Logistics trajectory optimization method and device for industrial chain and electronic equipment

Abstract

The embodiment of the invention discloses a logistics trajectory optimization method and device for an industrial chain and electronic equipment. A specific embodiment of the method comprises the following steps: carrying out no-fly zone detection on a flight zone of the flight logistics equipment; acquiring a flight state vector sequence; in response to the fact that the remaining flight range is larger than the preset distance value, executing the following trajectory optimization steps: determining predicted flight state features based on a trajectory optimization model and the no-fly zone information set; inputting the predicted flight state characteristics into a Sheng differential equation network to obtain optimized control parameters; controlling the flight logistics equipment according to the optimized control parameters, and determining a predicted flight state vector based on a flight dynamics equation; generating an updated flight state vector sequence according to the predicted flight state vector; and taking the updated flight state vector sequence as a flight state vector sequence to execute the trajectory optimization step again. According to the embodiment, the logistics transportation time can be shortened, and the flight energy consumption is reduced.

Description

Technical Field

[0001] The embodiments of the present disclosure relate to the field of computer technology, and in particular to a method, device, and electronic device for optimizing logistics trajectories in an industrial chain. Background Art

[0002] Avoiding no-fly zones is a core challenge in flight path planning for logistics transport aircraft. Currently, numerical methods are commonly used to optimize flight trajectories, calculating them by solving flight dynamics equations.

[0003] However, when using the above method to optimize the flight trajectory, a technical problem that often occurs is that the numerical method usually requires complex integral calculations and relies on manual experience to design the flight corridor, which consumes a lot of manpower and computing resources. In addition, due to the lack of a dynamic adjustment mechanism, it is difficult to achieve global optimal trajectory planning when circumventing no-fly zones, resulting in longer transportation time and increased flight energy consumption.

[0004] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure concept and therefore it may contain information that does not form the prior art that is already known in this country to a person of ordinary skill in the art. Summary of the Invention

[0005] The content of this disclosure is used to briefly introduce concepts that will be described in detail in the detailed description section below. The content of this disclosure is not intended to identify key features or essential features of the claimed technical solution, nor is it intended to limit the scope of the claimed technical solution.

[0006] Some embodiments of the present disclosure propose logistics trajectory optimization methods, devices and electronic devices for industrial chains to solve one or more of the technical problems mentioned in the above background technology section.

[0007] In a first aspect, some embodiments of the present disclosure provide a logistics trajectory optimization method for an industrial chain, the method comprising: performing a no-fly zone detection on the flight area of a flying logistics equipment to obtain a no-fly zone information set; obtaining a flight state vector sequence of the above-mentioned flying logistics equipment; in response to the remaining flight distance corresponding to the above-mentioned flying logistics equipment being greater than a preset distance value, executing the following trajectory optimization steps: based on a trajectory optimization model and the above-mentioned no-fly zone information set, determining predicted flight state characteristics corresponding to the flight state vector sequence, wherein the above-mentioned trajectory optimization model includes a long short-term memory network, a fully connected layer and a neural ordinary differential equation network; inputting the above-mentioned predicted flight state characteristics into the above-mentioned neural ordinary differential equation network to obtain optimized control parameters; controlling the above-mentioned flying logistics equipment according to the above-mentioned optimized control parameters, and determining the predicted flight state vector corresponding to the above-mentioned optimized control parameters based on a preset flight dynamics equation; generating an updated flight state vector sequence according to the above-mentioned predicted flight state vector and the above-mentioned flight state vector sequence; and executing the above-mentioned trajectory optimization steps again using the above-mentioned updated flight state vector sequence as the flight state vector sequence.

[0008] In a second aspect, some embodiments of the present disclosure provide a logistics trajectory optimization device for an industrial chain, the device comprising: a no-fly zone detection unit, configured to perform no-fly zone detection on the flight area of a flying logistics device to obtain a no-fly zone information set; an acquisition unit, configured to obtain a flight state vector sequence of the above-mentioned flying logistics device; an execution unit, configured to execute the following trajectory optimization steps in response to the remaining flight distance corresponding to the above-mentioned flying logistics device being greater than a preset distance value: determining the predicted flight state characteristics corresponding to the flight state vector sequence based on a trajectory optimization model and the above-mentioned no-fly zone information set, wherein the above-mentioned trajectory optimization model comprises a long short-term memory network, a fully connected layer and a neural ordinary differential equation network; inputting the above-mentioned predicted flight state characteristics into the above-mentioned neural ordinary differential equation network to obtain optimized control parameters; controlling the above-mentioned flying logistics device according to the above-mentioned optimized control parameters, and determining the predicted flight state vector corresponding to the above-mentioned optimized control parameters based on a preset flight dynamics equation; generating an updated flight state vector sequence according to the above-mentioned predicted flight state vector and the above-mentioned flight state vector sequence; a second execution unit, configured to execute the above-mentioned trajectory optimization step again using the above-mentioned updated flight state vector sequence as the flight state vector sequence.

[0009] In a third aspect, some embodiments of the present disclosure provide an electronic device comprising: one or more processors; a storage device on which one or more programs are stored, and when the one or more programs are executed by one or more processors, the one or more processors implement the method described in any implementation of the first aspect above.

[0010] In a fourth aspect, some embodiments of the present disclosure provide a computer-readable medium having a computer program stored thereon, wherein when the program is executed by a processor, the method described in any implementation of the first aspect is implemented.

[0011] The above-described embodiments of the present disclosure have the following beneficial effects: Through the logistics trajectory optimization method for an industrial chain, some embodiments of the present disclosure can reduce flight energy consumption. Specifically, the increased flight energy consumption is caused by the fact that numerical methods typically require complex integral calculations and rely on manual experience to design flight corridors, which consumes a lot of manpower and computing resources. Furthermore, the lack of a dynamic adjustment mechanism makes it difficult to achieve globally optimal trajectory planning when circumventing no-fly zones, resulting in extended transportation time. Based on this, the logistics trajectory optimization method for an industrial chain in some embodiments of the present disclosure first performs a no-fly zone detection on the flight area of a flying logistics device to obtain a no-fly zone information set. Second, a flight state vector sequence of the flying logistics device is obtained. Then, in response to the remaining flight range corresponding to the flying logistics device being greater than a preset distance value, the following trajectory optimization steps are performed: First, based on a trajectory optimization model and the no-fly zone information set, predicted flight state features corresponding to the flight state vector sequence are determined. The trajectory optimization model includes a long short-term memory network, a fully connected layer, and a neural ordinary differential equation network. Prediction using the trajectory optimization model trained through reinforcement learning can dynamically adjust the flight path. Reinforcement learning training enables the model to adjust the flight path based on real-time feedback, ensuring global optimal trajectory planning when circumventing no-fly zones, thereby reducing unnecessary flight time and energy consumption. Afterwards, the above-mentioned predicted flight state characteristics are input into the above-mentioned neural ordinary differential equation network to obtain optimized control parameters. In this way, more accurate optimized control parameters can be obtained, which can effectively adjust the control strategy of the flight equipment to ensure maximum energy efficiency under different flight conditions. Then, according to the above-mentioned optimized control parameters, the above-mentioned flight logistics equipment is controlled, and based on the preset flight dynamics equation, the predicted flight state vector corresponding to the above-mentioned optimized control parameters is determined. Then, based on the above-mentioned predicted flight state vector and the above-mentioned flight state vector sequence, an updated flight state vector sequence is generated. Finally, the above-mentioned trajectory optimization step is executed again using the above-mentioned updated flight state vector sequence as the flight state vector sequence. By dynamically updating the flight state vector sequence and re-executing the optimization step, the flight path can be adjusted in real time according to changes in the flight environment, avoiding the problem of decreased efficiency or increased energy consumption that may occur during the flight. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The above and other features, advantages, and aspects of the various embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and that components and elements are not necessarily drawn to scale.

