Unmanned aerial vehicle flight safety early warning method

Abstract

The invention discloses an unmanned aerial vehicle flight safety early warning method and belongs to the technical field of unmanned aerial vehicle flight safety. In order to improve the flight safety performance of the unmanned aerial vehicle, the method comprises the steps of collecting surrounding object information data, environmental factor data and unmanned aerial vehicle information data in the operation process of the unmanned aerial vehicle; performing standardization processing, and dividing the standardized data into three groups of factor data sets; constructing an unmanned aerial vehicle flight minimum safe distance calculation method, and calculating unmanned aerial vehicle minimum safe flight distances at different time points; constructing an unmanned aerial vehicle flight minimum safety distance prediction model based on different single environment factors, and calculating unmanned aerial vehicle flight minimum safety distances under different single environment factors; calculating an unmanned aerial vehicle flight comprehensive minimum safety prediction distance; and constructing a collision early warning method based on the unmanned aerial vehicle flight comprehensive minimum safety prediction distance. The method is accurate in early warning.

Description

Technical Field

[0001] The present invention belongs to the technical field of unmanned aerial vehicle flight safety, and in particular relates to a unmanned aerial vehicle flight safety early warning method. Background Art

[0002] The impact of different structures must be considered during the flight of drones, as these structures will have a direct impact on the flight stability, safety and efficiency of drones. High-rise buildings, bridges, trees, etc. not only differ in height, they may also change the surrounding airflow patterns and form complex wind flow situations, which puts higher requirements on the control accuracy and flight path planning of drones. For example, high-rise buildings can produce strong wind pressure effects, so that drones may encounter sudden airflow changes when flying near them, affecting flight stability. In addition, complex transportation infrastructure such as highway interchanges are not only complex in structure, but also have a dynamic and changeable surrounding environment, which requires drones to have a high degree of adaptability and immediate response capabilities to deal with possible collision risks.

[0003] If the impact of these structures is ignored during the flight of the drone, it may lead to a variety of flight safety issues. First, the drone may not be able to effectively avoid obstacles of different heights. For example, if the flight altitude is not adjusted in time to avoid tall buildings or tree beams, a collision accident may occur, causing damage to the drone or even harm to people and property on the ground. Secondly, ignoring the changes in airflow caused by the structure may make it difficult to control the drone, especially in strong wind conditions, where stability and maneuverability are greatly reduced, and the risk of flight mission failure increases. For example, when flying in a wind tunnel formed by tall buildings, the drone may encounter a sudden change in wind speed, resulting in control failure, affecting the accuracy of data collection or leading to more serious safety accidents.

[0004] The invention patent with application number 201711349951.X and invention name “Real-time safety warning device for drones” proposes a real-time safety warning device for drones, including an airborne warning device, integrating a flight controller, a status monitoring module (including a drone parameter acquisition module) and a warning module. This system can collect the flight and environmental parameters of drones and analyze these parameters through the warning module. If the parameters are abnormal, the safety warning instruction will be triggered and the safety mechanism will be activated, thereby improving the flight safety of the drone.

[0005] The invention patent with application number 202111300583.6 and invention name "A delivery drone, a delivery drone early warning method and device" discloses a delivery drone and its early warning method and device, which are operated by different power supply and sensor submodules. The early warning module and the control module evaluate the collected data to determine the abnormal state of the drone. When an abnormality is found, the early warning module will send an alarm to the surroundings and report to the server to prompt the surrounding people to avoid and quickly handle the abnormality, effectively reducing the risk of accidents.

[0006] However, existing studies have focused on real-time status monitoring and abnormal warning systems for drones, but none of them explicitly mention the consideration of safe distance issues during drone flight. In particular, the requirements for safe distances may vary under different facility types and environmental conditions. The lack of specific considerations for safe flight distances may affect the safe flight and collision avoidance capabilities of drones in complex environments, thereby limiting the breadth and safety of their applications. Summary of the invention

[0007] The problem to be solved by the present invention is to improve the safety performance of unmanned aerial vehicle flight and propose a unmanned aerial vehicle flight safety early warning method.

[0008] To achieve the above object, the present invention is implemented through the following technical solutions: A UAV flight safety early warning method comprises the following steps: S1. Collecting information data of surrounding objects, environmental factors, and drone information during the operation of the drone. The surrounding object information data includes object type, object height, and width. The environmental factors data include ambient wind speed, light intensity, and temperature. The drone information data includes drone length, drone width, and drone weight. S2. The data collected in step S1 is standardized, and the standardized data is divided into three groups of factor data sets to obtain a data set for temperature analysis, a data set for light intensity analysis, and a data set for ambient wind speed analysis, and set aside; S3. Construct a method to calculate the minimum safe flight distance of UAVs, and calculate the minimum safe flight distance of UAVs at different time points; S4. Construct a prediction model for the minimum safe distance of UAV flight based on different single environmental factors, and calculate the minimum safe distance of UAV flight under different single environmental factors; S5. Based on the minimum safe distance for UAV flight under different single environmental factors obtained in step S4, calculate the comprehensive minimum safe predicted distance for UAV flight; S6. Construct a collision warning method based on the comprehensive minimum safe prediction distance of UAV flight.

