A method for optimizing the flight path of a swarm intelligence logistics unmanned aerial vehicle for urban low altitude

Abstract

The present application relates to a kind of urban low-altitude-oriented swarm intelligence logistics unmanned aerial vehicle path optimization method, comprising the following steps: component distribution task mathematical model, unmanned aerial vehicle position, speed and fitness parameter initialization, introduce strategy gradient and action value evaluation real-time update unmanned aerial vehicle's position and speed, and calculate the fitness value of current unmanned aerial vehicle, carry out local calculation to obtain local solution and global optimal solution, whether the next path distribution node meets the requirement is judged by local solution and global optimal solution, repeat local calculation until reach maximum iteration number, finally obtain global optimal solution to determine the path route of unmanned aerial vehicle group and logistics distribution task.The path planning method of the present application can more efficiently process optimization problems in dynamic environment, especially when facing multi-objective optimization, nonlinear constraint and other complex tasks, it shows stronger robustness.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) trajectory planning technology, specifically to a method for optimizing the trajectory of swarm intelligent logistics UAVs for urban low-altitude environments. Background Technology

[0002] Currently, although drone swarm logistics and delivery has broad prospects, there are still many practical difficulties in achieving large-scale and routine commercial applications.

[0003] The existing technology mainly has the following technical problems:

[0004] 1. Insufficient environmental adaptability: The "skyscraper canyon" effect in cities and complex electromagnetic interference can affect the drone's accurate navigation and obstacle avoidance capabilities;

[0005] 2. Limitations in battery life and payload: Currently, the battery life and payload of mainstream logistics drones are limited. How can drones achieve multi-point delivery with limited weight?

[0006] Therefore, before the logistics delivery of drone swarms, it is necessary to optimize the delivery route of each drone based on the logistics delivery information in order to achieve the maximum delivery efficiency of drones. However, how to optimize the flight path delivery route is currently the difficulty of drone logistics delivery. Summary of the Invention

[0007] To address the shortcomings of the existing technologies, this invention provides a method for optimizing the flight paths of swarm intelligent logistics drones for low-altitude urban environments.

[0008] To achieve the above objectives, the present invention adopts the following technical solution:

[0009] A method for optimizing the flight paths of swarm intelligent logistics drones for low-altitude urban environments, including...

[0010] (1) Construct an initial mathematical model for the delivery task based on environmental parameters, trajectory planning cost parameters, and logistics delivery parameters, and simultaneously analyze the location set of the UAV swarm. and velocity set Initialize the program and calculate the fitness value of the current UAV; where, Represents the total quantity. positive integer;

[0011] (2) The drone swarm introduces strategy gradient and action value assessment during the flight to the predetermined destination based on the delivery task mathematical model, updates the drone's position and speed in real time, and calculates the current drone's fitness value.

[0012] (3) Based on the latest UAV position and speed, perform local calculations on the UAV based on its undelivered destination to obtain a local solution. And optimize the global optimal solution. ;

[0013] (4) The UAV solves the local solution and the global optimal solution Determine if the next delivery node on the flight path meets the requirements; if it does, proceed to the next step; if it does not, re-initialize the drone and the delivery task.

[0014] (5) Repeat steps (2), (3) and (4) to calculate the track delivery node until the logistics delivery task is completed or the maximum number of iterations is reached;

[0015] (6) Determine the flight path and logistics delivery task of the UAV swarm based on the final global optimal solution.

[0016] Furthermore, logistics and delivery parameters include the total weight of the delivered goods, the number of delivered goods, and the coordinates of the delivered goods. These parameters will affect the selection or planning of the drone's flight path.

[0017] in, For the exponential decay exploration rate, It follows a Gaussian distribution. The maximum number of iterations, This is the attenuation coefficient.

[0018] Furthermore, in step (3), the delivery task mathematical model is updated in real time during the local calculation process of the UAV. The parameters such as static obstacles, hidden radar, airspace restricted areas, and new targets in the new delivery task mathematical model are updated in real time through the information fed back by the UAV monitoring, so that the UAV's trajectory planning is closer to reality, dynamic, flexible and not rigid under deep reinforcement learning.