[0013] Figure 1 is a flow chart of some embodiments of the method for optimizing logistics trajectories for an industrial chain according to the present disclosure; Figure 2 is an example diagram of a flyable area according to some embodiments of the logistics trajectory optimization method for an industrial chain disclosed herein; Figure 3 is a flight trajectory diagram according to some embodiments of the logistics trajectory optimization method for an industrial chain disclosed herein; Figure 4 is a schematic structural diagram of some embodiments of the logistics trajectory optimization device for the industrial chain according to the present disclosure; Figure 5 It is a structural diagram of an electronic device suitable for implementing some embodiments of the present disclosure. DETAILED DESCRIPTION

[0014] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein. On the contrary, these embodiments are provided to provide a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are for illustrative purposes only and are not intended to limit the scope of protection of the present disclosure.

[0015] It should also be noted that, for ease of description, only the parts related to the invention are shown in the drawings. In the absence of conflict, the embodiments and features in the embodiments of the present disclosure may be combined with each other.

[0016] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0017] It should be noted that the modifications of "one" and "multiple" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, they should be understood as "one or more".

[0018] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0019] The present disclosure will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments.

[0020] Figure 1 The process 100 of some embodiments of the method for optimizing logistics trajectory for an industrial chain according to the present disclosure is shown. The method for optimizing logistics trajectory for an industrial chain includes the following steps: Step 101: Perform a no-fly zone detection on the flight area of the flying logistics equipment to obtain a no-fly zone information set.

[0021] In some embodiments, the execution entity of the logistics trajectory optimization method for an industrial chain can perform no-fly zone detection on the flight area of flying logistics equipment to obtain a no-fly zone information set. The execution entity can be a computing device. The flying logistics equipment can be an aircraft used for logistics transportation, such as a logistics transport aircraft.

[0022] In some optional implementations of some embodiments, the execution entity performs a no-fly zone detection on the flight area of the flying logistics equipment to obtain a no-fly zone information set, which may include the following steps: The first step is to obtain flight status information corresponding to the flight logistics equipment. This flight status information may include flight start coordinates, flight end coordinates, and aircraft parameter information. The flight start coordinates may be the longitude, latitude, and altitude of the starting point of the flight logistics equipment's current mission. The flight end coordinates may be the longitude, latitude, and altitude of the destination of the mission. The aircraft parameter information may include the maximum horizontal flight altitude (absolute ceiling) corresponding to the flight logistics equipment.

[0023] The second step is to generate the flight area information corresponding to the above-mentioned flying logistics equipment based on the above-mentioned flight situation information. Among them, the above-mentioned flight area information can represent the cylindrical area between the above-mentioned flight starting coordinates and the above-mentioned flight terminal coordinates. In practice, first, the midpoint between the above-mentioned flight starting coordinates and the above-mentioned flight terminal coordinates can be determined as the center of the bottom surface of the above-mentioned cylindrical area. Secondly, the distance value between the above-mentioned flight starting coordinates and the above-mentioned flight terminal coordinates can be determined as the bottom surface diameter of the above-mentioned cylindrical area. Afterwards, the maximum horizontal flight height of the above-mentioned flying logistics equipment can be determined as the height of the above-mentioned cylindrical area.

[0024] In the third step, based on the aforementioned flight zone information, no-fly information in the preset no-fly information set that meets preset conditions is determined as no-fly zone information, thereby obtaining a no-fly zone information set. Each no-fly information in the no-fly information set can represent a no-fly zone. The no-fly information can include information such as the coordinates of the center of mass of the no-fly zone and the radius of the no-fly zone. The no-fly information set can be obtained from navigation notices issued by aviation authorities. In practice, no-fly information in the no-fly information set that overlaps with the cylindrical area can be determined as no-fly zone information, thereby obtaining a no-fly zone information set.

[0025] Step 102: Obtain the flight state vector sequence of the flying logistics equipment.

[0026] In some embodiments, the above-mentioned execution entity can obtain the flight state vector sequence of the above-mentioned flying logistics equipment. In practice, first, the flight data of the above-mentioned flying logistics equipment at the same moment can be obtained (including: flight starting altitude, flight starting speed, flight starting longitude, flight starting latitude, flight terminal longitude, flight terminal latitude, aircraft geocentric distance, aircraft longitude, aircraft latitude, flight speed, aircraft track angle and aircraft heading angle), and the above-mentioned various flight data can be determined as vectors to obtain a flight state vector. Secondly, the flight state vector of a time period when the above-mentioned flying logistics equipment performs a flight mission can be determined as a flight state vector sequence. The above-mentioned flight state vector sequence can include flight state vectors of 5 time steps, which is not specifically limited here.

[0027] Step 103: In response to the remaining flight distance of the flight logistics equipment being greater than a preset distance value, the following trajectory optimization steps are performed: Step 1031 : Determine the predicted flight state features corresponding to the flight state vector sequence based on the trajectory optimization model and the no-fly zone information set.

[0028] In some embodiments, the execution entity may determine the predicted flight state characteristics corresponding to the flight state vector sequence based on the trajectory optimization model and the no-fly zone information set. The remaining flight distance may be the distance between the flight logistics equipment and the flight destination coordinates. The preset distance value may be a length value. For example, the distance value may be 500m, which is not specifically limited here. The trajectory optimization model includes a long short-term memory network, a fully connected layer, and a neural ordinary differential equation network. The trajectory optimization model may be the following trajectory optimization model after reinforcement training. The long short-term memory network may be a long short-term memory network trained through reinforcement learning, such as an LSTM network. The fully connected layer may be a fully connected layer (FC) trained through reinforcement learning. The neural ordinary differential equation network may be a neural ordinary differential equation network (Neural ODE) trained through reinforcement learning.

[0029] In practice, the current longitude, latitude, and altitude of the flight logistics device can be first obtained. Then, based on the current longitude, latitude, and altitude of the flight logistics device, the distance between the flight logistics device and the flight terminal coordinates can be determined to obtain the remaining flight range. Next, if the remaining flight range corresponding to the flight logistics device is greater than a preset distance value, the trajectory optimization step is executed.

[0030] In some optional implementations of some embodiments, the execution entity may determine the predicted flight state features corresponding to the flight state vector sequence based on the trajectory optimization model and the no-fly zone information set, which may include the following steps: The first step is to extract temporal features from the flight state vector sequence based on the LSTM network and the no-fly zone information set, generating a temporal feature vector set. In practice, the flight state vector sequence and the no-fly zone information set can be input into the reinforcement learning-trained LSTM network to generate a temporal feature vector set. Each flight state feature vector in the flight state vector sequence can correspond to a temporal feature vector.

[0031] In the second step, the time series feature vector corresponding to the last one in the above time series feature vector set is determined as the implicit feature vector.

[0032] The third step is to perform feature mapping on the latent feature vector based on the fully connected layer to obtain predicted flight state features. These predicted flight state features can represent the flight state at a future time. These predicted flight state features can be in vector form. In practice, the latent feature vector can be input into the fully connected layer to obtain the predicted flight state features.

[0033] Step 1032: Input the predicted flight state characteristics into the Neural Ordinary Differential Equation network to obtain the optimized control parameters.

[0034] In some embodiments, the execution entity may input the predicted flight state characteristics into the neural ordinary differential equation network to obtain optimized control parameters. The optimized control parameters may represent control parameters for the next time step. The optimized control parameters may include the aircraft's roll angle and the aircraft's angle of attack.