[0009] Furthermore, in step S1, the data of object height, width, ambient wind speed, light intensity, temperature, drone length, drone width, and drone weight at the same time point are collected, and the selected time point is set as t 1, t 2…, ti …, tn ,in, ti For the i A point in time, n is the total number of time points.

[0010] Furthermore, in step S2, the standardized data are divided into three groups of factor data sets, to obtain a data set for temperature analysis, a data set for light intensity analysis, and a data set for ambient wind speed analysis; The factors in the data set for temperature analysis include object type, object height, object width, ambient wind speed, light intensity, drone length, drone width, and drone weight; The factors in the data set for analyzing light intensity include object type, object height, object width, ambient wind speed, temperature, drone length, drone width, and drone weight; The factors in the data set for environmental wind speed analysis include object type, object height, object width, light intensity, temperature, drone length, drone width, and drone weight.

[0011] Furthermore, the specific implementation method of step S3 includes the following steps: S3.1. According to i Time point ti , the corresponding object type, object height, width, ambient wind speed, light intensity, temperature, drone length, drone width, and drone weight data are used to establish a simulated drone flight field. In the simulated drone flight field, the distance between the drone and the object is set to , m Refers to the total number of distances between the drone and the object when it is set to fly. Point to the i Time point ti, No. j The distance between the drone and the object when flying; S3.2. The height fluctuation amplitude of the UAV is measured by the airborne inertial measurement unit. The flight distance of the drone from detecting obstacles to responding using the vehicle-mounted GNSS navigation system is The noise generated by the drone is measured using a sound level meter placed in the surrounding environment. ,in, , , Respectively refer to i Time point ti No. j The distance between the drone and the object during flight The measured height fluctuation amplitude, the distance the drone flies from the time an obstacle is detected to the time it reacts, and the noise generated by the drone; S3.3. Based on , , , calculate the comprehensive evaluation index value of UAV safe flight , the calculation formula is: ; in, , , They are , , The corresponding weight value; On this basis, the distance between the drone and the object during flight is calculated. Corresponding comprehensive evaluation index value of UAV safe flight ; S3.4. Use the expert scoring method to set the threshold of comprehensive evaluation indicators for safe flight of drones , compare in turn and The size relationship of All greater than The comprehensive evaluation index value of the safe flight of the UAV is the condition, and then the obstacle corresponding to the minimum value that meets the condition is selected until the UAV flight distance when the reaction is made is the minimum safe flight distance of the UAV, which is recorded as ; S3.5. The calculated time point is t 1, t 2…, ti …, tn The corresponding minimum safe flight distance of the UAV is .