[0019] Furthermore, the speed update formula for the drone is:

[0020] ;

[0021] in, For policy gradient, For the action value evaluation function, It is inertial weight. and For adaptive weights, For the parameters of the policy network, For the first One drone location. For the first The probability of detection by a radar. The current time value;

[0022] The and The formula for calculating adaptive weights is:

[0023] ;

[0024] in, is the base of the natural logarithm, used to calculate exponential functions. To represent the objective function For parameters The norm of the gradient, i.e., the magnitude of the gradient.

[0025] Furthermore, the position update formula is:

[0026] ;

[0027] The local calculation mechanism is as follows:

[0028] ;

[0029] in, For the scope of local calculation, It is a random vector that follows a normal distribution;

[0030] The formula for the global optimal solution is:

[0031] ;

[0032] in, Indicates the state of the given point. Below, based on the current policy function Choose an action The expected value of the value is usually calculated by averaging over multiple simulations or training.

[0033] Furthermore, environmental parameters include static obstacles, areas of complex electromagnetic interference, and overall airspace confinement areas;

[0034] The functional representation of a static obstacle is:

[0035] ;

[0036] in, Indicates at given coordinates The height value at that location, This indicates the total number of obstacles. It is the first The center coordinates of the obstacles Indicates the first The height of the obstacle and These are the attenuation amounts along the x-axis and y-axis, respectively;

[0037] The complex electromagnetic interference region is represented by a hemispherical model, specifically as follows:

[0038] ;

[0039] in, Indicates the first A complex electromagnetic interference zone, ( ) is the first The center coordinates of a complex electromagnetic interference zone It is the first The coverage radius of a complex electromagnetic interference zone;

[0040] The overall airspace restricted area is represented as follows: ;

[0041] For the first Each airspace restricted region is modeled using a cylindrical region, specifically expressed as follows:

[0042] ;

[0043] in, Represents the spatial coordinates of the drone. The center of the cylinder's base is represented by r, and the radius is r. The drone's altitude range is... .

[0044] Furthermore, the fitness value function is expressed as:

[0045] ;

[0046] in, Represents the fitness value. This represents the cost of track length. Indicates the cost of electromagnetic interference. This represents the sum of yaw angle costs. , and These are the weighting constants for the sum of track length cost, electromagnetic interference cost, and yaw angle cost, respectively.

[0047] Furthermore, the cost of track length is expressed as:

[0048] ;

[0049] in, It is the first The coordinates of the interpolated track delivery node It represents the total number of track delivery nodes after interpolation.

[0050] Furthermore, the cost of electromagnetic interference is expressed as:

[0051] ;

[0052] in, This indicates the total number of complex electromagnetic interference zones. positive integer, This indicates the total number of delivery nodes on the flight path. Indicates that the drone is in The probability of a complex electromagnetic interference zone being interfered with;

[0053] Probability of being subjected to electromagnetic interference Represented as:

[0054] ;

[0055] in, and These represent the effective detection radius and the maximum detection radius, respectively. It is the first Center of a complex electromagnetic interference zone and the Each tracking delivery node The straight-line distance between them is calculated using the following formula:

[0056] .

[0057] Furthermore, the sum of yaw angle costs is expressed as:

[0058] ;

[0059] in, Indicates the first The yaw angle cost of each track delivery node; The function formula is:

[0060] ;

[0061] in, Indicates the yaw angle. The calculation function is:

[0062] ;

[0063] in, and They are the first and the +1 segment of the delivery node vector, It is the dot product of vectors. and It is the magnitude of the vector. Indicates the maximum turning angle.

[0064] In summary, the present invention has the following advantages:

[0065] 1. The trajectory optimization method of the present invention first simulates and calculates whether the UAV swarm can complete the logistics delivery task based on the initialized delivery task mathematical model, and plans the optimal route in the simulation process; if the logistics delivery task is not completed, the model is re-initialized.

