A trajectory tracking and safe obstacle avoidance method for high-speed fixed-wing unmanned aerial vehicles

Abstract

The present application relates to a kind of trajectory tracking and safety barrier avoidance method for high-speed fixed-wing unmanned aerial vehicle, belong to unmanned aerial vehicle aircraft control technical field.The present application is to solve the problem of high-speed fixed-wing unmanned aerial vehicle trajectory tracking, barrier avoidance difficult, traditional algorithm is safe, precision, real-time conflict.The steps of the present application include: S1: generating global reference path using artificial potential field method;S2: establish fixed-wing unmanned aerial vehicle three-dimensional kinematics discrete model;S3: build NMPC-CBF path tracking and safety barrier avoidance controller;S4: deploy the above trajectory tracking and safety barrier avoidance controller on fixed-wing unmanned aerial vehicle, so that unmanned aerial vehicle realizes accurate trajectory tracking and flight safety guarantee in barrier environment.

Description

Technical Field

[0001] This invention relates to a trajectory tracking and obstacle avoidance method for high-speed fixed-wing unmanned aerial vehicles (UAVs), belonging to the field of UAV control technology. Background Technology

[0002] Fixed-wing unmanned aerial vehicles (UAVs) are increasingly widely used in reconnaissance and surveillance, geographic mapping, and logistics transportation due to their advantages such as long endurance, high speed, and large payload. However, the actual flight environment of UAVs is becoming increasingly complex, and UAVs face multiple challenges such as nonlinear aerodynamic characteristics, actuator saturation, and obstacle collisions, which places higher demands on the accuracy and safety of trajectory tracking.

[0003] Currently, trajectory tracking and obstacle avoidance methods for fixed-wing UAVs are mainly divided into three categories: The first category is based on geometric guidance or vector field methods, such as nonlinear guidance laws and vector field path following. These methods are computationally simple and easy to implement in engineering, but they are difficult to explicitly handle the dynamic constraints of UAVs, such as minimum turning radius, speed range, and attitude angle limits. When avoiding obstacles on complex paths or at close range, they are prone to violating constraints, leading to trajectory deviation or collisions. The second category is based on model predictive control (MPC) methods, which solve for the optimal control sequence that satisfies the constraints online through rolling optimization. This can effectively handle multiple constraints and nonlinear characteristics. However, traditional MPC uses a "soft constraint" form. When path tracking accuracy and safe obstacle avoidance conflict, safety is often sacrificed to ensure optimal accuracy, causing the UAV to enter dangerous areas. The third category is a safety control method that combines control obstacle function (CBF). CBF can theoretically ensure that the system state is always in a safe area by constructing a safety set and applying inequality constraints. However, CBF alone lacks joint optimization of path tracking accuracy and system dynamic constraints, and the control commands may be too conservative or jittery. Summary of the Invention

[0004] This invention addresses the challenges of trajectory tracking and obstacle avoidance for high-speed fixed-wing UAVs, as well as the conflicts between security, accuracy, and real-time performance inherent in traditional algorithms. Therefore, it proposes a trajectory tracking and safe obstacle avoidance method for high-speed fixed-wing UAVs.

[0005] The technical solution adopted by the present invention to solve the above problems is as follows: The present invention includes the following steps: Step 1: Generate a global reference path based on the waypoint sequence and obstacle information in the flight environment, combined with the artificial potential field method; Step 2: Based on the global reference path and the kinematic characteristics of the fixed-wing UAV, establish a three-dimensional kinematic discrete model of the fixed-wing UAV; Step 3: Based on the three-dimensional kinematic discrete model, construct a nonlinear model predictive control optimization problem; Step 4: Based on the distance between the fixed-wing UAV and the obstacle, construct the control obstacle function constraint, and embed the control obstacle function constraint into the nonlinear model predictive control optimization problem to obtain the fusion optimization problem; Step 5: Based on the fusion optimization problem, solve for the control variables and apply them to the fixed-wing UAV.