[0035] Step 1033 , controlling the flight logistics equipment according to the optimized control parameters, and determining the predicted flight state vector corresponding to the optimized control parameters based on the preset flight dynamics equation.

[0036] In some embodiments, the execution entity can control the flight logistics equipment according to the optimized control parameters, and determine the predicted flight state vector corresponding to the optimized control parameters based on the preset flight dynamics equations. The predicted flight state vector can represent the flight state of the flight logistics equipment in the next time step. The numerical values in the predicted flight state vector can respectively represent the flight starting altitude, flight starting speed, flight starting longitude, flight starting latitude, flight terminal longitude, flight terminal latitude, aircraft geocentric distance in the next time step, aircraft longitude in the next time step, aircraft latitude in the next time step, flight speed in the next time step, aircraft track angle in the next time step, and aircraft heading angle in the next time step. In practice, first, the flight logistics equipment can be controlled according to the optimized control parameters so that the bank angle and angle of attack of the flight logistics equipment are equal to the aircraft bank angle and aircraft angle of attack contained in the optimized control parameters. Here, the flight logistics equipment can include a control component. The control component can adjust the aircraft bank angle and aircraft angle of attack of the flight logistics equipment according to the optimized control parameters. Afterwards, the predicted flight state vector corresponding to the optimized control parameters can be determined based on the following flight dynamics equations and the value of the last flight state vector in the flight state vector sequence: .

[0037] in, , , , , , They represent the aircraft's distance from the center of the earth, aircraft's longitude, aircraft's latitude, flight speed, aircraft's track angle, and aircraft's heading angle at the previous time step, respectively. , , , , , They represent the aircraft's distance from the center of the earth, aircraft's longitude, aircraft's latitude, flight speed, aircraft's track angle, and aircraft's heading angle at the next time step respectively. , They represent the lift and drag of the aircraft respectively, and can be determined based on the above-mentioned aircraft angle of attack, the flight speed of the previous time step and the formula in sub-step 5 of the following optional steps. Indicates the weight of the above-mentioned flight logistics equipment. Indicates the aircraft's roll angle. Indicates the time step, which can be 0.1s. represents the reference gravitational acceleration, which may be the gravitational acceleration of the earth's surface.

[0038] Step 1034: Generate an updated flight state vector sequence based on the predicted flight state vector and the flight state vector sequence.

[0039] In some embodiments, the execution entity may generate an updated flight state vector sequence based on the predicted flight state vector and the flight state vector sequence. In practice, the updated flight state vector sequence may be determined as a sequence consisting of each flight state vector in the flight state vector sequence other than the first flight state vector and the predicted flight state vector.

[0040] Step 104 : Using the updated flight state vector sequence as the flight state vector sequence, the trajectory optimization step is performed again.

[0041] In some embodiments, the execution entity may use the updated flight state vector sequence as a flight state vector sequence to perform the trajectory optimization step again.

[0042] Optionally, the trajectory optimization model can be trained by the following steps: The first step is to generate a sample training dataset using numerical methods. In practice, this can be generated using numerical methods based on the aircraft's flight dynamics equations. Each sample training data set in this sample training dataset can include a sequence of sample flight vectors and a sequence of true value vectors. Each sample flight vector sequence corresponds to a sequence of true value vectors. These sample flight vector sequences and the corresponding true value vector sequences can collectively represent the aircraft's flight state over a period of time.

[0043] The second step is to pre-train a preset initial trajectory optimization model based on the sample training dataset to obtain a pre-trained trajectory optimization model. The initial trajectory optimization model includes an initial long short-term memory network, an initial fully connected layer, and an initial neural ordinary differential equation network. The initial long short-term memory network can be an LSTM network. The initial fully connected layer can be an untrained fully connected layer (FC). The initial neural ordinary differential equation network can be a Neural ODE network. In practice, the sample flight vector sequence corresponding to each sample training data in the sample training dataset can be used as the input of the initial trajectory optimization model, and the true value vector sequence corresponding to each sample training data in the sample training dataset can be used as the expected output of the initial trajectory optimization model. The initial trajectory optimization model is then trained to obtain a pre-trained trajectory optimization model.

[0044] The third step is to perform reinforcement learning training on the pre-trained trajectory optimization model to obtain the trajectory optimization model. In practice, a reinforcement learning model training method can be used to perform reinforcement learning training on the pre-trained trajectory optimization model to obtain the trajectory optimization model.

[0045] In practice, when generating sample training data, a common technical problem is that when using integral algorithms such as the prediction-correction algorithm or the pseudospectral method to solve the numerical value of the aircraft's roll angle, it often consumes a lot of computing power and time, resulting in a waste of computing resources.

[0046] Optionally, the execution subject generates a sample training data set based on a numerical method, which may include the following steps: The first step is to obtain an aircraft trajectory dataset. Each aircraft trajectory data in the aircraft trajectory dataset may include a sample aircraft state quantity sequence. The elements in the sample aircraft state quantity sequence may be arranged in chronological order. Each sample aircraft state quantity in the sample aircraft state quantity sequence may include a sample flight starting altitude, a sample flight starting speed, a sample flight starting longitude, a sample flight starting latitude, a sample flight terminal longitude, a sample flight terminal latitude, a sample aircraft geocentric distance, a sample aircraft longitude, a sample aircraft latitude, a sample flight speed, a sample aircraft track angle, and a sample aircraft heading angle.

[0047] In the second step, for each aircraft trajectory data in the above aircraft trajectory dataset, the following steps are performed to generate sample training data to obtain a sample training dataset: The first sub-step is to perform the following steps on each sample aircraft state quantity in the sample aircraft state quantity sequence included in the aircraft trajectory data: Sub-step 1: Perform dimensionless processing on the sample aircraft earth center distance and sample flight speed included in the sample aircraft state quantities to obtain dimensionless altitude and dimensionless speed. In practice, the following formula can be used to perform dimensionless processing on the sample aircraft earth center distance and sample flight speed included in the sample aircraft state quantities to obtain dimensionless altitude and dimensionless speed.

[0048] .

[0049] Where z is the dimensionless altitude and r is the distance from the Earth's center to the sample aircraft. represents the reference length in the dimensionless processing, which can be the radius of the Earth. u represents the dimensionless velocity. V represents the sample flight velocity. Represents the reference velocity in dimensionless processing. Represents the reference gravitational acceleration. Represents dimensionless time, and t represents the length of the time step during the flight of the aircraft, for example 0.1s.

[0050] Sub-step 2: Determine the sample aircraft energy corresponding to the dimensionless altitude and the dimensionless velocity. In practice, the sample aircraft energy corresponding to the dimensionless altitude and the dimensionless velocity can be determined using the following formula.

[0051] .

[0052] in, Represents the sample aircraft energy. When z is the sample flight terminal altitude and u is the sample flight terminal velocity, the result of the above formula can represent the sample aircraft terminal energy. Here, the sample aircraft state quantity corresponds to a flight terminal. The flight terminal is the end point of the current flight mission of the sample aircraft. The flight terminal may correspond to a sample flight terminal longitude, a sample flight terminal latitude, a sample flight terminal altitude, and a sample flight terminal speed. The sample flight terminal altitude may be the altitude of the end point of the current flight mission of the sample aircraft. The sample flight terminal speed may be a preset speed corresponding to the end point of the current flight mission.

[0053] Sub-step three: Calculate the aircraft's atmospheric density based on the sample aircraft's distance from the Earth's center, corresponding to the sample aircraft's state. In practice, the difference between the sample aircraft's distance from the Earth's center and the Earth's radius can be used as the sample aircraft's altitude. Subsequently, the aircraft's atmospheric density corresponding to the sample aircraft's state can be determined using a pre-set altitude-to-atmospheric-density mapping table.

