Risk identification and emergency method for unmanned aerial vehicle flying in airspace adjacent to expressway

Abstract

The invention belongs to the field of unmanned aerial vehicle emergency control systems, and relates to a risk identification and emergency method when an unmanned aerial vehicle flies in an airspace adjacent to an expressway, and the method comprises the steps: obtaining the balance state of the unmanned aerial vehicle through constructing a six-degree-of-freedom dynamic model and incorporating environmental factors and control intentions; dynamically comparing the balance state with real-time data to obtain a state deviation, deducing the fault probability of the unmanned aerial vehicle according to the state deviation, and realizing sensitive early warning of possible faults of the unmanned aerial vehicle; besides, the method constructs a double-layer linkage framework of an electromagnetic safety boundary and a physical safety boundary, realizes safety risk identification in combination with a fault probability, and ensures that a safety area between the unmanned aerial vehicle and a motorcade on a highway can be represented by a risk value along with speed, motorcade density and interference intensity; the fault probability of the unmanned aerial vehicle flying in the air is considered, when the fault probability is larger than a preset threshold value, the out-of-control state of the unmanned aerial vehicle is judged, a corresponding control strategy is adopted, and the flight safety of the unmanned aerial vehicle during partial or comprehensive out-of-control is guaranteed.

Description

Technical Field

[0001] The invention belongs to the field of unmanned aerial vehicle emergency control systems, and specifically designs a risk identification and emergency response method when a unmanned aerial vehicle flies in the airspace adjacent to a highway. Background Art

[0002] In recent years, with the rapid development of quad-rotor drone technology, it has shown significant application value in the fields of navigation capture near highways and emergency monitoring. Compared with urban airspace, this scene has the geographical advantage of low obstacle density, but the flight safety of drones faces dual risks: on the one hand, drones need to avoid physical collisions with dynamic traffic flows and infrastructure; on the other hand, the dense communication signals on highways can easily cause the interruption of drone communication links or distortion of navigation signals, indirectly leading to collision accidents. In addition, if the drone's own power failure, control failure and other abnormal conditions are not warned in time, its crash trajectory may directly invade the driving area of the highway, which is very likely to cause catastrophic secondary accidents. However, in the above scenario, current research has the following defects: existing research focuses on collision avoidance between drones or between drones and static obstacles. Few technologies build composite risk models for dynamic scenes on highways, especially lacking modeling of electromagnetic interference and physical collisions. For example, Chinese patent CN 119577439 A discloses a drone fault diagnosis method and system that integrates random convolution data enhancement and deep learning. This method not only has the above problems, but also relies on the training of massive data samples. In addition, in the current research on drone risk identification, the possible fault states of drones are less considered. For example, Chinese patent CN 118968823 A discloses a method for identifying the risk and level of low-altitude drone operations, which collects drone operation data in real time, sets static alarm areas, and sets dynamic alarm areas based on the operation data. However, this method is difficult to deal with possible fault risks and cannot guarantee the flight safety when the drone is partially or completely out of control. Summary of the invention

[0003] In view of the disadvantages and deficiencies of the existing technology, the present invention provides a method for risk identification and emergency response of an unmanned aerial vehicle (UAV) flying in the airspace adjacent to a highway. This method constructs a six-degree-of-freedom dynamics model and incorporates environmental factors and control intentions, combines driving information to obtain the balance state of the UAV, and then dynamically compares the balance state with real-time data to obtain the state deviation of the position and attitude. The state deviation is used to deduce the failure probability of the UAV, and based on the numerical change of the failure probability, a sensitive early warning of possible UAV failures is realized. This method abandons the dependence on a large number of samples. In addition, this method separately considers electromagnetic communication and physical collision, abstracts the continuous high-speed vehicle fleet on the highway as a spatio-temporal extended safety boundary considering the reachable domain of the time window, constructs a double-layer linkage framework of the electromagnetic safety boundary and the physical safety boundary, and combines the failure margin and failure probability representing the UAV state deviation to realize safety risk identification, ensuring that the safety area between the UAV and the vehicle fleet on the highway can be characterized by a risk value according to the speed, vehicle fleet density, and interference intensity. It also considers the failure probability of the UAV during flight in the air. When it is determined that the failure probability of the UAV is greater than a preset threshold, the out-of-control state of the UAV is judged, and corresponding control countermeasures are taken to ensure the flight safety of the UAV in case of partial or complete out-of-control.

[0004] To achieve the above object, the present invention adopts the following technical solutions: A method for risk identification and emergency response of an unmanned aerial vehicle flying in the airspace adjacent to a highway, the method comprising the following steps: Step S1: Data collection, obtaining the ambient data around the UAV and the basic information of the vehicles traveling nearby; Step S2: Establish a six-degree-of-freedom dynamics model including wind force; convert the joystick signal at the operator to a change in kinematic state, calculate the balance state of the UAV based on the established six-degree-of-freedom dynamics model, compare the current state of the UAV obtained in real time with the balance state, obtain the state deviation between the balance state and the current state, and calculate the instantaneous failure margin of the position loop and the instantaneous failure margin of the attitude loop based on the state deviation, so as to obtain the power failure probability, attitude failure probability, and current comprehensive failure probability of the UAV; Step S3: Construct the physical safety boundary of the vehicle fleet and infrastructure; at the same time, perform real-time detection on the frequency bands emitted by the vehicles and infrastructure traveling on the road and compare them with the current communication frequency band of the UAV. If it coincides with the current communication frequency of the UAV, an electromagnetic safety boundary is generated; generate a global safety boundary based on the physical safety boundary and the electromagnetic safety boundary; Step S4: Establish a collision box for the UAV. Meanwhile, determine whether the comprehensive failure probability exceeds a preset threshold. If it does not exceed the preset threshold, the operator controls the UAV. Meanwhile, based on the UAV collision box and the global safety boundary in Step S3, identify the current safety risk of the UAV, and issue corresponding warnings to the operator according to the safety risk level. If the comprehensive failure probability exceeds the preset threshold, turn on the automatic control and execute Step S5; Step S5: Identify the type of UAV failure. Based on the UAV state at the current moment, the power failure probability of the UAV, and the attitude failure probability, calculate the comprehensive failure fall path domain of the UAV, and obtain the failure state risk value of the UAV in the case of failure. Then, determine the degree of UAV out-of-control according to the comprehensive failure probability. If it is partially out of control, control the flight direction of the UAV to make it move forward in the direction with a low failure state risk value. When the calculated failure state risk value is zero, the UAV is forced to land safely. If it is completely out of control, map the comprehensive failure fall path domain to the global coordinate system to form the final fall impact boundary. Based on the global safety boundary and the formed final fall impact boundary, issue avoidance instructions to the vehicles around the fall location, and perform protection if there is infrastructure.

