A fixed-wing unmanned aerial vehicle terrain following human-simulated intelligent control method
By employing a humanoid intelligent control method and utilizing the hyperbolic tangent function to construct altitude control primitives, the altitude tracking problem of fixed-wing UAVs under complex terrain and external disturbances was solved, achieving high-precision tracking and control stability while avoiding overshoot and vibration.
Patent Information
- Application Number
- CN202611519605.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-29
- Publication Date
- 2026-11-13
AI Technical Summary
Existing fixed-wing UAVs have insufficient control robustness in terrain-following flight, making it difficult to achieve high-precision altitude tracking under complex terrain and external disturbances. Furthermore, traditional control methods are prone to overshoot and vibration.
A humanoid intelligent control method is adopted, which constructs a height control primitive through a hyperbolic tangent function and combines it with a pitch angle control law. Under the humanoid intelligent control framework, hierarchical decision-making is carried out to realize the adaptive switching and parameter optimization of the height control primitive, thereby suppressing excessive overshoot and input vibration.
It achieves high-precision altitude tracking under complex terrain and external disturbances, avoids excessive oscillation of control input, and ensures flight safety and control quality.
Smart Images

Figure CN123089362A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of fixed-wing unmanned aerial vehicle (UAV) control technology, specifically relating to a humanoid intelligent control method for terrain following of a fixed-wing UAV. Background Technology
[0002] Fixed-wing unmanned aerial vehicles (UAVs) play a vital role in military reconnaissance, terrain mapping, agricultural plant protection, and power line inspection. Among these, terrain-following flight, as an important flight mode, requires UAVs to maintain a constant relative altitude to the ground, which is crucial for tasks such as low-altitude penetration and terrain-following mapping. Terrain-following flight not only helps UAVs evade radar detection but also enables them to acquire high-quality ground imagery data, allowing for safe flight in complex terrain environments.
[0003] However, terrain-following flight control of fixed-wing UAVs faces numerous challenges. First, terrain altitude changes are uncertain and non-smooth. In complex terrains such as mountains and hills, the surface altitude may change abruptly, and the reference trajectory tracked by the fixed-wing UAV may be non-smooth, leading to overshoot and vibration in the controller output. Second, in the presence of wind disturbances at low-altitude surfaces, conventional linear controllers cannot effectively suppress external disturbances, resulting in system instability.
[0004] Currently, altitude control for fixed-wing UAVs mainly employs methods such as PID control, backstepping control, and sliding mode control. PID controllers are simple in structure and easy to implement, but their linear feedback form has limited performance when handling nonlinear systems, and parameter tuning is difficult, making it hard to simultaneously guarantee speed and overshoot. While backstepping control can handle nonlinear systems, its virtual control law design is complex, and it typically uses linear feedback, making it prone to large overshoot and oscillations when tracking non-smooth desired trajectories. Sliding mode control is robust to parameter uncertainties and external disturbances, but it suffers from inherent chattering problems, which increase actuator wear and may excite unmodeled dynamics in the system. To mitigate chattering, researchers have proposed improved methods such as higher-order sliding mode and continuous sliding mode, but these methods often increase algorithm complexity.
[0005] In recent years, observer-based disturbance compensation methods have attracted attention. By designing a disturbance observer to estimate uncertainties and external disturbances in the system, and then performing feedforward compensation in the control law, the system's anti-interference capability can be improved. However, traditional linear observers require large observer gains when estimating rapidly changing disturbances, which may lead to large peaks and oscillations in the estimated signal, affecting control performance. Furthermore, most existing methods employ linear feedback in the control law design. When the system needs to track a reference trajectory with discontinuous derivatives or rapid changes (such as terrain abrupt changes in terrain-following flight), linear feedback can easily cause spikes in the control input, leading to actuator saturation.
[0006] Furthermore, to further improve the intelligence level of control systems under complex operating conditions, the academic community has proposed humanoid intelligent control methods. Humanoid intelligent control simulates the structure and behavior of the human body's control system on a macroscopic level. Through hierarchical information processing and decision-making mechanisms, feature models, feature identification, and feature memory, it realizes a secondary mapping process from "cognition-judgment-operation," enabling the control system to adaptively select control modes and control laws based on the controlled variable, deviation, and their changing trends.
