A high-precision anti-interference control method for a UAV holder under transient impact load

Abstract

This invention provides a high-precision anti-disturbance control method for UAV gimbals under transient impact loads, belonging to the field of precision servo control technology. It employs a cascade control architecture: the outer loop uses model predictive control (MPC) and introduces control increment constraints to plan an overshoot-free velocity trajectory; the inner loop uses linear active disturbance rejection control (LADRC) and establishes an extended state observer to estimate and compensate for the total disturbance of multi-stage transient impacts in real time. This invention, using the aforementioned high-precision anti-disturbance control method for UAV gimbals under transient impact loads, reduces the settling time to 0.18 s and lowers the maximum dynamic position error during firing to 0.0005 rad, achieving precise steady-state maintenance under high-impact environments.

Description

Technical Field

[0001] This invention relates to the field of precision servo control technology, and in particular to a high-precision anti-disturbance control method for a UAV gimbal with transient impact loads. Background Technology

[0002] In the field of precision servo control for unmanned aerial vehicles (UAVs), when an airborne gimbal carries a special load with transient high-pressure release and internal linear reciprocating motion characteristics, the system actuators will be subjected to extremely severe transient shocks. These shock disturbances not only have extremely large amplitudes but also extremely complex physical evolution processes, typically including multiple nonlinear stages such as high-pressure working fluid expansion and work, fluid pressure relief feedback, jet aftereffects at the nozzle, and high-frequency rigid collisions at the ends of mechanical moving components. This multi-stage superimposed high-frequency disturbance directly disrupts the steady-state configuration of the gimbal through rigid mechanical connections. Traditional PI (proportional-integral) control schemes, when facing such extremely narrow pulse width, large gradient nonlinear transient disturbances, suffer from slow system response, large overshoot, and poor steady-state recovery accuracy due to the inherent hysteresis of the integral element. This easily leads to overcurrent divergence in the motors, making it impossible to guarantee the high-precision attitude stability of the gimbal under continuous impact conditions. Summary of the Invention

[0003] The purpose of this invention is to provide a high-precision anti-disturbance control method for UAV gimbals under transient impact loads. It combines the advantages of Model Predictive Control (MPC) in handling multiple constraints and trajectory planning with the active estimation and compensation capability of Linear Active Disturbance Rejection Control (LADRC) for internal and external disturbances, thereby achieving high-precision steady-state maintenance and rapid disturbance suppression under transient impacts.

[0004] To achieve the above objectives, the present invention provides a high-precision anti-disturbance control method for a UAV gimbal under transient impact loads, comprising the following steps: Built on dq The mathematical model of the motor in the rotating coordinate system includes the stator voltage balance equation and flux linkage equation, the electromagnetic torque equation and control strategy, and the mechanical motion equation containing the total disturbance of multi-stage transient impact. Construct a piecewise mathematical model containing four typical physical stages The total disturbance of multi-stage transient impact is obtained. ; Based on dq A mathematical model of the motor in a rotating coordinate system is used to construct a predictive model, and an outer-loop predictive control trajectory is planned to obtain the outer-loop speed command. Based on the outer ring speed command, the total disturbance of multi-stage transient impacts is offset in real time, combined with... dq A mathematical model of the motor in a rotating coordinate system is used to design a linear active disturbance rejection controller, thereby achieving high-precision disturbance rejection control of the UAV gimbal.

[0005] Preferably, the stator voltage balance equation and flux linkage equation are established: (1); in, , The stator voltage is respectively at , Components on the axis, , The stator currents are respectively , Components on the axis, For stator winding resistance, The angular velocity of the electric rotor, and They are respectively , Shaft inductance, for surface-mount motors , It is a permanent magnet flux chain. for Shaft stator flux linkage components, for Shaft stator flux linkage components, This is the stator equivalent inductance.

[0006] Preferably, an electromagnetic torque equation and control strategy are established, and the following is adopted: Vector control strategy, electromagnetic torque and The shaft current is proportional to: (2); in, This represents the number of pole pairs of the motor.

[0007] The preferred mechanical motion equation, which includes the total disturbance of multi-stage transient impacts, is expressed as: (3); (4); in, The mechanical angular velocity of the motor is represented by J, which is the total moment of inertia of the gimbal system, including the inertia of the motor rotor, load, and gimbal frame. B The coefficient of viscous friction, For normal load torque, This refers to the total disturbance torque caused by nonlinear multi-stage transient impacts under special loads. It is the first derivative of angular velocity with respect to time t.