[0005] As a preference of the present invention, the six-degree-of-freedom dynamics model established in Step S2 is: ; Wherein, , , are the second-order differentials of the current center-of-gravity position of the UAV . , , are , , of the second-order time differentials, , are , of the first-order time differentials, is the rotation angle around the axis in the UAV body coordinate system, is the rotation angle around the axis in the UAV body coordinate system, is the rotation angle around the axis in the UAV body coordinate system, is the moment of inertia of the UAV around the axis in the body coordinate system, is the moment of inertia of the UAV around the axis in the body coordinate system, is the moment of inertia of the UAV around the axis in the body coordinate system, is the mass of the drone, is the total lift force of the four rotors; is the rotational torque of the front-back channel and is the pitch angle control amount; is the rotational torque of the left-right channel and is the roll angle control amount; is the rotational torque of the fuselage plane and is the yaw angle control amount, represents the disturbance generated by the wind near the highway on the drone in the global coordinate system, and the expression is: ; where, is the disturbing force of the wind on the th rotor, is the anti-torque disturbance on the th rotor, is the arm length of the drone.

[0006] As a preference of the present invention, after receiving the joystick signal of the operator in step S2, the received signal is normalized, and then through the transformation matrix , the joystick signal is converted into the change amount of the rotor speed required by the six-degree-of-freedom dynamics model. Then, based on the established six-degree-of-freedom dynamics model and the change amount of the rotor speed, the second-order differential change values of each state of the drone are calculated. By integrating the second-order differential change values of each state of the drone, the speed, position, attitude angular velocity and attitude angle of the drone in the balanced state are obtained.

[0007] As a preference of the present invention, the instantaneous fault margin of the position loop in step S2 is: ; where, represents the 2-norm in mathematics, represents the current position of the drone, that is, , is the position of the drone in the balanced state; is the range of velocities in all directions that the drone can reach, is the set of actual velocities of the drone at time, is the set of velocities of the drone in the balanced state at time; and , is the time decay factor; is the weight coefficient for controlling the importance of the term; The instantaneous fault margin of the attitude loop is: ; where, , Indicates the current attitude angle and attitude angular velocity of the UAV, satisfying , , , are the attitude angle and attitude angular velocity of the UAV in the balanced state, is the maximum attitude angle that the UAV can reach, is the attitude range in all directions that the UAV can reach; and , is the time decay factor, is the weight coefficient for controlling the importance of items; The power failure probability of the UAV and the attitude failure probability are expressed as: ; ; The comprehensive failure probability of the UAV is expressed as: ; Among them, and are the Rayleigh distribution scale parameters of the position and attitude angle respectively, and are the weight coefficients, with the value range of 0.2 - 0.7, has a value range of 0.2 - 0.7, and the sum of the two should be equal to 1.

[0008] As a preference of the present invention, the method for constructing the physical safety boundary of the vehicle fleet in step S3 is: obtaining the vehicle information of the vehicle fleet on the highway, predicting the trajectory within a future period according to the current state of the vehicle and future control inputs, and constructing a three-dimensional semi-ellipsoidal reachable domain for each vehicle , then for a vehicle fleet with a total of vehicles, the combined collision body is expressed as ; The method for constructing the physical safety boundary of the infrastructure is: the geometric center coordinates of the infrastructure are , the length, width and height of the infrastructure itself , and the safety boundary spatial domain constructed based on this is strictly defined by the three-dimensional coordinate constraint conditions as: ; Among them, represents the position coordinate parameter variable of the infrastructure; The infrastructure includes high-voltage wires, photovoltaic panels and highway toll stations.

[0009] As a preference of the present invention, the method for generating the electromagnetic safety boundary in step S3 is: obtaining the maximum power emitted during communication by the device with overlapping communication frequencies and setting the maximum allowable power at the UAV to be . Then, by inversely solving the free space path formula for communication propagation, the safety distance is obtained, and the expression is: ; In the formula, is the signal wavelength, is the reflection coefficient of the metal railing, and its value range is . The safety distance is the distance between the infrastructure emitting the communication frequency band and the UAV. Through the pointing direction of the safety distance, its projection is obtained through the rotation matrix to get the distances X g 、Y g 、Z g on the , , axes in the global coordinate system; Based on the obtained safety distance , a geometric body with as the length, as the width, and as the height is constructed, and the electromagnetic safety boundary obtained is: ; Among them, represents the position variable of the object with overlapping communication frequency band with the UAV, , , represent the projection lengths of the safety distance calculated at the moment of the geometric body on the three coordinate axes of the global coordinate system; All the established electromagnetic safety boundaries are synthesized and a union operation is performed. Suppose there are N devices with overlapping communication frequency bands with the UAV, then the electromagnetic safety boundary within the detectable range of the UAV at the current position is .

[0010] As a preference of the present invention, the method for establishing the UAV collision box in step S4 is to set the volume of the UAV as a cylinder. Suppose the maximum horizontal dimension of the UAV is , then the corresponding radius of the body cylinder is , and the height is , is half of the UAV height, , ​That is the radius of the smallest circumscribed sphere of the cylindrical body of the drone. The smallest circumscribed sphere is used as the collision box of the drone.

[0011] As an optimization of the present invention, the calculation method of the safety risk in step S4 is as follows: ; wherein, is the safety risk characterization value of the drone at the current position, represents the intersection volume of the drone collision box and the global safety boundary , is the volume of the drone collision box, is the weight factor; In step S4, set the level threshold of the safety risk of the drone and . When the safety risk characterization value is within , it is a low-level risk. When the safety risk characterization value , it is a medium-level risk. When the safety risk characterization value , it is a high-level risk; when the drone is at a low-level risk, the risk level is superimposed on the control interface layer and displayed in blue; when the drone is at a medium-level risk, the risk level is superimposed on the control interface layer and displayed in yellow, and at the same time, intermittent voice broadcasts are made to convey the risk level of the current area to the operator; when the drone is at a high-level risk, the risk level is superimposed on the control interface layer and displayed in red, and at the same time, intermittent voice broadcasts are made to convey the risk level of the current area to the operator. The operator actively adjusts the heading or altitude of the drone based on the voice prompt and visual feedback to ensure that it maintains a safe distance from the highway infrastructure and traffic flow without being affected.

[0012] As an optimization of the present invention, in step S5, the failure fall path domains caused by the power system failure and the control system failure of the drone are solved respectively. After obtaining the fall path domains of the power system failure and the control system failure of the drone, the convex hulls of the two different failure types are weighted and combined to obtain the comprehensive failure fall path domain of the drone at the current position; When solving the failure fall path domains caused by the power system failure and the control system failure of the drone, the parameter space ranges are set respectively. In the parameter space, N groups of parameter samples are generated using Latin hypercube sampling. For each group of parameters, based on the established six-degree-of-freedom dynamics model, the trajectory points of the drone are obtained using the fourth-order Runge-Kutta method, and then N trajectories are generated through the Monte Carlo sampling algorithm. Finally, the convex hull of the point set is calculated using the fast convex hull algorithm.