[0012] Furthermore, the specific implementation method of step S4 includes the following steps: S4.1. A deep learning model is used to build a temperature-based prediction model for the minimum safe distance of drone flight. The specific structure is as follows: Input layer: Set 8 input nodes, corresponding to the object type, object height, width, ambient wind speed, light intensity, drone length, drone width, and drone weight at each time point after data standardization; Hidden Layer: The first hidden layer: consists of 128 neurons and uses the ReLU activation function; The second hidden layer consists of 64 neurons and uses the ReLU activation function. The third hidden layer: consists of 32 neurons and uses the ReLU activation function; The fourth hidden layer: consists of 16 neurons and uses the ReLU activation function; Output layer: set 1 neuron corresponding to the minimum safe flight distance of the drone at each time point; The object type, object height, width, ambient wind speed, light intensity, drone length, drone width, drone weight, and drone minimum safe flight distance corresponding to each time point are input into the temperature-based drone minimum safe flight distance prediction model for training, and the square of the error between the drone minimum safe flight distance predicted by the temperature-based drone minimum safe flight distance prediction model and the calculated drone minimum safe flight distance is used as the loss function. , the expression is: ; in, The minimum safe flight distance of a UAV predicted by the temperature-based prediction model for the minimum safe flight distance of a UAV; The error back propagation method is used to optimize the parameters of the temperature-based UAV flight minimum safety distance prediction model, and the expression is: ; in, are model parameters, is the learning rate, is the loss function with respect to the parameter The gradient of Represents the update of model parameters; After optimization, the temperature-based prediction model for the minimum safe distance of UAV flight is obtained. P 1; S4.2. A deep learning model is used to build a prediction model for the minimum safe distance of drone flight based on light intensity. The specific structure is as follows: Input layer: Set 8 input nodes, corresponding to the object type, object height, width, ambient wind speed, temperature, drone length, drone width, and drone weight at each time point after data standardization; Hidden Layer: The first hidden layer: consists of 128 neurons and uses the ReLU activation function; The second hidden layer consists of 64 neurons and uses the ReLU activation function. The third hidden layer: consists of 32 neurons and uses the ReLU activation function; The fourth hidden layer: consists of 16 neurons and uses the ReLU activation function; Output layer: set 1 neuron corresponding to the minimum safe flight distance of the drone at each time point; The object type, object height, width, ambient wind speed, temperature, drone length, drone width, drone weight, and drone minimum safe flight distance corresponding to each time point are input into the drone minimum safe flight distance prediction model based on light intensity for training, and the square of the error between the drone minimum safe flight distance predicted by the drone minimum safe flight distance prediction model based on light intensity and the calculated safety distance is used as the loss function. , the expression is: ; in, The minimum safe flight distance of a UAV predicted by the prediction model of the minimum safe flight distance of a UAV based on light intensity; The error back propagation method is used to optimize the parameters of the model, and after optimization, the minimum safe distance prediction model for UAV flight based on light intensity is obtained. P 2; S4.3. A deep learning model is used to build a prediction model for the minimum safe distance of UAV flight based on ambient wind speed. The specific structure is as follows: Input layer: Set 8 input nodes, corresponding to the object type, object height, width, temperature, light intensity, drone length, drone width, and drone weight at each time point after data standardization; Hidden Layer: The first hidden layer: consists of 128 neurons and uses the ReLU activation function; The second hidden layer consists of 64 neurons and uses the ReLU activation function. The third hidden layer: consists of 32 neurons and uses the ReLU activation function; The fourth hidden layer: consists of 16 neurons and uses the ReLU activation function; Output layer: set 1 neuron corresponding to the minimum safe flight distance of the drone at each time point; The object type, object height, width, ambient wind speed, temperature, drone length, drone width, drone weight, and drone minimum safe flight distance corresponding to each time point are input into the drone minimum safe flight distance prediction model based on ambient wind speed for training, and the square of the error between the drone minimum safe flight distance predicted by the drone minimum safe flight distance prediction model based on ambient wind speed and the calculated safety distance is used as the loss function , the expression is: ; in, The minimum safe flight distance of a UAV predicted by the prediction model of the minimum safe flight distance of a UAV based on the ambient wind speed; The error back propagation method is used to optimize the parameters of the model, and after optimization, the minimum safe distance prediction model for UAV flight based on ambient wind speed is obtained. P 3; S4.4. Based on P 1. P 2. P 3. Predict the minimum safe distance for drone flight; By inputting the drone flight time, the corresponding object type, object height, width, ambient wind speed, light intensity, drone length, drone width, and drone weight data are P 1, and then predict P 1 The minimum safe predicted distance for drone flight ; By inputting the drone flight time, the corresponding object type, object height, width, ambient wind speed, temperature, drone length, drone width, and drone weight data are P 2, and then predict P 2. The minimum safe predicted distance for drone flight ; By inputting the drone flight time, the corresponding object type, object height, width, light intensity, temperature, drone length, drone width, and drone weight data into P3, the prediction is obtained. P 3. The minimum safe predicted distance for drone flight .

[0013] Furthermore, in step S5, the minimum safe predicted distance for UAV flight is calculated. The expression is: ; in, , , They are , , The weight coefficient of .

[0014] Furthermore, the specific implementation method of step S6 includes the following steps: S6.1. Using airborne laser radar, the drone collects the three-dimensional coordinates of surrounding objects while flying. , For surrounding objects x Axis coordinates, For surrounding objects y Axis coordinates, For surrounding objects z Axis coordinates; S6.2. Use the drone-mounted GNSS system to collect the drone’s three-dimensional coordinates , Space for drones x Axis coordinates, Space for drones y Axis coordinates, Space for drones z Axis coordinates; S6.3. From the three-dimensional coordinates of surrounding objects , filter and the spatial 3D coordinates of the drone Point cloud data with similar heights are retained. The specific screening method is to traverse all three-dimensional coordinates. ,Compare and Relationship: when When , the corresponding point cloud data is retained; when , delete the corresponding point cloud data; in, is the height deviation threshold; S6.4. From the retained point cloud data, arbitrarily select a point in the retained point cloud data. Then, by calculating and comparing the points around this point with The distance between the points, select the points adjacent to the T Nearest Neighbor ; Point-based and its adjacent T nearest neighbor points, fitting a plane PM, the corresponding plane equation is: ; in, , , refers to the component of the plane normal vector, , , Refers to the coordinates of any point in the retained point cloud data of the surrounding objects. is a constant; based on and its adjacent T Nearest Neighbor , using the least squares method, we get , , , , and then determine the plane equation; S6.5. Calculate the three-dimensional coordinates of the drone The distance Ds from the plane PM is calculated as: ; S6.6. Construct collision warning conditions and compare Ds with the minimum safe predicted distance of the UAV flight obtained in step S5 The relationship between: when When , it indicates that the distance between the drone and the surrounding objects exceeds the comprehensive minimum safety prediction distance, and no warning is required; when , it indicates that the distance between the drone and surrounding objects does not exceed the comprehensive minimum safety prediction distance, and an early warning is required.

[0015] Beneficial effects of the present invention: The UAV flight safety warning method described in the present invention takes into account the surrounding object information, environmental factors, UAV information, etc., establishes a UAV flight safety distance prediction model, and realizes the prediction of the minimum safe distance for UAV flying in different environments; supports the UAV's task execution in complex environments and expands the application scope of UAVs.