[0006] Furthermore, in step 1, the waypoint sequence includes the north and east coordinates and altitude of the UAV's starting point, intermediate point, and target point; obstacle information includes the intermediate position, radius, and set minimum safe distance. An artificial potential field is established based on the current position of the drone, waypoint positions, Euclidean distance between the drone and the waypoint, obstacle positions, and Euclidean distance between the obstacle and the drone. The artificial potential field method models obstacles and desired waypoints in the environment as repulsive and attractive points, respectively. The artificial potential field includes the gravitational field generated by the waypoint sequence and the repulsive field generated by the obstacles. The gravitational potential energy parameter in the gravitational field is... and The repulsive potential energy function in the repulsive field is: ; The net force and positional changes of the UAV in the environment are obtained based on the gravitational and repulsive fields; Based on the resultant force received by the UAV, a reference path is generated pointing to by the resultant force. The global reference path of the UAV is obtained by combining the UAV's desired speed, yaw angle, and pitch angle. ; The gravitational potential energy parameter is and for: (1); In formula (1), X The location of the drone. The location of the waypoint The distance between the drone and the waypoint is expressed in Euclidean form. This is the gravitational gain coefficient. Let represent the gravitational potential energy of the drone at its current location. The direction of gravity; Repulsive potential energy parameters for: (2); In formula (2), The repulsive gain coefficient is... q Indicates the location of the drone. Location of the obstacle. The distance between the drone and the obstacle. Minimum safe distance; The net force acting on the drone in the environment and its positional change are as follows: (3); In formula (3), The net force acting on the current position. To determine the location of the drone for the next step. The current location of the drone. Step size; Global reference path of drone for: (4); In formula (4), These are the expected speed, yaw angle, and pitch angle of the UAV, respectively.

[0007] Furthermore, step 2 specifically includes: The state vector is determined based on the UAV's three-dimensional position, ground speed, yaw angle, pitch angle, and acceleration in the inertial frame. With control quantity ; Based on the state vector, control quantity, and Euler discretization method, a three-dimensional kinematic discretization model is established, and the state constraints and input constraints of the model are defined. State vector With control quantity The expression is: (5); In formula (5), This represents the three-dimensional position of the UAV in an inertial frame. V For ground speed, These are the yaw angle and the pitch angle, respectively. a For acceleration; The three-dimensional kinematic discrete model is as follows: (6); In formula (6), , , , , and They are respectively k Real-time three-dimensional position, ground speed, yaw angle, and pitch angle of the drone. , and These are yaw rate, pitch rate, and acceleration, respectively. The dynamic differential equations of the UAV; The expression for the state constraint is: (7); In formula (7), and The minimum and maximum flight speeds of the drone. for k Flight speed at any moment Minimum yaw angle, for k Yaw angle at any moment For the maximum yaw angle, Minimum climb angle, for k The angle of ascent at any moment, Maximum climb angle, Minimum flight altitude for k Flight altitude at any given moment Maximum flight altitude; The expression for the input constraint is: (8); In formula (8), This represents the minimum pitch rate of the drone. This is the maximum pitch rate of the drone. Let yaw rate be the yaw rate of the drone. This represents the minimum yaw rate of the drone. This represents the maximum yaw rate of the drone. Let yaw rate be the yaw rate of the drone. The minimum acceleration of the drone. This represents the maximum acceleration of the drone.

[0008] Furthermore, step 3 specifically includes: Step 3.1: Set the prediction time domain of the NMPC controller for N Step, control time domain for M Step, and ; Step 3.2: Based on the prediction time domain and the control time domain, define a cost function to measure the trajectory tracking accuracy and control energy consumption. Based on the cost function, the three-dimensional kinematic discrete model, state constraints, and input constraints, construct a nonlinear model predictive control optimization problem. The expression for the cost function is: (9); In formula (9), for k+i The UAV state reference at time step, where Q is the state weight matrix, N is the prediction interval length, and R is the control input weight matrix. for k The later time prediction i The state at any given moment, for k+i The amount of control at any given moment; The nonlinear model predictive control optimization problem is as follows: (10); In formula (10), for k+M- Control quantity at time 1.