[0007] Humanoid intelligent control typically employs a high-order generative system structure consisting of a central command level, an organization and coordination level, and a unit control level. Within the unit control level, a real-time controller with a certain degree of intelligence is formed through an operational control level, a parameter self-tuning level, and a task adaptive level. Existing research on fixed-wing UAVs often uses humanoid intelligent control for higher-level problems such as mission planning, expert rule reasoning, or multi-aircraft collaboration. However, for continuous-time nonlinear control problems like altitude channels, it still mainly relies on traditional linear or single-mode control laws. There is a lack of systematic methods to encapsulate nonlinear control laws as "altitude control primitives" and perform multi-mode switching and adaptive tuning within a humanoid intelligent control structure.
[0008] Therefore, designing a ground-mimicking intelligent control method that can guarantee tracking performance, avoid excessive overshoot and input vibration, and incorporate a hierarchical decision-making mechanism and high-level control primitives to be robust to external disturbances is an important problem that needs to be solved in the field of fixed-wing UAV control. Summary of the Invention
[0009] To address the aforementioned shortcomings in existing technologies, this invention provides a humanoid intelligent control method for terrain following of fixed-wing UAVs, which solves the problems of insufficient robustness and lack of intelligent control mode switching and parameter adaptation capabilities based on flight status and terrain changes in traditional control methods.
[0010] To achieve the aforementioned objectives, the technical solution adopted by this invention is: a humanoid intelligent control method for terrain following of a fixed-wing unmanned aerial vehicle, comprising the following steps: S1. Obtain UAV flight status information and terrain altitude information; S2. Generate the desired flight altitude trajectory based on terrain height information; S3. Under the humanoid intelligent control framework, based on the desired flight altitude trajectory, design the altitude control primitive set and pitch angle control law through the hyperbolic tangent function, and generate pitch angle control commands. S4. Send pitch angle control commands to the flight attitude control system to complete flight altitude control.
[0011] Furthermore, in S1, the UAV flight status information includes the attitude information of the fixed-wing UAV, its flight altitude relative to the takeoff point, flight speed, and track angle.
[0012] Furthermore, S2 includes the following sub-steps: S21. Determine the desired relative flight altitude based on the UAV's flight speed and current position; S22. Based on the terrain height information and the desired relative flight altitude, generate the desired flight altitude trajectory.
[0013] Furthermore, in S3, under the humanoid intelligent control framework, the unit control level of the fixed-wing UAV's terrain-following flight control includes an operational control level, a parameter self-calibration level, and a mission adaptive level. The operational control level outputs pitch angle control commands, the parameter self-calibration level is used for online fine-tuning and learning updates of altitude control primitive parameters, and the mission adaptive level selects and switches altitude control primitives based on flight mission and environmental changes. S3 includes the following sub-steps: S31. Constructing virtual control terms and height control primitives based on the hyperbolic tangent function: The altitude error is calculated based on the expected flight altitude trajectory and the actual flight altitude, and a virtual control term is constructed using the hyperbolic tangent function. Using altitude error, altitude error change rate, track angle, and track angle change rate as characteristic quantities, an error phase plane is constructed and divided into several characteristic regions. Each characteristic region corresponds to an altitude control mode. Different parameters are configured according to different characteristic regions, and combined with the corresponding altitude control modes, the virtual control items are encapsulated into several altitude control primitives to form an altitude control primitive library. S32. Define the expanded error and construct the pitch angle control law: Based on the kinematic relationship of fixed-wing UAVs, the extended error is defined, and based on the altitude error, extended error and track angle, the pitch angle control law is constructed by selecting the Lyapunov function; S33. High-level control of primitive selection and switching based on feature recognition through task adaptive level: During flight, feature quantities are calculated in real time, and the feature region to which the current working point belongs is determined in the error phase plane. According to the pre-set production rules, the altitude control primitives that match the current feature region are selected and activated, so that the altitude control primitive parameters are adapted to the current flight state, terrain changes and mission requirements. When the working point crosses the feature region boundary, the frequent switching of altitude control primitives is suppressed by introducing hysteresis band and feature memory mechanism.
