An aircraft center of gravity adaptive control method, system, medium and product

Abstract

The application discloses a kind of aircraft gravity center adaptive control method, system, medium and product, belong to aircraft control field, the method is: obtaining the roll angle and roll angular velocity of aircraft current control period, and desired roll angle instruction, and calculate roll angle tracking error and roll angle error change rate;Based on the roll angle tracking error and roll angle error change rate, the sliding surface state quantity is constructed, and the upper bound of bounded disturbance in aircraft roll channel is recursively updated according to sliding surface state quantity;According to the desired roll angle instruction, roll angle error change rate, sliding surface state quantity and bounded disturbance upper bound, the control torque required for roll channel is calculated;According to control torque, the control signal of each motor of aircraft under current cycle is generated, and each described motor is driven to control the roll attitude of aircraft, and flight control is completed. By implementing the present application, the problem of attitude instability caused by the dramatic unknown change of gravity center during the operation of the aircraft in the prior art can be solved.

Description

Technical Field

[0001] This invention belongs to the field of aircraft control and relates to an adaptive control method, system, medium, and product for the center of gravity of an aircraft. Background Technology

[0002] Multirotor aircraft are widely used in agricultural plant protection, logistics transportation, power line inspection and other operational scenarios. When performing spraying, hoisting or dropping tasks, the mass distribution of the aircraft will change significantly due to load consumption or center of gravity shift, which places extremely high demands on the robustness of the flight control system.

[0003] Existing technologies typically employ linear controllers with fixed parameters (such as PID control) or static balancing before flight to address changes in the center of gravity. The former is prone to performance degradation or even instability when faced with unknown and rapidly changing model disturbances, while the latter is only applicable to preset limited operating conditions and cannot adapt to sudden or continuous changes in mass distribution during flight, making it difficult to guarantee the stability and safety of the aircraft in complex operating scenarios. Summary of the Invention

[0004] This application provides a method, system, medium, and product for adaptive control of the center of gravity of an aircraft, which can solve the problem of attitude instability caused by drastic and unknown changes in the center of gravity of an aircraft during operation in the prior art.

[0005] To achieve the above objectives, in a first aspect, the present invention provides an adaptive control method for the center of gravity of an aircraft, comprising: The system acquires the roll angle and roll rate of the aircraft in the current control cycle, as well as the desired roll angle command, and calculates the roll angle tracking error and the roll angle error rate of change. Based on the roll angle tracking error and the roll angle error change rate, a sliding surface state variable is constructed, and the bounded upper limit of the spacecraft roll channel is recursively updated according to the sliding surface state variable. Based on the desired roll angle command, the roll angle error rate of change, the sliding surface state quantity, and the bounded upper limit of the disturbance, the control torque required for the roll channel is calculated; wherein, the control torque includes a disturbance compensation term that is proportional to the bounded upper limit of the disturbance. Based on the control torque, control signals for each motor of the aircraft are generated in the current cycle, and each motor is driven to control the roll attitude of the aircraft to complete flight control.

[0006] Compared to existing technologies, the embodiments of this application have the following beneficial effects: They acquire the roll angle and roll angular velocity of the aircraft in the current control cycle, as well as the desired roll angle command, and calculate the roll angle tracking error and roll angle error rate of change. By constructing second-order error information containing attitude deviations and their rates of change, they provide complete state observation inputs for the high-order controller, enabling control decisions to simultaneously respond to current deviations and deviation development trends. Based on the roll angle tracking error and roll angle error rate of change, they construct sliding surface state variables and recursively update the bounded disturbance upper bound in the aircraft's roll channel according to the sliding surface state variables. The sliding surface state variables serve as the target manifold for sliding mode control, and their construction can guide the system state to a robust switching surface. The recursive update mechanism of the bounded disturbance upper bound utilizes the amplitude of the sliding surface state variables as a proxy indicator of disturbance intensity, achieving online quantitative perception of system uncertainty through cycle-by-cycle accumulation. Based on the desired roll angle command, the roll angular velocity, and the desired roll angle command, the rolling angular velocity, and the roll angle rate of change, they calculate the roll angle tracking error and roll angle error rate of change. The control torque required for the roll channel is calculated using the rate of change of the rotational error, the sliding surface state variables, and the bounded upper limit of the disturbance. This control torque, by explicitly introducing a compensation term proportional to the upper limit of the disturbance, allows the control output to actively offset the equivalent input uncertainty caused by operational disturbances such as center of gravity shifts, thus maintaining the effectiveness of the control channel even when the aircraft's dynamic parameters are unknown and time-varying. A control signal is generated based on the control torque to control the aircraft's roll attitude, converting algorithmic commands into physical execution signals, forming a closed-loop control link. These features work synergistically, defining a convergence target through the sliding surface, using its state variables to drive a recursive estimation of the bounded upper limit of the disturbance, and embedding this estimate into the control torque generation to achieve adaptive compensation. Since this mechanism does not rely on an accurate model but maintains stability through real-time perception and active offsetting of system uncertainties, it can solve the problem of attitude instability caused by unknown changes such as drastic changes in the center of gravity during operation.

[0007] In some embodiments of the first aspect of this application, obtaining the roll angle and roll angular velocity of the current control cycle of the aircraft includes: Raw roll angle and roll angular velocity data are acquired through an inertial measurement unit; Perform a Kalman filter operation on the original roll angle and roll velocity data to generate filtered roll angle and roll velocity.

[0008] Compared with the prior art, the above embodiments have the following beneficial effects: Kalman filtering effectively improves the signal-to-noise ratio of attitude and angular velocity information by fusing multi-source sensor data and suppressing high-frequency noise, thereby providing high-precision state feedback for subsequent error calculation and control law generation, and avoiding control chattering or performance degradation caused by the amplification of original sensor noise.

[0009] In some embodiments of the first aspect of this application, the calculation of the roll angle tracking error and the roll angle error change rate includes: The roll angle is subtracted from the desired roll angle command to generate the roll angle tracking error. Perform numerical differentiation on the desired roll angle command to generate the first derivative of the desired roll angle command; The roll angle velocity is subtracted from the first derivative of the desired roll angle command to generate the roll angle error change rate.

[0010] Compared with the prior art, the above embodiments have the following beneficial effects: by performing numerical differentiation on the desired roll angle command to obtain its first derivative, and subtracting it from the measured angular velocity to obtain the error change rate, the controller can accurately reconstruct the complete error dynamics required for the sliding surface, ensuring that the sliding surface reaching law effectively drives the system state to converge to the sliding surface, and avoiding the failure of the sliding surface definition or the decrease in convergence speed due to the lack of derivative.

[0011] In some embodiments of the first aspect of this application, the step of constructing the sliding surface state variables based on the roll angle tracking error and the roll angle error change rate includes: The roll angle tracking error is multiplied by a preset sliding coefficient using a scalar multiplication operation, and then added to the roll angle error change rate to generate a sliding surface state variable; wherein the preset sliding coefficient is a positive number.

[0012] Compared with the prior art, the above embodiments have the following beneficial effects: the sliding surface state variable is generated by multiplying the roll angle tracking error with the preset sliding mode coefficient and then superimposing the error change rate. This construction form corresponds to a stable first-order error dynamic system. When the sliding mode coefficient is positive, the error differential equation corresponding to the sliding surface has a negative real part eigenvalue, ensuring that the error decays at an exponential rate. This design enables the system to achieve fast and overshoot-free attitude tracking once it enters the sliding mode state, providing a theoretical guarantee for high dynamic performance.