[0066] 2. The trajectory optimization method of the present invention integrates the advantages of intelligent computing and reinforcement learning policy optimization. By introducing policy gradient and action value evaluation into the mathematical model, it makes full use of the delivery efficiency of the UAV swarm.

[0067] 3. The trajectory planning method of this invention provides a unified optimization framework that is flexible in structure, efficient in training, and robust, enabling UAV swarms to enhance environmental adaptability while maximizing efficiency (maximum range and payload) during logistics delivery. Attached Figure Description

[0068] Figure 1 This is a flowchart of a method for optimizing the flight path of swarm intelligent logistics drones for urban low-altitude environments, according to the present invention. Detailed Implementation

[0069] The present invention will now be described in further detail.

[0070] A method for optimizing the flight path of swarm intelligent logistics drones for low-altitude urban environments includes the following steps:

[0071] (1) Construct an initial mathematical model for the delivery task based on environmental parameters, trajectory planning cost parameters, and logistics delivery parameters, and simultaneously analyze the location set of the UAV swarm. and velocity set Initialization is performed. Each drone's initial position and velocity can be 0 or assigned by the drone's control system. After initialization, the fitness value of the current drone is calculated. This fitness value is used to evaluate whether the drone can complete the remaining logistics delivery tasks and return successfully. Represents the total quantity. Positive integers. Logistics and distribution parameters include the total weight of the delivered goods, the number of delivered goods, and the coordinates of the delivered goods.

[0072] (2) The drone swarm introduces strategy gradient and action value assessment during the flight to the predetermined destination according to the delivery task mathematical model to update the drone's position and speed in real time, and calculates the current drone's fitness value. This achieves fast convergence speed and does not require a lot of exploration in the early stage of training. In particular, it is not easy to have performance fluctuations in multi-logistics delivery or unstable environments.

[0073] (3) Based on the latest UAV position and speed, perform local calculations on the UAV based on its undelivered destination to obtain a local solution. And optimize the global optimal solution. The delivery task mathematical model is updated in real time during the drone's local flight path delivery node calculation process, avoiding the static nature of the delivery task mathematical model and better reflecting actual logistics delivery scenarios. The specific calculation and update logic involves obtaining new data through local calculations. Its fitness value is better than the current global optimum. If the algorithm approaches the optimal solution, it will update the global optimal solution. This strategy effectively avoids the algorithm from stagnating when it is close to the optimal solution, allowing it to escape from the trap of local optima and thus improve the accuracy of the global optimal solution.

[0074] (4) The UAV solves the local solution and the global optimal solution The system determines whether the next delivery node meets the requirements. If it does, the process proceeds to the next step; otherwise, the drone and delivery task are re-initialized. The local solution and the global optimal solution are determined by two main logics: if the local solution is greater than the global optimal solution, the drone uses the local solution to determine whether the next delivery node (the destination of the logistics task) meets the requirements; if the local solution is less than or equal to the global optimal solution, the drone uses the global optimal solution to determine whether the next delivery node meets the requirements. The requirements mentioned here are that the drone consumes the least amount of energy and has the best safety when flying to the next delivery node.

[0075] (5) Repeat steps (2), (3) and (4) to calculate the delivery node until the logistics delivery task is completed or the maximum number of iterations is reached; when the number of iterations reaches the maximum value, it means that the UAV can complete the logistics delivery task according to the predetermined delivery task mathematical model.

[0076] (6) Determine the flight path and logistics delivery task of the UAV swarm based on the final global optimal solution.

[0077] The initial delivery task mathematical model has preliminarily completed the delivery task allocation of the UAV swarm based on logistics delivery information. Steps (2) to (5) are to verify whether the UAVs can complete the initial delivery task and to plan the optimal flight path for the delivery task. If the UAVs cannot complete the initial delivery task, the uncompleted delivery task and the corresponding UAVs are re-initialized to construct a new delivery task mathematical model until all logistics delivery is completed, that is, the maximum number of iterations is reached. The initial delivery task is randomly allocated based on the total weight of the delivered goods, the number of delivered goods, and the coordinates of the delivered goods according to the principle of proximity and straight route.