[0008] Preferred, multi-stage transient impact total disturbance: (5); in, Let the lever arm be a given value, and introduce the force transmission coefficient. Then, the combined transient impact The piecewise mathematical model is as follows: (6); In the formula, F1(t) is the transient thrust generated by the expansion of the high-pressure fluid in the driving cavity, F2(t) is the forward reaction force generated by the damping piston during the aerodynamic feedback period, F3(t) is the jet recoil force generated by the residual high-pressure fluid ejecting from the nozzle during the after-effect period, and F4(t) is the rigid impact force generated by the reciprocating motion component inside the end rigid impact period. t g The moment when the linear motion load passes the bypass pressure relief hole. t local Relative time, , For the firing moment, t k The moment when the linear motion load leaves the pipe opening. t end This is the time when the effect of the discharge ends. At the moment of impact, This is the impact pulse width coefficient.

[0009] Preferred, comprehensive transient impact The piecewise mathematical model specifically includes: High voltage working period The high-pressure power-on period simulates the thrust generated by the transient release of high-pressure fluid within the drive cavity, and the pressure rise process within the cavity is described using a sinusoidal half-wave function: (7); in, S For the area of ​​force application, For the maximum peak driving pressure, This is the moment of peak pressure. Bypass feedback period During the bypass feedback period, after the simulated linear motion load passes the bypass pressure relief hole, some high-pressure fluid enters the feedback pipeline, pushing the damping piston to generate a forward reaction force, which offsets part of the transient impact. (8); in, The area of ​​the piston subjected to feedback damping is the force area. The fluid pressure attenuation coefficient, The initial discharge pressure at the pipe opening. The constant representing the pressure attenuation of the fluid inside the pipe; Post-leakage effect period After the discharge effect period, simulating the linear motion load leaving the pipe opening, the recoil force generated by the residual high-pressure fluid ejecting at high speed from the pipe opening exhibits an exponential decay characteristic: (9); in, The decay time constant; End-of-life rigid impact period After the fluid finishes performing work, the internal reciprocating motion component enters the inertial free sliding phase. Subsequently, when the component moves into position, the impact at the end of the buffer generates a rigid impact, which is modeled as a Gaussian impulse function: (10); in, This represents the peak impact force. This is the pulse width coefficient.

[0010] Preferred, based on dq The mathematical model of the motor in a rotating coordinate system is used to construct a predictive model and plan the outer-loop predictive control trajectory, including: Based on the mathematical model of the motor, at the current moment status Starting from this point, we can deduce the future. System output at each moment: (11); in, To predict the output sequence, The control input sequence to be solved is... For the current moment The rotor angular position, For the current moment The rotor angular velocity, where T is the matrix transpose symbol. This is the system state prediction matrix. To control the input prediction matrix, For prediction in the time domain; Design the cost function, and construct the following quadratic cost function. : (12); In the formula, For the future The reference angle position at that moment. Based on the current moment Predicted future Angular position at time, For Let be the squared Euclidean norm of the weight matrix. For Let be the squared Euclidean norm of the weight matrix. To control the time domain (and satisfy) (i) represents the prediction step size index, j represents the control step size index, and k represents the current discrete time. For the future The increment of the angular velocity control command at time t, where R and Q are both weight matrices; This indicates a penalty for position tracking error. This indicates an increase in the amount of punishment controlled; By introducing physical constraints and transforming the quadratic cost function into a standard quadratic programming problem, the optimal control sequence is obtained. According to the rolling optimization principle, only the first element of the optimal control sequence is applied to the system, and the outer loop output at the current moment is: (13); In the formula, For the current moment The optimal angular velocity control command increment obtained by solving the problem. This is the outer loop output at the current moment.

[0011] Preferred physical constraints include: Speed ​​constraints: ; Acceleration constraints: .

[0012] Preferably, based on the outer ring speed command, the total disturbance of multi-stage transient impact is offset in real time, combined with... dq The mathematical model of the motor in a rotating coordinate system is used to design a linear active disturbance rejection controller, specifically including: According to the speed command given by the outer ring To counteract the total disturbance caused by multi-stage transient impacts in real time, a second-order linear active disturbance rejection controller is designed based on the mechanical motion equations of the permanent magnet synchronous motor. The total disturbance of a system, including stage transient impacts, frictional forces, and parameter perturbations, is defined as the extended state variable. , For the total disturbance moment, the second-order LESO is established as follows: (14); in, This is an estimated value for angular velocity. This is the estimated total system disturbance. This is an estimated value for angular acceleration. Let be the rate of change of the total system disturbance, and let the observer bandwidth be . ,but , , For system control variables, For actual measured speed , For proportional gain, This is the integral gain; To offset the estimated total system disturbance The following control law is designed: (15); in, This is an estimated value for the system control gain. This is the disturbance compensation component; To control the input of the unperturbed integral cascade model, proportional control is adopted as follows: (16); in, For controller bandwidth Determined proportional gain, .