[0013] In some embodiments of the first aspect of this application, the recursive updating of the bounded upper bound of the spacecraft roll channel based on the sliding surface state variables includes: The absolute value of the sliding surface state quantity is multiplied by a preset adaptive gain coefficient to generate the upper boundary update quantity of the disturbance. The updated disturbance upper bound is summed with the estimated value of the bounded disturbance upper bound of the previous control period to generate the bounded disturbance upper bound of the current control period.

[0014] Compared with the prior art, the above embodiments have the following beneficial effects: the further the system deviates from the sliding surface, the stronger the disturbance or the current compensation is insufficient, and the upper bound estimation needs to be enhanced to improve the control strength. Therefore, the disturbance upper bound estimation uses the product of the absolute value of the sliding surface state variable and the adaptive gain as the update quantity, and is accumulated with the previous cycle estimate to avoid conservatism or undercompensation caused by a fixed upper bound, so that the disturbance suppression capability can be dynamically adjusted according to actual needs, and the control signal is prevented from being too aggressive while ensuring robustness.

[0015] In some embodiments of the first aspect of this application, calculating the control torque required for the roll channel based on the desired roll angle command, the roll angle error rate of change, the sliding surface state quantity, and the bounded disturbance upper limit includes: The equivalent control quantity is generated by subtracting the product of the roll angle error change rate and the preset sliding mode coefficient, and the product of the sliding mode surface state quantity and the preset reaching law gain from the second derivative of the desired roll angle command. Based on the sliding surface state variables and the preset boundary layer thickness parameters, the saturation function value of the sliding surface state variables is generated, and a scalar multiplication operation is performed with the bounded disturbance upper bound to generate a disturbance compensation term. Subtract the disturbance compensation term from the equivalent control quantity, and then divide by the nominal gain coefficient of the roll channel to generate the control torque required for the roll channel.

[0016] Compared to existing technologies, the above embodiments have the following advantages: By subtracting the product of the roll angle error change rate and the preset sliding mode coefficient from the second derivative of the desired roll angle command, and then subtracting the product of the sliding mode surface state quantity and the preset reaching law gain, an equivalent control quantity is generated to construct the reference control force required to maintain sliding mode dynamics under disturbance-free conditions; at the same time, a saturation function value is generated based on the sliding mode surface state quantity and the preset boundary layer thickness parameter, and multiplied by the bounded disturbance upper bound to form a disturbance compensation term, so that the control torque can be adaptively corrected in combination with the disturbance estimation result; finally, the equivalent control quantity is subtracted from the disturbance compensation term and divided by the nominal gain coefficient of the roll channel to obtain the executable roll channel control torque; this calculation process unifies ideal dynamic tracking, disturbance adaptive compensation, and physical input mapping into a single control law, so that the system can still output a control force that matches the current disturbance level under complex conditions of drastic changes in the center of gravity and coexistence of external disturbances, ensuring the robustness of attitude tracking and the executableness of control commands.

[0017] In some embodiments of the first aspect of this application, generating the saturation function value of the sliding surface state quantity based on the sliding surface state quantity and a preset boundary layer thickness parameter includes: The sliding surface state quantity is compared with the preset boundary layer thickness parameter. If the sliding surface state quantity is greater than the preset boundary layer thickness parameter, the saturation function value is one. If the sliding surface state quantity is less than the negative of the preset boundary layer thickness parameter, the saturation function value is negative one. If the absolute value of the sliding surface state quantity is less than or equal to the preset boundary layer thickness parameter, the saturation function value is the ratio of the sliding surface state quantity to the preset boundary layer thickness parameter.

[0018] Compared with the prior art, the above embodiments have the following beneficial effects: By comparing the magnitude relationship between the sliding surface state quantity and the preset boundary layer thickness parameter, the saturation function value is set to one when the sliding surface state quantity is greater than the boundary layer thickness, and set to negative one when it is less than the negative of the boundary layer thickness. When its absolute value does not exceed the boundary layer thickness, it is set to the ratio of the sliding surface state quantity to the boundary layer thickness parameter. This makes the saturation function exhibit a fixed amplitude sign switching outside the boundary layer to maintain a strong robust suppression capability against disturbances, and exhibit a continuous output proportional to the sliding surface state quantity inside the boundary layer to avoid discontinuous jumps in the control law. Thus, while ensuring that the system has sufficient compensation capability for complex disturbances such as center of gravity shift, it effectively weakens the control chattering caused by the high-frequency switching of the ideal sign function, so that the motor control signal transitions smoothly near the sliding surface, improving the response quality and long-term operational reliability of the actuator.

[0019] Secondly, the present invention also provides an adaptive control system for the center of gravity of an aircraft, comprising: a data acquisition module, an upper bound update module, a torque calculation module, and a control module; The data acquisition module is used to acquire the roll angle and roll angular velocity of the aircraft in the current control cycle, as well as the desired roll angle command, and to calculate the roll angle tracking error and the roll angle error change rate. The upper bound update module is used to construct the sliding surface state quantity based on the roll angle tracking error and the roll angle error change rate, and recursively update the bounded disturbance upper bound in the aircraft roll channel according to the sliding surface state quantity. The torque calculation module is used to calculate the control torque required for the roll channel based on the desired roll angle command, the roll angle error change rate, the sliding surface state quantity, and the bounded disturbance upper limit; wherein, the control torque includes a disturbance compensation term that is proportional to the bounded disturbance upper limit; The control module is used to generate control signals for each motor of the aircraft in the current cycle according to the control torque, and drive each motor to control the roll attitude of the aircraft to complete flight control.

[0020] Compared to existing technologies, the above embodiments of this application have the following beneficial effects: They acquire the roll angle and roll angular velocity of the aircraft in the current control cycle, as well as the desired roll angle command, and calculate the roll angle tracking error and roll angle error rate of change. By constructing second-order error information containing attitude deviations and their rates of change, they provide complete state observation inputs for the high-order controller, enabling control decisions to simultaneously respond to the current deviation and the deviation development trend. Based on the roll angle tracking error and the roll angle error rate of change, they construct sliding surface state variables and recursively update the bounded disturbance upper bound in the aircraft's roll channel according to the sliding surface state variables. The sliding surface state variables serve as the target manifold for sliding mode control, and their construction can guide the system state to a robust switching surface. The recursive update mechanism of the bounded disturbance upper bound utilizes the amplitude of the sliding surface state variables as a proxy index of disturbance intensity, achieving online quantitative perception of system uncertainty through cycle-by-cycle accumulation. Based on the desired roll angle command, The control torque required for the roll channel is calculated using the roll angle error rate of change, sliding surface state variables, and bounded upper limit of the disturbance. This control torque, by explicitly introducing a compensation term proportional to the upper limit of the disturbance, allows the control output to actively offset the equivalent input uncertainty caused by operational disturbances such as center of gravity shifts, thus maintaining the effectiveness of the control channel even when the aircraft's dynamic parameters are unknown and time-varying. A control signal is generated based on the control torque to control the aircraft's roll attitude, converting algorithmic commands into physical execution signals, forming a closed-loop control link. These features work synergistically, defining a convergence target through the sliding surface, using its state variables to drive a recursive estimation of the bounded upper limit of the disturbance, and embedding this estimate into the control torque generation for adaptive compensation. Since this mechanism does not rely on an accurate model but maintains stability through real-time perception and active offsetting of system uncertainties, it can solve the problem of attitude instability caused by unknown changes such as drastic changes in the center of gravity during operation.