[0013] As an optimization of the present invention, the expression of the failure state risk value in step S5 is: ; Among them, represents the intersection volume of the comprehensive failure fall path domain obtained in the state at moment and the global safety boundary, represents the volume of the comprehensive failure fall path domain obtained in the state at moment.

[0014] Advantages and beneficial effects of the present invention: (1) By abstracting a high-speed moving vehicle fleet and its surrounding infrastructure into a three-dimensional geometric body that evolves dynamically over time, the present invention retains the core features while filtering redundant detailed data, and realizes the dynamic monitoring of the vehicle fleet and infrastructure with simple expressions.

[0015] (2) The present invention constructs an electromagnetic-physical double-layer adaptive safety boundary framework for highway traffic flow and its surrounding infrastructure. At the physical layer, a three-dimensional dynamic risk cube of vehicles and infrastructure is generated in real time by constructing an envelope body; at the electromagnetic layer, based on frequency band conflict judgment, the electromagnetic safety distance is dynamically calculated by comparing the communication frequency bands of devices and drones; a double-layer risk coupling analysis mechanism is designed to realize a dynamic safety boundary that can be adaptively adjusted according to the vehicle flow speed, density and environmental interference intensity.

[0016] (3) The present invention establishes a theoretical benchmark model based on the six-degree-of-freedom dynamic characteristics of drones, real-time monitors flight data, and dynamically compares the real-time monitored flight data with the equilibrium state calculated based on the remote control signal conversion of the operator to accurately quantify the possible failures of drones.

[0017] (4) The present invention introduces a fault margin index to establish a mapping physical model relationship between state deviation and fault probability, accurately quantifies the possible failures of drones, and breaks through the limitations of traditional risk identification methods in data dependence.

[0018] (5) The present invention breaks through the limitations of traditional single-fault models, constructs a fault probability model that fuses position and attitude dual parameters based on the Rayleigh distribution, realizes the joint quantification of power attenuation and control failure, and calculates the fault probability of drones in real time. When the fault probability exceeds a preset threshold, the possible fall trajectory of the drone is calculated, and the safety risk in the state where the drone may have a fault is expressed by calculating the fault state risk value.

[0019] (6) The present invention constructs a hierarchical fault response system, which refines the out-of-control state of the UAV into two modes: partial out-of-control and complete out-of-control. Different coping strategies are adopted for different out-of-control states, so that when the UAV is not completely out of control, it can fly to find a location with lower risk for landing; when the UAV is completely out of control and loses its flight ability, calculate the probability of collision with the highway traffic flow, enable the emergency response mechanism, and mobilize relevant resources for timely handling to ensure that flight safety can still be guaranteed to a certain extent when a full out-of-control occurs.

[0020] (7) The present invention breaks through the core problems such as insufficient quantification of the interaction between the UAV and the spatial facilities near the highway, low sensitivity to environmental disturbances, and rigid safety boundaries, and provides a solution that emphasizes both high precision and high efficiency for airspace safety monitoring and risk warning. Brief Description of the Drawings

[0021] Figure 1 It is a flowchart of the risk identification and emergency method for the UAV of the present invention when flying in the airspace adjacent to the highway. Detailed Embodiment

[0022] To enable those skilled in the art to better understand the technical solutions and advantages of the present invention, the present application will be described in detail below with reference to the accompanying drawings, but it is not used to limit the protection scope of the present invention.

[0023] As Figure 1 shown, the risk identification and emergency method for the UAV provided in this embodiment when flying in the airspace adjacent to the highway includes the following steps: Step S1. Data acquisition: Monitor and obtain the UAV body data, including the current angular velocity, flight speed, body weight, current position of the UAV, rotor moment of inertia, and the moment of inertia of the three axes (roll axis, pitch axis, yaw axis) in the global coordinate system of the UAV body, as well as the real-time control instructions issued by the operator. At the same time, collect the surrounding environment information, including wind speed vector and electromagnetic field distribution data, and perform three-dimensional dynamic modeling on the highway surrounding environment based on the multi-modal sensors carried by the UAV (such as lidar, depth camera, millimeter wave radar), and calculate the relative distance between the UAV and the infrastructure and the moving vehicle fleet in real time.

[0024] Step S2. Obtain the UAV failure probability: Step S2.1. Establish a dynamic model of the quadrotor UAV: In this embodiment, the body coordinate system of the UAV is a spatial rectangular coordinate system established with the position of the UAV as the base point, and the body coordinate system B Taking the geometric center of gravity of the quadrotor UAV as the origin, and the two rigid frames of the quadrotor are respectively axis and For the z-axis, the direction of the aircraft body can be confirmed according to the right-hand rule. Define the global coordinate system E with the origin at this point, and the direction pointing due north with this point as the central axis as the z-axis, and the direction obtained by rotating the x-axis counterclockwise by 90° as the x-axis. For the y-axis, the direction of the aircraft body can be confirmed according to the right-hand rule. The direction of the y-axis is perpendicular to the aircraft body and upward.

[0025] Based on the body coordinate system of the UAV, use the rotation matrix to convert to the global coordinate system. In this embodiment, this process is completed using the rotation matrix, so that the global coordinate system and the UAV body coordinate system can be mutually converted. Through three rotations around the coordinate axes, the conversion between the two coordinate systems can be achieved; Specifically, the expression of the rotation matrix is: ; where, is the rotation angle around the z-axis in the UAV body coordinate system, that is, the pitch angle; is the rotation angle around the x-axis in the UAV body coordinate system, that is, the roll angle; is the rotation angle around the y-axis in the UAV body coordinate system, that is, the yaw angle.

[0026] In this embodiment, the dynamic model of the quadrotor aircraft obtained by the Newton-Euler method is used as the dynamic basis. Taking yaw, roll, and pitch as the attitude angles, a six-degree-of-freedom motion equation of the UAV based on the mutual coupling of the attitude angles is established, and the expression is: ; This formula describes the motion equation of the quadrotor UAV, where the center-of-gravity position of the UAV is represented by , is the second-order differential of , is the mass of the UAV, is the resultant external force matrix composed of the forces acting on the UAV body in the global coordinate system, which is mainly related to the lift , gravity and wind disturbance received by the UAV in the body coordinate system. The expressions of gravity and wind disturbance are: ; where, is the acceleration due to gravity, and its value is taken as 9.8; ; ； where, is the disturbing force of the wind on the th rotor, is the air density, is the lift coefficient of the UAV under the wind, and is set as , is the relative velocity between the UAV rotor and the wind, is regarded as the projected area of the UAV rotor disk on the crosswind, , is the radius of a single rotor; It is set that is the velocity of the wind in the global coordinate system, then the expression for the relative velocity between the UAV rotor and the wind is: ; where, is the set of the three-dimensional velocities of the UAV in the global coordinate system Xg , Yg, Zg axis.