[0054] Sub-step 4: Determine the sample aircraft angle of attack corresponding to the sample flight speed included in the sample aircraft state quantity based on the preset maximum angle of attack, the preset maximum lift-to-drag angle of attack, the preset maximum angle of attack velocity, the preset maximum lift-to-drag angle of attack velocity, and the preset terminal velocity. The sample aircraft angle of attack can represent the angle between the incoming wind and the wing chord, with a positive angle of attack when the wing is pointing upward and a negative angle of attack when the wing is pointing downward. In practice, the sample aircraft angle of attack corresponding to the sample aircraft state quantity can be determined using the following formula.

[0055] .

[0056] in, represents the sample vehicle angle of attack. Indicates the preset maximum angle of attack. Indicates the angle of attack for maximum lift-to-drag ratio. Indicates the lower bound of the maximum angle of attack velocity, which can be a pre-set velocity value at the maximum angle of attack. Indicates the upper limit of the maximum lift-to-drag ratio angle of attack speed, which can be a pre-set speed value at the maximum lift-to-drag ratio angle of attack. represents the sample flight terminal velocity.

[0057] In practice, the maximum angle of attack may be 20 degrees. The lower limit of the maximum angular velocity of attack may be 5000 meters per second. The maximum lift-to-drag ratio angle of attack may be 12 degrees. The upper limit of the maximum lift-to-drag ratio angle of attack may be 2500 meters per second.

[0058] Sub-step 5: Generate the lift and drag of the aircraft based on the atmospheric density of the aircraft, the sample aircraft angle of attack, the sample flight speed included in the sample aircraft state, and the preset aircraft characteristic area. In practice, the lift and drag of the aircraft can be generated using the following formula: .

[0059] Where L represents the lift force on the aircraft, and D represents the drag force on the aircraft. represents the lift coefficient. Represents the drag coefficient. Indicates the density of the atmosphere in which the aircraft is located. All represent coefficients and are not specifically limited here. e represents a natural constant. S represents the characteristic area of the aircraft, which represents the wind-exposed area of the sample aircraft (m 2 ).

[0060] Sub-step 6: Determine the distance to be flown corresponding to the sample aircraft state quantity based on the sample flight terminal longitude and sample flight terminal latitude included in the sample aircraft state quantity. The distance to be flown can be the arc formed by the current aircraft longitude and latitude and the terminal longitude and latitude. In practice, the distance to be flown corresponding to the sample aircraft state quantity can be determined using the following formula: .

[0061] in, Indicates the distance to be flown corresponding to the above sample aircraft state quantity. Indicates the sample aircraft longitude. Indicates the sample flight terminal longitude. Indicates the latitude of the sample aircraft. Indicates the sample flight terminal latitude.

[0062] Sub-step seven: Generate a sample aircraft roll angle corresponding to the sample aircraft state quantity based on the distance to be flown, the lift force on the aircraft, the drag force on the aircraft, the sample aircraft energy, and the dimensionless altitude. The sample aircraft roll angle may represent the angle of rotation about the aircraft's longitudinal axis during flight. In practice, the sample aircraft roll angle corresponding to the sample aircraft state quantity may be generated using the following formula.

[0063] .

[0064] in, Indicates the sample vehicle roll angle.

[0065] Secondly, the azimuth of the above sample aircraft can be determined by the following formula: .

[0066] in, Indicates the azimuth.

[0067] Next, when the difference between the azimuth angle and the heading angle corresponding to the sample aircraft equals a preset angle value, indicating that the sample aircraft has triggered a lateral motion boundary, the sign of the sample aircraft's roll angle can be flipped. That is, the sign of the sample aircraft's roll angle at this time step can be opposite to the sign of the sample aircraft's roll angle at the previous time step. The preset angle value can be a preset number of degrees and is not specifically limited here. The sample aircraft's roll angle at the first time step can be determined using an integral algorithm such as a predictor-corrector algorithm or pseudospectral method.

[0068] In practice, the flight process is in the process of energy decline. The bank angle value is solved by using the energy and the distance to be flown so that the remaining range is equal to the distance to be flown. The above remaining range can represent the actual distance that the sample aircraft can still fly at the current energy.

[0069] Sub-step eight: determining the sample aircraft angle of attack and the sample aircraft roll angle as sample aircraft control quantities corresponding to the sample aircraft state quantities. The sample aircraft control quantities may include the sample aircraft roll angle and the sample aircraft angle of attack.

[0070] In practice, the friction between the wings and the airflow during gliding generates aerodynamic heat. To ensure the overall safety of the fuselage, the maximum heat flux rate needs to be limited. In addition, lift, drag, and gravity act on the fuselage simultaneously. To prevent the combined force from exceeding the upper limit of the material strength, the maximum overload and dynamic pressure need to be limited. The above-mentioned dimensionless height and dimensionless speed need to meet the following constraints: .

[0071] in, Represents the heat flow rate. Indicates the maximum heat flow rate (kw / m 2 ). Indicates dynamic pressure. Indicates the maximum limiting dynamic pressure (kPa). Indicates overload. Indicates the maximum limit overload.

[0072] The second sub-step is to combine the above-mentioned sample aircraft state quantity sequence and the sample aircraft control quantity sequence corresponding to each determined sample aircraft control quantity into sample training data.

[0073] The first and second steps and their related content, as an inventive feature of the embodiments of the present disclosure, address the aforementioned technical issue of "waste of computing resources." This issue is often caused by the following factors: Using integral algorithms such as the predictive-correction algorithm or the pseudospectral method to calculate the aircraft's roll angle often consumes a significant amount of computational effort and time. Addressing this issue can reduce flight duration and energy consumption. To achieve this, first, for each sample aircraft state quantity in the sample aircraft state quantity sequence, the altitude, velocity, and other parameters included in the sample aircraft state quantity can be dimensionlessly processed to reduce workload. Then, the sample aircraft's angle of attack, lift, and drag can be determined according to a preset formula. Subsequently, the sample aircraft's energy at its current position can be determined based on the dimensionless altitude and velocity corresponding to the current sample aircraft. This allows for an energy-based analytical prediction of the remaining range. Next, the arc between the longitude and latitude of the current sample aircraft state quantity and the longitude and latitude of the terminal can be determined as the distance to be flown. Secondly, during flight, due to the continuous energy consumption phase, the optimal bank angle that matches the remaining range with the distance to be flown can be derived based on the constraint relationship between the current remaining energy and the distance to be flown, combined with the lift and drag acting on the aircraft. Compared to integral algorithms such as the predictive correction algorithm and the pseudospectral method, the bank angle calculated using the energy method significantly reduces the amount of computation and computational time, thereby reducing the waste of computing resources.

[0074] Optionally, the execution subject pre-trains a preset initial trajectory optimization model based on the sample training data set to obtain a pre-trained trajectory optimization model, which may include the following steps: The first step is to preprocess each sample training data in the sample training data set to generate preprocessed sample training data, thereby obtaining a preprocessed sample training data set. Each preprocessed sample training data in the preprocessed sample training data set includes a preprocessed sample state quantity sequence and a preprocessed control quantity true value. In practice, first, for each sample training data in the sample training data set, a sequence consisting of a plurality of consecutive sample aircraft state quantities in the sample aircraft state quantity sequence corresponding to the sample training data can be determined as the preprocessed sample state quantity sequence, and the sample aircraft control quantity of the next time step corresponding to the plurality of consecutive sample aircraft state quantities can be determined as the preprocessed control quantity true value.