[0021] Thirdly, the present invention also provides a computer program product, including a computer program or instructions, characterized in that, when the computer program or instructions are executed, they implement any one of the adaptive center of gravity control methods for aircraft of the present invention.

[0022] Fourthly, embodiments of this application also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements any one of the adaptive center-of-gravity control methods for aircraft of the present invention. Attached Figure Description

[0023] Figure 1 This is a flowchart illustrating an adaptive center of gravity control method for an aircraft provided in some embodiments of the present invention.

[0024] Figure 2This is a schematic diagram of the structure of an adaptive control system for the center of gravity of an aircraft provided in some embodiments of the present invention. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] Example 1: With the rapid development of drone technology, its application scenarios have expanded from traditional aerial photography and reconnaissance to more complex aerial operations. Integrated aerial work drones, a product of this trend, typically refer to specialized drones that integrate a flight platform with specific operational units (such as robotic arms, slings, spraying systems, and cargo delivery devices). These aircraft are widely used in logistics transportation, high-altitude operations, precision agriculture, and emergency rescue.

[0027] Unlike traditional drones, the system parameters of an integrated aerial work platform undergo drastic and unpredictable changes during mission execution. For example, when the robotic arm extends, grabs, or releases objects in the air, the suspended load swings during flight, or the liquid mass of an agricultural drone continuously decreases during spraying, these operations cause real-time and significant dynamic changes in the aircraft's total mass and center of gravity.

[0028] The stability and maneuverability of an aircraft are extremely sensitive to its center of gravity position. A shift in the center of gravity generates additional disturbance torques, directly disrupting the aircraft's original dynamic balance. In severe cases, this can lead to attitude instability, decreased control precision, and even loss of control and crashes. Therefore, ensuring that integrated flight operations aircraft can safely, stably, and accurately complete flight and operational tasks under these dynamic, strongly coupled, and uncertain system characteristics is currently a core technological bottleneck and key challenge in this field.

[0029] Regarding the issue of changes in the center of gravity of aircraft, the industry and academia currently offer the following main solutions: 1. Passive Trim and Pre-planning Scheme: Before the flight mission begins, ballast or component installation positions are pre-configured through ground tests or simulation calculations to trim one or more typical center of gravity positions. For scenarios with fixed mission procedures, flight trajectories and attitudes can also be pre-planned to mitigate the impact of center of gravity shifts to some extent. However, this approach lacks flexibility and adaptability: pre-trim schemes are only effective for a few preset operating conditions and cannot cope with unknown or continuous changes in the center of gravity during flight. Once the mission scenario changes, tedious ground adjustments and tests are required, completely lacking autonomous adaptability.

[0030] 2. Active Mechanical Balancing Systems: This type of technology actively adjusts the aircraft's physical center of gravity by adding movable mechanical components, such as moving mass blocks, battery rails, or deformable arms. For example, patent CN111942577A proposes a method for balancing the center of gravity of a UAV through mechanical structural adjustments. Furthermore, research has specifically designed operational flying robots with center of gravity adjustment mechanisms, using dynamic calculations and moving counterweights to compensate for center of gravity shifts in real time. However, this approach increases the system burden: introducing mechanical balancing devices undoubtedly increases the structural complexity of the aircraft, the total system weight, and power consumption, thus shortening flight time. Simultaneously, the additional mechanical components introduce new points of failure, reducing the overall reliability of the system. More importantly, the response speed of mechanical adjustment is limited; its compensation effect is often delayed for instantaneous center of gravity impacts caused by rapid movements of robotic arms.

[0031] 3. Gain Scheduling Scheme Based on Classical Control Algorithms: This is currently the most widely used software control scheme. Its core idea is to use a traditional proportional-integral-derivative (PID) controller or its variants. Since PID controller parameters are tuned at a fixed operating point, their control performance significantly degrades when the system model (e.g., center of gravity position) changes. To address this, a gain scheduling strategy is typically employed, which involves offline tuning of multiple sets of PID parameters based on different load conditions or flight phases, switching between them during flight based on sensor information or preset logic. However, this scheme lacks robustness: the PID controller is essentially a linear controller, and its performance deteriorates severely in nonlinear, time-varying systems. While the gain scheduling scheme improves adaptability to some extent, it is a piecewise approximation approach and cannot achieve smooth, real-time adaptation to continuously changing system parameters. At parameter switching points, the control output may jump; for unpreset operating conditions, controller performance cannot be guaranteed. Furthermore, PID parameter tuning itself is very time-consuming and usually only optimizes a single objective, making it difficult to balance speed and stability. In strongly coupled aircraft systems, changes in the parameters of one attitude loop can degrade the performance of other loops.

[0032] 4. Advanced Model-Based Control Algorithms: Some research attempts to utilize more advanced control theories such as Model Predictive Control (MPC) and Linear Quadratic Regulators (LQR). These methods rely on relatively accurate system dynamics models. For example, some studies have designed controllers to eliminate instability by establishing dynamic models that include center-of-gravity shifts. However, these methods typically involve large computational loads, demanding high computing power from airborne processors, and real-time performance may become a bottleneck. Furthermore, this approach has a high implementation threshold: model-based control methods, such as model predictive control, while theoretically superior, are extremely dependent on accurate system models. For aerial work platforms, the geometry, mass, and grasping position of the load may be unknown, making it nearly impossible to establish an accurate model covering all operating conditions. Moreover, the enormous online computational demands place stringent requirements on the performance of airborne computers, limiting their widespread application on cost- and power-sensitive UAV platforms.

[0033] To address the limitations and shortcomings of existing technologies in practical applications, this invention provides a flight control solution based purely on software algorithms, requiring no additional mechanical leveling devices. This solution enables high-precision and robust attitude tracking control of the aircraft even when key dynamic parameters such as the mass, moment of inertia, and center of gravity of the integrated flight operation vehicle undergo unknown, rapid, and significant changes. It also effectively suppresses interference from the external environment (such as wind disturbance), thereby ensuring flight stability and operational safety throughout the entire dynamic operation process. The technical challenges to be solved by this invention can be broken down into the following aspects: 1. Uncertainty Modeling and Suppression: How to model the complex disturbances, which are composed of center of gravity shift, mass change and external wind disturbance, as an uncertainty in the system dynamics model, and design a controller to effectively suppress them.

[0034] 2. Controller Adaptability: How to enable the controller to estimate the upper bound of the uncertainty in real time online and automatically adjust the strength of the control law based on the estimation results, thereby avoiding the performance degradation or instability caused by the fixed gain of traditional controllers.

[0035] 3. The balance between control performance and robustness: How to ensure that the algorithm provides strong robustness while also guaranteeing excellent dynamic performance, such as fast convergence speed, small tracking error and small overshoot, and avoiding high-frequency jitter in the control signal.