[0027] In this embodiment, is the set of the attitude angular velocities of the UAV, that is , is the first-order differential of, is the third-order identity matrix, , is used to describe the moments of inertia of the UAV about three directions in the body coordinate system, where is the moment of inertia of the UAV about the axis in the body coordinate system, is the moment of inertia of the UAV about the axis in the body coordinate system, is the moment of inertia of the UAV about the axis in the body coordinate system; represents the disturbance generated by the wind near the highway in the global coordinate system on the UAV, and is calculated by the following formula: ; where, is the anti-torque disturbance on the th rotor, and the expression is: ; where, is the torque coefficient of the UAV under the wind, and is set as , is the arm length of the UAV.

[0028] Considering gravity, the lift force on the UAV in the body coordinate system can be expressed as , where F1 to F4 respectively represent the lift forces provided by the four rotors. The total lift force is converted into the lift force in the global coordinate system. Define the rotational speeds of the rotors as the rotational speed of the rotor located at the front left of the body , the rotational speed of the rotor at the front right , the rotational speed of the rotor at the rear right , and the rotational speed of the rotor at the rear left . Then, the lift forces on the three axes of Xg , Yg, Zg in the global coordinate system are obtained , and its mathematical formula is expressed as: ; where is the lift coefficient, and its range is ; Based on the above-established dynamic model, the resultant external moment on the UAV in the body coordinate system can be expressed as: ; where represents the gyroscopic moments generated by the gyroscopic effects of the four rotors of the UAV on the roll, pitch, and yaw axes, which are set as , , in sequence, and its expression is ; is the moment of inertia of the UAV rotor (assuming that the moments of inertia of the four rotors are the same), is the sum of the total rotational speeds of the four rotors, represents the set of moments generated by the lift force provided by the UAV rotor along the , , axes of the body coordinate system; based on the UAV dynamics, the moment can be explicitly expressed as ; and , , are respectively the angular velocities about the , , three axes in the body coordinate system, , , are respectively , , the differentials of the angular velocities about the axes. The superscript T represents the transpose symbol. Then, the moment in the body coordinate system is converted into the moment in the global coordinate system ; Define the set , is the total lift force of the four rotors, and thus serves as the position altitude control variable. is the rotational torque of the front and rear channels. is the rotational torque of the left and right channels, and thus can serve as the pitch angle control variable and the roll angle control variable. is the rotational torque of the fuselage plane and can serve as the yaw angle control variable. The actual information of the UAV can be obtained through the rotor speeds of the UAV after a time window .

[0029]

[0030]

[0031]

[0032]

[0033] Among them, is the thrust coefficient of the UAV , is the anti-torque coefficient .

[0034] Therefore, the expression of the resultant external force matrix F composed of the forces acting on the UAV body in the global coordinate system is: ; Among them, is the disturbing force of the wind on the th rotor, is the acceleration due to gravity, and its value is taken as 9.8; In summary, the six-degree-of-freedom dynamic model of the quadrotor UAV can be expressed as: ; Among them, , , are the second-order differentials of the current center-of-gravity position of the UAV, , , are the second-order time differentials of , , , , are the first-order time differentials of , .

[0035] Step S2.2. Obtain the equilibrium state of the UAV: Based on the obtained input set, calculate the desired state of the drone for the operator or signal transmission, and set the time window The balanced state after is: ; In this embodiment, it is considered that the allowable delay for signal transmission is , that is, the drone can receive instructions at time . The moment is the current moment, is a very short period of time, with a value range of 0.05s - 0.2s, is the defined time window; after the drone receives the joystick signal from the operator, the received signal is normalized to obtain ; where is the vertical joystick (throttle), controlling the total lift , is the pitch joystick (front and back), controlling the pitch moment, is the roll joystick (left and right), controlling the roll moment, is the yaw joystick (rotation), controlling the yaw moment. Then, through linear mapping and physical constraints, the joystick signal is converted into the control quantity required by the dynamic model: Define the transformation matrix . The purpose of setting this matrix is to convert the joystick signal received from the operator into the matrix required for the rotor speed of the drone after receiving it, and it should be a 4×4 matrix; satisfying ; where is the square of the change in the rotational speed of the th rotor, that is, the change amount that should occur after receiving the operator's signal at time ; the transformation matrix should be calibrated for different drones.

[0036] Substitute the square of the above change in rotor speed into the six-degree-of-freedom dynamic model in step S2.1 to obtain the second-order differential change values of each state of the drone , , , , , .

[0037] Integrate the obtained parameters to obtain the speed of the drone in the balanced state:

[0038]

[0039]

[0040] The equilibrium position of the drone in the global coordinate system is:

[0041] where is the initial position of the drone at time and is the initial velocity in the global coordinate system; is the equilibrium position coordinates at time ; according to the built-in GPS positioning system of the drone, the position coordinates of the drone after time are recorded as and the velocity is ; then , , are integrated to obtain the angular velocity in the equilibrium state: where , , are the attitude angular velocities of the drone at time ; after deriving the attitude angles in the equilibrium state, the expression:

[0042] where is the set of attitude angles of the drone at time ; the equilibrium state of the drone in the global coordinate system at time is recorded as ; according to the gyroscope installed on the drone, the set of attitude angles of the drone in the global coordinate system at time is obtained as .

[0043] Step S2.3. Fault margin: To further quantify the relationship between the state deviation between the equilibrium state and the current state and the drone failure probability, in this embodiment, based on the real-time state deviation between the equilibrium state and the current state, a new risk judgment index, namely the fault margin, is introduced.

[0044] The instantaneous fault margin of the position loop is defined as: ; where represents the 2-norm in mathematics, represents the current position of the drone, that is , is the position of the drone in the equilibrium state; is the range of velocities in all directions that the drone can reach, is the drone at The set of actual speeds at a moment, is the set of speeds of the UAV in a balanced state at moment; and , is the time decay factor ( ); is the weight coefficient for controlling the importance degree of item, and its value range is ( ); The instantaneous fault margin of the attitude loop defined is: ; Among them, , represent the current attitude angle and attitude angular velocity of the UAV, and the parameters in the formula satisfy , , , are the attitude angle and attitude angular velocity of the UAV in a balanced state, is the maximum attitude angle that the UAV can reach, is the attitude range in all directions that the UAV can reach; and , is the time decay factor ( ), is the weight coefficient for controlling the importance degree of item, and its value range is ( ); Calculate the global fault margin ; among them, , both represent weight coefficients, and their values are greater than zero and should satisfy .