[0009] Furthermore, step 4 specifically includes: Step 4.1: Establish an obstacle control function based on the real-time position of the fixed-wing UAV, the position of the obstacle, the threat radius, and the safe distance; Step 4.2: Embed the obstacle control function into the NMPC controller to obtain the fusion optimization problem; The obstacle control function is: (11); In formula (11), p This is the real-time location of the drone. For the first m The location of the obstacle. For the first m The threat radius of an obstacle For the first m Safe distance from each obstacle For state space dimension; The expression for the fusion optimization problem is: (12); In formula (12), , A safety set for the state and constraints of unmanned aerial vehicles (UAVs). for t Current state for t Time Prediction t+k The state at any given moment, for t+k The amount of control at any given moment The state at the current moment. For the boundary convergence factor of the safe set, For the present t+k The convergence of the safety boundary at any given time. for t+k The safety boundary is predicted at any time.

[0010] Furthermore, step 5 specifically includes: Step 5.1: Solve the optimal control sequence obtained from the previous sampling time. Discard the first control variable and shift the remaining control sequence one position to the left, repeat the last control variable at the end, and obtain the initial value for hot start at the current sampling time; Step 5.2: Apply control barrier function constraints only to the current sampling time, and do not apply control barrier function constraints to other times in the prediction time domain, to obtain the simplified fusion optimization problem; Step 5.3: Based on the initial values ​​of the hot start, solve the simplified fusion optimization problem to obtain the control variables and apply them to the fixed-wing UAV; Step 5.4: Repeat steps 3, 4 and 5.1-5.3 to change the constraint strength and controller parameters of the UAV in an obstacle environment, and observe and verify the trajectory tracking accuracy and obstacle avoidance safety of the controller. The expression for the optimal control sequence is: (13); The initial values ​​for warm-start at the current sampling time are: (14); In formulas (13) and (14), for k The applied control variable at time -1 for k Application control variables at any time for k The applied control variable at time +1 for k+M The applied control variable at time -1.

[0011] The beneficial effects of this invention are: 1) Safety assurance: This invention embeds a distance-based obstacle control function as a hard constraint into NMPC, forcing the UAV to maintain a distance greater than a safety threshold from obstacles, thus solving the problem of "soft constraints being easily violated" in traditional NMPC; 2) The unity of tracking accuracy and safe flight: NMPC is responsible for minimizing path tracking error, while CBF is responsible for safety fallback, breaking the contradiction between "accuracy and safety" in traditional methods.

[0012] 3) Path smoothness: The reference path generated by the artificial potential field method in this invention is continuous and smooth, which avoids the NMPC outputting drastic control commands due to abrupt changes in the reference path, and is beneficial to attitude stability under high-speed flight.

[0013] 4) Real-time optimization: The present invention adopts a hot start strategy and a simplified method of applying CBF constraints only to the current step, which greatly reduces the online calculation time of NMPC and meets the real-time control requirements of high-speed fixed-wing UAVs.

[0014] 5) Parameter adjustability: This invention can flexibly balance conservative obstacle avoidance and aggressive tracking by adjusting the CBF attenuation factor, and can adapt to different mission scenarios by changing the flight speed constraint. Attached Figure Description

[0015] Figure 1 A flowchart of a trajectory tracking and safe obstacle avoidance method for high-speed fixed-wing UAVs; Figure 2 A diagram showing the results of three-dimensional path planning using the artificial potential field method; Figure 3 A 3D path tracing result diagram for classic NMPC; Figure 4 This is a graph showing the minimum distance change during classic NMPC path tracing. Figure 5 A comparison chart of NMPC-CBF path tracing results under different constraint strengths; Figure 6 A comparison chart of the minimum distance variation during the NMPC-CBF path tracing process under different constraint strengths; Figure 7 A comparison chart of path tracking results at different flight speeds; Figure 8 A comparison chart of minimum distance changes during path tracking at different flight speeds; Figure 9 A graph showing the solution time of a traditional NMPC-CBF controller during path tracing. Figure 10 A graph showing the solution time recording process of the NMPC-CBF controller during path tracing. Detailed Implementation