[0014] S34. Generate pitch angle control commands based on the selected altitude control primitives through the operation control level: Based on the selected altitude control primitive, the pitch angle is calculated according to the altitude control primitive parameters and the pitch angle control law, and then the pitch angle control command is output through the running control level. S35. Adaptive optimization of height control element parameters through parameter self-calibration stage: The parameter self-calibration stage performs online fine-tuning and learning updates of the parameters of each height control element based on height error, overshoot, control input peak value, and the indicators of different height control elements.
[0015] The beneficial effects of the above-mentioned further solutions are as follows: the present invention achieves height control and adaptive selection of control modes through two-layer hierarchical control and multiple height control primitives under the humanoid intelligent control structure.
[0016] Furthermore: In S31, virtual control items The specific expression is: In the formula, This is the first gain parameter. is the hyperbolic tangent function. For the first proportional gain, This is for height error; In the formula, For flight altitude, This represents the desired flight altitude.
[0017] Furthermore: In S32, the expression for the pitch angle control law is as follows: In the formula, The pitch angle, The lift acceleration per unit angle of attack, , air density, For wing reference area, This is the derivative of the lift coefficient with respect to the angle of attack. For the mass of the aircraft, For virtual control items The derivative, , Let be the derivative of the altitude error with respect to time. t For time, This is the second gain parameter. This is the second proportional gain. The damping coefficient is... To expand the error, For the track angle, It is a hyperbolic secant function; In the formula, This refers to flight speed.
[0018] Furthermore, S4 includes the following sub-steps: S41. Send the pitch angle control command to the flight attitude control system, and execute the pitch angle control through the flight attitude control system, and output the servo drive command. S43. Input the servo drive command into the servo system to drive the elevator to deflect. S44. After the elevator deflects, the attitude and motion of the UAV are changed through aerodynamics to complete the flight altitude control.
[0019] The beneficial effects of this invention are as follows: (1) This invention provides a humanoid intelligent control method for terrain following of fixed-wing UAVs. Under the hierarchical structure framework of humanoid intelligent control, an altitude control feature model and control element library are constructed. The nonlinear hierarchical control law constructed by the hyperbolic tangent function is encapsulated into several altitude control elements. Through feature identification and feature memory mechanism, the online selection of control elements and parameter self-correction are realized. Thus, in the presence of parameter uncertainty and external disturbance, high-precision tracking of terrain altitude changes is achieved. At the same time, excessive overshoot and control input peak are suppressed, actuator saturation is avoided, flight safety and control quality are guaranteed. This solves the problems of insufficient robustness and lack of intelligent control mode switching and parameter self-adaptation capability according to flight status and terrain changes in traditional control methods.
[0020] (2) Under the high-order generative system structure of humanoid intelligent control, this invention decomposes the altitude control problem into a two-layer hierarchical structure and multiple altitude control primitives. The first layer uses altitude error and its variation trend as characteristic quantities, constructs virtual control quantities using the hyperbolic tangent function, and forms a library of altitude control primitives with different control modes and parameter configurations. The second layer, based on the dynamics of the flight path angle, combines the virtual control quantities and their derivatives to design actual pitch angle control commands, and realizes online selection and tuning of control primitives through the task adaptive level and parameter self-correction level. The entire control law, while ensuring smooth continuity, natural saturation, and robustness, supports multimodal control and intelligent decision-making, and can effectively cope with complex terrain and external disturbances.
[0021] (3) Experimental verification shows that the control input of the present invention does not have high-frequency oscillation during the switching of multiple altitude control elements. The ground-following flight intelligent control method of the present invention can achieve accurate altitude tracking and ensure control quality. Attached Figure Description
[0022] Figure 1 This is a flowchart of a humanoid intelligent control method for terrain following of a fixed-wing UAV according to the present invention.
[0023] Figure 2 The numerical simulation results show the altitude tracking of the intelligent control method for ground-following flight of the fixed-wing UAV of this invention.
[0024] Figure 3 The numerical simulation results are for the control input of the intelligent control method for ground-following flight of the fixed-wing UAV of the present invention. Detailed Implementation
[0025] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.