[0016] The UAV flight safety warning method described in the present invention proposes a collision warning method based on the comprehensive minimum safety prediction distance of the UAV flight, which realizes the collision warning of the UAV flight. Through the accurate analysis of the surrounding environment of the UAV and the flight data, potential risk factors can be discovered in time, and risk avoidance measures can be taken in advance to prevent possible accidents. Not only can the damage caused by collision or loss of control of the UAV be avoided, and the high cost of repairing or replacing equipment can be reduced, but also liability disputes or legal proceedings caused by accidents can be prevented. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 The present invention is a flowchart of a UAV flight safety early warning method. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solution and advantages of the present invention more clear, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention, that is, the specific embodiments described are only part of the embodiments of the present invention, rather than all of the specific embodiments. The components of the specific embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations, and the present invention can also have other embodiments.

[0019] Therefore, the following detailed description of the specific embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents the selected specific embodiments of the present invention. Based on the specific embodiments of the present invention, all other specific embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present invention.

[0020] In order to further understand the content, features and effects of the present invention, the following specific implementation methods are given as examples, and the attached Figure 1 The detailed instructions are as follows:

[0021] Embodiment 1: A UAV flight safety early warning method, characterized in that it includes the following steps: S1. Collecting information data of surrounding objects, environmental factors, and drone information during the operation of the drone. The surrounding object information data includes object type, object height, and width. The environmental factors data include ambient wind speed, light intensity, and temperature. The drone information data includes drone length, drone width, and drone weight. Furthermore, in step S1, the data of object height, width, ambient wind speed, light intensity, temperature, drone length, drone width, and drone weight at the same time point are collected, and the selected time point is set as t 1, t 2…, ti …, tn ,in, ti For the i A point in time, n is the total number of time points.

[0022] Furthermore, the data collected for object height, width, ambient wind speed, light intensity, temperature, drone length, drone width, and drone weight correspond and are the same, that is, they are all data information at the same time. The selected time point is t 1, t 2…, ti …, tn , the object heights corresponding to each time point areA 1, A 2, ..., An ; The object widths corresponding to each time point are B 1, B 2, ..., Bn ; The corresponding ambient wind speeds at each time point are C 1, C 2, ..., Cn ; The light intensity corresponding to each time point is D 1, D 2, ..., Dn The temperatures at each time point are E 1, E 2, ..., En ; The length of the drone corresponding to each time point is F 1, F 2, ..., Fn ; The width of the drone corresponding to each time point is G 1, G 2, ..., Gn ; The weight of the drone at each time point is H 1, H 2, ..., Hn ; The above data are obtained from public channels; S2. The data collected in step S1 is standardized, and the influence of a single environmental factor data on the prediction of safe flight of the UAV is extracted and analyzed. The standardized data is divided into three groups of factor data sets to obtain a data set for temperature analysis, a data set for light intensity analysis, and a data set for environmental wind speed analysis, and the data set is set aside; Furthermore, in step S2, the standardized data are divided into three groups of factor data sets, to obtain a data set for temperature analysis, a data set for light intensity analysis, and a data set for ambient wind speed analysis; The factors in the data set for temperature analysis include object type, object height, object width, ambient wind speed, light intensity, drone length, drone width, and drone weight; The factors in the data set for analyzing light intensity include object type, object height, object width, ambient wind speed, temperature, drone length, drone width, and drone weight; The factors in the data set for environmental wind speed analysis include object type, object height, object width, light intensity, temperature, drone length, drone width, and drone weight.

[0023] Furthermore, after data standardization, the object heights corresponding to each time point are ; After data standardization, the object widths corresponding to each time point are ; After data standardization, the ambient wind speeds corresponding to each time point are ; After data standardization, the light intensity corresponding to each time point is ; After data standardization, the temperatures corresponding to each time point are ; After data standardization, the drone lengths corresponding to each time point are ; After data standardization, the width of the drone corresponding to each time point is ; After data standardization, the weight of the drone at each time point is ; S3. Construct a method to calculate the minimum safe flight distance of UAVs, and calculate the minimum safe flight distance of UAVs at different time points; Furthermore, the specific implementation method of step S3 includes the following steps: S3.1. According to i Time point ti , the corresponding object type, object height, width, ambient wind speed, light intensity, temperature, drone length, drone width, and drone weight data are used to establish a simulated drone flight field. In the simulated drone flight field, the distance between the drone and the object is set to , m Refers to the total number of distances between the drone and the object when it is set to fly. Point to the i Time point ti, No. j The distance between the drone and the object when flying; S3.2. The height fluctuation amplitude of the UAV is measured by the airborne inertial measurement unit. The flight distance of the drone from detecting obstacles to responding using the vehicle-mounted GNSS navigation system is The noise generated by the drone is measured using a sound level meter placed in the surrounding environment. ,in, , , Respectively refer to i Time point ti No. j The distance between the drone and the object during flight The measured height fluctuation amplitude, the distance the drone flies from the time an obstacle is detected to the time it reacts, and the noise generated by the drone; S3.3. Based on , , , calculate the comprehensive evaluation index value of UAV safe flight , the calculation formula is: ; in, , , They are , , The corresponding weight value; On this basis, the distance between the drone and the object during flight is calculated. Corresponding comprehensive evaluation index value of UAV safe flight ; S3.4. Use the expert scoring method to set the threshold of comprehensive evaluation indicators for safe flight of drones , compare in turn and The size relationship of All greater than The comprehensive evaluation index value of the safe flight of the UAV is the condition, and then the obstacle corresponding to the minimum value that meets the condition is selected until the UAV flight distance when the reaction is made is the minimum safe flight distance of the UAV, which is recorded as ; S3.5. The calculated time point is t 1, t 2…, ti …, tn The corresponding minimum safe flight distance of the UAV is .

