Fixed-wing unmanned aerial vehicle cruise phase control method based on NMPC and INDI cascade control

Abstract

The application provides a fixed-wing unmanned aerial vehicle cruise control method based on NMPC and INDI cascade control, and comprises the following steps: constructing a cascade hierarchical control architecture composed of an outer loop NMPC controller and an inner loop INDI controller; the outer loop NMPC controller generates an optimal control sequence containing thrust instructions, pitch angle instructions and roll angle instructions based on a simplified unmanned aerial vehicle nonlinear model, in combination with a preset target function and flight constraint conditions; the inner loop INDI controller takes the pitch angle instructions and the roll angle instructions as inputs, and calculates rudder deflection instructions in combination with unmanned aerial vehicle state information fed back by real-time sensors; and the unmanned aerial vehicle cruise track is tracked and controlled according to the thrust instructions and the rudder deflection instructions. The application combines the predictive optimization of NMPC with the fast robustness of INDI, forms a collaborative mechanism of'slow optimization + fast response', solves the problem of heavy calculation burden of single NMPC, makes up for the defect of lack of foresight of INDI, and significantly improves the cruise control performance.

Description

Technical Field

[0001] This invention relates to the field of aircraft control technology, and in particular to a cruise segment control method for fixed-wing unmanned aerial vehicles based on NMPC and INDI cascade control. Background Technology

[0002] Fixed-wing unmanned aerial vehicles (UAVs) are typical nonlinear, multivariable, and strongly coupled dynamic systems. In existing technologies, cruise control for fixed-wing UAVs often adopts classical control theory based on PID or linearized model design (such as linear quadratic regulators (LQR) and linear active disturbance rejection control (LADRC)). These methods perform well under finite flight conditions near the equilibrium point. However, in actual cruise missions, traditional linear control methods struggle to explicitly handle these problems due to continuous atmospheric disturbances, uncertainties in their own model parameters, and complex path and flight state constraints (such as airspace restrictions, attitude angle restrictions, rudder deflection angle and angular rate restrictions, and throttle restrictions). They often exhibit poor adaptability, decreased control accuracy, and even deterioration in stability, failing to meet the requirements of modern UAVs for high-performance autonomous flight.

[0003] Nonlinear Model Predictive Control (NMPC) is an advanced control strategy that predicts future dynamics by directly utilizing the system's nonlinear model through online rolling optimization. It explicitly incorporates system state and constraints into the optimization problem, significantly improving its resistance to UAV interference and control accuracy. However, NMPC suffers from high computational burden and high model dependence, which limits its application in engineering.

[0004] Nonlinear Dynamic Inverse (INDI) is a robust control method based on sensor feedback. It utilizes real-time sensor measurement data to calculate the control increment through a control performance model, thus exhibiting strong robustness to model uncertainties and external disturbances. However, INDI is essentially an instantaneous control method, lacking foresight and unable to predict the future behavior of the system.

[0005] In summary, existing cruise control methods for fixed-wing UAVs struggle to simultaneously address the explicit handling of complex flight constraints, the robustness to model uncertainties and external disturbances, and the real-time computational efficiency of embedded platforms, resulting in insufficient control accuracy and stability in complex environments. Summary of the Invention

[0006] This invention proposes a cruise segment control method for fixed-wing UAVs based on NMPC and INDI cascade control, which solves the problem of insufficient control accuracy and stability of existing fixed-wing UAV cruise control methods in complex environments.

[0007] The technical solution of this invention is implemented as follows: This invention provides a cruise segment control method for a fixed-wing unmanned aerial vehicle (UAV) based on NMPC and INDI cascade control, comprising the following steps: Construct a cascaded hierarchical control architecture consisting of an outer-loop NMPC controller and an inner-loop INDI controller; The outer-loop NMPC controller is based on a simplified UAV nonlinear model. It combines a preset objective function and flight constraints to generate an optimal control sequence that includes thrust, pitch, and roll commands through rolling optimization. The inner loop INDI controller takes pitch angle and roll angle commands as inputs, and combines them with the UAV status information fed back by real-time sensors to calculate the control surface deflection command through an incremental control law. The tracking and control of the UAV's cruise trajectory is achieved based on thrust commands and control surface deflection commands.