[0078] During the drone trajectories exploration process, the position and velocity of each drone represent a candidate solution in the problem's solution space. Through repeated iterations, the drone swarm gradually converges toward the optimal solution.

[0079] The speed update formula for the drone is:

[0080] ;

[0081] in, For policy gradient, For the action value evaluation function, It is inertial weight. and For adaptive weights, For the first One drone location. For the first The probability of detection by a radar. The current time value;

[0082] The and The formula for calculating adaptive weights is:

[0083] ;

[0084] in, is the base of the natural logarithm, used to calculate exponential functions. To represent the objective function For parameters The norm of the gradient, i.e., the magnitude of the gradient.

[0085] The position update formula is as follows:

[0086] ;

[0087] in, For the exponential decay exploration rate, The distribution follows a Gaussian pattern, and T is the maximum number of iterations. This is the attenuation coefficient.

[0088] The dynamic adjustment of UAV speed and position incorporates policy gradient and action value assessment. It achieves a balance between directionality and adaptability by combining the Hadamard product with the empirical difference between the individual UAV and the global optimum. For speed updates, a Softmax mechanism is used for dynamic adjustment, relying on both the policy gradient norm and action value assessment. For position updates, a Gaussian noise perturbation term is introduced to enhance local computational capabilities, combined with an exponentially decaying exploration rate. This enables the transition from breadth-based computation to stable convergence.

[0089] The introduction of local computation mechanism not only improves accuracy but also enhances the algorithm's adaptability to complex problems, particularly in multi-logistics delivery task optimization problems, where it can more accurately approximate the global optimum. The local computation mechanism is as follows:

[0090] ;

[0091] in, The scope of local calculation is usually small, controlling the degree of disturbance; It is a random vector that follows a normal distribution, and determines the direction and magnitude of the disturbance.

[0092] The formula for the global optimal solution is:

[0093] ;

[0094] in, Indicates the state of the given point. Below, based on the current policy function The expected value of the Q-value for the selected action is usually calculated by averaging over multiple simulations or training sessions.

[0095] In the process of calculating the global optimal solution, actions are generated through the Actor network and evaluated by the Critic network, thereby realizing a "value-driven" optimal selection mechanism. The improvement embeds the experience evaluation and policy guidance mechanism in reinforcement learning into the calculation of UAV swarm, forming a deeply integrated optimization structure.

[0096] The construction of a mathematical model for delivery tasks can better simulate the logistics delivery tasks of drone swarms, helping drones plan safer and more efficient flight paths for delivery nodes. The environmental parameters include static obstacles, areas with complex electromagnetic interference, and overall airspace restrictions.

[0097] The static obstacle is represented functionally as follows:

[0098] ;

[0099] in, Indicates at given coordinates The height value at that location, This indicates the total number of obstacles. It is the first The center coordinates of the obstacles Indicates the first The height of the obstacle and These represent the attenuation along the x-axis and y-axis, respectively. This method allows for the flexible construction of complex terrains by superimposing multiple obstacles, helping to simulate environmental obstacle changes in real flight environments.

[0100] Radar in cities often interferes with the signal reception of drones, thus affecting their control. Therefore, drones must avoid radar detection range. Consequently, an electromagnetic interference region needs to be introduced into the delivery task mathematical model. This electromagnetic interference region is represented by a hemispherical model, specifically:

[0101] ;

[0102] in, Indicates the first The detection area of ​​each radar It is the first The coordinates of the radar center It is the first The radius of the radar model. When a UAV enters this range, its detection probability increases significantly. To adapt to practical applications, the maximum detection radius also needs to be considered in areas with complex electromagnetic interference. and effective detection radius These two parameters are used to limit the effective detection range of the radar and prevent drones from entering dangerous areas.