[0024] S4. Construct a prediction model for the minimum safe distance of UAV flight based on different single environmental factors, and calculate the minimum safe distance of UAV flight under different single environmental factors; Furthermore, the specific implementation method of step S4 includes the following steps:

[0025] S4.1. A deep learning model is used to build a temperature-based prediction model for the minimum safe distance of drone flight. The specific structure is as follows: Input layer: Set 8 input nodes, corresponding to the object type, object height, width, ambient wind speed, light intensity, drone length, drone width, and drone weight at each time point after data standardization; Hidden Layer: The first hidden layer: consists of 128 neurons and uses the ReLU activation function; The second hidden layer consists of 64 neurons and uses the ReLU activation function. The third hidden layer: consists of 32 neurons and uses the ReLU activation function; The fourth hidden layer: consists of 16 neurons and uses the ReLU activation function; Output layer: set 1 neuron corresponding to the minimum safe flight distance of the drone at each time point; The object type, object height, width, ambient wind speed, light intensity, drone length, drone width, drone weight, and drone minimum safe flight distance corresponding to each time point are input into the temperature-based drone minimum safe flight distance prediction model for training, and the square of the error between the drone minimum safe flight distance predicted by the temperature-based drone minimum safe flight distance prediction model and the calculated drone minimum safe flight distance is used as the loss function. , the expression is: ; in, The minimum safe flight distance of a UAV predicted by the temperature-based prediction model for the minimum safe flight distance of a UAV; The error back propagation method is used to optimize the parameters of the temperature-based UAV flight minimum safety distance prediction model, and the expression is: ; in, are model parameters, is the learning rate, is the loss function with respect to the parameter The gradient of Represents the update of model parameters; After optimization, the temperature-based prediction model for the minimum safe distance of UAV flight is obtained. P 1; S4.2. A deep learning model is used to build a prediction model for the minimum safe distance of drone flight based on light intensity. The specific structure is as follows: Input layer: Set 8 input nodes, corresponding to the object type, object height, width, ambient wind speed, temperature, drone length, drone width, and drone weight at each time point after data standardization; Hidden Layer: The first hidden layer: consists of 128 neurons and uses the ReLU activation function; The second hidden layer consists of 64 neurons and uses the ReLU activation function. The third hidden layer: consists of 32 neurons and uses the ReLU activation function; The fourth hidden layer: consists of 16 neurons and uses the ReLU activation function; Output layer: set 1 neuron corresponding to the minimum safe flight distance of the drone at each time point; The object type, object height, width, ambient wind speed, temperature, drone length, drone width, drone weight, and drone minimum safe flight distance corresponding to each time point are input into the drone minimum safe flight distance prediction model based on light intensity for training, and the square of the error between the drone minimum safe flight distance predicted by the drone minimum safe flight distance prediction model based on light intensity and the calculated safety distance is used as the loss function. , the expression is: ; in, The minimum safe flight distance of a UAV predicted by the prediction model of the minimum safe flight distance of a UAV based on light intensity; The error back propagation method is used to optimize the parameters of the model, and after optimization, the minimum safe distance prediction model for UAV flight based on light intensity is obtained. P 2; S4.3. A deep learning model is used to build a prediction model for the minimum safe distance of UAV flight based on ambient wind speed. The specific structure is as follows: Input layer: Set 8 input nodes, corresponding to the object type, object height, width, temperature, light intensity, drone length, drone width, and drone weight at each time point after data standardization; Hidden Layer: The first hidden layer: consists of 128 neurons and uses the ReLU activation function; The second hidden layer consists of 64 neurons and uses the ReLU activation function. The third hidden layer: consists of 32 neurons and uses the ReLU activation function; The fourth hidden layer: consists of 16 neurons and uses the ReLU activation function; Output layer: set 1 neuron corresponding to the minimum safe flight distance of the drone at each time point; The object type, object height, width, ambient wind speed, temperature, drone length, drone width, drone weight, and drone minimum safe flight distance corresponding to each time point are input into the drone minimum safe flight distance prediction model based on ambient wind speed for training, and the square of the error between the drone minimum safe flight distance predicted by the drone minimum safe flight distance prediction model based on ambient wind speed and the calculated safety distance is used as the loss function , the expression is: ; in, The minimum safe flight distance of a UAV predicted by the prediction model of the minimum safe flight distance of a UAV based on the ambient wind speed; The error back propagation method is used to optimize the parameters of the model, and after optimization, the minimum safe distance prediction model for UAV flight based on ambient wind speed is obtained. P 3; S4.4. Based on P 1. P 2. P 3. Predict the minimum safe distance for drone flight; By inputting the drone flight time, the corresponding object type, object height, width, ambient wind speed, light intensity, drone length, drone width, and drone weight data are P 1, and then predict P 1 The minimum safe predicted distance for drone flight ; By inputting the drone flight time, the corresponding object type, object height, width, ambient wind speed, temperature, drone length, drone width, and drone weight data are P 2, and then predict P 2. The minimum safe predicted distance for drone flight ; By inputting the drone flight time, the corresponding object type, object height, width, light intensity, temperature, drone length, drone width, and drone weight data into P3, the prediction is obtained. P 3. The minimum safe predicted distance for drone flight .

