region, based on the air-to-ground channel model, grid. To the monitoring station The path loss is expressed as: ； in, The Euclidean distance between the grid center and the monitoring station. The location of the grid center. At the speed of light, For carrier frequency, and Let represent the average additional loss of the LoS and NLoS links, respectively. The expected path loss of the mesh is: ； Assuming the drone's transmission power is The receiving antenna gain is Then at time Grid The signal-to-interference-plus-noise ratio (SIR) of an internal drone is expressed as: ； in, For noise power spectral density, To monitor signal bandwidth, As the effective interference factor, aggregated interference power Consider drone signals from the same grid and adjacent grids, i.e.: ； Indicates time Occupy adjacent grid The number of drones, when the signal-to-interference-plus-noise ratio of the drones is higher than a preset threshold. If the signal is received successfully, it is considered that the monitoring signal has been successfully received; otherwise, it is considered that a communication blind spot has occurred, affecting the reliability of the overall airspace monitoring.

5. The UAV dynamic path planning method based on airspace capacity and communication quality constraints according to claim 4, characterized in that, In step 4, the dynamically arriving flight mission flow is considered, denoted as... Each task Grid cells consisting of start and end points Arrival time and deadline As defined, upon arrival of a task, the control center needs to decide whether to approve the task; if the task is approved, the system allocates a feasible spatiotemporal path for it, assuming a binary variable. Indicates task At any moment Whether it is approved, i.e. ； Each approved task Assign a grid sequence path ,in For the first on the path One grid cell, The path length is a fixed value, and the flight time of each drone within a single grid cell is a fixed value. This duration includes the nominal flight time as well as additional time reserved for safe maneuvers and operational adjustments. Total flight time is , To achieve proactive and safety-aware path planning, a predictive occupancy map needs to be constructed to estimate future airspace occupancy based on currently executed tasks. Tasks are divided into two categories: active tasks. This corresponds to the flight missions that have already been executed along the assigned path, and the missions that have already arrived. Then wait for path allocation. For each task to be planned Predicting the occupancy map in its feasible planning time domain Internal build, in which For the current moment, For the task The deadline for activated tasks Its allocation path is Then for the grid At any moment The predicted occupancy contribution is: ； in, For the first on the path One grid, For mission takeoff time, For flight duration, mission The total predicted usage is calculated by summing the contributions of all activated tasks: ； The complete predicted occupancy chart is denoted as: 。 6. The UAV dynamic path planning method based on airspace capacity and communication quality constraints according to claim 5, characterized in that, In step 5, spatiotemporal path planning is performed for the newly arriving task, generating a corresponding grid sequence flight path. The spatial state of each search node is represented as follows: This indicates that the drone is at any time. Occupy grid State transitions only allow the UAV to move from one grid to an adjacent grid in a fixed time increment. Specifically, from state... Once launched, the drone can move to any nearby location. ,in Indicates and For adjacent grid sets, the state expansion during the search process is based on the total estimated cost function: ； in, Indicates from the initial state to The cumulative cost, For heuristic estimation, reflecting from The remaining cost to reach the goal, During the search process, from the state Transferred to The cost takes into account three objectives, namely ； in, , and These are non-negative weights used to control the relative importance of path length, congestion penalty, and signal quality penalty. Basic cost Assigning a unit cost to each grid transfer incentivizes choosing shorter paths; the congestion penalty is defined as: ； in ； To predict occupancy rates and suppress drones from passing through congested areas, a penalty is imposed on states with low predicted signal-to-interference-plus-noise ratios to ensure communication quality. For state The predicted signal-to-interference-plus-noise ratio (SINNR) is then defined as follows: ； in, and These represent the signal-to-interference-plus-noise ratio (SIR) threshold and the ideal level, respectively. and By controlling the penalty magnitude, this design enables the planner to comprehensively consider path length, airspace congestion, and communication quality during the search process, thereby generating a safe, efficient flight path with stable signal. Flight path planning is performed by the control center for each new arrival mission. Calculate feasible paths, specifically, search from the initial state. Add to open list Initially, the list stores candidate states to be explored. In each iteration, the method starts from... Extracting from the options has the minimum total cost. The state in which The cumulative cost to reach the current state. For the Manhattan distance heuristic estimation, for the selected state Check each adjacent grid And generate candidate next states. The following condition must be met throughout the entire transfer period: the arrival time must not exceed the task deadline. And the prediction occupancy of the next grid does not exceed the capacity limit. All feasible state transitions are evaluated using the defined multi-objective cost function; if a lower-cost path to s' is found, the state is updated. and All newly discovered or costly improvements will be re-inserted into the state. For future expansion, the search terminates upon reaching the target mesh, at which point the final path is generated using predecessor mapping. If no feasible path exists, the task is rejected.

Images