[0036] 4. Practicality of the algorithm: How to ensure that the computational complexity of the algorithm is moderate, so that it can run in real time on current mainstream embedded flight control hardware, and realize the leap from theory to engineering application.

[0037] Specifically, the design of this invention is as follows: I. Establishment of the dynamic model of the integrated flight operation aircraft: First, a dynamic model of the roll channel of a multi-rotor UAV considering center of gravity shift and external disturbances is established. The dynamic equation of the roll motion can be expressed as: ;in: φ is the roll angle of the aircraft.

[0038] d²φ / dt² is the roll acceleration.

[0039] Ixx is the moment of inertia of the aircraft about the x-axis. During operation, Ixx is time-varying.

[0040] τ_φ is the rolling control torque generated by the motor, which is our control input.

[0041] M_φ_dist is the total roll disturbance torque.

[0042] M_φ_dist is a composite term that includes all uncertainties and disturbances, and it is the key term that this algorithm needs to handle: M_φ_dist = M_cg + M_gyro + M_wind; Here, M_cg is the disturbance torque generated by the shift in the center of gravity. When the center of gravity G deviates from the geometric center O, the total weight mg of the aircraft will generate a torque r_cg × mg, where r_cg is the vector from the geometric center O to the center of gravity G. The component of this torque acts on the roll axis, which is M_cg. Due to load variations, m and r_cg are both unknown and time-varying.

[0043] M_gyro is the torque generated by coupling terms such as the gyro effect, and it is also affected by changes in system parameters.

[0044] M_wind is the random disturbance torque generated by the external wind field.

[0045] To simplify controller design, we rewrite the original dynamic equations in state-space form. We define the state variables as x1 = φ (roll angle) and x2 = dφ / dt (roll angular velocity). The system equations are then: dx1 / dt = x2; dx2 / dt = (τ_φ + M_φ_dist) / Ixx; Let b = 1 / Ixx, f(x) = M_φ_dist / Ixx. Since Ixx and M_φ_dist are both time-varying and uncertain, b is an uncertain control gain, and f(x) is an unknown function that incorporates all model uncertainties and external disturbances. The system can be rewritten as: ; Assume b = b0 + Δb, where b0 is the known part calculated from the aircraft's nominal parameters, and Δb is the uncertain part. Then the equation becomes: ; make Let d(t) be the total unknown composite disturbance of the system. Our goal is to design a controller τ_φ that, given the unknown d(t), makes the roll angle φ accurately track the desired command φ_d.

[0046] II. Design of Adaptive Control Law for Sliding Mode Variable Structure: The core of this invention lies in designing an adaptive sliding mode controller. The design process consists of three steps: defining the sliding surface, designing the adaptive sliding mode control law, and proving its stability.

[0047] Step 1: Design the sliding surface: The function of a sliding surface is to define a desired system dynamic. Once the system state reaches the sliding surface, it will slide steadily along that surface towards the equilibrium point.

[0048] The tracking error of the roll angle is defined as e = φ - φ_d.

[0049] We design a linear sliding surface s, whose expression is: ; Here, c is a design constant greater than zero. The value of c determines the convergence rate of the system on the sliding surface; the larger the value of c, the faster the convergence. When the system state is forcibly constrained to the sliding surface s = 0, it means... This is a stable first-order differential equation, and its solution is... This indicates that the tracking error e will converge to zero rapidly in an exponential manner.

[0050] Step 2: Design the adaptive reaching law and control law: The goal of designing the control law τ_φ is to ensure that, regardless of the initial state of the system, it reaches the sliding surface s = 0 within a finite amount of time and remains there. This process is called the approaching motion, and is detailed below: Differentiate with respect to the sliding surface s: ; To make s approach zero, we expect ds / dt to have the opposite sign to s. Therefore, we design a sliding mode reaching law, an improved form of the exponential reaching law, to reduce chattering: ; Where: k and ε are design constants that are greater than zero. This term guarantees a faster approach speed when s is large. This term guarantees that even when s approaches zero, the speed at which the system state crosses the sliding surface remains constant, ensuring finite-time convergence.

[0051] sat(s) is a saturation function, defined as follows: sat(s) = 1, if s > δ; sat(s) = -1, if s < -δ; sat(s) = s / δ, if |s| ≤ δ; where δ is a small positive number representing the boundary layer thickness. Using a saturation function instead of the traditional sign function sign() allows for smooth control switching within the boundary layer, thereby significantly suppressing the inherent high-frequency chattering problem of sliding mode control and protecting actuators such as motors.

[0052] By simultaneously solving the two expressions for ds / dt, the control law τ_φ can be obtained: ; The results were: ; This control law consists of two parts: Equivalent control section: This is the control quantity required to maintain the system on the sliding surface s=0, assuming no disturbance (d(t)=0).

[0053] Switching control section: This part is used to overcome the uncertainty d(t) and push the system state toward the sliding surface.

[0054] The problem is that the composite disturbance d(t) is unknown, so the above control law cannot be directly implemented. Therefore, this application introduces an adaptive law to estimate the upper bound of d(t) online.

[0055] Suppose the composite perturbation d(t) is bounded, i.e., |d(t)| ≤ D, but the magnitude of D is unknown. Our goal is to design an adaptive law to estimate D, denoted as D_hat.

[0056] Designing ε as an adaptive term in the switching control section, let ε = D_hat. Then the final adaptive sliding mode control law is: ; Now, we need to design the update rule for D_hat, i.e., the adaptive law. Based on Lyapunov stability theory, our designed adaptive law is as follows: ; Here, γ is a positive-zero adaptive gain that determines the update rate of the estimate. The physical meaning of this adaptive law is: when the system state deviates from the sliding surface (|s|>0), the estimate of the upper bound of the disturbance, D_hat, is increased, thereby enhancing the controller's suppression capability; when the system state is on the sliding surface (s tends to 0 within the boundary layer), the growth of D_hat slows down or stops. This ensures that the control force is "distributed on demand," effectively suppressing disturbances without generating excessive control output.

[0057] Step 3: Proving System Stability To prove that the designed control system is stable and effective, we use Lyapunov's second method.

[0058] Define the Lyapunov candidate function V: ; Where D is the true (but unknown) upper bound of the perturbation d(t). V is positive definite because it consists of the sum of two squares.

[0059] Find the time derivative dV / dt with respect to V: ; Substitute the expression for ds / dt and the adaptive law into: ; Substitute the expression for the control law τ_φ into the equation and simplify: ; ; ; Considering Outside the boundary layer, it equals |s|; inside the boundary layer, it is s² / δ, and... We can analyze the sign of dV / dt.

[0060] Outside the sliding surface boundary layer, |s|>δ, at which point sat(s) = sgn(s), so Substituting into the above formula: ; ; because ,therefore: ; ; Since k > 0, dV / dt ≤ 0. According to Barbalat's lemma, this guarantees that s will eventually converge to the boundary layer |s| ≤ δ. This means that the tracking errors e and de / dt are eventually bounded, and the size of the bounds can be adjusted by designing the parameters c and δ, making them arbitrarily small.

[0061] Therefore, this application designs an adaptive sliding mode controller that can ensure the stability of the closed-loop system without knowing the system dynamic parameters and external disturbances, and makes the attitude tracking error of the aircraft converge to a preset, sufficiently small neighborhood.