[0008] Specifically, the simplified UAV nonlinear model uses a north-facing position. Eastward position ,high ,airspeed Track inclination and track azimuth For the system state, in terms of thrust Pitch angle and roll angle To control the input, its state equation is as follows: ; in, For the quality of drones, It is the acceleration due to gravity; and The lift and drag of the drone are calculated using the following formulas: ; in, For dynamic pressure, Atmospheric density, For the reference area of ​​the drone, For the angle of attack of the drone, This is the aerodynamic coefficient.

[0009] Furthermore, the objective function of the outer-loop NMPC controller Defined as prediction time domain Weighted sum of internal state tracking error and control quantity: ; in, for Target height at any time Target airspeed and target azimuth The reference state constituted; for Actual height at any moment Actual airspeed and actual azimuth The real-time status of the system; Here is the state error weight matrix. To control the weight matrix; for Control parameters for the time period.

[0010] Furthermore, the outer-loop NMPC controller employs a sequential quadratic programming algorithm to solve for the optimal control sequence, with specific steps including: a. Obtain the current measured status of the system. and initial value of optimal control sequence If this is the first time solving the problem, then Take the initial guess value; otherwise, take the rolling optimization result from the previous time step. b, at the current work point At this point, a first-order Taylor expansion is performed on the nonlinear model, and a second-order Taylor expansion is performed on the objective function, constructing the following quadratic programming subproblem: ; in, This represents the current iteration number of the quadratic programming algorithm. and They represent the first The system state sequence and control sequence at each iteration; For the first The Hessian matrix of the Lagrange function at the nth iteration For the first The gradient of the objective function at the nth iteration For the first The control increment is solved in the next iteration; c. Linearize the system state constraints and control constraints at the current point to form linear equality constraints and linear inequality constraints for the quadratic programming subproblem: ; in, and These are the linearized equality constraint matrix and the right-hand side terms, respectively. and These are the linearized inequality constraint matrix and the right-hand side terms, respectively. Solving the quadratic programming subproblem yields the optimal control increment. ; d. Determine the step size using the line search method. And update the decision variables: ; e. Determine whether the convergence condition is met. Preset convergence accuracy or If the value approaches zero or reaches the set maximum number of iterations, the iteration stops, and the optimal control sequence for the current prediction time domain is output. Otherwise, let Repeat steps b through e to continue iterative calculation.

[0011] Furthermore, the flight constraints include system state constraints and control quantity constraints; The system state constraints include at least one of the following: angle of attack, sideslip angle, pitch angle, roll angle, pitch rate, roll rate, and airspeed. The constraint objects of the control quantity constraint conditions include at least one of the following: pitch angle command change rate, roll angle command change rate, motor thrust, thrust change rate, motor speed, servo yaw angle, and servo yaw angular velocity.

[0012] Furthermore, based on the tracking error at the current sampling time Using an exponential function to weight the state error matrix Perform adaptive adjustments: ; in, These are the initial values ​​of the state error weight matrix. For adaptive coefficients, This represents the maximum value of the state error weight matrix.

[0013] Specifically, the outer-loop NMPC controller solves for the prediction time domain at each sampling time. Optimal control sequence within Then, extract the first set of values. pitch angle command and roll angle command , as the input to the inner loop INDI controller; At the same time, the second to the third in the optimal control sequence Group control sequence and the first The group control sequences are concatenated to obtain , which serves as the initial value for the outer loop NMPC controller to perform rolling optimization at the next sampling time; The inner loop INDI controller operates according to the input pitch angle command. and roll angle command Combined with real-time measured pitch angle Roll angle and pitch angular velocity Roll angular velocity Calculate the angular velocity command and the angular acceleration command in sequence: ; ; in, and These are the pitch rate command and the roll rate command, respectively. and These are the pitch acceleration command and the roll acceleration command, respectively. , , , To control the gain; Based on the error between the actual angular acceleration at the current moment and the angular acceleration command, and combined with the control quantity from the previous moment, the control commands for the pitch and roll channels at the current moment are calculated.