[0045] Step S2.4. Obtain the fault probability: Obtain the possible fault probability of the UAV according to different fault margins. In this embodiment, it is considered that the possible fault probability of the UAV follows a Rayleigh distribution. Obtain the current historical operation state information of the UAV through the on-board equipment of the UAV, extract the characteristic values, and obtain the Rayleigh distribution scale parameters and of the position and attitude angle by fitting the historical fault data to characterize the fault characteristic values of the UAV's driving behavior. The power fault probability and the attitude fault probability of the UAV can be expressed as:

[0046]

[0047] Based on the above two fault probabilities, perform weighted accumulation to obtain the comprehensive fault probability of the UAV , the expression is: ; Among them, and are weight coefficients, and the value range is 0.2 - 0.7. The value range is 0.2 - 0.7, and the sum of the two should be equal to 1.

[0048] Step S3. Construct the global safety boundary: Step S3.1. Generate the physical safety boundary: Based on the vehicle positions near the highway and the central coordinate positions of the highway surrounding infrastructure (toll stations, high-voltage wire towers, photovoltaic panels, etc.) at time draw up a collision box; ; ; Among them, refers to the vehicles and infrastructure on the highway (mainly high-voltage wires, photovoltaic panels, and facilities belonging to highway toll stations). Based on the acquired data, obtain the set of attribute coordinates of the collision box , where are the length, width, and height of the vehicle fleet or infrastructure.

[0049] Modeling and analysis of the motion state of the highway vehicle fleet; in this embodiment, obtain the vehicle information at the vehicle fleet end near the drone on the highway. The vehicle information includes the trajectory prediction of the autonomous vehicle for the next time, the position information of the vehicle, and the vehicle parameters; under the autonomous vehicle control system, the vehicle will predict the trajectory within a certain period in the future according to the current state (such as speed, position, and heading angle), as well as the future control inputs (such as acceleration and steering angle).

[0050] For example: Through the current state quantity of the th vehicle and the control input quantity , the motion trajectory at the future time can be obtained, and then the reachable domain model of the vehicle at time is constructed through numerical integration; among them, is the current ( time) position of the vehicle , is the heading angle of the vehicle at time (i.e., the angle of the vehicle driving direction relative to the global coordinate system), is the speed at time ; at The acceleration of time, For vehicles exist Front wheel steering angle at the moment.

[0051] In this embodiment, the time distance between the vehicle heads within the monitoring range of the drone is less than the threshold value. The continuous vehicles are defined as a cooperative convoy, and their future motion reachable domain is assumed to be a semi-ellipsoidal shape. The three-dimensional semi-ellipsoidal reachable domain of a vehicle is mathematically represented as: ; in, For the A car in The position coordinates at the time, is the square of the maximum achievable distance of the vehicle along the driving direction, ,in is the maximum acceleration of the vehicle in the time window, For vehicles in The speed of time, is the maximum distance that the vehicle can reach when turning, which is obtained from the turning radius of the vehicle, that is, ,in For the The wheelbase of the vehicle, For the A car in The front wheel steering angle at time is the design height of the vehicle, ,Right now For the The design height of the vehicle, Representative vehicle The location coordinate parameter variable.

[0052] The safety boundary of a single vehicle is composed of the possible vehicle trajectory vectors. The joint collision volume of the convoy of vehicles can be expressed as .

[0053] For the collision detection modeling of stationary infrastructure along the highway, its spatial representation adopts a double-layer cuboid structure system based on the global coordinate system. First, the core parameters of infrastructure collision are defined as the geometric center coordinates, i.e. The length, width and height of the facility , the safety boundary space domain constructed based on this is strictly defined by the three-dimensional coordinate constraints: ; in, A parameter variable representing the location coordinates of the infrastructure.

[0054] Step S3.2. Generate an electromagnetic safety boundary: Infrastructure near highways can also have a certain electromagnetic impact on the flight and communication of drones. Since the impacts caused by most infrastructure can be ignored, only the impacts of high-voltage power lines, photovoltaic panels, and toll station facilities around highways on drone communication are considered here.

[0055] In this embodiment, before generating the electromagnetic boundary, the following conditions should be met: The drone flies along the highway while simultaneously monitoring in real time the frequency bands emitted by vehicles traveling on the road, high-voltage power lines, photovoltaic panels, and facilities belonging to the highway toll stations beside the highway, and comparing them with the current communication frequency band of the drone.

[0056] Specifically, it should be able to identify various wireless communication or electromagnetic radiation frequency bands that may be used by the above four sources and conduct a comparative analysis with the communication frequency band of the drone itself; when it is found that there is an overlap in the communication frequency band of a certain device or infrastructure inside the vehicle, that is, there is a potential interference risk, an electromagnetic safety boundary is established with the infrastructure having the same communication frequency band as the drone as the geometric center.

[0057] Obtain the device with overlapping communication frequencies The maximum power emitted during communication , set the maximum power allowable at the drone to be , according to the free space path formula for communication propagation: ; where is the signal wavelength, is the reflection coefficient of the metal railing, and its value range is , the phase difference between the direct wave and the reflected wave, is the direct path distance, and the unknowns of the equation are obtained through experimental fitting by simulating the drone flight environment.

[0058] By solving the above equation inversely, the safety distance can be obtained, and the expression is: ; In the formula, the safety distance is the distance from the infrastructure emitting the communication frequency band to the drone. Through the pointing direction of this distance, it is projected by a rotation matrix to obtain the X g 、Y g 、Z g distance along the , , ; Based on the obtained safety distance , construct a geometric body with as the length, as the width, as the height, and obtain the electromagnetic safety boundary as: ; Among them, represents the position variable of the object that coincides with the communication frequency band of the UAV, , , represent the projected lengths of the safety distance calculated by the geometric body at time on the three coordinate axes of the global coordinate system; Synthesize all the established electromagnetic safety boundaries and perform a union operation. Suppose there are N devices that coincide with the communication frequency band of the UAV, then the electromagnetic safety boundary within the detectable range of the UAV at the current position is .

[0059] Step S4. UAV safety risk identification: Establish a UAV collision box. For the convenience of calculation, the volume of the UAV is assumed to be a cylinder. Let the maximum horizontal dimension of the UAV be , then the corresponding radius of the cylinder of its body is , and the height is , is half of the UAV height. According to solid geometry knowledge, , is the radius of the smallest circumscribed sphere of the UAV body cylinder. In subsequent calculations, the smallest circumscribed sphere is used as the UAV collision box.