[0062] Furthermore, in practical implementation, the algorithm is deployed in the embedded flight control system of the integrated flight operation machine for real-time cyclic operation, and the control is as follows: Please refer to Figure 1 To address the problem of attitude instability caused by drastic and unknown changes in the center of gravity during operation in existing technologies, an embodiment of the present invention provides an adaptive control method for the center of gravity of an aircraft, comprising steps S1 to S4: Step S1: Obtain the roll angle and roll rate of the aircraft in the current control cycle, as well as the desired roll angle command, and calculate the roll angle tracking error and the roll angle error change rate.

[0063] Furthermore, step S1 can be implemented through the following preferred embodiments, including steps S11-S15, as follows: S11: Acquire raw roll angle and roll angular velocity data through the inertial measurement unit; S12: Perform Kalman filtering on the original roll angle and roll velocity data to generate filtered roll angle and roll velocity.

[0064] In this preferred embodiment, Kalman filtering effectively improves the signal-to-noise ratio of attitude and angular velocity information by fusing multi-source sensor data and suppressing high-frequency noise, thereby providing high-precision state feedback for subsequent error calculation and control law generation, and avoiding control chattering or performance degradation caused by the amplification of original sensor noise.

[0065] S13: Subtract the roll angle from the desired roll angle command to generate a roll angle tracking error; S14: Perform numerical differentiation on the desired roll angle command to generate the first derivative of the desired roll angle command; S15: Subtract the first derivative of the roll angular velocity from the desired roll angle command to generate the roll angle error change rate.

[0066] In this preferred embodiment, the first derivative of the desired roll angle command is obtained by performing numerical differentiation, and the error change rate is obtained by subtracting it from the measured angular velocity. This enables the controller to accurately reconstruct the complete error dynamics required for the sliding surface, ensuring that the sliding surface reaching law effectively drives the system state to converge toward the sliding surface, and avoiding the failure of the sliding surface definition or the decrease in convergence speed due to the lack of derivative.

[0067] Step S2: Based on the roll angle tracking error and the roll angle error change rate, construct the sliding surface state variables, and recursively update the bounded upper limit of the spacecraft roll channel according to the sliding surface state variables.

[0068] Furthermore, step S2 can be implemented through the following preferred embodiments, including steps S21-S23, as follows: S21: Perform a scalar multiplication operation between the roll angle tracking error and the preset sliding mode coefficient, and add the result to the roll angle error change rate to generate the sliding mode surface state quantity; wherein, the preset sliding mode coefficient is a positive number.

[0069] In this preferred embodiment, the rolling angle tracking error is multiplied by a preset sliding mode coefficient and then superimposed with the error change rate to generate the sliding surface state variable. This construction corresponds to a stable first-order error dynamic system. When the sliding mode coefficient is positive, the error differential equation corresponding to the sliding surface has a negative real part eigenvalue, ensuring that the error decays at an exponential rate. This design enables the system to achieve fast and overshoot-free attitude tracking once it enters the sliding mode state, providing a theoretical guarantee for high dynamic performance.

[0070] S22: Perform a scalar multiplication operation between the absolute value of the sliding surface state quantity and the preset adaptive gain coefficient to generate the upper boundary update quantity of the disturbance; S23: The updated disturbance upper bound is added to the estimated value of the bounded disturbance upper bound of the previous control period to generate the bounded disturbance upper bound of the current control period.

[0071] In this preferred embodiment, the further the system deviates from the sliding surface, the stronger the disturbance or the current compensation is insufficient. It is necessary to enhance the upper bound estimation to improve the control strength. Therefore, the disturbance upper bound estimation uses the product of the absolute value of the sliding surface state variable and the adaptive gain as the update quantity, and adds it to the previous cycle estimate to avoid conservatism or undercompensation caused by a fixed upper bound. This allows the disturbance suppression capability to be dynamically adjusted according to actual needs, ensuring robustness while preventing the control signal from being too aggressive.

[0072] Step S3: Calculate the control torque required for the roll channel based on the desired roll angle command, the roll angle error change rate, the sliding surface state quantity, and the bounded disturbance upper limit; wherein the control torque includes a disturbance compensation term that is proportional to the bounded disturbance upper limit.

[0073] Furthermore, step S3 can be implemented through the following preferred embodiments, including steps S31-S33, as follows: S31: Subtract the product of the roll angle error change rate and the preset sliding mode coefficient, and the product of the sliding mode surface state quantity and the preset reaching law gain from the second derivative of the desired roll angle command to generate an equivalent control quantity. S32: Based on the sliding surface state variables and the preset boundary layer thickness parameters, generate the saturation function value of the sliding surface state variables, and perform scalar multiplication with the bounded disturbance upper bound to generate a disturbance compensation term; S33: Subtract the disturbance compensation term from the equivalent control quantity, and then divide by the nominal gain coefficient of the roll channel to generate the control torque required for the roll channel.

[0074] In this preferred embodiment, an equivalent control quantity is generated by subtracting the product of the roll angle error rate of change and the preset sliding mode coefficient from the second derivative of the desired roll angle command, and then subtracting the product of the sliding surface state quantity and the preset reaching law gain. This generates the baseline control force required to maintain sliding mode dynamics under disturbance-free conditions. Simultaneously, a saturation function value is generated based on the sliding surface state quantity and the preset boundary layer thickness parameter, and multiplied by the bounded upper limit of the disturbance to form a disturbance compensation term. This allows the control torque to be adaptively corrected based on the disturbance estimation results. Finally, the equivalent control quantity is subtracted from the disturbance compensation term and divided by the nominal gain coefficient of the roll channel to obtain the executable roll channel control torque. This calculation process unifies ideal dynamic tracking, disturbance adaptive compensation, and physical input mapping into a single control law. This ensures that the system can still output a control force that matches the current disturbance level under complex conditions where the center of gravity changes drastically and external disturbances coexist, guaranteeing the robustness of attitude tracking and the executability of control commands.

[0075] Furthermore, the saturation function value of S32 can be generated through the following preferred embodiment, as follows: The sliding surface state quantity is compared with the preset boundary layer thickness parameter. If the sliding surface state quantity is greater than the preset boundary layer thickness parameter, the saturation function value is one. If the sliding surface state quantity is less than the negative of the preset boundary layer thickness parameter, the saturation function value is negative one. If the absolute value of the sliding surface state quantity is less than or equal to the preset boundary layer thickness parameter, the saturation function value is the ratio of the sliding surface state quantity to the preset boundary layer thickness parameter.

[0076] In this preferred embodiment, by comparing the magnitude relationship between the sliding surface state quantity and the preset boundary layer thickness parameter, the saturation function value is set to one when the sliding surface state quantity is greater than the boundary layer thickness, and set to negative one when it is less than the negative of the boundary layer thickness. When its absolute value does not exceed the boundary layer thickness, it is set to the ratio of the sliding surface state quantity to the boundary layer thickness parameter. This allows the saturation function to exhibit a fixed amplitude sign switching outside the boundary layer to maintain strong robust suppression of disturbances, while inside the boundary layer it exhibits a continuous output proportional to the sliding surface state quantity to avoid discontinuous jumps in the control law. Thus, while ensuring that the system has sufficient compensation capability for complex disturbances such as center of gravity shift, it effectively weakens the control chattering caused by high-frequency switching of the ideal sign function, making the motor control signal transition smoothly near the sliding surface, and improving the response quality and long-term operational reliability of the actuator.