[0026] S5. Based on the minimum safe distance for UAV flight under different single environmental factors obtained in step S4, calculate the comprehensive minimum safe predicted distance for UAV flight; Furthermore, in step S5, the minimum safe predicted distance for UAV flight is calculated. The expression is: ; in, , , They are , , The weight coefficient of .

[0027] S6. Construct a collision warning method based on the comprehensive minimum safe prediction distance of UAV flight; Furthermore, the specific implementation method of step S6 includes the following steps:

[0028] S6.1. Using airborne laser radar, the drone collects the three-dimensional coordinates of surrounding objects while flying. , For surrounding objects x Axis coordinates, For surrounding objects y Axis coordinates, For surrounding objects z Axis coordinates; S6.2. Use the drone-mounted GNSS system to collect the drone’s three-dimensional coordinates , Space for drones x Axis coordinates, Space for drones y Axis coordinates, Space for drones z Axis coordinates; S6.3. From the three-dimensional coordinates of surrounding objects , filter and the spatial 3D coordinates of the drone Point cloud data with similar heights are retained. The specific screening method is to traverse all three-dimensional coordinates. ,Compare and Relationship: when When , the corresponding point cloud data is retained; when , delete the corresponding point cloud data; in, is the height deviation threshold; S6.4. From the retained point cloud data, arbitrarily select a point in the retained point cloud data. Then, by calculating and comparing the points around this point with The distance between the points, select the points adjacent to the T Nearest Neighbor ; Point-based and its adjacent T nearest neighbor points, fitting a plane PM, the corresponding plane equation is: ; in, , , refers to the component of the plane normal vector, , , Refers to the coordinates of any point in the retained point cloud data of the surrounding objects. is a constant; based on and its adjacent T Nearest Neighbor , using the least squares method, we get , , , , and then determine the plane equation; S6.5. Calculate the three-dimensional coordinates of the drone The distance Ds from the plane PM is calculated as: ; S6.6. Construct collision warning conditions and compare Ds with the minimum safe predicted distance of the UAV flight obtained in step S5 The relationship between: when When , it indicates that the distance between the drone and the surrounding objects exceeds the comprehensive minimum safety prediction distance, and no warning is required; when , it indicates that the distance between the drone and surrounding objects does not exceed the comprehensive minimum safety prediction distance, and an early warning is required.

[0029] The key points and intended protection points of the present invention are: (1) A comprehensive method for predicting the safe flight distance of UAVs.

[0030] (2) A collision warning method based on the comprehensive minimum safe prediction distance of UAV flight.

[0031] It should be noted that relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0032] Although the present application has been described above with reference to specific embodiments, various modifications may be made thereto and parts thereof may be replaced with equivalents without departing from the scope of the present application. In particular, as long as there is no structural conflict, the various features in the specific embodiments disclosed in the present application may be used in combination with each other in any manner, and the fact that these combinations are not exhaustively described in this specification is only for the sake of omitting space and saving resources. Therefore, the present application is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A UAV flight safety early warning method, characterized in that: The steps include: S1. Collecting information data of surrounding objects, environmental factors, and drone information during the operation of the drone. The surrounding object information data includes object type, object height, and width. The environmental factors data include ambient wind speed, light intensity, and temperature. The drone information data includes drone length, drone width, and drone weight. S2. The data collected in step S1 is standardized, and the standardized data is divided into three groups of factor data sets to obtain a data set for temperature analysis, a data set for light intensity analysis, and a data set for ambient wind speed analysis, and set aside; S3. Construct a method to calculate the minimum safe flight distance of UAVs, and calculate the minimum safe flight distance of UAVs at different time points; S4. Construct a prediction model for the minimum safe distance of UAV flight based on different single environmental factors, and calculate the minimum safe distance of UAV flight under different single environmental factors; S5. Based on the minimum safe distance for UAV flight under different single environmental factors obtained in step S4, calculate the comprehensive minimum safe predicted distance for UAV flight; S6. Construct a collision warning method based on the comprehensive minimum safe prediction distance of UAV flight.

