Unmanned aerial vehicle suspended cargo safety evaluation method based on low-altitude wind direction deflection prediction
By constructing a multi-dimensional safety assessment model based on cross-entropy time-series prediction, the problem of insufficient wind direction deflection prediction in the safety assessment of UAV-suspended cargo was solved. This model enables accurate prediction and multi-dimensional safety assessment of low-altitude wind direction deflection, thereby improving the safety and applicability of UAV-suspended cargo operations.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- FUZHOU UNIV
- Filing Date
- 2026-03-03
- Publication Date
- 2026-05-12
AI Technical Summary
Existing methods for assessing the safety of cargo suspended by drones lack the ability to predict wind direction deflection, have limited assessment indicators, and have weak model generalization capabilities. As a result, they are unable to accurately predict the impact of low-altitude wind direction deflection on suspended cargo, leading to frequent safety accidents.
Data is collected by equipping a meteorological sensing module and a suspension status monitoring module to construct a safety-related dataset. The cross-entropy time series prediction algorithm is used to predict wind direction deflection. A multi-dimensional safety assessment model is constructed by combining the dynamic characteristics of the UAV. The model parameters are optimized by training with historical data, multi-level assessment indicators and safety thresholds are set, and a comprehensive safety assessment value is calculated.
It enables accurate prediction of low-altitude wind direction deflection, multi-dimensional assessment of suspended cargo safety, adaptability to different operating scenarios, reduction of safety risks caused by wind interference, and improvement of the safety and applicability of drone-borne cargo operations.
Smart Images
Figure CN121766546B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of unmanned aerial vehicle (UAV) application technology, specifically relating to a method for safety assessment of UAV-suspended cargo based on low-altitude wind direction deflection prediction. Background Technology
[0002] With the rapid iteration of drone technology, its application scenarios have expanded from traditional aerial surveying and inspection to complex suspended cargo operations such as material transportation, engineering hoisting, emergency rescue, and offshore wind power maintenance. Drone-based suspended cargo operations, with their advantages of flexibility, efficiency, and lack of terrain limitations, are playing an irreplaceable role in scenarios such as cargo transfer from high-rise buildings in cities, emergency cargo delivery in mountainous areas, and equipment hoisting on offshore platforms. However, the high proportion of safety accidents caused by wind interference has become a core bottleneck restricting the large-scale implementation of this technology.
[0003] Low-altitude environments are characterized by variable wind speeds and frequent turbulence. The coupling effect of wind direction deflection and atmospheric turbulence can directly lead to drone attitude instability and violent sling swings, resulting in serious safety accidents such as cargo falling off, collisions with obstacles, and drone crashes. For example, in cargo slinging operations in urban high-rise buildings, the "narrowing effect" between tall buildings can cause local wind speeds to increase sharply by 3-5 times. Sudden wind direction deflections can easily cause slings to collide with building exteriors, resulting in damage to or falling of the suspended cargo. In offshore wind power maintenance scenarios, irregular pulsations and turbulent interference from sea winds can cause sling swing angles exceeding 30°, which can easily cause suspended cargo to fall, seriously threatening the safety of personnel and equipment. In mountain emergency rescue scenarios, the alternating day and night wind direction changes in valleys often cause drones to lose attitude control when slinging rescue cargo, leading to cargo delivery deviations or falls, delaying critical rescue opportunities.
[0004] Existing methods for assessing the safety of drone-borne cargo have significant limitations: First, they rely heavily on real-time meteorological data for static assessments, failing to predict low-altitude wind direction deflection patterns and making it difficult to mitigate potential risks during cargo slinging. Second, the assessment indicators are relatively singular, focusing primarily on drone attitude or wind speed parameters, neglecting the coupled effects of sling status, cargo sway amplitude, and wind direction deflection. Third, the models lack generalization ability, failing to incorporate historical cargo slinging data from different operational scenarios for training and optimization, resulting in assessment accuracy that cannot match actual operational needs. Therefore, there is an urgent need for a drone-borne cargo safety assessment method that can accurately predict low-altitude wind direction deflection and consider multiple dimensions of cargo safety influencing factors to overcome existing technological bottlenecks and provide effective technical support for the safe operation of drone-borne cargo slinging. Summary of the Invention
[0005] The purpose of this invention is to address the shortcomings of existing drone-borne cargo safety assessment methods, such as the lack of wind direction deflection prediction capabilities, the simplification of assessment indicators, and the weak generalization ability of models. This invention provides a drone-borne cargo safety assessment method based on low-altitude wind direction deflection prediction, which helps to accurately quantify the safety risks of drones suspending cargo at low altitudes and improve the safety assurance level of drone suspension.