[0077] Step S4: Based on the control torque, generate control signals for each motor of the aircraft in the current cycle, and drive each motor to control the roll attitude of the aircraft to complete flight control.

[0078] Referring to the aforementioned design process, in practical implementation, after the algorithm is deployed, each round of the loop is as follows: Step 1: Collect real-time roll angle φ, roll angular velocity dφ / dt, and the desired roll angle command φ_d issued by the ground station or mission system through flight control sensors (gyroscope, accelerometer, magnetometer).

[0079] First, the Kalman filter algorithm is used to denoise the raw sensor data and remove measurement noise. Simultaneously, the desired command φ_d is parsed, and its first and second derivatives dφ_d / dt and d²φ_d / dt² are calculated using a numerical differentiation algorithm. Finally, the denoised roll angle φ_clean, roll angular velocity (dφ / dt)_clean, and the desired command derivatives (dφ_d / dt) and (d²φ_d / dt²) are obtained.

[0080] Step 2: Calculate the roll angle tracking error using the formula e = φ_clean - φ_d; The rate of change of error is calculated using the formula de / dt = (dφ / dt)_clean - dφ_d / dt. Then apply the sliding surface formula Calculate the sliding surface state variables (c is a preset positive number that needs to be calibrated in advance).

[0081] Finally, the tracking error e, the error change rate de / dt, and the sliding surface state quantity s are obtained.

[0082] Step 3: Obtain the upper bound estimate of the perturbation stored in the previous cycle, D_hat_prev, and then apply the adaptive law formula. Perform numerical integration (using Euler integration, with the integration step size equal to the flight control cycle period T) to calculate the estimated upper bound of the current cycle disturbance. (γ is the preset adaptive gain, which needs to be calibrated in advance).

[0083] Step 4: Calculate sat(s) according to the saturation function formula (sat(s)=1, if s>δ; sat(s)=-1, if s<-δ; sat(s)=s / δ, if |s|≤δ); then calculate the roll control torque τ_φ according to the control law formula τ_φ = (1 / b0)(d²φ_d / dt² - cde / dt - ks -D_hat_currsat(s)) (k is a preset positive number and needs to be calibrated in advance).

[0084] Where d²φ_d / dt² is the second derivative of the desired command; e is the tracking error; de / dt is the error rate of change; s is the sliding surface state variable; D_hat_curr is the estimated upper bound of the disturbance; b0 is the known control gain calculated from the nominal parameters of the aircraft; and k and δ are preset design constants.

[0085] Step 5: Using the motor torque distribution algorithm, map the roll control torque command τ_φ to the corresponding motor speed command (calculate the speed corresponding to the torque to be output by each motor based on the multi-rotor motor layout); then convert the speed command into the PWM control signal of each corresponding motor (calculated according to the preset speed-PWM mapping table).

[0086] Step 6: The flight control system sends the PWM signals of each motor to the corresponding motor driver through the output interface, driving the motor to adjust the speed and generate actual rolling torque to counteract the combined disturbances such as center of gravity shift and wind disturbance.

[0087] In summary, compared with the prior art, the above embodiments of this application have the following beneficial effects: They acquire the roll angle and roll angular velocity of the aircraft in the current control cycle, as well as the desired roll angle command, and calculate the roll angle tracking error and roll angle error rate of change. By constructing second-order error information containing attitude deviations and their rates of change, they provide complete state observation inputs for the high-order controller, enabling control decisions to simultaneously respond to the current deviation and the deviation development trend. Based on the roll angle tracking error and the roll angle error rate of change, they construct sliding surface state variables and recursively update the bounded disturbance upper bound in the aircraft's roll channel according to the sliding surface state variables. The sliding surface state variables serve as the target manifold for sliding mode control, and their construction can guide the system state to a highly robust switching surface. The recursive update mechanism of the bounded disturbance upper bound utilizes the amplitude of the sliding surface state variables as a proxy index of disturbance intensity, achieving online quantitative perception of system uncertainty through cycle-by-cycle accumulation. Based on the desired roll angle command... The control torque required for the roll channel is calculated by considering the roll angle error rate of change, the sliding surface state variables, and the bounded upper limit of the disturbance. This control torque is achieved by explicitly introducing a compensation term proportional to the upper limit of the disturbance, enabling the control output to actively offset the equivalent input uncertainty caused by operational disturbances such as center of gravity shifts. This maintains the effectiveness of the control channel even when the aircraft's dynamic parameters are unknown and time-varying. A control signal is generated based on the control torque to control the aircraft's roll attitude, transforming algorithmic commands into physical execution signals and forming a closed-loop control link. These features work synergistically, defining a convergence target through the sliding surface, using its state variables to drive a recursive estimation of the bounded upper limit of the disturbance, and embedding this estimate into the control torque generation for adaptive compensation. Since this mechanism does not rely on an accurate model but maintains stability through real-time sensing and active offsetting of system uncertainties, it can solve the problem of attitude instability caused by unknown changes such as drastic changes in the center of gravity during operation.

[0088] Example 2: Please refer to Figure 2 Based on the same inventive concept, the present invention discloses an adaptive control system for the center of gravity of an aircraft, comprising: a data acquisition module M1, an upper bound update module M2, a torque calculation module M3, and a control module M4; The data acquisition module M1 is used to acquire the roll angle and roll angular velocity of the aircraft in the current control cycle, as well as the desired roll angle command, and to calculate the roll angle tracking error and the roll angle error change rate.

[0089] Furthermore, the data acquisition module M1 includes: a data acquisition unit and a preprocessing unit; The acquisition unit is used to acquire raw roll angle and roll angular velocity data through an inertial measurement unit. The preprocessing unit is used to perform Kalman filtering on the original roll angle and roll velocity data to generate filtered roll angle and roll velocity.

[0090] In this preferred embodiment, Kalman filtering effectively improves the signal-to-noise ratio of attitude and angular velocity information by fusing multi-source sensor data and suppressing high-frequency noise, thereby providing high-precision state feedback for subsequent error calculation and control law generation, and avoiding control chattering or performance degradation caused by the amplification of original sensor noise.

[0091] Furthermore, the data acquisition module M1 also includes: an error calculation unit, a differential unit, and a rate of change calculation unit; The error calculation unit is used to subtract the roll angle from the desired roll angle command to generate a roll angle tracking error. The differential unit is used to perform numerical differentiation operations on the desired roll angle command to generate the first derivative of the desired roll angle command; The rate of change calculation unit is used to subtract the first derivative of the roll angular velocity from the desired roll angle command to generate the roll angle error rate of change.

[0092] In this preferred embodiment, the first derivative of the desired roll angle command is obtained by performing numerical differentiation, and the error change rate is obtained by subtracting it from the measured angular velocity. This enables the controller to accurately reconstruct the complete error dynamics required for the sliding surface, ensuring that the sliding surface reaching law effectively drives the system state to converge toward the sliding surface, and avoiding the failure of the sliding surface definition or the decrease in convergence speed due to the lack of derivative.