[0060] When the current comprehensive failure probability of the UAV does not exceed the preset threshold, based on the physical safety boundary and electromagnetic safety boundary obtained in steps S3.2 - S3.3, compare the physical and electromagnetic safety boundaries respectively, and use the safety boundary obtained after the union operation as the global safety boundary ， It is expressed as: ， ， ; Introduce the safety risk characterization value to calculate the risk of the UAV at the current position. The expression is , represents the intersection volume of the UAV collision box and the global safety boundary , is the volume of the UAV collision box, is the weight factor, which characterizes the acceptable range of the UAV risk.

[0061] In this embodiment, a level threshold for the safety risk characterization value of the UAV is set and , that is, when the safety risk characterization value is in , it is a low-level risk. When the safety risk characterization value , it is a medium-level risk. When the safety risk characterization value , it is a high-level risk.

[0062] In this embodiment, when the comprehensive failure probability of the UAV during flight ≤ the preset threshold, the control right is in the hands of the operator. When the UAV is in a low-level risk, the risk level is superimposed on the control interface layer and displayed in blue. When the UAV is in a medium-level risk, the risk level is superimposed on the control interface layer and displayed in yellow. At this time, the system performs intermittent voice broadcasts every seconds to transmit the risk level of the current area to the operator; when the UAV is in a high-level risk, the risk level is superimposed on the control interface layer and displayed in red. At this time, the system performs intermittent voice broadcasts every seconds to transmit the risk level of the current area to the operator. The operator actively adjusts the UAV's heading or altitude based on the voice prompt and visual feedback to ensure a safe distance from the highway infrastructure and traffic flow without being affected; among them, the time parameters , should both be determined based on expert opinions in actual applications, while ensuring the control logic of the present invention and realizing the safe control of the UAV by the operator.

[0063] If the comprehensive failure probability of the UAV is greater than the preset threshold, automatic control is enabled and step S5 is executed; Step S5. Identify the failure type of the UAV, calculate the comprehensive failure fall path domain of the UAV at the current position based on the state of the UAV at the current moment, the power failure probability of the UAV, and the attitude failure probability, and obtain the failure state risk value of the UAV in the case of failure; then judge the degree of UAV out-of-control. If it is partially out of control, control the flight direction of the UAV to make it move forward in the direction with a lower failure state risk value. When the calculated failure state risk value is zero, the UAV is forced to land safely; if it is completely out of control, calculate the fall path domain of the current state, issue a risk warning to the vehicles near the landing location, and send a notice to the relevant department if there is infrastructure.

[0064] In this embodiment, a failure threshold should be determined in advance. When the comprehensive failure probability of the UAV itself exceeds the failure threshold at a certain time t, it is considered that the UAV may have a failure, track the UAV state, and obtain the failure type ratio and probability.

[0065] Specifically, in this embodiment, the fault types are divided into two categories, namely power system faults and control system faults. Drone crash models caused by these two types of faults are respectively constructed, and the comprehensive fault crash path domain is solved based on the current position of the drone.

[0066] When describing power system faults, it is assumed that the pitch, roll, and yaw angles of the drone can reach an equilibrium state at time However, in the case of power system faults, the thrust of the four rotors of the drone will experience a certain attenuation, resulting in a gradual failure of the drone's motion state.

[0067] The continuous domain of the crash path caused by power system faults of the drone is expressed as , that is, all possible power fault crash trajectories of the drone from to within the time; where is the crash time of the drone, and is the parameter set.

[0068] In this embodiment, the continuous domain of the crash path is calculated based on some parameters at the current time ( time), including the real-time state at the current time ( time) , environmental parameters: , where is the air density, is the component in the X , Y g , Z g , Z g axes of the global coordinate system at time , where is the power attenuation rate of the drone's power system after a fault in the global coordinate system; this power attenuation rate will affect the control input, that is , where , is the control input considering power attenuation; define the parameter space with a range of ; Based on the set parameter space range, N groups of parameter samples are generated using Latin Hypercube Sampling (LHS) within this parameter space: ; For each group of parameters, the dynamic differential equation is solved to obtain the drone trajectory route under this group of parameters, is the trajectory point, is the drone trajectory point under the parameter , including the state of the drone such as position and speed, where It is the six - degree - of - freedom dynamic equation (including position, velocity, and angular velocity equations) of the UAV in step S2, ; specifically, the fourth - order Runge - Kutta method is used for numerical solution; through this iterative process, the state of the UAV from to different time steps within the time can be obtained.

[0069] In this embodiment, based on the fourth - order Runge - Kutta method in the above steps, the paths of the UAV at different time steps are obtained, and trajectories are obtained through Monte Carlo simulation. The positions of the trajectories simulated in the current state (the state at time) are expressed as: , and the QuickHull algorithm is used to calculate the convex hull of the point set, and the mathematical representation is as follows: ; where is the vertex of the convex hull, is the weight coefficient, is the number of vertices of the convex hull, is the position of the trajectories simulated, that is, the convex hull parameters of the power system failure.

[0070] In this embodiment, through the above convex hull calculation, the spatial coverage area of the falling path obtained in the current state (the state at time) is obtained, that is, the failure falling path domain caused by the power system failure of the UAV. Let .

[0071] The control system failure of the UAV causes the attitude angles (pitch angle, roll angle, and yaw angle) of the UAV to be disordered, thus triggering the falling phenomenon of the UAV. The control system failure makes the UAV unable to maintain its original flight attitude, resulting in the flight trajectory gradually deviating from the original orbit and forming a falling path different from that of the power system failure when it cannot be restored. When considering the falling of the UAV due to control system failure, it is assumed that the UAV maintains its current direction and speed unchanged.

[0072] To quantify the trajectory range of such out - of - control motion, the continuous domain of the falling path caused by the control system failure is also defined as , that is, all possible control - failure falling trajectories of the UAV from to within the time; where is the parameter set, including the real - time state at time, environmental parameters: , failure parameters: . Among them is the power attenuation rate after the failure of the UAV control system in the global coordinate system; this power attenuation rate will affect the control input, that is , among which , is the control input considering power attenuation; define the parameter space The range of is , that is, the range between the current state of the UAV state and fault parameters and the equilibrium state.

[0073] Based on the set parameter space range, N groups of parameter samples are generated using Latin Hypercube Sampling (LHS) within this parameter space: ; Similarly, for each group of parameters, solve the differential equation to obtain the UAV trajectory route under this group of parameters, and also use the fourth-order Runge-Kutta method for numerical solution; obtain trajectories through Monte Carlo simulation. Similarly, here is , the position representation of trajectories simulated under the current state (the state at moment): , use the QuickHull algorithm to calculate the convex hull of the point set, and the mathematical representation is as follows: ; Among them, is the vertex of the convex hull, is the weight coefficient, is the number of vertices of the convex hull, is the position of trajectories simulated under the current state (the state at moment), that is, the convex hull parameters of the control system failure.