[0014] Furthermore, based on the current angular velocity and the angular velocity at the previous moment, the actual angular acceleration at the current moment is obtained through differential calculation and filtering: ; in, and These are the actual roll acceleration and pitch acceleration at the current moment, respectively. The filter time constant is This represents the system sampling time.

[0015] Furthermore, the formula for the inner loop INDI controller to calculate the control commands for the pitch and roll channels at the current moment is as follows: ; ; in, and These represent the control commands for the pitch and roll channels at the current moment, respectively. and These represent the control commands for the pitch and roll channels at the previous moment, respectively. For the control effectiveness matrix, For dynamic pressure, For the reference area of ​​the drone, For the drone exhibition, The average aerodynamic chord length of the UAV. Let be the moment of inertia of the UAV rotating around its x-axis. Let be the moment of inertia of the UAV rotating around its y-axis; This is the roll rudder effectiveness coefficient. This is the pitch control coefficient.

[0016] Furthermore, for flying-wing UAVs, channel control commands will be... and Mapped to the actual left and right aileron deflection angle commands required by the actuators. and : .

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) This invention constructs a cascade hierarchical control architecture consisting of an outer-loop NMPC controller and an inner-loop INDI controller, which organically combines the long-term prediction, rolling optimization and constraint handling capabilities of NMPC with the fast response and strong robustness advantages of INDI. The outer-loop NMPC generates the optimal attitude command based on a simplified model, while the inner-loop INDI uses real-time sensor feedback to quickly track the command and suppress disturbances, forming a collaborative control mechanism of "slow optimization + fast response". This cascade architecture not only solves the problem of heavy computational burden and difficulty in meeting real-time requirements of a single NMPC, but also makes up for the lack of foresight and inability to handle complex constraints of a single INDI, realizing the unity of high-precision trajectory tracking and strong robustness in complex flight environments, and significantly improving the comprehensive control performance of fixed-wing UAVs during the cruise phase. (2) The present invention uses a sequential quadratic programming algorithm to iteratively solve constrained nonlinear optimization problems. By linearization approximation and successive approximation of quadratic programming subproblems, the solution efficiency is significantly improved while ensuring optimization accuracy, enabling NMPC to achieve real-time rolling optimization on embedded platforms. (3) This invention incorporates flight constraints into the optimization framework of NMPC in the form of linearized equations and inequalities. Through explicit processing of constraints, it ensures that the generated instructions always meet the flight safety and actuator capability boundaries, avoiding the risk of performance degradation or instability caused by constraint violation under extreme conditions in traditional control methods, and improving the flight safety of UAVs in complex environments. (4) This invention adopts an incremental control law in the inner loop INDI controller, obtains the current angular acceleration by combining the attitude angular velocity information fed back by real-time sensors with differential filtering, and calculates the control surface deflection increment by using the control efficiency matrix. This method has natural robustness to model errors and external disturbances and can maintain high-precision attitude tracking even under model mismatch conditions. At the same time, a hybrid control allocation matrix is ​​designed for flying wing UAVs to map channel commands to actual control surface deflection angles, realizing effective allocation of control quantities and ensuring accurate execution of inner loop control. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram of the NMPC+INDI cascade control architecture in an embodiment of the present invention.

[0020] Figure 2 This is a schematic diagram of the control flow of the NMPC controller in an embodiment of the present invention.

[0021] Figure 3 This is a schematic diagram of the control flow of the INDI controller in an embodiment of the present invention.

[0022] Figure 4 This is a schematic diagram of the horizontal position trajectory curve of the UAV in an embodiment of the present invention.

[0023] Figure 5 This is a schematic diagram of the altitude curve of the UAV in an embodiment of the present invention.

[0024] Figure 6 This is a schematic diagram of the airspeed curve of the UAV in an embodiment of the present invention.

[0025] Figure 7 This is a schematic diagram of the UAV roll angle command following curve in an embodiment of the present invention.