[0075] The second step is to input the preprocessed sample state quantity sequence included in each preprocessed sample training data set in the above-mentioned preprocessed sample training data set into the preset initial trajectory optimization model to obtain a set of predicted control quantities. Each predicted control quantity in the above-mentioned predicted control quantity set can represent the aircraft control parameters for one time step in the future. The above-mentioned predicted control quantities can represent the roll angle and angle of attack of the aircraft. In practice, each preprocessed sample training data set in the above-mentioned preprocessed sample training data set can be input into the initial long short-term memory network and the initial fully connected layer included in the above-mentioned initial trajectory optimization model to obtain a sample prediction feature set. Afterwards, each sample prediction feature in the above-mentioned sample prediction feature set is input into the initial neural ordinary differential equation network included in the above-mentioned initial trajectory optimization model to obtain a set of sample control parameter sequences.

[0076] In practice, the initial neural ordinary differential equation network may include a preset fitting algorithm. The fitting algorithm may be an Euler algorithm or a Runge-Kutta algorithm. The fitting algorithm may approximate the original function of the aircraft's roll angle over time through an iterative method. The original function of the aircraft's roll angle over time is shown in the following formula: .

[0077] in, The primitive function that represents the change of the aircraft's roll angle with time t. represents the roll angle differential. represents the derivative of the roll angle over time, as represented by the neural network. The neural network may be a fully connected network. p represents a trainable parameter of the neural network.

[0078] In practice, after the original function is calculated by the above fitting algorithm, the roll angle value at the required time point can be arbitrarily specified to achieve continuous prediction.

[0079] The third step is to generate a total loss value based on the loss value between each predicted control variable in the predicted control variable set and the true value of the preprocessed control variable included in the corresponding preprocessed sample training data in the preprocessed sample training data set. In practice, a preset loss algorithm can be used to first determine the loss value between each predicted control variable in the predicted control variable set and the true value of the preprocessed control variable included in the corresponding preprocessed sample training data in the preprocessed sample training data set. The loss algorithm can be a mean squared error (MSE) algorithm. The total loss value can then be determined by summing the determined loss values.

[0080] In the fourth step, the parameters of the initial trajectory optimization model are updated by backpropagation according to the total loss value to obtain the pre-trained trajectory optimization model.

[0081] In practice, when encountering a no-fly zone, trajectory optimization methods based on numerical calculations often need to solve a highly complex and mutually coupled set of differential equations of the dynamic system, which is computationally intensive and time-consuming. At the same time, it is difficult to effectively avoid the no-fly zone while shortening the detour path as much as possible, resulting in a longer flight route for the flight equipment, thereby increasing the flight time and energy consumption during the flight.

[0082] Optionally, the execution subject performs reinforcement learning training on the pre-trained trajectory optimization model to obtain the trajectory optimization model, which may include the following steps: The first step is to determine the sample flight starting point and sample flight endpoint corresponding to the sample aircraft. The sample aircraft may be a flight simulator in a simulation environment. The sample flight starting point may be the starting point of the sample aircraft's flight mission and may be characterized by longitude, latitude, and altitude. The sample flight endpoint may be the end point of the sample aircraft's flight mission and may be characterized by longitude, latitude, and altitude.

[0083] The second step is to generate a sample no-fly zone set based on the above-mentioned sample flight starting point and the above-mentioned sample flight end point. The above-mentioned sample no-fly zone set may include at least one sample no-fly zone, and the above-mentioned sample no-fly zone may be characterized by the latitude, longitude and radius of the centroid. In practice, first, at least one coordinate point may be randomly generated in the boundary area between the above-mentioned sample flight starting point and the above-mentioned sample flight end point. Each coordinate point in the above-mentioned at least one coordinate point corresponds to a radius. Afterwards, the sample no-fly zone may be determined based on each coordinate point in the above-mentioned at least one coordinate point and its corresponding radius to obtain a sample no-fly zone set. Here, the boundary area between the above-mentioned sample flight starting point and the above-mentioned sample flight end point may be a circular area in a two-dimensional bird's-eye view with the line segment between the above-mentioned sample flight starting point and the above-mentioned sample flight end point as the diameter.

[0084] The third step is to perform the following repeated training steps based on the preset rounds: The first sub-step involves generating a sequence of sample flight state vectors based on the sample flight starting point and the sample flight endpoint. The rounds may refer to a pre-set number of repeated training cycles. The pre-set number of rounds may be 200, but this is not specifically limited here. The sequence of sample flight state vectors may be a sequence of sample flight state vectors from multiple consecutive time steps. The values in each sample flight state vector in the sequence of sample flight state vectors may represent the sample flight starting point altitude, sample flight starting point speed, sample flight starting point longitude, sample flight starting point latitude, sample flight endpoint longitude, sample flight endpoint latitude, sample aircraft geocentric distance, sample aircraft longitude, sample aircraft latitude, sample flight speed, sample aircraft track angle, and sample aircraft heading angle. In practice, the initial sample flight state vector may be determined. Specifically, the sample flight starting point altitude, sample flight starting point longitude, sample flight starting point latitude, sample flight endpoint longitude, and sample flight endpoint latitude corresponding to the initial sample flight state vector may be determined based on the longitude, latitude, and altitude corresponding to the sample flight starting point and the longitude, latitude, and altitude corresponding to the sample flight endpoint. Secondly, the sample aircraft geocentric distance, sample aircraft longitude and sample aircraft latitude corresponding to the above-mentioned initial sample flight state vector can be the same as the values of the sample flight starting point altitude, sample flight starting point longitude and sample flight starting point latitude. Then, the angle between the line segment between the above-mentioned sample flight starting point and the above-mentioned sample flight end point and the line segment in the north direction on the ground plane can be determined as the sample aircraft track angle corresponding to the above-mentioned initial sample flight state vector. The value of the sample aircraft heading angle corresponding to the above-mentioned initial sample flight state vector can be the same as the value of the above-mentioned sample aircraft track angle. The sample flight starting point speed corresponding to the above-mentioned initial sample flight state vector can be a preset speed value. The value of the sample flight speed corresponding to the above-mentioned initial sample flight state vector can be the same as the value of the sample flight starting point speed. Afterwards, a sample flight state vector sequence can be generated based on the above-mentioned initial sample flight state vector. Specifically, the above-mentioned initial sample flight state vector can be copied to obtain multiple sample flight state vectors to generate a sample flight state vector sequence.

[0085] In a second sub-step, in response to determining that the last sample flight state vector in the sequence of sample flight state vectors does not meet a preset termination condition, the following enhanced training steps are performed based on the pre-trained trajectory optimization model: Sub-step 1: Determine the sample prediction control quantity corresponding to the sample flight state vector sequence based on the pre-trained trajectory optimization model. The termination conditions may include no-fly constraints, boundary constraints, flyable area constraints, and normal termination conditions. The no-fly constraints may be that the longitude and latitude corresponding to the sample aircraft (i.e., the longitude and latitude of the sample aircraft corresponding to the sample flight state vector) are not within each sample no-fly area in the sample no-fly area set, which can be expressed by the following formula: .

[0086] in, Indicates the latitude of the centroid of the no-fly zone. Indicates the longitude of the centroid of the no-fly zone. Indicates the radius of the no-fly zone.