[0026] like Figure 1 As shown, in one embodiment of the present invention, a humanoid intelligent control method for terrain following of a fixed-wing unmanned aerial vehicle includes the following steps: S1. Obtain UAV flight status information and terrain altitude information; S2. Generate the desired flight altitude trajectory based on terrain height information; S3. Under the humanoid intelligent control framework, based on the desired flight altitude trajectory, design the altitude control primitive set and pitch angle control law through the hyperbolic tangent function, and generate pitch angle control commands. S4. Send pitch angle control commands to the flight attitude control system to complete flight altitude control.
[0027] In S1, the UAV flight status information includes the attitude information of the fixed-wing UAV, its flight altitude relative to the takeoff point, flight speed, and flight path angle.
[0028] In this embodiment, an airborne inertial measurement unit (IMU) is used to acquire the attitude information of the fixed-wing UAV. The IMU typically includes a three-axis gyroscope and a three-axis accelerometer. The attitude angles of the UAV are calculated in real time using a sensor fusion algorithm. (Pitch angle) Defined as the tilt angle of the body coordinate system relative to the horizontal plane, with upward being positive.
[0029] The flight altitude of a fixed-wing UAV is obtained using a barometric altimeter and GPS positioning. In practical applications, flight altitude Altitude is typically defined as the vertical distance relative to the takeoff point. A barometric altimeter calculates altitude by measuring atmospheric pressure; GPS provides absolute altitude information. By fusing data from both sensors using an extended Kalman filter, stable altitude information relative to the takeoff point of the fixed-wing UAV is obtained.
[0030] The flight speed of a fixed-wing drone is obtained through an airspeed sensor. An airspeed sensor calculates the speed relative to the air by measuring the difference between dynamic and static pressure.
[0031] Calculating the flight path angle using the flight speed information of a fixed-wing UAV The track angle is defined as the angle between the velocity vector and the horizontal plane, reflecting the aircraft's climb or descent trend. The track angle can be calculated as the ratio of the rate of change of altitude to the horizontal velocity.
[0032] The terrain ranging sensor is used to obtain the elevation information of the terrain ahead. The terrain ranging sensor can be a lidar sensor. The elevation distance is measured by the slope distance from the sensor to the ground. and installation angle By combining the current flight altitude and attitude information, the absolute altitude distribution of the terrain ahead can be calculated.
[0033] S2 includes the following steps: S21. Determine the desired relative flight altitude based on the UAV's flight speed and current position; S22. Based on the terrain height information and the desired relative flight altitude, generate the desired flight altitude trajectory.
[0034] In S21, the desired relative flight altitude is determined by combining the UAV's flight speed and current position. The core objective of terrain-following flight is to maintain a constant relative altitude between the UAV and the ground; this altitude is called the desired altitude above the ground. (Above Ground Level), typically set at 30-100 meters. Desired flight altitude. Equal to terrain height Desired ground height The sum of .
[0035] In S22, the desired flight altitude trajectory is generated based on the terrain undulations. The trajectory generator extracts the terrain profile along the flight path from the terrain altitude distribution data, and then generates the corresponding desired altitude trajectory. The generated desired trajectory serves as the reference input for the control system and is stored in the form of a time series.
[0036] In S3, under the humanoid intelligent control framework, the terrain-following flight control of fixed-wing UAVs is divided into a central command level, an organization and coordination level, and a unit control level. The unit control level includes an operational control level, a parameter self-calibration level, and a mission adaptive level. The operational control level outputs pitch angle control commands. The parameter self-calibration level is used for online fine-tuning and learning updates of altitude control primitive parameters. The mission adaptive level selects and switches altitude control primitives based on flight mission and environmental changes. S3 includes the following steps: S31. Constructing virtual control terms and height control primitives based on the hyperbolic tangent function: The altitude error is calculated based on the expected flight altitude trajectory and the actual flight altitude, and a virtual control term is constructed using the hyperbolic tangent function. Using altitude error, altitude error change rate, track angle, and track angle change rate as characteristic quantities, an error phase plane is constructed and divided into several characteristic regions. Each characteristic region corresponds to an altitude control mode. Different parameters are configured according to different characteristic regions, and combined with the corresponding altitude control modes, the virtual control items are encapsulated into several altitude control primitives to form an altitude control primitive library. S32. Define the expanded error and construct the pitch angle control law: Based on the kinematic relationship of fixed-wing UAVs, the extended error is defined, and based on the altitude error, extended error and track angle, the pitch angle control law is constructed by selecting the Lyapunov function; S33. High-level control of primitive selection and switching based on feature recognition through task adaptive level: During flight, characteristic quantities are calculated in real time. The characteristic region to which the current operating point belongs is determined in the error phase plane. Based on pre-defined production rules, altitude control primitives matching the current characteristic region are selected and activated from among the altitude control primitives. This ensures that the altitude control primitive parameters adapt to the current flight state, terrain changes, and mission requirements. The altitude control primitive parameters include... , , , and When the working point crosses the boundary of the feature region, frequent switching of the height control primitive is suppressed by introducing hysteresis band and feature memory mechanism.