[0026] Figure 8 This is a schematic diagram of the drone pitch angle command following curve in an embodiment of the present invention. Detailed Implementation

[0027] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. 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 of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0028] This invention provides a cruise segment control method for a fixed-wing unmanned aerial vehicle (UAV) based on NMPC and INDI cascade control, comprising the following steps: Construct a cascade hierarchical control architecture consisting of an outer-loop NMPC controller and an inner-loop INDI controller, such as Figure 1 As shown; The outer-loop NMPC controller is based on a simplified UAV nonlinear model. It combines a preset objective function and flight constraints to generate an optimal control sequence that includes thrust, pitch, and roll commands through rolling optimization. The inner loop INDI controller takes pitch angle and roll angle commands as inputs, and combines them with the UAV status information fed back by real-time sensors to calculate the control surface deflection command through an incremental control law. The tracking and control of the UAV's cruise trajectory is achieved based on thrust commands and control surface deflection commands.

[0029] The cascade architecture of this invention combines predictive control with incremental robust control to achieve complementary advantages: the outer loop NMPC is responsible for forward planning and constraint processing to generate optimal attitude commands; the inner loop INDI is responsible for fast tracking of commands and suppression of disturbances, so that the overall control system can not only meet the optimal performance under complex constraints, but also has strong robustness to model uncertainty and external disturbances, effectively solving the problems of heavy computational burden or lack of foresight of single controllers.

[0030] Specifically, the simplified UAV nonlinear model uses a north-facing position. Eastward position ,high ,airspeed Track inclination and track azimuth For the system state, in terms of thrust Pitch angle and roll angle To control the input, its state equation is as follows: ; in, For the quality of drones, It is the acceleration due to gravity; and The lift and drag of the drone are calculated using the following formulas: ; in, For dynamic pressure, This is atmospheric density (obtained by interpolation based on altitude). For the reference area of ​​the drone, For the angle of attack of the drone, It is the aerodynamic coefficient (which can be obtained through simulation or experiment).

[0031] The simplified model described above, designed for the characteristics of cruise flight, retains only six state variables closely related to trajectory tracking, significantly reducing the computational dimension of NMPC online optimization. This allows for a longer prediction time domain (e.g., N=10), thus meeting the real-time requirements of the embedded platform while ensuring prediction accuracy. The lift and drag calculations employ linear or polynomial fitting of aerodynamic coefficients, which can accurately reflect the aerodynamic characteristics of the cruise phase.

[0032] Furthermore, the objective function of the outer-loop NMPC controller Defined as prediction time domain The weighted sum of internal state tracking error and control quantity, taking constant altitude, constant airspeed, and straight-line cruise as examples, is given by the following objective function: ; in, for Target height at any time Target airspeed and target azimuth The reference state (desired track) constituted; for Actual height at any moment Actual airspeed and actual azimuth The real-time status of the system; Here is the state error weight matrix. To control the weight matrix; for Control parameters at any given time; taking pitch angle into account Track inclination and attack angle The relationship is The present invention selects As the optimized control objective of the NMPC controller, thrust is one of the following. and pitch angle Control altitude and airspeed, roll angle Controlling yaw, the three components work together to complete the trajectory tracking and airspeed control of the UAV during its cruise phase.

[0033] The objective function, through a weighted sum, penalizes both the state tracking error and the magnitude of the control input, ensuring that the optimization process pursues accurate tracking while considering the consumption of control energy, avoiding excessively large or drastic changes in the control input. In a specific embodiment, the following can be selected: , As initial weights, they are used to balance the tracking priorities and control costs of altitude, airspeed, and azimuth.