[0087] The above-mentioned boundary constraint condition may be that the longitude and latitude corresponding to the above-mentioned sample aircraft cannot be outside the above-mentioned boundary area. The above-mentioned flyable area constraint condition may be that the longitude and latitude corresponding to the above-mentioned sample aircraft cannot be outside the preset flyable area. The above-mentioned flyable area may be the area between the line segment between the above-mentioned sample flight starting point and the above-mentioned sample flight end point and the preset parallel line. Specifically, the line segment between the above-mentioned sample flight starting point and the above-mentioned sample flight end point may be determined as the target line segment. The above-mentioned preset parallel line may be a line segment that is parallel to the above-mentioned target line segment and whose distance from each sample no-fly area in the above-mentioned sample no-fly area set is higher than a preset distance threshold. The above-mentioned distance threshold may be 1 km, which is not specifically limited here. The example diagram corresponding to the above-mentioned flyable area is as follows: Figure 2 As shown, the area between the two parallel lines is the flyable area. The normal termination condition can be that the distance between the sample aircraft and the sample flight endpoint is less than a preset termination distance threshold. The termination distance threshold can be 0.5 km, which is not specifically limited here.

[0088] In practice, first, for each sample flight state vector in the above-mentioned sample flight state vector sequence, the above-mentioned sample flight state vector and each no-fly zone in the above-mentioned sample no-fly zone set can be combined into a state space vector to generate a state space vector sequence. Particularly, each state space vector in the above-mentioned state space vector sequence can include a sample flight state vector, the center of mass coordinates and the radius of each no-fly zone. Afterwards, the above-mentioned state space vector sequence can be input into the above-mentioned pre-trained trajectory optimization model to obtain a sample prediction control quantity. Particularly, the above-mentioned sample prediction control quantity can include a predicted roll angle and a predicted angle of attack. Here, the above-mentioned end condition can also include a roll angle constraint condition. The above-mentioned roll angle constraint condition can limit the range of the roll angle to between -50 degrees and 50 degrees.

[0089] Sub-step 2: Determine the next-moment flight state vector corresponding to the sample predicted control variable based on the aforementioned flight dynamics equations. In practice, the next-moment flight state vector corresponding to the sample predicted control variable can be determined based on the aforementioned flight dynamics equations using the content of step 1033. This will not be further elaborated here.

[0090] Sub-step three: Determine the reward value corresponding to the next moment flight state vector, and update the parameters of the pre-trained trajectory optimization model based on the reward value to obtain an updated trajectory optimization model. The reward value can be the sum of the distance reward, the end reward, and the time reward. The distance reward can be expressed by the following formula: .

[0091] in, Indicates distance bonus. Indicates the distance to fly. The distance to be flown may be the distance between the sample aircraft and the sample flight destination. The total flight distance may be the distance between the sample flight start point and the sample flight destination.

[0092] The end reward may be a preset end reward value determined as the end reward when the distance between the sample aircraft and the sample flight endpoint is less than the end distance threshold. The end reward value may be a positive number and is not specifically limited herein.

[0093] The above time bonus can be expressed by the following formula: .

[0094] in, Indicates time reward. Represents the time reward coefficient. Represents the time step corresponding to the flight state vector at the next moment. Represents the preset maximum flight time step. This maximum flight time step can be 15,000 steps. The time bonus coefficient can be a natural number, and its specific value can be determined through cross-validation experiments and is not specifically limited here. This time bonus can prevent the reinforcement learning model from wasting time aiming to obtain a distance reward, thereby reducing training time.

[0095] In practice, first, the distance to be flown between the sample aircraft and the sample flight endpoint can be generated based on the sample aircraft's geocentric distance, longitude, and latitude corresponding to the next-moment flight state vector, and the longitude, latitude, and altitude corresponding to the sample flight endpoint. Then, a distance reward can be obtained based on the formula for generating the distance reward. Secondly, a time reward can be generated based on the time step corresponding to the next-moment flight state vector, using the formula for generating the time reward. Next, when the distance to be flown is less than the end distance threshold, the end reward value is determined as the end reward; otherwise, the end reward is 0. Then, the sum of the distance reward, time reward, and end reward can be determined as the reward value. Finally, the parameters of the pre-trained trajectory optimization model can be updated using a preset parameter update technique to obtain an updated trajectory optimization model. Here, the parameter update technique can be a value function method (Q-learning) or a policy gradient method (REINFORCE).

[0096] Sub-step 4: Generate a second flight state vector sequence based on the next-moment flight state vector and the sample flight state vector sequence. In practice, the second flight state vector sequence can be determined as a sequence consisting of each sample flight state vector in the sample flight state vector sequence other than the first sample flight state vector and the next-moment flight state vector.

[0097] Sub-step five, in response to determining that the flight state vector at the next moment does not satisfy the above-mentioned termination condition, the above-mentioned updated trajectory optimization model is used as the pre-trained trajectory optimization model, and the above-mentioned second flight state vector sequence is used as the sample flight state vector sequence, and the above-mentioned reinforcement training step is performed again. In practice, first, it is possible to determine whether the above-mentioned flight state vector at the next moment satisfies the above-mentioned termination condition based on the sample aircraft geocentric distance, sample aircraft longitude, and sample aircraft latitude included in the above-mentioned flight state vector at the next moment. Secondly, when the above-mentioned flight state vector at the next moment does not satisfy the above-mentioned termination condition, the above-mentioned updated trajectory optimization model can be used as the pre-trained trajectory optimization model, and the above-mentioned second flight state vector sequence can be used as the sample flight state vector sequence, and the above-mentioned reinforcement training step can be performed again.

[0098] The third sub-step is to terminate the intensive training step and re-execute the repeated training step in response to determining that the next-moment flight state vector satisfies the termination condition and the current round is not equal to the preset round. In practice, when the next-moment flight state vector satisfies the termination condition and the current round of training is not equal to the preset round, the round corresponding to the current intensive training step may be terminated. Thereafter, the repeated training step may be re-executed (i.e., a new round of training).

[0099] The fourth step is to determine the obtained updated trajectory optimization model as the trajectory optimization model. In practice, the updated trajectory optimization model obtained in the last round of training can be determined as the trajectory optimization model.

[0100] Optionally, the above execution entity may further perform the following steps: The first step is to obtain flight endpoint information, no-fly zone information, and at least one flight starting point information. Each of the at least one flight starting point information may include the starting longitude, starting latitude, and starting altitude of the flight starting point. Each of the at least one flight starting point information may be a randomly sampled coordinate point. The flight endpoint information may include the destination longitude, destination latitude, and destination altitude of the flight destination. The no-fly zone information may include the centroid longitude, centroid latitude, and no-fly radius of the no-fly zone.

[0101] In the second step, for each piece of the at least one piece of flight starting point information, perform the following steps: Based on the above trajectory optimization model, a trajectory point sequence is generated according to the above flight starting point information, the above flight endpoint information and the above no-fly zone information. In practice, first, a flight state vector sequence can be generated according to the above flight starting point information and the above flight endpoint information in the manner of generating a sample flight state vector sequence in the first sub-step of the above optional steps. No further details will be given here. Secondly, according to the operation in the above step 103, a trajectory optimization step can be performed based on the above trajectory optimization model and the above no-fly zone information. For each of the predicted flight state vectors determined in the above trajectory optimization step, a trajectory point can be generated according to the corresponding aircraft longitude and aircraft latitude in the above predicted flight state vector. Finally, the sequence composed of the above-generated trajectory points can be determined as a trajectory point sequence.

[0102] The third step is to render each of the at least one trajectory point sequence generated according to the preset flight map to obtain a flight trajectory diagram. The flight map may be a bird's-eye view covering the flight destination and each of the at least one flight starting point. In practice, for each of the at least one trajectory point sequence generated, the trajectory points may be drawn on the flight map according to the longitude and latitude of each trajectory point in the trajectory point sequence to obtain a flight trajectory diagram. The flight trajectory diagram is as follows: Figure 3 As shown in the (a) sub-figure in Figure 3 In the figure, the trajectories in subgraph (b) are obtained by the numerical algorithm. It can be seen that compared with the flight trajectory obtained by the numerical algorithm, the flight trajectory obtained by the trajectory optimization model is shorter, thereby reducing energy consumption during the flight.