[0037] S34. Generate pitch angle control commands based on the selected altitude control primitives through the operation control level: Based on the selected altitude control primitive, the pitch angle is calculated according to the altitude control primitive parameters and the pitch angle control law, and then the pitch angle control command is output through the running control level. S35. Adaptive optimization of height control element parameters through parameter self-calibration stage: The parameter self-calibration stage performs online fine-tuning and learning updates of the parameters of each altitude control element based on altitude error, overshoot, peak control input, and the indicators of different altitude control elements. This enables the pitch angle control command to have experience accumulation and adaptive capabilities similar to those of a human operator.
[0038] In this embodiment, the initial values of the height control primitive parameters are tuned according to the following principles: (1) , Tuning: Parameters The maximum climb / descent speed is determined based on the aircraft's maximum climb / descent capability, typically taken as 60-80% of the maximum vertical speed to ensure sufficient control margin. For a typical small fixed-wing UAV, the maximum climb speed is approximately 5-8 m / s, therefore, in this embodiment... Choose 3-6. Parameters The speed at which the height error converges is determined by its magnitude. It can improve the system's sensitivity and convergence speed within a small error range, but at the same time it will amplify measurement noise and increase the high-frequency components of the control input. Therefore, its value needs to be a trade-off between altitude error convergence speed and noise immunity. A good compromise value should be selected based on the actual flight test results.
[0039] (2) , Tuning: Parameters Related to the maneuverability of the aircraft, it can be estimated based on the maximum normal acceleration. For small fixed-wing UAVs, the maximum normal acceleration is approximately 1.5-2.5 times the acceleration due to gravity, corresponding to... The value ranges from 5 to 10. (Parameter) The tuning method and Similarly, an initial value of 1 is recommended, adjusted based on actual results. Larger values... It can improve tracking accuracy, but may cause high-frequency oscillations.
[0040] (3) Damping coefficient Tuning: Damping coefficient Used to suppress system oscillations, a higher value results in a stronger damping effect but reduces response speed. For small fixed-wing UAVs, it is recommended to start from... Starting with a value of 0.2, if the system exhibits continuous oscillations, the value can be gradually increased. If the system is too sluggish, the value can be reduced appropriately.
[0041] In S31, virtual control items The specific expression is: In the formula, This is the first gain parameter. is the hyperbolic tangent function. For the first proportional gain, For height error, , ; In the formula, For flight altitude, This represents the desired flight altitude.
[0042] In this embodiment, different parameters are configured according to different feature regions. And in conjunction with the corresponding control mode, the parameters and activation conditions will be... It is encapsulated into several height control primitives to form a height control primitive library.
[0043] In S32, based on height error Expanded error For dynamic trajectory angles, the Lyapunov function is selected. Using the Lyapunov direct method, a pitch angle control law is designed such that the time derivative of the Lyapunov function is negative definite. The specific expression of the pitch angle control law is as follows: In the formula, The pitch angle, The lift acceleration per unit angle of attack, , air density, For wing reference area, This is the derivative of the lift coefficient with respect to the angle of attack. For the mass of the aircraft, For virtual control items The derivative, , Let be the derivative of the altitude error with respect to time. t For time, obtained through numerical differentiation, This is the second gain parameter. , This is the second proportional gain. , The damping coefficient is... , To expand the error, For the track angle, It is a hyperbolic secant function; In this embodiment, the kinematic relationships of the fixed-wing UAV are considered. Define the expanded error : In the formula, For flight speed, expand error The physical meaning is the actual vertical velocity. With virtual control items The deviation between them.