[0013] Therefore, the present invention employs the above-mentioned high-precision anti-disturbance control method for UAV gimbals under transient impact loads, and the technical effects are as follows: Strong anti-disturbance capability and high attitude maintenance accuracy: For nonlinear, high-amplitude transient impacts generated under special loads, this invention utilizes the Extended State Observer (LESO) in LADRC to observe and estimate the "total disturbance," including high-voltage power impacts and mechanical impacts, in real time, and actively cancels it through a feedforward channel. Simulation data shows that after applying a single transient impact disturbance at 0.5s, the maximum dynamic position error generated by traditional PI (proportional-integral) control is as high as 0.025 rad, while the method of this invention reduces this error to within 0.0005 rad, an error reduction of over 98%, effectively ensuring the attitude stability of the gimbal under continuous high-frequency impact conditions.

[0014] Fast response and no overshoot: The outer-loop MPC controller uses a rolling optimization algorithm to handle the physical constraints of the control quantity, planning a smooth and fast speed command, thus solving the overshoot problem caused by traditional PID (proportional-integral-derivative) controllers in pursuit of fast response. Experiments show that after experiencing a mechanical impact, this system can fully recover to steady state (error converges to zero) in only 0.18s, which is about 55% shorter than the settling time of traditional control methods, greatly improving efficiency.

[0015] High robustness and adaptability to complex working conditions: This cascade architecture exhibits strong robustness to changes in system parameters, such as changes in rotational inertia and friction caused by load replacement. LADRC treats internal parameter perturbations as disturbances for compensation, decoupling the controller's dependence on an accurate model. This makes the algorithm easily portable to different models of UAV gimbals subjected to transient impact loads, demonstrating high engineering application value. Attached Figure Description

[0016] Figure 1 This is a block diagram of the principle of a UAV gimbal cascade control system based on MPC-LADRC; Figure 2 This is a block diagram of the simulation control model; Figure 2 (a) is a diagram of the internal structure of the position-velocity composite disturbance rejection control unit; Figure 2 (b) is a diagram of the internal structure of the FOC vector drive and SVPWM (space vector pulse width modulation) modulation unit; Figure 2 (c) is a diagram of the internal structure of the gimbal under electromechanical dynamics and transient impact. Figure 3 The graph shows the total disturbance mathematical model of multi-stage transient impact. Figure 4 This is a schematic diagram of the outer loop MPC controller structure; Figure 5 This is a schematic diagram of the inner loop LADRC controller structure; Figure 6 This is a schematic diagram of the Linear Extended State Observer (LESO) structure. Figure 7 The tracking performance of the LESO observer on composite disturbance torque; Figure 8 This is a comparison chart of the dynamic tracking performance of different control algorithms under strong impact conditions. Detailed Implementation

[0017] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0018] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.

[0019] Example 1 like Figure 1 , Figure 2 , Figure 3 , Figure 4 , Figure 5 , Figure 6 As shown, a high-precision anti-disturbance control method for a UAV gimbal under transient impact load includes the following steps: Step 1: Experimental Platform and Motor Parameter Settings This embodiment takes a certain type of UAV gimbal servo system as the research object. The actuator is a three-phase permanent magnet synchronous motor (PMSM), and its specific physical parameters are shown in Table 1.

[0020] Table 1 Physical parameters of permanent magnet synchronous motor (PMSM)

[0021] Step 2: Establish a mathematical model of the motor containing disturbances. Assuming the magnetic linearity of the motor and neglecting core saturation and eddy current losses, in A mathematical model of the motor is established in a rotating coordinate system.

[0022] First, establish the stator voltage balance equation and the flux linkage equation.

[0023] ; in, , and , Stator voltage and current respectively , Components on the axis, For stator winding resistance, The angular velocity of the electric rotor, and for , Shaft inductance, for surface-mount motors , It is a permanent magnet flux linkage.

[0024] Establish electromagnetic torque equations and control strategies, and adopt... Vector control strategy, electromagnetic torque and The shaft current is proportional to: ; in, This represents the number of pole pairs of the motor.