2. The UAV flight safety early warning method according to claim 1, characterized in that: In step S1, data on object height, width, ambient wind speed, light intensity, temperature, drone length, drone width, and drone weight are collected at the same time point, and the selected time point is set as t 1, t 2…, ti …, tn ,in, ti For the i A point in time, n is the total number of time points.

3. The UAV flight safety early warning method according to claim 2 is characterized in that: In step S2, the standardized data are divided into three groups of factor data sets, to obtain a data set for temperature analysis, a data set for light intensity analysis, and a data set for ambient wind speed analysis; The factors in the data set for temperature analysis include object type, object height, object width, ambient wind speed, light intensity, drone length, drone width, and drone weight; The factors in the data set for analyzing light intensity include object type, object height, object width, ambient wind speed, temperature, drone length, drone width, and drone weight; The factors in the data set for environmental wind speed analysis include object type, object height, object width, light intensity, temperature, drone length, drone width, and drone weight.

4. The UAV flight safety early warning method according to claim 3 is characterized in that: The specific implementation method of step S3 includes the following steps: S3.

1. According to i Time point ti , the corresponding object type, object height, width, ambient wind speed, light intensity, temperature, drone length, drone width, and drone weight data are used to establish a simulated drone flight field. In the simulated drone flight field, the distance between the drone and the object is set to , m Refers to the total number of distances between the drone and the object when it is set to fly. Point to the i Time point ti, No. j The distance between the drone and the object when flying; S3.

2. The height fluctuation amplitude of the UAV is measured by the airborne inertial measurement unit. The flight distance of the drone from detecting obstacles to responding using the vehicle-mounted GNSS navigation system is The noise generated by the drone is measured using a sound level meter placed in the surrounding environment. ,in, , , Respectively refer to i Time point ti, No. j The distance between the drone and the object during flight The measured height fluctuation amplitude, the distance the drone flies from the time an obstacle is detected to the time it reacts, and the noise generated by the drone; S3.

3. Based on , , , calculate the comprehensive evaluation index value of UAV safe flight , the calculation formula is: ； in, , , They are , , The corresponding weight value; On this basis, the distance between the drone and the object during flight is calculated. Corresponding comprehensive evaluation index value of UAV safe flight ; S3.

4. Use the expert scoring method to set the threshold of comprehensive evaluation indicators for safe flight of drones , compare in turn and The size relationship of All greater than The comprehensive evaluation index value of the safe flight of the UAV is the condition, and then the obstacle corresponding to the minimum value that meets the condition is selected until the UAV flight distance when the reaction is made is the minimum safe flight distance of the UAV, which is recorded as ; S3.

5. The calculated time point is t 1, t 2…, ti …, tn The corresponding minimum safe flight distance of the UAV is .

5. The UAV flight safety early warning method according to claim 4 is characterized in that: The specific implementation method of step S4 includes the following steps: S4.

1. A deep learning model is used to build a temperature-based prediction model for the minimum safe distance of drone flight. The specific structure is as follows: Input layer: Set 8 input nodes, corresponding to the object type, object height, width, ambient wind speed, light intensity, drone length, drone width, and drone weight at each time point after data standardization; Hidden Layer: The first hidden layer: consists of 128 neurons and uses the ReLU activation function; The second hidden layer consists of 64 neurons and uses the ReLU activation function. The third hidden layer: consists of 32 neurons and uses the ReLU activation function; The fourth hidden layer: consists of 16 neurons and uses the ReLU activation function; Output layer: set 1 neuron corresponding to the minimum safe flight distance of the drone at each time point; The object type, object height, width, ambient wind speed, light intensity, drone length, drone width, drone weight, and drone minimum safe flight distance corresponding to each time point are input into the temperature-based drone minimum safe flight distance prediction model for training, and the square of the error between the drone minimum safe flight distance predicted by the temperature-based drone minimum safe flight distance prediction model and the calculated drone minimum safe flight distance is used as the loss function. , the expression is: ； in, The minimum safe flight distance of a UAV predicted by the temperature-based prediction model for the minimum safe flight distance of a UAV; The error back propagation method is used to optimize the parameters of the temperature-based UAV flight minimum safety distance prediction model, and the expression is: ； in, are model parameters, is the learning rate, is the loss function with respect to the parameter The gradient of Represents the update of model parameters; After optimization, the temperature-based prediction model for the minimum safe distance of UAV flight is obtained. P 1; S4.