[0034] Furthermore, based on the given objective function and constraints, the outer-loop NMPC controller employs a Sequence Quadratic Programming (SQP) algorithm to solve for the optimal control sequence in the prediction time domain, such as... Figure 2 As shown, the specific steps include: a. Obtain the current measured status of the system. and initial value of optimal control sequence If this is the first time solving the problem, then Take the initial guess value; otherwise, take the rolling optimization result from the previous time step. b, at the current work point At this point, a first-order Taylor expansion is performed on the nonlinear model, and a second-order Taylor expansion is performed on the objective function, constructing the following quadratic programming subproblem: ; in, This represents the current iteration number of the quadratic programming algorithm. and They represent the first The system state sequence and control sequence at each iteration; For the first The Hessian matrix of the Lagrange function at the nth iteration For the first The gradient of the objective function at the nth iteration For the first The control increment is solved in the next iteration; c. Linearize the system state constraints and control constraints at the current point to form linear equality constraints and linear inequality constraints for the quadratic programming subproblem: ; in, and These are the linearized equality constraint matrix and the right-hand side terms, respectively. and These are the linearized inequality constraint matrix and the right-hand side terms, respectively. Solving the quadratic programming subproblem yields the optimal control increment. ; d. Determine the step size using the line search method. And update the decision variables: ; e. Determine whether the convergence condition is met. Preset convergence accuracy ( )or Approaching zero, or reaching the set maximum number of iterations ( If the condition is met, then stop iterating and output the optimal control sequence in the current prediction time domain. Otherwise, let Repeat steps b through e to continue iterative calculation.

[0035] The SQP solution process described above transforms the original nonlinear optimization problem into a series of quadratic programming subproblems that are solved iteratively. Each iteration obtains the search direction (control increment) through linearization. The step size is then determined by line search to ensure convergence. This method can effectively handle nonlinear constraints, and using the solution from the previous time step as the initial value (hot start) can significantly reduce the number of iterations and improve online computation efficiency. The settings for convergence accuracy and maximum number of iterations ensure the real-time performance and accuracy of the solution.

[0036] Furthermore, the flight constraints include system state constraints and control quantity constraints; The system state constraints include at least one of the following: angle of attack, sideslip angle, pitch angle, roll angle, pitch rate, roll rate, and airspeed. The constraint objects of the control quantity constraint conditions include at least one of the following: pitch angle command change rate, roll angle command change rate, motor thrust, thrust change rate, motor speed, servo yaw angle, and servo yaw angular velocity.

[0037] These constraints cover the safety boundaries of the UAV's cruise flight and the physical limitations of its actuators; specific values ​​can be set according to the UAV model and mission requirements. For example, in one embodiment, the following constraints can be set: -15°≤Angle of attack≤20°; -5°≤Side slip angle≤5°; -30°≤Pitch angle≤30°; -30°≤roll angle≤30°; -100° / s≤pitch angular velocity≤100° / s; -150° / s≤roll angular velocity≤150° / s; 10m / s ≤ airspeed ≤ 25m / s; |Pitch angle command change rate| ≤ 5° / Ts; |Roll angle command change rate|≤5° / Ts; 0.1N ≤ Thrust ≤ 28N; |Thrust variation rate|≤5N / Ts; Motor speed ≤7500RPM; |Servo deflection angle| ≤ 40°; |Servo yaw rate| ≤ 360° / s; Where Ts is the NMPC sampling step size, by explicitly incorporating these constraints into the optimization problem, the instructions generated by NMPC are always within a safe range, avoiding the constraint violation problems that may occur in traditional control methods under extreme conditions.

[0038] Furthermore, based on the tracking error at the current sampling time Using an exponential function to weight the state error matrix Perform adaptive adjustments: ; in, These are the initial values ​​of the state error weight matrix. For adaptive coefficients, This represents the maximum value of the state error weight matrix.

[0039] This adaptive mechanism can dynamically adjust the weights according to the current tracking error: when the error increases, the weights increase exponentially, strengthening the controller's ability to correct deviations; when the error decreases, the weights return to normal levels, avoiding over-adjustment. This allows the controller to maintain appropriate control strength under different flight phases and disturbance conditions, improving its adaptability to complex operating conditions and tracking accuracy.