[0103] The above-mentioned optional steps and their related content, as an inventive feature of the embodiments of the present disclosure, address the aforementioned technical issue of "increased flight time and energy consumption during flight." The factors contributing to this technical issue are often as follows: When encountering a no-fly zone, numerically-based trajectory optimization methods often require solving a highly complex and interconnected set of differential equations for the dynamical system, which is computationally intensive and time-consuming. Furthermore, it is difficult to effectively avoid the no-fly zone while minimizing the detour path, resulting in a longer flight path for the aircraft. Addressing these factors can reduce flight time and energy consumption. To achieve this, sample flight start and end points can be randomly determined, and no-fly zones can be randomly generated. This improves the effectiveness of reinforcement learning training. After generating the no-fly zone coordinates, multiple rounds of reinforcement training are repeated, based on preset reward values and termination conditions. This allows the trajectory optimization model to generate the optimal trajectory while maintaining airborne constraints, reducing flight time. Reinforcement learning training utilizes a pre-trained trajectory optimization model. Leveraging a combination of long short-term memory networks and neural ordinary differential equations, this architecture can more effectively process the continuous changes in time series data during flight and make real-time decisions, enabling the generated trajectory to better adapt to the environment and effectively avoid no-fly zones. Finally, by plotting the flight trajectory determined by the trajectory optimization model, a flight trajectory graph is generated. The trajectory optimization model effectively avoids no-fly zones in all flight trajectories generated based on different flight starting points, and the resulting paths are shorter, effectively reducing flight time and energy consumption.

[0104] The above-described embodiments of the present disclosure have the following beneficial effects: Through the logistics trajectory optimization method for an industrial chain, some embodiments of the present disclosure can reduce flight energy consumption. Specifically, the increased flight energy consumption is caused by the fact that numerical methods typically require complex integral calculations and rely on manual experience to design flight corridors, which consumes a lot of manpower and computing resources. Furthermore, the lack of a dynamic adjustment mechanism makes it difficult to achieve globally optimal trajectory planning when circumventing no-fly zones, resulting in extended transportation time. Based on this, the logistics trajectory optimization method for an industrial chain in some embodiments of the present disclosure first performs a no-fly zone detection on the flight area of a flying logistics device to obtain a no-fly zone information set. Second, a flight state vector sequence of the flying logistics device is obtained. Then, in response to the remaining flight range corresponding to the flying logistics device being greater than a preset distance value, the following trajectory optimization steps are performed: First, based on a trajectory optimization model and the no-fly zone information set, predicted flight state features corresponding to the flight state vector sequence are determined. The trajectory optimization model includes a long short-term memory network, a fully connected layer, and a neural ordinary differential equation network. Prediction using the trajectory optimization model trained through reinforcement learning can dynamically adjust the flight path. Reinforcement learning training enables the model to adjust the flight path based on real-time feedback, ensuring the global optimal trajectory planning when circumventing the no-fly zone, thereby reducing unnecessary flight time and energy consumption. Afterwards, the above-mentioned predicted flight state characteristics are input into the above-mentioned neural ordinary differential equation network to obtain the optimized control parameters. In this way, more accurate optimized control parameters can be obtained, which can effectively adjust the control strategy of the flight equipment to ensure that the energy efficiency is maximized under different flight conditions. Then, according to the above-mentioned optimized control parameters, the above-mentioned flight logistics equipment is controlled, and based on the preset flight dynamics equation, the predicted flight state vector corresponding to the above-mentioned optimized control parameters is determined. Finally, based on the above-mentioned predicted flight state vector and the above-mentioned flight state vector sequence, an updated flight state vector sequence is generated, and the above-mentioned trajectory optimization step is executed again using the above-mentioned updated flight state vector sequence as the flight state vector sequence. By dynamically updating the flight state vector sequence and re-executing the optimization step, the flight path can be adjusted in real time according to changes in the flight environment, avoiding the problem of decreased efficiency or increased energy consumption that may occur during the flight.

[0105] Further references Figure 4 As an implementation of the methods shown in the above figures, the present disclosure provides some embodiments of a logistics trajectory optimization device for an industrial chain. These device embodiments are similar to Figure 1 Corresponding to the method embodiments shown, the device can be specifically applied to various electronic devices.

[0106] like Figure 4As shown, a logistics trajectory optimization device 400 for an industrial chain in some embodiments includes: a no-fly zone detection unit 401 , an acquisition unit 402 , and an execution unit 403 . Among them, the no-fly zone detection unit 401 is configured to perform no-fly zone detection on the flight area of the flying logistics equipment and obtain a no-fly zone information set; the acquisition unit 402 is configured to obtain the flight state vector sequence of the above-mentioned flying logistics equipment; the execution unit 403 is configured to perform the following trajectory optimization steps in response to the remaining flight distance corresponding to the above-mentioned flying logistics equipment being greater than a preset distance value: based on the trajectory optimization model and the above-mentioned no-fly zone information set, determine the predicted flight state characteristics corresponding to the flight state vector sequence, wherein the above-mentioned trajectory optimization model includes a long short-term memory network, a fully connected layer and a neural ordinary differential equation network; input the above-mentioned predicted flight state characteristics into the above-mentioned neural ordinary differential equation network to obtain optimized control parameters; according to the above-mentioned optimized control parameters, control the above-mentioned flying logistics equipment, and determine the predicted flight state vector corresponding to the above-mentioned optimized control parameters based on the preset flight dynamics equation; generate an updated flight state vector sequence based on the above-mentioned predicted flight state vector and the above-mentioned flight state vector sequence; the second execution unit 404 is configured to use the above-mentioned updated flight state vector sequence as the flight state vector sequence to execute the above-mentioned trajectory optimization step again.

[0107] It is understood that the units described in the device 400 are similar to those in the reference Figure 1 Therefore, the operations, features and beneficial effects described above for the method are also applicable to the device 400 and the units included therein, and will not be repeated here.

[0108] Reference below Figure 5 , which shows a structural schematic diagram of an electronic device (eg, a computing device) 500 suitable for implementing some embodiments of the present disclosure. Figure 3 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0109] like Figure 5 As shown, electronic device 500 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 501, which can perform various appropriate actions and processes based on programs stored in read-only memory 502 or programs loaded from storage device 508 into random access memory 503. Random access memory 503 also stores various programs and data required for the operation of electronic device 500. Processing device 501, read-only memory 502, and random access memory 503 are connected to each other via bus 504. Input / output interface 505 is also connected to bus 504.

[0110] Typically, the following devices may be connected to the I / O interface 505: an input device 506 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 507 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 508 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 509. The communication device 509 may allow the electronic device 500 to communicate with other devices wirelessly or by wire to exchange data. Figure 5 The electronic device 500 is shown with various devices, but it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed instead. Figure 5 Each block shown in the figure may represent one device, or may represent multiple devices as needed.

[0111] In particular, according to some embodiments of the present disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, some embodiments of the present disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program comprising program code for executing the method shown in the flowchart. In some such embodiments, the computer program can be downloaded and installed from a network via the communication device 509, or installed from the storage device 508, or installed from the read-only memory 502. When the computer program is executed by the processing device 501, the above-mentioned functions defined in the method of some embodiments of the present disclosure are performed.