[0074] In this embodiment, through the above convex hull calculation, the spatial coverage area of the falling path obtained under the current state (the state at moment) is obtained, that is, the fault falling path domain caused by the UAV control system failure. Let .

[0075] After obtaining the falling path domains of the UAV power system failure and the control system failure, the convex hulls of the two different fault types are weighted and combined to obtain the comprehensive fault falling path domain of the UAV at the current position. The expression is: ; Among them, is at the current state ( The comprehensive failure fall path domain obtained under the state at a moment 、 are respectively and variables in [[ ]] (vertices of the convex hull) and and are weight coefficients, and the weight coefficients should satisfy , 。

[0076] In this embodiment, when the comprehensive failure probability of the UAV exceeds the preset threshold, based on the UAV collision box constructed in step S4, the failure state risk value is calculated , represents the intersection volume of the volume of the comprehensive failure fall path domain obtained under the current state ( state at a moment) and the global safety boundary, represents the volume of the comprehensive failure fall path domain obtained under the current state ( state at a moment).

[0077] In this embodiment, when the UAV executes the flight mission controlled by the operator near the highway, it will continuously monitor and calculate its own state deviation, dynamically evaluate the current power failure probability and attitude failure probability of the UAV based on this state deviation, and calculate the current comprehensive failure probability; when the comprehensive failure probability exceeds the preset threshold ,and continuously exceeds , judge its failure degree. When the comprehensive failure probability is less than , greater than , it is considered that the UAV is out of control at this time but has the function of continuing to fly, that is, in a partial out-of-control state. At this time, calculate the failure fall path domain of the UAV under the possible failure, and judge the failure state risk value of the UAV under the possible failure; after obtaining the failure state risk value, control the flight direction of the UAV to make it move forward in the direction with a lower failure state risk value. When the failure state risk value is zero, force the UAV to land safely.

[0078] When the comprehensive failure probability of the UAV is greater than or equal to , it is considered that the power of the UAV is completely out of order. Based on the real-time state parameters and the failure fall path domain containing all possible fall trajectories generated by the dynamic model, map this area (the comprehensive failure fall path domain) to the global coordinate system to form the final fall impact boundary. Then calculate the overlapping volume of this dynamically updated boundary and the global safety boundary. If it is detected that there is a geometric intersection between the fall area and the safety boundary of any vehicle or infrastructure, immediately send an avoidance instruction to the surrounding vehicles and start the relevant protection mechanism for the key infrastructure, and continuously iteratively update the prediction result of the fall area until a collision occurs or the emergency measure takes effect.

[0079] In this embodiment, the above parameters (range 0 - 1 s), (range 0 - 0.3), (range 0.3 - 0.6) should all be formulated based on expert opinions under actual application conditions, while ensuring the control logic of the present invention and achieving safe control of the drone.

[0080] The above uses specific examples to elaborate on the present invention, which is only used to help understand the present invention and is not intended to limit the present invention. For those skilled in the technical field to which the present invention pertains, several simple deductions, deformations, or substitutions can also be made according to the idea of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claimed rights.

Claims

1. Risk identification and emergency method for drones flying in the airspace adjacent to expressways, characterized in that, The method includes the following steps: Step S1: Data collection, obtaining the ambient data around the UAV and the basic information of the vehicles traveling nearby; Step S2: Establish a six-degree-of-freedom dynamic model including wind force; convert the joystick signals at the operator's position into changes in kinematic states, calculate the equilibrium state of the UAV based on the established six-degree-of-freedom dynamic model, compare the current state of the UAV obtained in real time with the equilibrium state, obtain the state deviation between the equilibrium state and the current state, calculate the instantaneous fault margin of the position loop and the instantaneous fault margin of the attitude loop based on the state deviation, so as to obtain the power fault probability, attitude fault probability and the current comprehensive fault probability of the UAV; Step S3: Construct the physical safety boundaries of the vehicle fleet and infrastructure; at the same time, perform real-time detection on the frequency bands emitted by the vehicles and infrastructure traveling on the road and compare them with the current communication frequency band of the UAV. If it coincides with the current communication frequency of the UAV, an electromagnetic safety boundary is generated; generate a global safety boundary based on the physical safety boundary and the electromagnetic safety boundary; Step S4: Establish a UAV collision box, and at the same time determine whether the comprehensive fault probability exceeds a preset threshold. If it does not exceed the preset threshold, the operator controls the UAV, and at the same time identifies the current safety risk of the UAV based on the UAV collision box and the global safety boundary in Step S3, and gives corresponding warnings to the operator according to the safety risk level; if the comprehensive fault probability exceeds the preset threshold, execute Step S5; Step S5: Identify the fault type of the UAV, calculate the comprehensive fault fall path domain of the UAV at the current position based on the state of the UAV at the current moment, the power fault probability of the UAV, and the attitude fault probability, and obtain the fault state risk value of the UAV in the case of a fault; then judge the degree of UAV out-of-control according to the comprehensive fault probability. If it is partially out of control, control the flight direction of the UAV to make it move forward in the direction with a low fault state risk value. When the calculated fault state risk value is zero, the UAV is forced to land safely; if it is completely out of control, map the comprehensive fault fall path domain to the global coordinate system to form the final fall impact boundary, and based on the global safety boundary and the formed final fall impact boundary, issue an avoidance instruction to the vehicles around the fall location, and perform protection if there is infrastructure.

2. The risk identification and emergency response method for the drone flying in the airspace adjacent to the highway according to claim 1, characterized in that The six-degree-of-freedom dynamic model established in Step S2 is: ； Among them, , , are the current center of gravity positions of the UAV, is the second-order differential of , , is , , is the second-order time differential of , is , is the first-order time differential of is the rotation angle around the axis in the UAV body coordinate system, is the rotation angle around the axis in the UAV body coordinate system, is the rotation angle around the axis in the UAV body coordinate system, is the moment of inertia of the UAV around the axis in the body coordinate system, is the moment of inertia of the UAV around the axis in the body coordinate system is the moment of inertia of the UAV around the axis in the body coordinate system, is the mass of the UAV, is the total lift of the four rotors; is the rotational torque of the front and rear channels, which is the pitch angle control quantity; is the rotational torque of the left and right channels, which is the roll angle control quantity; is the rotational torque of the fuselage plane, which is the yaw angle control quantity, represents the disturbance generated by the wind near the highway on the UAV in the global coordinate system, and the expression is: ； Among them, is the disturbing force of the wind on the th rotor, is the anti-torque disturbance on the th rotor, is the arm length of the UAV.