[0040] Specifically, the control flow of the inner loop INDI controller is as follows: Figure 3 As shown, the outer-loop NMPC controller solves for the prediction time domain at each sampling time. Optimal control sequence within Then, extract the first set of values. pitch angle command and roll angle command , as the input to the inner loop INDI controller; At the same time, the second to the third in the optimal control sequence Group control sequence and the first The group control sequences are concatenated to obtain The remaining predicted sequence is shifted and filled in as the initial value for the rolling optimization of the outer loop NMPC controller at the next sampling time. This achieves "hot start", which greatly reduces the iterative computation of NMPC and ensures real-time performance. The inner loop INDI controller operates according to the input pitch angle command. and roll angle command Combined with real-time measured pitch angle Roll angle and pitch angular velocity Roll angular velocity Calculate the angular velocity command and the angular acceleration command in sequence: ; ; in, and These are the pitch rate command and the roll rate command, respectively. and These are the pitch acceleration command and the roll acceleration command, respectively. , , , To control the gain; Based on the error between the actual angular acceleration at the current moment and the angular acceleration command, and combined with the control quantity from the previous moment, the control commands for the pitch and roll channels at the current moment are calculated.

[0041] Furthermore, based on the current angular velocity and the angular velocity at the previous moment, the actual angular acceleration at the current moment is obtained through differential calculation and filtering: ; in, and These are the actual roll acceleration and pitch acceleration at the current moment, respectively. The filter time constant is This represents the system sampling time.

[0042] Since angular acceleration is usually not directly measurable, this method obtains a smooth angular acceleration estimate by angular velocity difference and supplementing it with a first-order low-pass filter; the filter time constant... The smoothness of the estimate is determined by a trade-off between noise suppression and response speed.

[0043] Furthermore, the formula for the inner loop INDI controller to calculate the control commands for the pitch and roll channels at the current moment is as follows: ; ; in, and These represent the control commands for the pitch and roll channels at the current moment, respectively. and These represent the control commands for the pitch and roll channels at the previous moment, respectively. For the control effectiveness matrix, For dynamic pressure, For the reference area of ​​the drone, For the drone exhibition, The average aerodynamic chord length of the UAV. Let be the moment of inertia of the UAV rotating around its x-axis. Let be the moment of inertia of the UAV rotating around its y-axis; This is the roll rudder effectiveness coefficient. This is the pitch control coefficient.

[0044] The required rudder deflection increment is obtained by multiplying the error between the desired angular acceleration and the actual angular acceleration by the control effectiveness matrix. This increment is then superimposed with the rudder deflection from the previous moment to obtain the current rudder deflection command. The control effectiveness matrix is ​​derived based on the UAV torque equation and reflects the ability of the control surface deflection to generate angular acceleration, thus offsetting the nonlinear characteristics of the system. Due to the use of an incremental form, this method is insensitive to model errors, has strong robustness, and can effectively suppress external disturbances and model uncertainties.

[0045] Furthermore, for flying-wing UAVs, elevons are typically used as control surfaces, meaning that simultaneous deflection of the left and right ailerons generates pitch torque, while differential deflection generates roll torque. For flying-wing UAVs, this invention uses channel control commands... and Mapped to the actual left and right aileron deflection angle commands required by the actuators. and : .

[0046] The aforementioned hybrid control matrix converts pitch and roll channel commands into the actual deflection angles of the left and right ailerons, thus achieving the correct allocation of control quantities.

[0047] To verify the effectiveness of the method of the present invention, a simulation experiment was conducted, and the experimental results are as follows: Figures 4 to 8 As shown.

[0048] Figure 4 This is a schematic diagram of the horizontal position trajectory curve of the UAV in an embodiment of the present invention. The figure shows the flight trajectory of the UAV in the horizontal plane. As can be seen from the figure, the horizontal flight trajectory of the UAV is relatively smooth and there is no large trajectory deviation, indicating that the method of the present invention can achieve accurate horizontal position tracking.

[0049] Figure 5 This is a schematic diagram of the altitude curve of the UAV in an embodiment of the present invention. The figure shows the UAV's ability to maintain altitude during the cruise phase. As can be seen from the figure, the UAV can maintain a stable set cruise altitude with small altitude fluctuations, demonstrating the excellent performance of the method of the present invention in altitude control.