2. A deep learning model is used to build a prediction model for the minimum safe distance of drone flight based on light intensity. The specific structure is as follows: Input layer: Set 8 input nodes, corresponding to the object type, object height, width, ambient wind speed, temperature, drone length, drone width, and drone weight at each time point after data standardization; Hidden Layer: The first hidden layer: consists of 128 neurons and uses the ReLU activation function; The second hidden layer consists of 64 neurons and uses the ReLU activation function. The third hidden layer: consists of 32 neurons and uses the ReLU activation function; The fourth hidden layer: consists of 16 neurons and uses the ReLU activation function; Output layer: set 1 neuron corresponding to the minimum safe flight distance of the drone at each time point; The object type, object height, width, ambient wind speed, temperature, drone length, drone width, drone weight, and drone minimum safe flight distance corresponding to each time point are input into the drone minimum safe flight distance prediction model based on light intensity for training, and the square of the error between the drone minimum safe flight distance predicted by the drone minimum safe flight distance prediction model based on light intensity and the calculated safety distance is used as the loss function. , the expression is: ； in, The minimum safe flight distance of a UAV predicted by the prediction model of the minimum safe flight distance of a UAV based on light intensity; The error back propagation method is used to optimize the parameters of the model, and after optimization, the minimum safe distance prediction model for UAV flight based on light intensity is obtained. P 2; S4.

3. A deep learning model is used to build a prediction model for the minimum safe distance of UAV flight based on ambient wind speed. The specific structure is as follows: Input layer: Set 8 input nodes, corresponding to the object type, object height, width, temperature, light intensity, drone length, drone width, and drone weight at each time point after data standardization; Hidden Layer: The first hidden layer: consists of 128 neurons and uses the ReLU activation function; The second hidden layer consists of 64 neurons and uses the ReLU activation function. The third hidden layer: consists of 32 neurons and uses the ReLU activation function; The fourth hidden layer: consists of 16 neurons and uses the ReLU activation function; Output layer: set 1 neuron corresponding to the minimum safe flight distance of the drone at each time point; The object type, object height, width, ambient wind speed, temperature, drone length, drone width, drone weight, and drone minimum safe flight distance corresponding to each time point are input into the drone minimum safe flight distance prediction model based on ambient wind speed for training, and the square of the error between the drone minimum safe flight distance predicted by the drone minimum safe flight distance prediction model based on ambient wind speed and the calculated safety distance is used as the loss function , the expression is: ； in, The minimum safe flight distance of a UAV predicted by the prediction model of the minimum safe flight distance of a UAV based on the ambient wind speed; The error back propagation method is used to optimize the parameters of the model, and after optimization, the minimum safe distance prediction model for UAV flight based on ambient wind speed is obtained. P 3; S4.

4. Based on P 1. P 2. P 3. Predict the minimum safe distance for drone flight; By inputting the drone flight time, the corresponding object type, object height, width, ambient wind speed, light intensity, drone length, drone width, and drone weight data are P 1, and then predict P 1 The minimum safe predicted distance for drone flight ; By inputting the drone flight time, the corresponding object type, object height, width, ambient wind speed, temperature, drone length, drone width, and drone weight data are P 2, and then predict P 2. The minimum safe predicted distance for drone flight ; By inputting the drone flight time, the corresponding object type, object height, width, light intensity, temperature, drone length, drone width, and drone weight data into P3, the prediction is obtained. P 3. The minimum safe predicted distance for drone flight .

6. The UAV flight safety early warning method according to claim 5 is characterized in that: Step S5 calculates the minimum safe predicted distance for UAV flight The expression is: ； in, , , They are , , The weight coefficient of .

7. The UAV flight safety early warning method according to claim 6, characterized in that: The specific implementation method of step S6 includes the following steps: S6.

1. Using airborne laser radar, the drone collects the three-dimensional coordinates of surrounding objects while flying. , For surrounding objects x Axis coordinates, For surrounding objects y Axis coordinates, For surrounding objects z Axis coordinates; S6.

2. Use the drone-mounted GNSS system to collect the drone’s three-dimensional coordinates , Space for drones x Axis coordinates, Space for drones y Axis coordinates, Space for drones z Axis coordinates; S6.

3. From the three-dimensional coordinates of surrounding objects , filtering and the spatial 3D coordinates of the drone Point cloud data with similar heights are retained. The specific screening method is to traverse all three-dimensional coordinates. ,Compare and Relationship: when When , the corresponding point cloud data is retained; when , delete the corresponding point cloud data; in, is the height deviation threshold; S6.

4. From the retained point cloud data, arbitrarily select a point in the retained point cloud data. Then, by calculating and comparing the points around this point with The distance between the points, select the points adjacent to the T Nearest Neighbor ; Point-based and its adjacent T nearest neighbor points, fitting a plane PM, the corresponding plane equation is: ； in, , , refers to the component of the plane normal vector, , , Refers to the coordinates of any point in the retained point cloud data of the surrounding objects. is a constant; based on and its adjacent T Nearest Neighbor , using the least squares method, we get , , , , and then determine the plane equation; S6.

5. Calculate the three-dimensional coordinates of the drone The distance Ds from the plane PM is calculated as: ； S6.

6. Construct collision warning conditions and compare Ds with the minimum safe predicted distance of the UAV flight obtained in step S5 The relationship between: when When , it indicates that the distance between the drone and the surrounding objects exceeds the comprehensive minimum safety prediction distance, and no warning is required; when , it indicates that the distance between the drone and surrounding objects does not exceed the comprehensive minimum safety prediction distance, and an early warning is required.

Images