[0006] To achieve the above objectives, the technical solution of the present invention is: a method for safety assessment of unmanned aerial vehicle (UAV) suspended cargo based on low-altitude wind direction deflection prediction, comprising:
[0007] By using the meteorological sensing module and the suspension status monitoring module carried by the UAV, low-altitude meteorological data of the suspension area, real-time attitude parameters of the UAV and sling status parameters are collected. After preprocessing the collected data, a safety association dataset D is constructed.
[0008] Based on the safety association dataset D, the cross-entropy time series prediction algorithm is used to predict the wind direction deflection, and combined with the dynamic characteristics of UAVs, a multi-dimensional suspended cargo safety assessment model is constructed.
[0009] The multi-dimensional suspended cargo safety assessment model was trained using historical safety correlation datasets from different suspension scenarios, and the model parameters were adjusted through iterative optimization.
[0010] A multi-level assessment index system was established, including the duration of risk and the swing angle of the sling, and the safety thresholds of each level of index and their influence weight in the multi-dimensional suspended cargo safety assessment model were clarified.
[0011] Based on the optimized evaluation model and the set evaluation indicators, the comprehensive safety assessment value M of the drone suspending cargo is calculated, and the safety level is divided according to the range of M values.
[0012] Furthermore, the suspension area is the low-altitude airspace where the UAV is suspended, ranging from 0 to 1000 meters above the ground; the low-altitude meteorological data for the suspension area includes: instantaneous low-altitude wind speed v(t), instantaneous wind direction θ(t), atmospheric turbulence intensity I(t), and turbulence fluctuation frequency f(t); the real-time attitude parameters of the UAV include: UAV pitch angle α(t), roll angle β(t), yaw angle γ(t), and takeoff and landing speed v. uav (t); Sling state parameters include: sling swing angle (t), the angular velocity of the sling swing ω(t); after preprocessing the collected data, the constructed safety association dataset D is represented as:
[0013]
[0014] In the formula, For the first Each sampling time, This represents the total number of samples taken.
[0015] Furthermore, based on the safety association dataset D, the cross-entropy time series prediction algorithm is used to predict the wind direction deflection, specifically including:
[0016] Define the standard wind direction for drones carrying cargo. Calculate real-time wind direction deflection The calculation formula is:
[0017]
[0018] in, For the first Real-time wind direction deflection, I( ) is the first Atmospheric turbulence intensity at any given time, with 0.1 representing the atmospheric turbulence influence coefficient;
[0019] The cross-entropy time series prediction algorithm is used to predict real-time wind direction deflection. Perform a prediction and output the wind direction deflection prediction sequence for the next time period T. }, T=10~30 seconds, the core prediction formula is:
[0020]
[0021] in, For the future Predicted wind direction deflection at any time to Here are the weighting coefficients of the prediction model, and ε(τ) is the turbulence disturbance error term;
[0022] In the statistical prediction sequence ≥ The continuous duration, as the duration of risk T r ,in The threshold for wind direction deflection warning.
[0023] Furthermore, combining the dynamic characteristics of UAVs, the expression for the multi-dimensional suspended cargo safety assessment model is as follows:
[0024]
[0025] Where M is the comprehensive security assessment value; The influencing factor of wind direction deflection; For the sling state factor; This is the attitude stability factor; Duration of the risk; , Let be the weighting coefficient, satisfying The values range from 0.2 to 0.5.
[0026] Furthermore, wind direction deflection influencing factors The formula for calculation is:
[0027]
[0028] In the formula, For the future The average value of the predicted wind direction deflection over a period of time. This represents the extreme wind direction deflection threshold.
[0029] Furthermore, the sling state factor The formula for calculation is:
[0030]
[0031] in, , , This is the standard swing angle of the sling. The standard angular velocity of the sling swing. This represents the extreme swing angle of the sling.