[0093] The upper bound update module M2 is used to construct the sliding surface state variables based on the roll angle tracking error and the roll angle error change rate, and recursively update the bounded disturbance upper bound in the aircraft roll channel according to the sliding surface state variables.

[0094] Furthermore, the upper bound update module M2 includes: a state variable generation unit; The state quantity generation unit is used to perform a scalar multiplication operation between the roll angle tracking error and a preset sliding mode coefficient, and add the result to the roll angle error change rate to generate a sliding mode surface state quantity; wherein the preset sliding mode coefficient is a positive number.

[0095] In this preferred embodiment, the rolling angle tracking error is multiplied by a preset sliding mode coefficient and then superimposed with the error change rate to generate the sliding surface state variable. This construction corresponds to a stable first-order error dynamic system. When the sliding mode coefficient is positive, the error differential equation corresponding to the sliding surface has a negative real part eigenvalue, ensuring that the error decays at an exponential rate. This design enables the system to achieve fast and overshoot-free attitude tracking once it enters the sliding mode state, providing a theoretical guarantee for high dynamic performance.

[0096] Furthermore, the upper bound update module M2 also includes: an update amount calculation unit and an accumulation unit; The update calculation unit is used to perform a scalar multiplication operation between the absolute value of the sliding surface state quantity and a preset adaptive gain coefficient to generate the upper boundary update quantity of the disturbance. The accumulation unit is used to accumulate the updated disturbance upper bound with the estimated value of the bounded disturbance upper bound of the previous control period to generate the bounded disturbance upper bound of the current control period.

[0097] The further the system deviates from the sliding surface, the stronger the disturbance or the insufficient current compensation. It is necessary to enhance the upper bound estimation to improve the control strength. Therefore, the disturbance upper bound estimation uses the product of the absolute value of the sliding surface state variable and the adaptive gain as the update quantity, and adds it to the previous cycle estimate to avoid conservatism or undercompensation caused by a fixed upper bound. This allows the disturbance suppression capability to be dynamically adjusted according to actual needs, ensuring robustness while preventing the control signal from being too aggressive.

[0098] The torque calculation module M3 is used to calculate the control torque required for the roll channel based on the desired roll angle command, the roll angle error change rate, the sliding surface state quantity, and the bounded disturbance upper limit; wherein, the control torque includes a disturbance compensation term that is proportional to the bounded disturbance upper limit.

[0099] Furthermore, the torque calculation module M3 includes: an equivalent control quantity calculation unit, a disturbance compensation term calculation unit, and a torque generation unit; The equivalent control quantity calculation unit is used to subtract the product of the roll angle error change rate and the preset sliding mode coefficient from the second derivative of the desired roll angle command, and the product of the sliding mode surface state quantity and the preset reaching law gain to generate the equivalent control quantity. The disturbance compensation term calculation unit is used to generate the saturation function value of the sliding surface state quantity based on the sliding surface state quantity and the preset boundary layer thickness parameter, and perform scalar multiplication operation with the bounded disturbance upper bound to generate the disturbance compensation term. The torque generation unit is used to subtract the disturbance compensation term from the equivalent control quantity and then divide it by the nominal gain coefficient of the roll channel to generate the control torque required for the roll channel.

[0100] In this preferred embodiment, an equivalent control quantity is generated by subtracting the product of the roll angle error rate of change and the preset sliding mode coefficient from the second derivative of the desired roll angle command, and then subtracting the product of the sliding surface state quantity and the preset reaching law gain. This generates the baseline control force required to maintain sliding mode dynamics under disturbance-free conditions. Simultaneously, a saturation function value is generated based on the sliding surface state quantity and the preset boundary layer thickness parameter, and multiplied by the bounded upper limit of the disturbance to form a disturbance compensation term. This allows the control torque to be adaptively corrected based on the disturbance estimation results. Finally, the equivalent control quantity is subtracted from the disturbance compensation term and divided by the nominal gain coefficient of the roll channel to obtain the executable roll channel control torque. This calculation process unifies ideal dynamic tracking, disturbance adaptive compensation, and physical input mapping into a single control law. This ensures that the system can still output a control force that matches the current disturbance level under complex conditions where the center of gravity changes drastically and external disturbances coexist, guaranteeing the robustness of attitude tracking and the executability of control commands.

[0101] Furthermore, the disturbance compensation term calculation unit includes: a saturation function value generation subunit; The saturation function value generation subunit is used to compare the sliding surface state quantity with the preset boundary layer thickness parameter. If the sliding surface state quantity is greater than the preset boundary layer thickness parameter, the saturation function value is one. If the sliding surface state quantity is less than the negative of the preset boundary layer thickness parameter, the saturation function value is negative one. If the absolute value of the sliding surface state quantity is less than or equal to the preset boundary layer thickness parameter, the saturation function value is the ratio of the sliding surface state quantity to the preset boundary layer thickness parameter.

[0102] In this preferred embodiment, by comparing the magnitude relationship between the sliding surface state quantity and the preset boundary layer thickness parameter, the saturation function value is set to one when the sliding surface state quantity is greater than the boundary layer thickness, and set to negative one when it is less than the negative of the boundary layer thickness. When its absolute value does not exceed the boundary layer thickness, it is set to the ratio of the sliding surface state quantity to the boundary layer thickness parameter. This allows the saturation function to exhibit a fixed amplitude sign switching outside the boundary layer to maintain strong robust suppression of disturbances, while inside the boundary layer it exhibits a continuous output proportional to the sliding surface state quantity to avoid discontinuous jumps in the control law. Thus, while ensuring that the system has sufficient compensation capability for complex disturbances such as center of gravity shift, it effectively weakens the control chattering caused by high-frequency switching of the ideal sign function, making the motor control signal transition smoothly near the sliding surface, and improving the response quality and long-term operational reliability of the actuator.

[0103] The control module M4 is used to generate control signals for each motor of the aircraft in the current cycle according to the control torque, and drive each motor to control the roll attitude of the aircraft to complete flight control.

[0104] In summary, compared with the prior art, the embodiments of this application have the following beneficial effects: They acquire the roll angle and roll angular velocity of the aircraft in the current control cycle, as well as the desired roll angle command, and calculate the roll angle tracking error and the roll angle error rate of change. By constructing second-order error information containing attitude deviations and their rates of change, they provide complete state observation inputs for the high-order controller, enabling control decisions to simultaneously respond to the current deviation and the deviation development trend. Based on the roll angle tracking error and the roll angle error rate of change, they construct sliding surface state variables and recursively update the bounded disturbance upper bound in the aircraft's roll channel according to the sliding surface state variables. The sliding surface state variables serve as the target manifold for sliding mode control, and their construction can guide the system state to a switching surface with strong robustness. The recursive update mechanism of the bounded disturbance upper bound uses the amplitude of the sliding surface state variables as a proxy index of disturbance intensity, achieving online quantitative perception of system uncertainty through cycle-by-cycle accumulation. Based on the desired roll angle command... The control torque required for the roll channel is calculated by considering the roll angle error rate of change, the sliding surface state variables, and the bounded upper limit of the disturbance. This control torque, through the explicit introduction of a compensation term proportional to the upper limit of the disturbance, allows the control output to actively offset the equivalent input uncertainty caused by operational disturbances such as center of gravity shifts, thus maintaining the effectiveness of the control channel even when the aircraft's dynamic parameters are unknown and time-varying. A control signal is generated based on the control torque to control the aircraft's roll attitude, converting algorithmic commands into physical execution signals, forming a closed-loop control link. These features work synergistically, defining a convergence target through the sliding surface, using its state variables to drive a recursive estimation of the bounded upper limit of the disturbance, and embedding this estimate into the control torque generation for adaptive compensation. Since this mechanism does not rely on an accurate model but maintains stability through real-time perception and active offsetting of system uncertainties, it can solve the problem of attitude instability caused by unknown changes such as drastic changes in the center of gravity during operation.