[0016] like Figure 1 As shown, the steps of the trajectory tracking and obstacle avoidance method for high-speed fixed-wing UAVs described in this embodiment include: S1: Generate a global reference path using the artificial potential field method; S101: Acquire waypoint sequence and static obstacle information in the flight environment. The waypoint sequence includes the north coordinates, east coordinates, and altitude of the starting point, intermediate points, and target point; the information for each obstacle includes its center position, radius, and set minimum safe distance.

[0017] S102: Construct an artificial potential field. The waypoint sequence generates a gravitational field. The gravitational potential energy function is set as follows: (1); In formula (1), X The location of the drone. The location of the waypoint The distance between the drone and the waypoint is expressed in Euclidean form. This is the gravitational gain coefficient. Let represent the gravitational potential energy of the drone at its current location. The direction of gravity.

[0018] The obstacle generates a repulsive field, and the repulsive potential energy function is set as follows: (2); In formula (2), The repulsive gain coefficient is... q Indicates the location of the drone. Location of the obstacle. The distance between the drone and the obstacle. This is the minimum safe distance.

[0019] The net force and positional change of a drone in the environment can be expressed as: (3); In formula (3), The net force acting on the current position. To determine the location of the drone for the next step. The current location of the drone. The step size.

[0020] S103: Generate a reference trajectory sequence. Obstacles and desired waypoints in the environment are modeled as repulsive and attractive points using the artificial potential field method, generating a reference path pointed to by the resultant force. , where is the desired three-dimensional position of the drone, and Let the desired speed, yaw angle, and pitch angle of the drone be given, where the angles can be defined with zero initial value, and the speed is a fixed value. The drone's state reference can then be obtained as follows: (4); In formula (4), These are the expected speed, yaw angle, and pitch angle of the UAV, respectively.

[0021] S2: Establish a three-dimensional kinematic discrete model of a fixed-wing UAV; S201: The state vector and control variables of the UAV kinematics are selected as follows: (5); In formula (5), This represents the three-dimensional position of the UAV in an inertial frame. V For ground speed, These are the yaw angle and the pitch angle, respectively. a For acceleration; S202: Based on the kinematic equations of the fixed-wing UAV and the Euler discretization method, the discrete three-dimensional kinematic model is obtained as follows: (6); In formula (6), , , , , and They are respectively k Real-time three-dimensional position, ground speed, yaw angle, and pitch angle of the drone. , and These are yaw rate, pitch rate, and acceleration, respectively. The dynamic differential equations of the UAV; S203: Define model constraints, including state constraints and input constraints. The state constraints are as follows: (7); In formula (7), and The minimum and maximum flight speeds of the drone. for k Flight speed at any moment Minimum yaw angle, for k Yaw angle at any moment For the maximum yaw angle, Minimum climb angle, for k The angle of ascent at any moment, Maximum climb angle, Minimum flight altitude for k Flight altitude at any given moment Maximum flight altitude; The input constraints are as follows: (8); In formula (8), This represents the minimum pitch rate of the drone. This is the maximum pitch rate of the drone. Let yaw rate be the yaw rate of the drone. This represents the minimum yaw rate of the drone. This represents the maximum yaw rate of the drone. Let yaw rate be the yaw rate of the drone. The minimum acceleration of the drone. This represents the maximum acceleration of the drone.

[0022] S3: Construct the NMPC-CBF path tracing and obstacle avoidance controller; S301: Design NMPC controller parameters, including setting the prediction time domain and control time domain. (Setting the prediction time domain...) for N Step, control time domain for M Step, to satisfy .