[0032] Furthermore, attitude stability factor The formula for calculation is:
[0033]
[0034] in, , , , This is the deviation of the extreme attitude angle. The standard attitude angle for drones suspending cargo.
[0035] Furthermore, the constructed multi-dimensional suspended cargo safety assessment model is trained using historical safety correlation datasets from different suspension scenarios. The model parameters are iteratively optimized and adjusted, specifically including:
[0036] Historical datasets covering different suspended scenarios in plains, mountains, cities, and seas are divided into training and validation sets according to proportions.
[0037] The training set is input into the multi-dimensional suspended cargo safety assessment model for iterative training, with the goal of minimizing the assessment error. The weight coefficients of the prediction model are adjusted using the gradient descent method. to and evaluation model weight coefficients , ;
[0038] Training stops when the change in the validation set evaluation error over multiple consecutive iterations is less than a set threshold, or when the number of iterations reaches the upper limit.
[0039] Verify that the prediction accuracy of the trained model is ≥90% and the evaluation error is ≤5%. If the criteria are not met, readjust the model parameters and retrain.
[0040] Furthermore, weighting coefficients , The risk level is dynamically determined based on the real-time risk level of its corresponding factor, where:
[0041] Weighting coefficients of wind direction deflection influencing factors The risk level is determined based on the impact factor of wind direction deflection.
[0042] Weighting coefficients of sling state factor The risk level is determined based on the sling condition factor.
[0043] Weighting coefficients of attitude stability factor The risk level of the attitude stability factor is determined accordingly.
[0044] Weighting coefficient of risk duration factor The risk level is determined based on the duration of the risk.
[0045] Furthermore, security levels are classified based on the comprehensive security assessment value M, where:
[0046] When M < 0.3, the risk level is "safe";
[0047] When 0.3 ≤ M < 0.6, the risk level is "low risk";
[0048] When 0.6 ≤ M < 0.8, the risk level is "medium risk";
[0049] When M ≥ 0.8, the risk level is "high risk".
[0050] Compared with the prior art, the present invention has the following beneficial effects:
[0051] 1. By using the cross-entropy time series prediction algorithm to predict future wind direction deflection patterns and identify the duration of risks in advance, the problem of static assessment and delayed risk prediction in traditional methods is solved.
[0052] 2. Construct a multi-dimensional assessment model that includes wind direction deflection, sling status, UAV attitude, and duration of risk, comprehensively consider the coupled impact of wind interference on the safety of suspended cargo, and achieve higher assessment accuracy;
[0053] 3. The model is trained and optimized by combining historical data from different work scenarios, making it suitable for various work environments such as plains, mountains, cities, and seas, with a wider range of applications;
[0054] 4. By clearly defining the safety thresholds and risk weights of each indicator, a refined classification of risk levels can be achieved, and targeted response strategies can be matched, which can significantly reduce the risk of accidents caused by wind interference.
[0055] 5. Applicable to various low-altitude operation drones, with strong compatibility and wide application scenarios, providing technical support for the safe development of the low-altitude economy. Attached Figure Description
[0056] Figure 1 A flowchart illustrating the overall process of a method for assessing the safety of unmanned aerial vehicle (UAV) cargo suspension based on low-altitude wind direction deflection prediction, provided in this embodiment of the invention.
[0057] Figure 2 This is a detailed flowchart of data acquisition and preprocessing in an embodiment of the present invention;
[0058] Figure 3 This is a detailed flowchart illustrating the construction of the wind direction deflection prediction and multi-dimensional evaluation model according to an embodiment of the present invention.
[0059] Figure 4 This is a flowchart illustrating the model training and optimization process of this invention, as described in an embodiment of the invention.
[0060] Figure 5 This invention sets multi-level evaluation index safety thresholds and risk weights for embodiments of the invention, and provides a flowchart for classifying safety risk levels. Detailed Implementation
[0061] The technical solution of the present invention will now be described in detail with reference to the accompanying drawings.
[0062] It should be noted that the following detailed descriptions are illustrative and intended to provide further explanation of this application. Unless otherwise specified, all technical and scientific terms used in this specification have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.
[0063] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.
[0064] This invention provides a method for safety assessment of unmanned aerial vehicle (UAV) suspended cargo based on low-altitude wind direction deflection prediction, comprising:
[0065] By using the meteorological sensing module and the suspension status monitoring module carried by the UAV, low-altitude meteorological data of the suspension area, real-time attitude parameters of the UAV and sling status parameters are collected. After preprocessing the collected data, a safety association dataset D is constructed.