[0044] S4 includes the following steps: S41. Send the pitch angle control command to the flight attitude control system, and execute the pitch angle control through the flight attitude control system, and output the servo drive command. In this embodiment, the pitch angle control command is sent to the attitude control module of the flight attitude control system. The pitch angle of the pitch angle control command is compared with the current actual pitch angle of the fixed-wing UAV, and the control outputs the servo drive command to drive the elevator to deflect, so as to achieve the target pitch angle.
[0045] S43. Input the servo drive command into the servo system to drive the elevator to deflect. The deflection of the elevator alters the aerodynamic lift distribution at the tail, generating a pitching moment at the aircraft's center of gravity. According to Newton's second law and rigid body dynamics, the pitching moment causes angular acceleration, which in turn changes the pitch rate and pitch angle. Changes in the pitch angle further affect the angle of attack, which in turn alters the overall lift distribution, affecting the flight path angle and trajectory, ultimately changing the flight altitude of the fixed-wing UAV.
[0046] S44. After the elevator deflects, the attitude and motion of the UAV are changed through aerodynamics to complete the flight altitude control.
[0047] To verify the effectiveness of the present invention, the following experimental data are provided: In this example, numerical simulation is used for verification. After steps S1 and S2, the desired flight altitude is obtained. In this embodiment, the altitude is taken as a sinusoidal variation to allow the aircraft to perform ground-following flight tracking. The relevant parameters are: amplitude 1m, offset 12m, and frequency 2. (radians per second). Air density The value is 1.255 kg / , is the wing reference area Set it to 0.5 , The derivative of the lift coefficient with respect to the angle of attack is taken as 5, and the mass... Take 4kg as the weight, and the aircraft speed. 15 The initial height is set to 10. The initial state is level flight, i.e., the track angle is... ; for the initial values of the height control primitive parameters After adjustment, it is taken as , , , , Additionally, in this example, a bounded external uncertainty disturbance input is added to simulate real-world conditions, with a maximum value of 1, acting on the rate of change of the track angle. superior.
[0048] In this embodiment, step S3 involves selecting and activating a suitable altitude control primitive within the humanoid intelligent control framework, calculating the control law, and substituting the calculated pitch angle control command into the dynamic equations of the fixed-wing UAV described in this invention to finally calculate the current altitude. . Figure 2 The numerical simulation results show the altitude tracking of the fixed-wing UAV's terrain-following intelligent control method according to the present invention. (This can be seen from...) Figure 2 The upper coordinate axis shows that the fixed-wing UAV can track the desired altitude well; from Figure 2 As can be seen from the lower center of the coordinate axis, even with bounded external disturbances, the height tracking error can still be suppressed to a small range near zero.
[0049] In this embodiment, Figure 3 The figures show the numerical simulation results of the control input for the terrain-following intelligent control method for fixed-wing UAVs according to the present invention. As can be seen from the figures, there is no high-frequency oscillation in the control input during the switching of various altitude control primitives. In summary, this terrain-following intelligent control method can achieve accurate altitude tracking and ensure control quality.
[0050] In the description of this invention, the above are merely preferred embodiments and are not intended to limit the scope of protection of this invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this invention should be included within the scope of protection of this invention.
Claims
1. A humanoid intelligent control method for terrain following of a fixed-wing unmanned aerial vehicle, characterized in that, Includes the following steps: S1. Obtain UAV flight status information and terrain altitude information; S2. Generate the desired flight altitude trajectory based on terrain height information; S3. Under the humanoid intelligent control framework, based on the desired flight altitude trajectory, design the altitude control primitive set and pitch angle control law through the hyperbolic tangent function, and generate pitch angle control commands. S4. Send pitch angle control commands to the flight attitude control system to complete flight altitude control.