[0112] It should be noted that the computer-readable medium described in some embodiments of the present disclosure may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. Computer-readable storage media may include, for example, but not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or components, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection having one or more conductors, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In some embodiments of the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device, or component. Furthermore, in some embodiments of the present disclosure, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. This propagated data signal may take a variety of forms, including, but not limited to, electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wire, optical cable, RF (radio frequency), or any suitable combination thereof.

[0113] In some embodiments, the client and server can communicate using any currently known or later developed network protocol, such as HTTP (Hypertext Transfer Protocol), and can be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), an internet (e.g., the Internet), and a peer-to-peer network (e.g., an ad hoc peer-to-peer network), as well as any currently known or later developed network.

[0114] The computer-readable medium may be included in the electronic device, or may exist separately and not incorporated into the electronic device. The computer-readable medium carries one or more programs. When executed by the electronic device, the one or more programs cause the electronic device to: perform no-fly zone detection on the flight area of the flight logistics equipment to obtain a no-fly zone information set; obtain a flight state vector sequence of the flight logistics equipment; and, in response to the remaining flight distance corresponding to the flight logistics equipment being greater than a preset distance value, perform the following trajectory optimization steps: determining predicted flight state features corresponding to the flight state vector sequence based on a trajectory optimization model and the no-fly zone information set, wherein the trajectory optimization model includes a long short-term memory network, a fully connected layer, and a neural ordinary differential equation network; inputting the predicted flight state features into the neural ordinary differential equation network to obtain optimized control parameters; controlling the flight logistics equipment based on the optimized control parameters, and determining a predicted flight state vector corresponding to the optimized control parameters based on preset flight dynamics equations; generating an updated flight state vector sequence based on the predicted flight state vector and the flight state vector sequence; and performing the trajectory optimization step again using the updated flight state vector sequence as the flight state vector sequence.

[0115] Computer program code for performing the operations of some embodiments of the present disclosure may be written in one or more programming languages, or a combination thereof, including object-oriented programming languages such as Java, Smalltalk, C++, and conventional procedural programming languages such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0116] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, program segment, or a part of code, and the module, program segment, or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of the boxes in the block diagram and / or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0117] The units described in some embodiments of the present disclosure may be implemented in software or hardware. The units described may also be provided in a processor. For example, they may be described as follows: a processor including a no-fly zone detection unit, an acquisition unit, an execution unit, and a second execution unit. The names of these units do not, in some cases, constitute limitations on the units themselves. For example, the acquisition unit may also be described as a "unit for acquiring the flight state vector sequence of the above-mentioned flight logistics equipment."

[0118] The functions described above herein may be performed, at least in part, by one or more hardware logic components. For example, and without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0119] The above descriptions are merely some preferred embodiments of the present disclosure and illustrate the underlying technical principles. Those skilled in the art should understand that the scope of the invention encompassed by the embodiments of the present disclosure is not limited to technical solutions formed by specific combinations of the aforementioned technical features. It also encompasses other technical solutions formed by any combination of the aforementioned technical features or their equivalents, without departing from the aforementioned inventive concept. For example, a technical solution formed by replacing the aforementioned features with (but not limited to) technical features with similar functions disclosed in the embodiments of the present disclosure.

Claims

1. A logistics trajectory optimization method for an industrial chain, comprising: Perform no-fly zone detection on the flight area of flying logistics equipment to obtain a no-fly zone information set; Obtaining a flight state vector sequence of the flight logistics equipment; In response to the remaining flight distance corresponding to the flight logistics equipment being greater than a preset distance value, the following trajectory optimization steps are performed: Determining predicted flight state features corresponding to a flight state vector sequence based on a trajectory optimization model and the no-fly zone information set, wherein the trajectory optimization model includes a long short-term memory network, a fully connected layer, and a neural ordinary differential equation network; Inputting the predicted flight state characteristics into the neural ordinary differential equation network to obtain optimized control parameters; Controlling the flight logistics equipment according to the optimized control parameters, and determining a predicted flight state vector corresponding to the optimized control parameters based on a preset flight dynamics equation; generating an updated flight state vector sequence according to the predicted flight state vector and the flight state vector sequence; The trajectory optimization step is performed again using the updated flight state vector sequence as the flight state vector sequence.

2. The method according to claim 1, wherein The no-fly zone detection is performed on the flight area of the flying logistics equipment to obtain a no-fly zone information set, including: Obtaining flight status information corresponding to the flight logistics equipment, wherein the flight status information includes flight starting point coordinates, flight end point coordinates, and aircraft parameter information; generating flight area information corresponding to the flight logistics equipment according to the flight situation information; According to the flight zone information, the no-fly information that meets preset conditions in the preset no-fly information set is determined as the no-fly zone information to obtain the no-fly zone information set.

3. The method according to claim 1, wherein The step of determining predicted flight state features corresponding to the flight state vector sequence based on the trajectory optimization model and the no-fly zone information set includes: Based on the long short-term memory network and the no-fly zone information set, extracting time series features from the flight state vector sequence to obtain a time series feature vector set; Determining the last time series feature vector in the time series feature vector set as a hidden feature vector; Based on the fully connected layer, feature mapping is performed on the implicit feature vector to obtain predicted flight state features.

4. The method according to claim 1, wherein The trajectory optimization model is trained by the following steps: Generate sample training data sets based on numerical methods; Pre-training a preset initial trajectory optimization model according to the sample training data set to obtain a pre-trained trajectory optimization model; Reinforcement learning training is performed on the pre-trained trajectory optimization model to obtain a trajectory optimization model.

5. The method according to claim 4, wherein The method of pre-training a preset initial trajectory optimization model based on the sample training data set to obtain a pre-trained trajectory optimization model includes: Preprocessing each sample training data in the sample training data set to generate preprocessed sample training data, thereby obtaining a preprocessed sample training data set, wherein each preprocessed sample training data in the preprocessed sample training data set includes a preprocessed sample state quantity sequence and a preprocessed control quantity true value; Inputting a preprocessed sample state quantity sequence included in each preprocessed sample training data set into a preset initial trajectory optimization model to obtain a predicted control quantity set, wherein the initial trajectory optimization model includes an initial long short-term memory network, an initial fully connected layer, and an initial neural ordinary differential equation network; generating a total loss value according to a loss value between each predicted control variable in the predicted control variable set and a true value of the preprocessed control variable included in the corresponding preprocessed sample training data in the preprocessed sample training data set; Back-propagation parameters of the initial trajectory optimization model are updated according to the total loss value to obtain a pre-trained trajectory optimization model.

6. A logistics trajectory optimization device for an industrial chain, comprising: a no-fly zone detection unit, configured to perform no-fly zone detection on a flight area of the flying logistics equipment and obtain a no-fly zone information set; an acquisition unit configured to acquire a flight state vector sequence of the flight logistics equipment; The execution unit is configured to execute the following trajectory optimization steps in response to the remaining flight distance corresponding to the flight logistics equipment being greater than a preset distance value: Determining predicted flight state features corresponding to a flight state vector sequence based on a trajectory optimization model and the no-fly zone information set, wherein the trajectory optimization model includes a long short-term memory network, a fully connected layer, and a neural ordinary differential equation network; Inputting the predicted flight state characteristics into the neural ordinary differential equation network to obtain optimized control parameters; Controlling the flight logistics equipment according to the optimized control parameters, and determining a predicted flight state vector corresponding to the optimized control parameters based on a preset flight dynamics equation; generating an updated flight state vector sequence according to the predicted flight state vector and the flight state vector sequence; The second execution unit is configured to execute the trajectory optimization step again using the updated flight state vector sequence as a flight state vector sequence.

7. An electronic device comprising: one or more processors; a storage device having one or more programs stored thereon; When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 5.

8. A computer-readable medium having a computer program stored thereon, wherein: When the program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images