3. The risk identification and emergency method of the drone according to claim 2 during flight in the airspace adjacent to the highway, characterized in that, After receiving the joystick signal from the operator in step S2, the received signal is normalized, and then through the transformation matrix , the joystick signal is converted into the change amount of the rotor speed required by the six-degree-of-freedom dynamics model. Then, based on the established six-degree-of-freedom dynamics model and the change amount of the rotor speed, the second-order differential change values of each state of the UAV are calculated. By integrating the second-order differential change values of each state of the UAV, the speed, position, attitude angular velocity, and attitude angle of the UAV in the equilibrium state are obtained.

4. The risk identification and emergency response method for the drone flying in the airspace adjacent to the highway according to claim 3, characterized in that, The instantaneous fault margin of the position loop in Step S2 is: ； Among them, represents the 2-norm in mathematics, represents the current position of the UAV, that is , is the position of the UAV in the balanced state; is the range of velocities in all directions that the UAV can reach, is the set of actual velocities of the UAV at time, is the set of velocities of the UAV in the balanced state at time; and , is the time decay factor; is the weight coefficient for controlling the importance degree of terms. The instantaneous fault margin of the attitude loop is: ； Among them, , represent the current attitude angle and attitude angular velocity of the UAV, satisfying , , , are the attitude angle and attitude angular velocity of the UAV in the balanced state, is the maximum attitude angle that the UAV can reach, is the attitude range in all directions that the UAV can reach; and , is the time decay factor, is the weight coefficient for controlling the importance of the item; Power failure probability of the UAV and attitude failure probability are expressed as: ； ； The comprehensive fault probability of the UAV is expressed as: ； Among them, and are respectively the Rayleigh distribution scale parameters of the position and the attitude angle, and are weight coefficients, and the value range is 0.2 - 0.

7. The value range is 0.2 - 0.7, and the sum of the two should be equal to 1.

5. The risk identification and emergency method of the drone according to claim 4 during flight in the airspace adjacent to the highway, characterized in that, The construction method of the physical safety boundary of the vehicle fleet in Step S3 is: Obtain the vehicle information of the vehicle fleet on the highway, predict the trajectory within a future period according to the current state of the vehicle and future control inputs, and construct a three-dimensional semi-ellipsoidal reachable domain for each vehicle respectively , then for a vehicle fleet with a total of vehicles, the combined collision body is represented as ; The construction method of the physical security boundary of the infrastructure is as follows: The geometric center coordinates of the infrastructure are , the length, width and height of the infrastructure itself , and the security boundary space domain constructed based on this is strictly defined by the three-dimensional coordinate constraint conditions as: ； Among them, represents the location coordinate parameter variable of the infrastructure; The infrastructure includes high-voltage wires, photovoltaic panels and highway toll stations.

6. The risk identification and emergency method for the drone flying in the airspace adjacent to the highway according to claim 5, characterized in that The method for generating the electromagnetic safety boundary in step S3 is as follows: Obtain the device with overlapping communication frequencies The maximum power emitted during communication , set the maximum allowable power at the UAV as , and reverse-solve the communication propagation free space path formula to obtain the safety distance , and the expression is: ； In the formula, is the signal wavelength, is the reflection coefficient of the metal railing, and its value range is , and the safety distance is the distance between the drones pointing to the infrastructure that emits the communication frequency band. Through the pointing direction of the safety distance, its projection is obtained through a rotation matrix to obtain the X g 、Y g 、Z g distance on the , , ; Based on the obtained safety distance , construct a geometric body with as the length, as the width, as the height. The electromagnetic safety boundary obtained is as follows: ； Among them, represents the position variable of an object whose communication frequency band overlaps with that of the UAV, , , represent the projected lengths of the safety distance calculated at time on the three coordinate axes of the global coordinate system; Synthesize all the established electromagnetic safety boundaries and perform a union operation. Suppose there are N devices whose communication frequency bands overlap with that of the drone. Then the electromagnetic safety boundary within the detectable range of the drone at the current position is .

7. The risk identification and emergency method for the drone flying in the airspace adjacent to the highway according to claim 6, wherein The method for establishing the collision box of the drone in step S4 is to set the volume of the drone as a cylinder. Let the maximum horizontal dimension of the drone be , then the corresponding radius of the cylinder of the fuselage is , and the height is , is half of the height of the drone, , is the radius of the smallest circumscribed sphere of the cylinder of the drone fuselage. Take the smallest circumscribed sphere as the collision box of the drone.

8. The risk identification and emergency response method for the drone flying in the airspace adjacent to the highway according to claim 7, characterized in that, The calculation method of the safety risk in Step S4 is: ； Among them, is the safety risk characterization value of the UAV at the current position, represents the intersection volume of the UAV collision box and the global safety boundary ; is the volume of the UAV collision box, is the weight factor; Set the level threshold of the safety risk of the drone in step S4 and , when the safety risk characterization value is in , it is a low-level risk. When the safety risk characterization value , it is a medium-level risk. When the safety risk characterization value , it is a high-level risk; when the drone is at a low-level risk, superimpose the risk level on the control interface layer and display it in blue; when the drone is at a medium-level risk, superimpose the risk level on the control interface layer and display it in yellow, and at the same time conduct intermittent voice broadcasts to convey the risk level of the current area to the operator; when the drone is at a high-level risk, superimpose the risk level on the control interface layer and display it in red, and at the same time conduct intermittent voice broadcasts to convey the risk level of the current area to the operator. The operator actively adjusts the heading or altitude of the drone according to the voice prompt and visual feedback to ensure that it maintains a safe distance from the highway infrastructure and traffic flow without being affected.

9. The risk identification and emergency response method for the drone flying in the airspace adjacent to the highway according to claim 8, wherein In Step S5, the fault fall path domains caused by the power system fault and the control system fault of the UAV are solved respectively. After obtaining the fall path domains of the UAV power system fault and the control system fault, the convex hulls of the two different fault types are weighted and combined to obtain the comprehensive fault fall path domain of the UAV at the current position; When solving the failure fall path domain caused by the power system failure and the control system failure of the UAV, the parameter space range is set respectively, and N groups of parameter samples are generated by Latin hypercube sampling in the parameter space. For each group of parameters, based on the established six-degree-of-freedom dynamic model, the trajectory points of the UAV are obtained by using the fourth-order Runge-Kutta method, and then N trajectories are generated by using the Monte Carlo sampling algorithm. Finally, the convex hull of the point set is calculated by using the fast convex hull algorithm.

10. The risk identification and emergency response method for the drone flying in the airspace adjacent to the highway according to claim 9, wherein, The expression of the failure state risk value in step S5 is: ； Among them, represents the intersection volume of the comprehensive failure fall path domain obtained in the state at time and the global safety boundary, represents the volume of the comprehensive failure fall path domain obtained in the state at time.

Images