[0103] Airspace restricted areas refer to three-dimensional spaces that drones are prohibited from entering, defined by geographical location or other means. In real-world scenarios, airspace restricted areas are typically set around important transportation hubs or in height-restricted areas. Airspace restricted areas are usually defined based on factors such as urban roads and the outward expansion of sensitive locations, combined with flight altitude restrictions to form a three-dimensional airspace restricted area. To better manage and plan air traffic, airspace restricted areas are modeled as cylindrical regions. The cylindrical shape not only simplifies calculations but also offers higher efficiency and accuracy in air traffic management.

[0104] In a task scenario, there will be multiple spatially restricted regions. Each spatially restricted region can be represented by a cylindrical model. Therefore, the overall spatially restricted region is represented as follows:

[0105] ;

[0106] The No. Each airspace restricted region is modeled using a cylindrical region, specifically expressed as follows:

[0107] ;

[0108] in, Represents the spatial coordinates of the drone. The center of the cylinder's base is represented by r, and the radius is r. The drone's altitude range is... This approach provides a clear and concise mathematical description of the airspace confinement area within the mission space, while also facilitating calculations and optimizations during subsequent trajectory planning.

[0109] The delivery task mathematical model employs obstacle functions to simulate static obstacles (tall buildings), a hemispherical model to represent complex electromagnetic interference areas, and a cylindrical model to represent airspace confinement areas. This provides a more realistic environmental mathematical model for UAV trajectory planning. This modeling method not only simplifies the handling of complex environments but also accurately simulates obstacles and electromagnetic interference areas encountered during flight missions, providing a valuable reference for UAV trajectory planning.

[0110] The fitness value directly determines the efficiency and quality of the trajectory optimization process. By quantifying the merits of trajectories, the fitness value helps the algorithm evaluate and select the optimal trajectory scheme, ensuring that the trajectory not only meets the requirements of flight efficiency and safety, but also avoids obstacles and electromagnetic interference areas as much as possible, thus ensuring the successful completion of the flight mission.

[0111] The fitness value is formed by weighting the track length cost, electromagnetic interference cost, and yaw angle cost. The fitness value function is expressed as follows:

[0112] ;

[0113] in, Represents the fitness value. This represents the cost of track length. Indicates the cost of electromagnetic interference. This represents the sum of yaw angle costs. , and These are the weight constants for the sum of track length cost, electromagnetic interference cost, and yaw angle cost, respectively. Since the units and ranges of the three costs are different, they need to be standardized before weighting to normalize their value range to the [0,1] interval, so that the weighting process is more reasonable and avoids one cost from overdoing the optimization process. Therefore, it is necessary to increase the weight constants to adjust this.

[0114] The cost of flight path length directly impacts flight efficiency. Generally, it's desirable to minimize flight path length to reduce flight time and energy consumption. By optimizing the cost of flight path length, it's possible to ensure that the UAV completes its mission within the designated mission area using the shortest path, reducing flight time and energy consumption. The cost of flight path length is expressed as:

[0115] ;

[0116] in, It is the first The coordinates of the interpolated track delivery node; It is the total number of track delivery nodes after interpolation, when When the length is large, the sum of the distances of the individual straight line segments can be used to approximate the length of the track.

[0117] The electromagnetic interference zone primarily considers the potential interference from urban radar signals emitted by the UAV during flight. If the UAV enters the effective detection range of the radar, it may be interfered with by radar signals. Therefore, avoiding entering the radar's detection range is a crucial factor in flight path planning. The distance between the UAV and the radar directly affects the probability of detection. The electromagnetic interference cost is expressed as:

[0118] ;

[0119] in, Indicates the total number of radars. =W∈positive integer, This represents the total number of track delivery nodes, and P is the probability of being affected by electromagnetic interference.

[0120] The probability P of being subjected to electromagnetic interference is expressed as:

[0121] ;

[0122] in, and These represent the effective detection radius and the maximum detection radius, respectively. It is the first Center of a complex electromagnetic interference zone and the Each tracking delivery node The straight-line distance between them is calculated using the following formula:

[0123] .