[0023] S302: Define the cost function and the optimization problem. The cost function is used to measure trajectory tracking accuracy and control energy consumption. (9); In formula (9), for k+i The UAV state reference at time step, where Q is the state weight matrix, N is the prediction interval length, and R is the control input weight matrix. for k The later time prediction i The state at any given moment, for k+i The amount of control at any given moment; For each sampling time k Based on the current measured state The optimization problem is as follows: (10); In formula (10), for k+M- Control quantity at time 1.

[0024] And it satisfies the state and input constraints in S202.

[0025] S303: Construct a control barrier function and embed it into NMPC. For the first... m For each obstacle, first define the safety set and the obstacle function: (11); In formula (11), p This is the real-time location of the drone. For the first m The location of the obstacle. For the first m The threat radius of an obstacle For the first m The safe distance from each obstacle.

[0026] Embedding CBF constraints into the NMPC controller creates the NMPC-CBF optimization problem: (12); In formula (12), , A safety set for the state and constraints of unmanned aerial vehicles (UAVs). for t Current state for t Time Prediction t+k The state at any given moment, for t+k The amount of control at any given moment The state at the current moment. For the boundary convergence factor of the safe set, For the present t+k The convergence of the safety boundary at any given time. for t+k The safety boundary is predicted at any time.

[0027] S303: Optimize the controller startup strategy to meet the application requirements of high-speed flight of fixed-wing UAVs. First, the problem iterative solution adopts a hot-start initial value strategy. Let the previous sampling time be... The optimal control sequence obtained by solving is: (13); The current moment k The initial value guess for optimization is set as follows: (14); In formulas (13) and (14), for k The applied control variable at time -1 for k Application control variables at any time for k The applied control variable at time +1 for k+M The applied control variable at time -1.

[0028] This involves discarding the first control variable from the previous time step, shifting the remaining sequence one position to the left, and repeating the last value at the end to maintain the sequence length. This initial value is close to the optimal solution, which can significantly reduce the number of iterations.

[0029] Secondly, the current CBF constraints for problem solving are simplified. The traditional NMPC-CBF method applies CBF constraints to every time step in the entire prediction time domain: (15); This implementation method retains only the following: Constraints are applied only at the current step, and not at other times. Due to the nature of rolling optimization, a CBF constraint is applied at each step, which is equivalent to guaranteeing the safety of the entire trajectory. This simplification reduces the number of constraints from... Down to .

[0030] S304: Repeat S301 to S303, change the constraint strength and controller parameters of the UAV in an obstacle environment, and observe and verify the trajectory tracking accuracy and obstacle avoidance safety of the controller.

[0031] S4: Deploy the above trajectory tracking and obstacle avoidance controllers on fixed-wing UAVs to enable UAVs to achieve accurate trajectory tracking and flight safety in obstacle environments.

[0032] Example To verify the technical effects of this invention, the following simulation verification was conducted: (1) Experimental environment This invention employs digital simulation to verify its correctness and rationality. First, a flight control simulation environment for the UAV's kinematic model is constructed in the MATLAB / Simulink environment. This environment contains four spherical obstacles and four target waypoints. The obstacles are positioned along the UAV's desired path, and the safe distance between the UAV and each obstacle is set to 100m. For the parameter constraints of the model and controller, the prediction interval is defined as 10, the control interval as 3, and the state weight matrix is... The input weight matrix is The simulation step size is 0.05s; the UAV's flight speed is 50m / s, with a speed range of 40-60m / s; the initial pitch angle is 0, with a constraint range of -10 to 20°; the initial pitch rate is 0, with a constraint range of -20 to 20° / s; the initial roll angle is 0, with a constraint range of -12 to 12°; the initial roll rate is 0, with a constraint range of -12 to 12° / s; the initial yaw rate is 0, with a constraint range of -12 to 12° / s; and the initial altitude is 500m, with a constraint range of 400-700m.