2. The humanoid intelligent control method for terrain following of a fixed-wing UAV according to claim 1, characterized in that, In S1, the UAV flight status information includes the attitude information of the fixed-wing UAV, its flight altitude relative to the takeoff point, flight speed, and flight path angle.
3. The humanoid intelligent control method for terrain following of a fixed-wing UAV according to claim 1, characterized in that, S2 includes the following steps: S21. Determine the desired relative flight altitude based on the UAV's flight speed and current position; S22. Based on the terrain height information and the desired relative flight altitude, generate the desired flight altitude trajectory.
4. The humanoid intelligent control method for terrain following of a fixed-wing UAV according to claim 1, characterized in that, In S3, under the humanoid intelligent control framework, the unit control level of the fixed-wing UAV's terrain-following flight control includes an operational control level, a parameter self-calibration level, and a mission adaptive level. The operational control level outputs pitch angle control commands, the parameter self-calibration level is used for online fine-tuning and learning updates of altitude control primitive parameters, and the mission adaptive level selects and switches altitude control primitives based on flight mission and environmental changes. S3 includes the following sub-steps: S31. Constructing virtual control terms and height control primitives based on the hyperbolic tangent function: The altitude error is calculated based on the expected flight altitude trajectory and the actual flight altitude, and a virtual control term is constructed using the hyperbolic tangent function. Using altitude error, altitude error change rate, track angle, and track angle change rate as characteristic quantities, an error phase plane is constructed and divided into several characteristic regions. Each characteristic region corresponds to an altitude control mode. Different parameters are configured according to different characteristic regions, and combined with the corresponding altitude control modes, the virtual control items are encapsulated into several altitude control primitives to form an altitude control primitive library. S32. Define the expanded error and construct the pitch angle control law: Based on the kinematic relationship of fixed-wing UAVs, the extended error is defined, and based on the altitude error, extended error and track angle, the pitch angle control law is constructed by selecting the Lyapunov function; S33. High-level control of primitive selection and switching based on feature recognition through task adaptive level: During flight, feature quantities are calculated in real time, and the feature region to which the current working point belongs is determined in the error phase plane. According to the pre-set production rules, the altitude control primitives that match the current feature region are selected and activated, so that the altitude control primitive parameters are adapted to the current flight state, terrain changes and mission requirements. When the working point crosses the feature region boundary, the frequent switching of altitude control primitives is suppressed by introducing hysteresis band and feature memory mechanism. S34. Generate pitch angle control commands based on the selected altitude control primitives through the operation control level: Based on the selected altitude control primitive, the pitch angle is calculated according to the altitude control primitive parameters and the pitch angle control law, and then the pitch angle control command is output through the running control level. S35. Adaptive optimization of height control element parameters through parameter self-calibration stage: The parameter self-calibration stage performs online fine-tuning and learning updates of the parameters of each height control element based on height error, overshoot, control input peak value, and the indicators of different height control elements.
5. The humanoid intelligent control method for terrain following of a fixed-wing UAV according to claim 4, characterized in that, In S31, virtual control items The specific expression is: In the formula, This is the first gain parameter. is the hyperbolic tangent function. For the first proportional gain, This is for height error; In the formula, For flight altitude, This represents the desired flight altitude.
6. The humanoid intelligent control method for terrain following of a fixed-wing UAV according to claim 5, characterized in that, In S32, the expression for the pitch angle control law is as follows: In the formula, The pitch angle, The lift acceleration per unit angle of attack, , air density, For wing reference area, This is the derivative of the lift coefficient with respect to the angle of attack. For the mass of the aircraft, For virtual control items The derivative, , Let be the derivative of the altitude error with respect to time. t For time, This is the second gain parameter. This is the second proportional gain. The damping coefficient is... To expand the error, For the track angle, It is a hyperbolic secant function; In the formula, This refers to flight speed.
7. The humanoid intelligent control method for terrain following of a fixed-wing UAV according to claim 1, characterized in that, S4 includes the following steps: S41. Send the pitch angle control command to the flight attitude control system, and execute the pitch angle control through the flight attitude control system, and output the servo drive command. S43. Input the servo drive command into the servo system to drive the elevator to deflect. S44. After the elevator deflects, the attitude and motion of the UAV are changed through aerodynamics to complete the flight altitude control.