[0050] Figure 6This is a schematic diagram of the airspeed curve of the UAV in an embodiment of the present invention. The figure shows the airspeed changes of the UAV during the cruise phase. As can be seen from the figure, the airspeed can be stabilized near the target value with small fluctuations. It is less affected by attitude adjustment and external disturbances, indicating that the thrust control and attitude control work well together and achieve precise maintenance of the target airspeed.

[0051] Figure 7 This is a schematic diagram of the roll angle command following curve of the UAV in an embodiment of the present invention. The figure shows the following effect of the roll angle command. As can be seen from the figure, the actual roll angle can quickly and accurately follow the command change, with rapid dynamic response and small steady-state error, which proves that the INDI controller has excellent command tracking capability and anti-disturbance performance.

[0052] Figure 8 This is a schematic diagram of the pitch angle command following curve of the UAV in an embodiment of the present invention. The figure shows the command following effect of the pitch angle. As can be seen from the figure, the actual pitch angle is highly consistent with the command value, the response speed is fast, and the overshoot is small, which further verifies the effectiveness of the inner loop INDI controller.

[0053] comprehensive Figures 4 to 8 The simulation results show that the cruise segment control method for fixed-wing UAVs based on NMPC and INDI cascade control proposed in this invention can achieve high-precision trajectory tracking, fast attitude response and stable airspeed maintenance. It has good control performance and robustness under complex flight conditions, and effectively solves the problems of insufficient control accuracy and stability in the prior art.

[0054] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A cruise segment control method for a fixed-wing unmanned aerial vehicle based on NMPC and INDI cascade control, characterized in that, Includes the following steps: Construct a cascaded hierarchical control architecture consisting of an outer-loop NMPC controller and an inner-loop INDI controller; The outer-loop NMPC controller is based on a simplified UAV nonlinear model. It combines a preset objective function and flight constraints to generate an optimal control sequence that includes thrust, pitch, and roll commands through rolling optimization. The inner loop INDI controller takes pitch angle and roll angle commands as inputs, and combines them with the UAV status information fed back by real-time sensors to calculate the control surface deflection command through an incremental control law. The tracking and control of the UAV's cruise trajectory is achieved based on thrust commands and control surface deflection commands.

2. The cruise segment control method for a fixed-wing UAV based on NMPC and INDI cascade control as described in claim 1, characterized in that, The simplified UAV nonlinear model uses a north-facing position. Eastward position ,high ,airspeed Track inclination and track azimuth For the system state, in terms of thrust Pitch angle and roll angle To control the input, its state equation is as follows: ； in, For the quality of drones, It is the acceleration due to gravity; and The lift and drag of the drone are calculated using the following formulas: ； in, For dynamic pressure, Atmospheric density, For the reference area of ​​the drone, For the angle of attack of the drone, This is the aerodynamic coefficient.

3. The cruise segment control method for a fixed-wing UAV based on NMPC and INDI cascade control as described in claim 2, characterized in that, The objective function of the outer loop NMPC controller Defined as prediction time domain Weighted sum of internal state tracking error and control quantity: ； in, for Target height at any time Target airspeed and target azimuth The reference state constituted; for Actual height at any moment Actual airspeed and actual azimuth The real-time status of the system; Here is the state error weight matrix. To control the weight matrix; for Control parameters for the time period.

4. The cruise segment control method for a fixed-wing UAV based on NMPC and INDI cascade control as described in claim 3, characterized in that, The outer-loop NMPC controller uses a sequential quadratic programming algorithm to solve for the optimal control sequence. The specific steps include: a. Obtain the current measured status of the system. and initial value of optimal control sequence If this is the first time solving the problem, then Take the initial guess value; otherwise, take the rolling optimization result from the previous time step. b, at the current work point At this point, a first-order Taylor expansion is performed on the nonlinear model, and a second-order Taylor expansion is performed on the objective function, constructing the following quadratic programming subproblem: ； in, This represents the current iteration number of the quadratic programming algorithm. and They represent the first The system state sequence and control sequence at each iteration; For the first The Hessian matrix of the Lagrange function at the nth iteration For the first The gradient of the objective function at the nth iteration For the first The control increment is solved in the next iteration; c. Linearize the system state constraints and control constraints at the current point to form linear equality constraints and linear inequality constraints for the quadratic programming subproblem: ； in, and These are the linearized equality constraint matrix and the right-hand side terms, respectively. and These are the linearized inequality constraint matrix and the right-hand side terms, respectively. Solving the quadratic programming subproblem yields the optimal control increment. ; d. Determine the step size using the line search method. And update the decision variables: ； e. Determine whether the convergence condition is met. Preset convergence accuracy or If the value approaches zero or reaches the set maximum number of iterations, the iteration stops, and the optimal control sequence for the current prediction time domain is output. Otherwise, let Repeat steps b through e to continue iterative calculation.