[0105] Example 3: This invention also provides a computer program product, including a computer program or instructions, capable of running on a computing device or stored in any available medium. When the computer program product is run on at least one computing device, it causes the at least one computing device to execute any of the aircraft center of gravity adaptive control methods of this invention.

[0106] Example 4: This invention also provides a computer-readable storage medium storing at least one executable instruction that, when executed on an aircraft center of gravity adaptive control system, causes the aircraft center of gravity adaptive control system to perform one of the aircraft center of gravity adaptive control methods described in any of the above method embodiments.

[0107] Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of this application may be practiced without these specific details. Similarly, for the purpose of simplification and aiding understanding of one or more aspects of the invention, in the above description of exemplary embodiments of this application, various features of the embodiments are sometimes grouped together in a single embodiment, figure, or description thereof. The claims, which follow the detailed description, are hereby expressly incorporated into that detailed description, wherein each claim itself is a separate embodiment of this application.

[0108] Those skilled in the art will understand that the modules in the system of the embodiments can be adaptively changed and placed in one or more systems different from that embodiment. Modules, units, or components in the embodiments can be combined into a single module, unit, or component, and further, they can be divided into multiple sub-modules, sub-units, or sub-components, except that at least some of such features and / or processes or units are mutually exclusive.

Claims

1. An adaptive control method for the center of gravity of an aircraft, characterized in that, include: The system acquires the roll angle and roll rate of the aircraft in the current control cycle, as well as the desired roll angle command, and calculates the roll angle tracking error and the roll angle error rate of change. Based on the roll angle tracking error and the roll angle error change rate, a sliding surface state variable is constructed, and the bounded upper limit of the spacecraft roll channel is recursively updated according to the sliding surface state variable. Based on the desired roll angle command, the roll angle error rate of change, the sliding surface state quantity, and the bounded upper limit of the disturbance, the control torque required for the roll channel is calculated; wherein, the control torque includes a disturbance compensation term that is proportional to the bounded upper limit of the disturbance. Based on the control torque, control signals for each motor of the aircraft are generated in the current cycle, and each motor is driven to control the roll attitude of the aircraft to complete flight control.

2. The adaptive control method for the center of gravity of an aircraft as described in claim 1, characterized in that, The acquisition of the roll angle and roll rate of the aircraft in the current control cycle includes: Raw roll angle and roll angular velocity data are acquired through an inertial measurement unit; Perform a Kalman filter operation on the original roll angle and roll velocity data to generate filtered roll angle and roll velocity.

3. The adaptive control method for the center of gravity of an aircraft as described in claim 2, characterized in that, The calculation of the roll angle tracking error and the roll angle error change rate includes: The roll angle is subtracted from the desired roll angle command to generate the roll angle tracking error. Perform numerical differentiation on the desired roll angle command to generate the first derivative of the desired roll angle command; The roll angle velocity is subtracted from the first derivative of the desired roll angle command to generate the roll angle error change rate.

4. The adaptive control method for the center of gravity of an aircraft as described in claim 1, characterized in that, The construction of sliding surface state variables based on the roll angle tracking error and the roll angle error change rate includes: The roll angle tracking error is multiplied by a preset sliding coefficient using a scalar multiplication operation, and then added to the roll angle error change rate to generate a sliding surface state variable; wherein the preset sliding coefficient is a positive number.

5. The adaptive control method for the center of gravity of an aircraft as described in claim 1, characterized in that, The step of recursively updating the bounded upper limit of the disturbance in the spacecraft roll channel based on the sliding surface state variables includes: The absolute value of the sliding surface state quantity is multiplied by a preset adaptive gain coefficient to generate the upper boundary update quantity of the disturbance. The updated disturbance upper bound is summed with the estimated value of the bounded disturbance upper bound of the previous control period to generate the bounded disturbance upper bound of the current control period.

6. The adaptive control method for the center of gravity of an aircraft as described in claim 1, characterized in that, The step of calculating the control torque required for the roll channel based on the desired roll angle command, the roll angle error rate of change, the sliding surface state quantity, and the bounded upper limit of the disturbance includes: The equivalent control quantity is generated by subtracting the product of the roll angle error change rate and the preset sliding mode coefficient, and the product of the sliding mode surface state quantity and the preset reaching law gain from the second derivative of the desired roll angle command. Based on the sliding surface state variables and the preset boundary layer thickness parameters, the saturation function value of the sliding surface state variables is generated, and a scalar multiplication operation is performed with the bounded disturbance upper bound to generate a disturbance compensation term. Subtract the disturbance compensation term from the equivalent control quantity, and then divide by the nominal gain coefficient of the roll channel to generate the control torque required for the roll channel.

7. The adaptive control method for the center of gravity of an aircraft as described in claim 6, characterized in that, The step of generating saturation function values ​​for the sliding surface state quantities based on the sliding surface state quantities and preset boundary layer thickness parameters includes: The sliding surface state quantity is compared with the preset boundary layer thickness parameter. If the sliding surface state quantity is greater than the preset boundary layer thickness parameter, the saturation function value is one. If the sliding surface state quantity is less than the negative of the preset boundary layer thickness parameter, the saturation function value is negative one. If the absolute value of the sliding surface state quantity is less than or equal to the preset boundary layer thickness parameter, the saturation function value is the ratio of the sliding surface state quantity to the preset boundary layer thickness parameter.

8. An adaptive control system for the center of gravity of an aircraft, characterized in that, include: Data acquisition module, upper bound update module, torque calculation module, and control module; The data acquisition module is used to acquire the roll angle and roll angular velocity of the aircraft in the current control cycle, as well as the desired roll angle command, and to calculate the roll angle tracking error and the roll angle error change rate. The upper bound update module is used to construct the sliding surface state quantity based on the roll angle tracking error and the roll angle error change rate, and recursively update the bounded disturbance upper bound in the aircraft roll channel according to the sliding surface state quantity. The torque calculation module is used to calculate the control torque required for the roll channel based on the desired roll angle command, the roll angle error change rate, the sliding surface state quantity, and the bounded disturbance upper limit; wherein, the control torque includes a disturbance compensation term that is proportional to the bounded disturbance upper limit; The control module is used to generate control signals for each motor of the aircraft in the current cycle according to the control torque, and drive each motor to control the roll attitude of the aircraft to complete flight control.

9. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or instructions are executed, they implement an adaptive control method for the center of gravity of an aircraft as described in any one of claims 1-7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements an adaptive control method for the center of gravity of an aircraft as described in any one of claims 1-7.

Images