[0033] The simulation test software environment of this invention is Windows 10 + Matlab2022b, and the hardware environment is AMD Ryzen 7 7840HS w / Radeon 780M Graphics 3.80 GHz.

[0034] (2) Experimental results and analysis This invention first verifies the smoothness of the reference path generated by the artificial potential field method and its adaptability to the flight of fixed-wing UAVs. Figure 2 The paper demonstrates the global reference path of a fixed-wing UAV generated by setting potential field method parameters in an environment with given waypoints and obstacles. As shown in the figure, the global reference path generated by the artificial potential field method has good smoothness and meets the safe distance requirements near obstacles, making it suitable for UAV tracking in obstacle environments.

[0035] Figure 3 The figure shows the trajectory obtained by the path tracking of the UAV controlled by the classic NMPC controller. As can be seen from the figure, when describing the cost function of path tracking, the classic NMPC algorithm only considers the accuracy of path tracking and does not constrain the threat of obstacles in the environment, i.e., flight safety factors. Figure 4 This is a record of the minimum distance between the drone and the obstacle during the process. As can be seen from the figure, the drone exceeded the minimum distance under the control of NMPC, resulting in flight safety issues.

[0036] Figure 5 The image shows a comparison of trajectories using the classic NMPC and the NMPC-CBF controller under different constraint strengths. Figure 6 This is a comparison chart showing the minimum distance between the UAV and obstacles during path tracking, under the control of each controller. As shown in the chart above, in... At that time, due to the relatively loose constraints of CBF, the minimum distance between the drone and the obstacle was 50.96m, which is less than the safe distance; when At that time, the constraints of CBF became more stringent, and mandatory. The slow decay allows the drone to stay away from obstacles at all times. At that time, the minimum distance to the obstacle was 146.12m and 101.07m, which is greater than the minimum safe distance, and the trajectory is relatively ideal; when At that time, the minimum distance was 153.33m, which is much greater than the minimum safe distance.

[0037] Figure 7 The image shows a comparison of the NMPC-CBF controller's trajectory tracking at different flight speeds under specific constraint strengths. Figure 8 This is a comparison of the minimum distances between the UAV and obstacles during path tracking under the control of each controller. As shown in the figure, when the UAV tracks its trajectory at different speeds, under the constraint of the CBF distance function, it can balance trajectory tracking accuracy and flight safety, and the minimum safe distance of 100m is never exceeded throughout the entire process. This indicates that the UAV effectively tracks its trajectory under the control of the NMPC and effectively ensures flight safety under the CBF distance constraint.

[0038] Figure 9 The figure shows the optimization solution time record under the traditional NMPC-CBF control of the UAV. Figure 10The figure shows the optimization solution time record under the improved NMPC-CBF control of the UAV. As can be seen from the figure, under the condition of high-speed flight of the fixed-wing UAV, the average solution time of the traditional NMPC with CBF constraint is 0.138s, compared with a sampling period of 0.05s, indicating that this controller does not meet the strict real-time solution requirements in this computing environment. The average solution time of the improved NMPC with CBF constraint is 0.02s, which is less than the sampling period of 0.05s, indicating that the real-time performance of this method can meet the real-time solution requirements in this computing environment and is suitable for actual flight control applications of UAVs.

[0039] The results above demonstrate that, for the trajectory tracking and flight safety assurance problems of fixed-wing UAVs in obstacle environments at high speeds, this invention proposes an improved NMPC-CBF flight control method. The NMPC algorithm defines the model and environmental constraints of the fixed-wing UAV, ensuring adequate lead time for the actuators. The CBF algorithm defines the distance constraints between obstacles and the UAV and embeds them into the NMPC controller for solution. By defining the optimization problem, the flight safety of the UAV is satisfied. Regarding the real-time performance of the NMPC algorithm, a hot-start initial value strategy and a current-step constraint calculation strategy are proposed, significantly reducing the algorithm's time requirement and meeting the real-time solution requirements of fixed-wing UAVs in high-speed flight.