[0124] By using the electromagnetic interference cost function, it is possible to ensure that the UAV avoids the radar detection range as much as possible during flight, minimizing signal interference. The calculation of the electromagnetic interference cost is not only related to the radar detection range, but also closely related to the distance between the UAV and the radar. By optimizing the electromagnetic interference cost, the UAV can achieve stealthy flight, avoiding exposure to radar signal interference.

[0125] Yaw angle is a crucial constraint that cannot be ignored during drone flight. Excessive turning can lead to flight instability, even exceeding the drone's turning capabilities, causing loss of control or collisions. Therefore, limiting the yaw angle during flight is essential. Between two adjacent track delivery nodes, the yaw angle can be calculated by the angle between the track vectors. By limiting the yaw angle, excessive turning during flight can be effectively prevented, ensuring a smooth and safe flight path and avoiding flight instability problems caused by excessive turns.

[0126] The sum of the yaw angle costs is expressed as:

[0127] ;

[0128] in, Indicates the first The yaw angle cost of each track delivery node; The function formula is:

[0129] ;

[0130] in, Indicates the yaw angle. The calculation function is:

[0131] ;

[0132] in, and They are the first and the +1 segment of the delivery node vector; It is the dot product of vectors; and It is the magnitude of the vector; This indicates the maximum turning angle. If the yaw angle is greater than the set maximum turning angle... If so, a penalty will be imposed on that turn.

[0133] This invention achieves a synergistic effect through the deep integration of three mechanisms: policy gradient-guided directional exploration, action value assessment-driven merit-based judgment, and a local perturbation escape mechanism. The policy gradient ensures that the UAV swarm converges rapidly toward high-value areas in the early iterations, while the action value assessment provides refined action quality feedback at each decision step. The adaptive weights of these two mechanisms mean that the switching between exploration and utilization no longer depends on manual parameter tuning but is automatically adjusted according to environmental feedback. At the same time, the local computation mechanism applies controllable random perturbations near the global optimum, effectively avoiding the problem of traditional particle swarm optimization stagnating at local optima. This perturbation, guided by the Q-value, does not diverge blindly, ultimately achieving a balance of fast convergence, high accuracy, and strong robustness. This synergistic mechanism enables the UAV swarm to adapt to different urban geographical scenarios and delivery task distributions without retraining, and can simultaneously balance flight safety, energy economy, and task completion rate in complex scenarios with multiple obstacles and interference zones in urban low-altitude areas.

[0134] The above embodiments are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above embodiments. Any changes, modifications, substitutions, combinations, or simplifications made without departing from the spirit and principle of the present invention shall be considered equivalent substitutions and shall be included within the protection scope of the present invention.

Claims

1. A method for optimizing the flight path of swarm intelligent logistics drones for urban low-altitude environments, characterized by: include (1) Construct an initial mathematical model for the delivery task based on environmental parameters, trajectory planning cost parameters, and logistics delivery parameters, and simultaneously analyze the location set of the UAV swarm. and velocity set Initialize the program and calculate the fitness value of the current UAV; where, Represents the total quantity. positive integer; (2) The drone swarm introduces strategy gradient and action value assessment during the flight to the predetermined destination based on the delivery task mathematical model, updates the drone's position and speed in real time, and calculates the current drone's fitness value. (3) Based on the latest UAV position and speed, perform local calculations on the UAV based on its undelivered destination to obtain a local solution. And optimize the global optimal solution. ; (4) The UAV solves the local solution and the global optimal solution Determine if the next delivery node on the flight path meets the requirements; if it does, proceed to the next step; if it does not, re-initialize the drone and the delivery task. (5) Repeat steps (2), (3) and (4) to calculate the track delivery node until the logistics delivery task is completed or the maximum number of iterations is reached; (6) Determine the flight path and logistics delivery task of the UAV swarm based on the final global optimal solution.

2. The method for optimizing the flight path of swarm intelligent logistics drones for urban low-altitude environments according to claim 1, characterized in that, Logistics and distribution parameters include the total weight of the delivered goods, the number of delivered goods, and the coordinates of the delivered goods.