[0066] Based on the safety association dataset D, the cross-entropy time series prediction algorithm is used to predict the wind direction deflection, and combined with the dynamic characteristics of UAVs, a multi-dimensional suspended cargo safety assessment model is constructed.
[0067] The multi-dimensional suspended cargo safety assessment model was trained using historical safety correlation datasets from different suspension scenarios, and the model parameters were adjusted through iterative optimization.
[0068] A multi-level assessment index system was established, including the duration of risk and the swing angle of the sling, and the safety thresholds of each level of index and their influence weight in the multi-dimensional suspended cargo safety assessment model were clarified.
[0069] Based on the optimized evaluation model and the set evaluation indicators, the comprehensive safety assessment value M of the drone suspending cargo is calculated, and the safety level is divided according to the range of M values.
[0070] The following is a detailed implementation process of the present invention.
[0071] like Figure 1 As shown, this invention proposes a method for assessing the safety of unmanned aerial vehicle (UAV) suspension based on low-altitude wind direction deflection prediction, comprising the following steps:
[0072] (1) Using the meteorological sensing module and suspension status monitoring module mounted on the UAV, meteorological data of the suspension area, real-time attitude of the UAV and sling status parameters are collected, and a safety association dataset is constructed after preprocessing; the flowchart of this step is as follows. Figure 2 As shown;
[0073] The suspended airspace refers to the low-altitude airspace where UAVs are suspended at an altitude of 0-1000 meters above the ground.
[0074] The suspended airspace meteorological data includes: low-altitude instantaneous wind speed v(t), m / s, instantaneous wind direction θ(t), °, atmospheric turbulence intensity I(t), m² / s³, and turbulence fluctuation frequency f(t), Hz;
[0075] The real-time attitude parameters of the UAV include: UAV pitch angle α(t), °, roll angle β(t), °, yaw angle γ(t), °. The sling state parameters include: sling swing angle. (t), °, angular velocity of sling swing ω(t), ° / s;
[0076] The security association dataset is as follows:
[0077]
[0078] In the formula, For the first Each sampling time, This represents the total number of samples, with a sampling frequency of 10~50Hz;
[0079] The data preprocessing includes removing outliers, filling in missing values using linear interpolation, and applying a moving average filter to the data. The filter formula is as follows:
[0080]
[0081] In the formula, v( is the filtered data at time i;) ) represents the raw data at time j; k is the sliding window size, ranging from 5 to 10.
[0082] (2) Using the cross-entropy time series prediction algorithm, the low-altitude wind direction deflection pattern is extracted from the dataset, and combined with the dynamic characteristics of the UAV, a multi-dimensional suspended cargo safety assessment model is constructed; the flowchart of this step is as follows. Figure 3 As shown;
[0083] 1) Define the standard wind direction for drones to suspend cargo. Calculate real-time wind direction deflection The formula is:
[0084] ,
[0085] In the formula, 0.1 represents the atmospheric turbulence influence coefficient. For real-time atmospheric turbulence intensity, For the first Real-time wind deflection;
[0086] The standard wind direction for drone cargo suspension is a baseline wind direction pre-set before the drone begins suspension, taking into account the requirements of the suspension mission and the environmental characteristics of the suspension area.
[0087] 2) The cross-entropy time series prediction algorithm is used to predict wind direction deflection, outputting a predicted wind direction deflection sequence for the next T time period (T=10~30s). The core prediction formula is:
[0088]
[0089] In the formula, For the future Predicted wind direction deflection at any time For the weighting coefficients of the prediction model, This is the turbulence interference error term;
[0090] 3) In the statistical prediction series ≥ The continuous duration of the risk is the duration T of the risk. r The The wind direction deflection warning threshold is set at 15°~30°.
[0091] 4) Construct a multi-dimensional safety assessment model for suspended cargo, the expression of which is:
[0092] In the formula, This is a comprehensive safety assessment value; The influencing factor of wind direction deflection; For the sling state factor; This is the attitude stability factor; Duration of the risk; , Let be the weighting coefficient, satisfying The values range from 0.2 to 0.5.