[0040] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent substitutions, and improvements made to the above embodiments without departing from the scope of the present invention, based on the technical essence of the present invention and within the spirit and principles of the present invention, shall still fall within the protection scope of the present invention.

Claims

1. A method for trajectory tracking and safe obstacle avoidance for high-speed fixed-wing unmanned aerial vehicles (UAVs), characterized in that, include: Step 1: Generate a global reference path based on the waypoint sequence and obstacle information in the flight environment, combined with the artificial potential field method; Step 2: Based on the global reference path and the kinematic characteristics of the fixed-wing UAV, establish a three-dimensional kinematic discrete model of the fixed-wing UAV; Step 3: Based on the three-dimensional kinematic discrete model, construct a nonlinear model predictive control optimization problem; Step 4: Based on the distance between the fixed-wing UAV and the obstacle, construct the control obstacle function constraint, and embed the control obstacle function constraint into the nonlinear model predictive control optimization problem to obtain the fusion optimization problem; Step 5: Based on the fusion optimization problem, solve for the control variables and apply them to the fixed-wing UAV.

2. The trajectory tracking and obstacle avoidance method for high-speed fixed-wing unmanned aerial vehicles according to claim 1, characterized in that, In step 1, the waypoint sequence includes the north coordinates, east coordinates, and altitude of the UAV's starting point, intermediate point, and target point; Obstacle information includes the center position, radius, and set minimum safe distance; An artificial potential field is established based on the current position of the drone, waypoint positions, the Euclidean distance between the drone and the waypoint, obstacle positions, and the Euclidean distance between the obstacle and the drone. The artificial potential field method models obstacles and desired waypoints in the environment as repulsive and attractive points, respectively. The artificial potential field includes the attractive field generated by the waypoint sequence and the repulsive field generated by the obstacles. The gravitational potential energy parameter in the attractive field is... and The repulsive potential energy function in the repulsive field is: ; The net force and positional changes of the UAV in the environment are obtained based on the gravitational and repulsive fields; Based on the resultant force received by the UAV, a reference path is generated pointing to by the resultant force. The global reference path of the UAV is obtained by combining the UAV's desired speed, yaw angle, and pitch angle. ; The gravitational potential energy parameter is and for: （1）； In formula (1), X The location of the drone. The location of the waypoint This represents the Euclidean distance between the drone and the waypoint. This is the gravitational gain coefficient. Let represent the gravitational potential energy of the drone at its current location. The direction of gravity; Repulsive potential energy parameters for: （2）； In formula (2), The repulsive force gain coefficient, q Indicates the location of the drone. Location of the obstacle. The distance between the drone and the obstacle. Minimum safe distance; The net force acting on the drone in the environment and its positional change are as follows: （3）； In formula (3), The net force acting on the current position. To determine the location of the drone for the next step. The current location of the drone. Step size; Global reference path of drone for: （4）； In formula (4), These are the expected speed, yaw angle, and pitch angle of the UAV, respectively.

3. The trajectory tracking and obstacle avoidance method for high-speed fixed-wing unmanned aerial vehicles according to claim 1, characterized in that, Step 2 specifically includes: The state vector is determined based on the UAV's three-dimensional position, ground speed, yaw angle, pitch angle, and acceleration in the inertial frame. With control quantity ; Based on the state vector, control quantity, and Euler discretization method, a three-dimensional kinematic discretization model is established, and the state constraints and input constraints of the model are defined. State vector With control quantity The expression is: （5）； In formula (5), This represents the three-dimensional position of the UAV in an inertial frame. V For ground speed, These are the yaw angle and the pitch angle, respectively. a For acceleration; The three-dimensional kinematic discrete model is as follows: （6）； In formula (6), , , , , and They are respectively k Real-time three-dimensional position, ground speed, yaw angle, and pitch angle of the drone. , and These are yaw rate, pitch rate, and acceleration, respectively. The dynamic differential equations of the UAV; The expression for the state constraint is: （7）； In formula (7), and The minimum and maximum flight speeds of the drone. for k Flight speed at any moment Minimum yaw angle, for k Yaw angle at any moment For the maximum yaw angle, Minimum climb angle, for k The angle of ascent at any moment, Maximum climb angle, Minimum flight altitude for k Flight altitude at any given moment Maximum flight altitude; The expression for the input constraint is: （8）； In formula (8), This represents the minimum pitch rate of the drone. This is the maximum pitch rate of the drone. Let yaw rate be the yaw rate of the drone. This represents the minimum yaw rate of the drone. This represents the maximum yaw rate of the drone. Let yaw rate be the yaw rate of the drone. The minimum acceleration of the drone. This represents the maximum acceleration of the drone.