3. The method for optimizing the flight path of swarm intelligent logistics drones for urban low-altitude environments according to claim 1, characterized in that, The speed update formula for drones is: ； in, For policy gradient, For the action value evaluation function, It is inertial weight. and For adaptive weights, For the parameters of the policy network, For the first One drone location. For the first The probability of detection by a radar. The current time value; The and The calculation formula is: ； in, is the base of the natural logarithm, used to calculate exponential functions. To represent the objective function For parameters The norm of the gradient, i.e., the magnitude of the gradient.

4. The method for optimizing the flight path of swarm intelligent logistics drones for urban low-altitude environments according to claim 3, characterized in that, The position update formula is: ； in, For the exponential decay exploration rate, It follows a Gaussian distribution. The maximum number of iterations, This is the attenuation coefficient.

5. A method for optimizing the flight path of swarm intelligent logistics drones for urban low-altitude environments according to any one of claims 1-4, characterized in that, The local calculation mechanism is as follows: ； in, For the scope of local calculation, It is a random vector that follows a normal distribution; The formula for the global optimal solution is: ； in, Indicates the state of the given point. Below, based on the current policy function Choose an action The expected value of the value.

6. The method for optimizing the flight path of swarm intelligent logistics drones for urban low-altitude environments according to claim 5, characterized in that, Environmental parameters include static obstacles, areas of complex electromagnetic interference, and overall airspace restrictions. The functional representation of a static obstacle is: ； in, Indicates at given coordinates The height value at that location, This indicates the total number of obstacles. It is the first The center coordinates of the obstacles Indicates the first The height of the obstacle and These are the attenuation amounts along the x-axis and y-axis, respectively; The complex electromagnetic interference region is represented by a hemispherical model, specifically as follows: ； in, Indicates the first A complex electromagnetic interference zone, ( ) is the first The center coordinates of a complex electromagnetic interference zone It is the first The coverage radius of a complex electromagnetic interference zone; The overall airspace restricted area is represented as follows: ; For the first Each airspace restricted region is modeled using a cylindrical region, specifically expressed as follows: ； in, Represents the spatial coordinates of the drone. The center of the cylinder's base is represented by r, and the radius is r. The drone's altitude range is... .

7. The method for optimizing the flight path of swarm intelligent logistics drones for urban low-altitude environments according to claim 1, characterized in that, The fitness value function is expressed as: ； in, Represents the fitness value. This represents the cost of track length. Indicates the cost of electromagnetic interference. This represents the sum of yaw angle costs. , and These are the weighting constants for the sum of track length cost, electromagnetic interference cost, and yaw angle cost, respectively.

8. The method for optimizing the flight path of swarm intelligent logistics drones for urban low-altitude environments according to claim 7, characterized in that, The cost of track length is expressed as: ； in, It is the first The coordinates of the interpolated track delivery node It represents the total number of track delivery nodes after interpolation.

9. A method for optimizing the flight path of swarm intelligent logistics drones for urban low-altitude environments according to claim 7, characterized in that, The cost of electromagnetic interference is expressed as: ； in, This indicates the total number of complex electromagnetic interference zones. positive integer, This indicates the total number of track delivery nodes. Indicates that the drone is in The probability of a complex electromagnetic interference zone being interfered with; Probability of being subjected to electromagnetic interference Represented as: ； in, and These represent the effective detection radius and the maximum detection radius, respectively. It is the first Center of a complex electromagnetic interference zone and the Each tracking delivery node The straight-line distance between them is calculated using the following formula: 。 10. A method for optimizing the flight path of swarm intelligent logistics drones for urban low-altitude environments according to claim 7, characterized in that, The sum of yaw angle costs is expressed as: ； in, Indicates the first The yaw angle cost of each track delivery node; The function formula is: ； in, Indicates the yaw angle. The calculation function is: ； in, and They are the first and the +1 segment of the delivery node vector, It is the dot product of vectors. and It is the magnitude of the vector. Indicates the maximum turning angle.

Images