[0093] Among them, wind direction deflection influencing factors The formula for calculation is:
[0094]
[0095] In the formula, For the future The average value of the predicted wind direction deflection over a period of time. The extreme wind direction deflection threshold is set to 40°;
[0096] Sling state factor The formula for calculation is:
[0097]
[0098] In the formula, ,
[0099] The The standard swing angle of the sling is taken as 0°. The standard angular velocity of the sling swing is taken as 0° / s. The ultimate swing angle of the sling is set to 25°.
[0100] Attitude stability factor The formula for calculation is:
[0101]
[0102] In the formula, , , , The extreme attitude angle deviation is taken as 18°.
[0103] The The standard attitude angle for the drone suspending cargo is taken as 0°.
[0104] (3) Train the model using historical data from different suspension scenarios and iteratively optimize the parameters; the flowchart for this step is as follows. Figure 4 As shown;
[0105] 1) The historical suspension safety association datasets for different suspension scenarios are divided into training and validation sets in a 7:3 ratio. The datasets cover sample data under different wind speeds, wind directions, turbulence intensities, and sling load conditions. The different suspension scenarios include plains, mountains, cities, and offshore suspension scenarios.
[0106] 2) Input the training set into the evaluation model for iterative training, aiming to minimize the evaluation error, and adjust the model weight coefficients using the gradient descent method. and ;
[0107] 3) Training is stopped when the change in validation set evaluation error is less than 0.001 after 100 consecutive iterations, or when the number of iterations reaches 500.
[0108] 4) The verification indicators are prediction accuracy ≥ 90% and evaluation error ≤ 5%. If the indicators are not met, return to step 2) in (3) to readjust the parameters for training.
[0109] (4) Set multi-level assessment indicators such as risk duration and sling swing angle, and clarify their safety thresholds and impact weights; based on the above assessment model and assessment indicators, calculate the safety level of the UAV during cargo suspension. The flowchart of this step is as follows: Figure 5 As shown;
[0110] The multi-level evaluation indicators adopt a hierarchical quantitative design, and the specific indicator system, safety threshold, and risk weights are as follows:
[0111] The weighting coefficient λ1 of the wind direction deflection factor is determined based on the risk level of the wind direction deflection factor:
[0112] Security level correspondence <15°, weight is 0.1;
[0113] Low risk level corresponds to 15°≤ <30°, weight is 0.2;
[0114] Medium risk level corresponds to 30°≤ <45°, weight is 0.3;
[0115] High-risk level corresponds to ≥45°, weight is 0.4.
[0116] The weighting coefficient λ2 of the sling condition factor is determined based on the risk level of the sling condition factor:
[0117] Security level correspondence The weight is 0.1;
[0118] Low risk level corresponds to 5°≤ The weight is 0.2;
[0119] Medium risk level corresponds to 12°≤ The weight is 0.3;
[0120] High-risk level corresponds to The weight is 0.4.
[0121] The weighting coefficient λ3 of the attitude response factor is determined based on the risk level of the attitude stability factor:
[0122] Security level correspondence <5°, weight is 0.1;
[0123] Low risk level corresponds to 5°≤ <10°, weight is 0.2;
[0124] Medium risk level corresponds to 10°≤ <15°, weight is 0.3;
[0125] High-risk level corresponds to ≥15°; weight is 0.4.
[0126] The weighting coefficient λ4 for the duration of risk is determined based on the risk level of the duration of risk.
[0127] Security level corresponding to T r <5s, weight is 0.1;
[0128] Low risk level corresponds to 5s≤T r <10s, weight is 0.2;
[0129] Medium risk level corresponds to 10s≤T r <20s, weight is 0.3;
[0130] High-risk level corresponds to T r ≥20s; weight is 0.4.
[0131] Furthermore, the specific process of step S5 is as follows:
[0132] Based on the value range of M, the risk level of drone suspension is accurately classified, and targeted safety decision-making solutions are matched to ensure the safety and reliability of suspension operations. The specific classification criteria and response strategies are as follows:
[0133] Risk levels are determined based on the value of M, where M is the comprehensive safety assessment value:
[0134] when When the value is less than 0.3, the risk level is "safe".
[0135] When 0.3 ≤ M < 0.6, the risk level is "low risk";
[0136] When 0.6 ≤ M < 0.8, the risk level is "medium risk";
[0137] when When the value is ≥0.8, the risk level is "high risk".