4. The trajectory tracking and obstacle avoidance method for high-speed fixed-wing UAVs according to claim 1, characterized in that, Step 3 specifically includes: Step 3.1: Set the prediction time domain of the NMPC controller for N Step, control time domain for M Step, and ; Step 3.2: Based on the prediction time domain and the control time domain, define a cost function to measure the trajectory tracking accuracy and control energy consumption. Based on the cost function, the three-dimensional kinematic discrete model, state constraints, and input constraints, construct a nonlinear model predictive control optimization problem. The expression for the cost function is: （9）； In formula (9), for k+i The UAV state reference at time step, where Q is the state weight matrix, N is the prediction interval length, and R is the control input weight matrix. for k The later time prediction i The state at any given moment, for k+i The amount of control at any given moment; The nonlinear model predictive control optimization problem is as follows: （10）； In formula (10), for k+M- Control quantity at time 1.

5. The trajectory tracking and obstacle avoidance method for high-speed fixed-wing UAVs according to claim 4, characterized in that, Step 4 specifically includes: Step 4.1: Establish an obstacle control function based on the real-time position of the fixed-wing UAV, the position of the obstacle, the threat radius, and the safe distance; Step 4.2: Embed the obstacle control function into the NMPC controller to obtain the fusion optimization problem; The obstacle control function is: （11）； In formula (11), p This is the real-time location of the drone. For the first m The location of the obstacle. For the first m The threat radius of an obstacle. For the first m Safe distance from each obstacle For state space dimension; The expression for the fusion optimization problem is: （12）； In formula (12), , A safety set for the state and constraints of unmanned aerial vehicles. for t Current state for t Time Prediction t+k The state at any given moment, for t+k The amount of control at any given moment The state at the current moment. For the boundary convergence factor of the safe set, For the present t+k The convergence of the safety boundary at any given time. for t+k The safety boundary is predicted at any time.

6. A trajectory tracking and obstacle avoidance method for high-speed fixed-wing unmanned aerial vehicles according to claim 1, characterized in that... Step 5 specifically includes: Step 5.1: Solve the optimal control sequence obtained from the previous sampling time. Discard the first control variable and shift the remaining control sequence one position to the left, repeat the last control variable at the end, and obtain the initial value for hot start at the current sampling time; Step 5.2: Apply control barrier function constraints only to the current sampling time, and do not apply control barrier function constraints to other times in the prediction time domain, to obtain the simplified fusion optimization problem; Step 5.3: Based on the initial hot start value, solve the simplified fusion optimization problem to obtain the control quantity and apply it to the fixed-wing UAV; Step 5.4: Repeat steps 3, 4 and 5.1-5.3 to change the constraint strength and controller parameters of the UAV in an obstacle environment, and observe and verify the trajectory tracking accuracy and obstacle avoidance safety of the controller. The expression for the optimal control sequence is: （13）； The initial values ​​for the warm start at the current sampling time are: （14）； In formulas (13) and (14), for k The applied control variable at time -1 for k Application control variables at any time for k The applied control variable at time +1 for k+M The applied control variable at time -1.

Images