5. The cruise segment control method for a fixed-wing UAV based on NMPC and INDI cascade control as described in claim 4, characterized in that, The flight constraints include system state constraints and control quantity constraints; The system state constraints include at least one of the following: angle of attack, sideslip angle, pitch angle, roll angle, pitch rate, roll rate, and airspeed. The constraint objects of the control quantity constraint conditions include at least one of the following: pitch angle command change rate, roll angle command change rate, motor thrust, thrust change rate, motor speed, servo yaw angle, and servo yaw angular velocity.

6. The cruise segment control method for a fixed-wing UAV based on NMPC and INDI cascade control as described in claim 3, characterized in that, Based on the tracking error at the current sampling time Using an exponential function to weight the state error matrix Perform adaptive adjustments: ； in, These are the initial values ​​of the state error weight matrix. For adaptive coefficients, This represents the maximum value of the state error weight matrix.

7. The cruise segment control method for a fixed-wing UAV based on NMPC and INDI cascade control as described in claim 1, characterized in that, The outer-loop NMPC controller solves for the prediction time domain at each sampling time. Optimal control sequence within Then, extract the first set of values. pitch angle command and roll angle command , as the input to the inner loop INDI controller; At the same time, the second to the third in the optimal control sequence Group control sequence and the first The group control sequences are concatenated to obtain , which serves as the initial value for the outer loop NMPC controller to perform rolling optimization at the next sampling time; The inner loop INDI controller operates according to the input pitch angle command. and roll angle command Combined with real-time measured pitch angle Roll angle and pitch angular velocity Roll angular velocity Calculate the angular velocity command and the angular acceleration command in sequence: ； ； in, and These are the pitch rate command and the roll rate command, respectively. and These are the pitch acceleration command and the roll acceleration command, respectively. , , , To control the gain; Based on the error between the actual angular acceleration at the current moment and the angular acceleration command, and combined with the control quantity from the previous moment, the control commands for the pitch and roll channels at the current moment are calculated.

8. The cruise segment control method for a fixed-wing UAV based on NMPC and INDI cascade control as described in claim 7, characterized in that, Based on the current angular velocity and the angular velocity at the previous moment, the actual angular acceleration at the current moment is obtained through differential calculation and filtering: ； in, and These are the actual roll acceleration and pitch acceleration at the current moment, respectively. The filter time constant is This represents the system sampling time.

9. The cruise segment control method for a fixed-wing UAV based on NMPC and INDI cascade control as described in claim 8, characterized in that, The formula for the inner loop INDI controller to calculate the control commands for the pitch and roll channels at the current moment is as follows: ； ； in, and These represent the control commands for the pitch and roll channels at the current moment, respectively. and These represent the control commands for the pitch and roll channels at the previous moment, respectively. For the control effectiveness matrix, For dynamic pressure, For the reference area of ​​the drone, For the drone exhibition, The average aerodynamic chord length of the UAV. Let be the moment of inertia of the UAV rotating around its x-axis. Let be the moment of inertia of the UAV rotating around its y-axis; This is the roll rudder effectiveness coefficient. This is the pitch control coefficient.

10. The cruise segment control method for a fixed-wing UAV based on NMPC and INDI cascade control as described in claim 9, characterized in that, For flying-wing UAVs, channel control commands will be implemented. and Mapped to the actual left and right aileron deflection angle commands required by the actuators. and : 。

Images