[0138] In summary, the UAV suspension safety assessment method based on low-altitude wind deflection prediction of this invention collects meteorological data of the suspension area, real-time attitude of the UAV, and sling status parameters through the meteorological perception module and suspension status monitoring module on the UAV. After preprocessing, a safety association dataset is constructed. The low-altitude wind deflection pattern is extracted from the dataset using the cross-entropy time series prediction algorithm. Combined with the dynamic characteristics of the UAV, a multi-dimensional suspension cargo safety assessment model is constructed. The model is trained using historical data of different suspension scenarios, and the parameters are iteratively optimized. Multi-level assessment indicators such as risk duration and sling swing angle are set, and their safety thresholds and influence weights are clarified. Finally, the safety level of the UAV during cargo suspension is calculated. The design method of this invention enables dynamic and accurate assessment of the safety of UAV-suspended cargo under complex wind field conditions in low-altitude airspace. It fills the technical gaps in this field, such as the lack of wind direction deflection prediction capabilities, the lack of coupling of assessment dimensions with sling state, and the impact of atmospheric turbulence and cargo dynamic response. It avoids the drawbacks of existing technologies that rely on static judgment based on real-time data and cannot predict in advance the risk of wind interference causing severe sling swing, UAV attitude instability, or cargo detachment and damage. At the same time, it saves testing costs and reduces the safety hazards of cargo detachment or collision with obstacles during the test compared with traditional live-flight suspended cargo testing. It also makes up for the shortcomings of existing technical solutions that only consider a single wind parameter or UAV attitude, and do not combine key factors such as the duration of risk and sling tension response and cargo swing threshold, making it difficult to propose refined cargo safety risk response strategies from multiple dimensions.
[0139] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0140] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0141] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0142] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0143] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.
[0144] This patent is not limited to the above-described preferred embodiment. Anyone can derive other various forms of energy-saving and emission-reduction evaluation methods for pure electric vehicles under complex road alignment conditions based on the teachings of this patent. All equivalent variations and modifications made within the scope of this patent application should be covered by this patent. This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0145] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0146] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0147] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.
[0148] This patent is not limited to the above-described preferred embodiments. Anyone can derive other various forms of energy-saving and emission-reduction evaluation methods for pure electric vehicles under complex road alignment conditions based on the inspiration of this patent. All equivalent changes and modifications made within the scope of this patent application shall fall within the scope of this patent.
Claims
1. A method for safety assessment of unmanned aerial vehicle (UAV) suspended cargo based on low-altitude wind direction deflection prediction, characterized in that, include: By using the meteorological sensing module and the suspension status monitoring module carried by the UAV, low-altitude meteorological data of the suspension area, real-time attitude parameters of the UAV and sling status parameters are collected. After preprocessing the collected data, a safety association dataset D is constructed. Based on the safety association dataset D, the cross-entropy time series prediction algorithm is used to predict the wind direction deflection, and combined with the dynamic characteristics of UAVs, a multi-dimensional suspended cargo safety assessment model is constructed. The multi-dimensional suspended cargo safety assessment model was trained using historical safety correlation datasets from different suspension scenarios, and the model parameters were adjusted through iterative optimization. A multi-level assessment index system was established, including the duration of risk and the swing angle of the sling, and the safety thresholds of each level of index and their influence weight in the multi-dimensional suspended cargo safety assessment model were clarified. Based on the optimized evaluation model and the set evaluation indicators, the comprehensive safety assessment value M of the drone suspending cargo is calculated, and the safety level is divided according to the range of M values. The suspension area is the low-altitude airspace for drone suspension, ranging from 0 to 1000 meters above the ground; Low-altitude meteorological data for the suspended area includes: low-altitude instantaneous wind speed. Instantaneous wind direction θ Atmospheric turbulence intensity I turbulent fluctuation frequency f The real-time attitude parameters of the UAV include: UAV pitch angle α. Roll angle β Yaw angle γ Takeoff and landing speed v uav The sling status parameters include: sling swing angle. Angular velocity ω of the sling swing After preprocessing the collected data, the constructed security association dataset D is represented as follows: In the formula, For the first Each sampling time, This represents the total number of samples. Based on the safety association dataset D, the cross-entropy time series prediction algorithm is used to predict wind direction deflection, specifically including: Define the standard wind direction for drones carrying cargo. Calculate real-time wind direction deflection The calculation formula is: in, For the first Real-time wind direction deflection, I( ) is the first Atmospheric turbulence intensity at any given time, with 0.1 representing the atmospheric turbulence influence coefficient; The cross-entropy time series prediction algorithm is used to predict real-time wind direction deflection. Perform a prediction and output the wind direction deflection prediction sequence for the next time period T. }, T=10~30 seconds, the core prediction formula is: in, For the future Predicted wind direction deflection at any time to Here are the weighting coefficients of the prediction model, and ε(τ) is the turbulence disturbance error term; In the statistical prediction sequence ≥ The continuous duration, as the duration of risk T r ,in The threshold for wind direction deflection warning.
2. The method for safety assessment of UAV-suspended cargo based on low-altitude wind direction deflection prediction according to claim 1, characterized in that, Combining the dynamic characteristics of UAVs, the expression for constructing a multi-dimensional suspended cargo safety assessment model is as follows: Where M is the comprehensive security assessment value; The influencing factor of wind direction deflection; For the sling state factor; This is the attitude stability factor; Duration of the risk; , These are the weighting coefficients for the wind direction deflection factor, the sling state factor, the attitude stability factor, and the risk duration factor, respectively, satisfying... The values range from 0.2 to 0.
5.
3. The method for safety assessment of UAV-suspended cargo based on low-altitude wind direction deflection prediction according to claim 2, characterized in that, Wind deflection influencing factors The formula for calculation is: In the formula, For the future The average value of the predicted wind direction deflection over a period of time. This represents the extreme wind direction deflection threshold.
4. The method for safety assessment of UAV-suspended cargo based on low-altitude wind direction deflection prediction according to claim 2, characterized in that, Sling state factor The formula for calculation is: in, , , This is the standard swing angle of the sling. The standard angular velocity of the sling swing. This represents the extreme swing angle of the sling.
5. The method for safety assessment of UAV-suspended cargo based on low-altitude wind direction deflection prediction according to claim 2, characterized in that, Attitude stability factor The formula for calculation is: in, , , , This is the deviation of the extreme attitude angle. The standard attitude angle for drones suspending cargo.
6. The method for safety assessment of UAV-suspended cargo based on low-altitude wind direction deflection prediction according to claim 2, characterized in that, The multi-dimensional suspended cargo safety assessment model was trained using historical safety correlation datasets from different suspension scenarios. The model parameters were iteratively optimized and adjusted, specifically including: Historical datasets covering different suspended scenarios in plains, mountains, cities, and seas are divided into training and validation sets according to proportions. The training set is input into the multi-dimensional suspended cargo safety assessment model for iterative training, with the goal of minimizing the assessment error. The weight coefficients of the prediction model are adjusted using the gradient descent method. to and evaluation model weight coefficients , ; Training stops when the change in the validation set evaluation error over multiple consecutive iterations is less than a set threshold, or when the number of iterations reaches the upper limit. Verify that the prediction accuracy of the trained model is ≥90% and the evaluation error is ≤5%. If the criteria are not met, readjust the model parameters and retrain.
7. The method for safety assessment of UAV-suspended cargo based on low-altitude wind direction deflection prediction according to any one of claims 2-6, characterized in that, Weighting coefficient , The risk level is dynamically determined based on the real-time risk level of its corresponding factor, where: Weighting coefficients of wind direction deflection influencing factors The risk level is determined based on the impact factor of wind direction deflection. Weighting coefficients of sling state factor The risk level is determined based on the sling condition factor. Weighting coefficients of attitude stability factor The risk level of the attitude stability factor is determined accordingly. Weighting coefficient of risk duration factor The risk level is determined based on the duration of the risk.
8. The method for safety assessment of UAV-suspended cargo based on low-altitude wind direction deflection prediction according to claim 1, characterized in that, Safety levels are classified based on the comprehensive safety assessment value M, where: When M < 0.3, the risk level is "safe"; When 0.3 ≤ M < 0.6, the risk level is "low risk"; When 0.6 ≤ M < 0.8, the risk level is "medium risk"; When M ≥ 0.8, the risk level is "high risk".