[0025] A mechanical motion equation incorporating multi-stage transient impact disturbances is established. For the special working condition of transient effects from specific UAV loads, the mechanical motion equilibrium equation of the gimbal motor is expressed as follows, based on Newton's second law: ; Writing Differential Form ; in, The mechanical angular velocity of the motor, in units of ω. , This refers to the total rotational inertia of the gimbal system, including the inertia of the motor rotor, load, and gimbal frame, measured in units of... , The coefficient of viscous friction is expressed in units of 1000 ppm. , For normal load torque, The total disturbance moment of nonlinear multi-stage transient impact generated under special loads, in units of , It is the first derivative of angular velocity with respect to time t.

[0026] Step 3: Controller Design and Parameter Tuning The system adopts a dual closed-loop cascade architecture of "outer loop MPC position control + inner loop LADRC speed control", and the key parameters of each link are tuned as follows: Outer ring MPC design: Sampling period set to This forms a multi-rate control system with the inner loop. A prediction time domain is defined. Control time domain The longer prediction time domain ensures the stability of the system operation. To balance the smoothness of position tracking speed and control input, a state weight matrix is ​​selected. Control weight matrix Control Quantity Weight Constraint Setting The virtual angular acceleration is used as a limit to force MPC to plan a smooth S-shaped velocity curve, preventing mechanical shocks during start-up and braking.

[0027] Table 2 MPC Parameter Settings

[0028] Inner Ring LADRC Design: To address high-frequency, multi-stage transient impact disturbances, the observer bandwidth was significantly increased. (Observer bandwidth settings are then defined.) Controller bandwidth This allows LESO to capture extremely high-frequency impact components. (Considering the system's rotational inertia...) In order to enhance the anti-disturbance "thrust", the system gain was adjusted. Scaling factor introduced That is, the actual use This significantly improves the transient compensation capability against static friction and impact.

[0029] Table 3 LADRC Parameter Settings

[0030] To verify the disturbance rejection performance of the proposed Linear Extended State Observer (LESO) under complex operating conditions, simulation tests were conducted in the Simulink environment. Figure 7 This demonstrates the disturbance estimates output by LESO during system operation. The results are compared with the actual disturbance torque experienced by the system. For example... Figure 7 As shown in the figure, the black solid line represents the actual disturbance, and the red solid line represents the LESO observation value. Analysis reveals that the LESO estimate exhibits low-pass filtering characteristics consistent with physical laws—the amplitude shows a certain attenuation, and the pulse widens accordingly, with a slight phase lag. However, the integral area of ​​the estimated value curve (i.e., the disturbance impulse) remains highly equivalent to the actual impact energy. This characteristic precisely demonstrates that the designed anti-integral saturation mechanism can effectively prevent the observer from falling into numerical divergence when facing extremely narrow, over-limit strong impacts. At 0.56s: the internal reciprocating motion component moves into position, generating a mechanical impact disturbance. Since the frequency component of this disturbance falls within the effective bandwidth of the observer designed in this invention, LESO responds rapidly, achieving accurate tracking.

[0031] Experimental results show that the observer parameters (bandwidth) designed in this invention are... The selected parameters are appropriate, possessing the unified observation capability for both "internal friction" and "external strong shocks." Accurate disturbance estimates are required. This provides a reliable basis for subsequent control law compensation and achieves active cancellation of the total disturbance.

[0032] Step 4: Loading the multi-stage transient impact physics model (perturbation input) To verify the algorithm's disturbance rejection performance, a transient physical impact model with multi-stage characteristics was loaded into the simulation, with the parameters set as follows: Maximum intracavitary peak pressure The impact time is extremely short. The peak impact force at which the internal reciprocating component reaches its position is set at 3500N, and the pulse duration is approximately 4ms. At this time, the disturbance is introduced to verify whether the controller can limit the position deviation to within a certain range. Within.

[0033] Table 4 Transient shock model parameters

[0034] Simulation Result Analysis To verify the disturbance rejection performance of the proposed MPC-LADRC cascade control strategy under strong impact conditions, a full-process firing test was conducted in the MATLAB / Simulink simulation platform, and the simulation results were compared with those of the optimized traditional PI control strategy. Figure 8 As shown.

[0035] The experiment was set up so that the system was initially in steady-state hovering mode, and the target position was... During simulation time The model continuously injects multi-stage transient impact disturbance signals. This disturbance model accurately reproduces the characteristics of real transient impacts, including... The high-pressure fluid work impact during the period, and approximately The rigid impact pulse generated when the internal reciprocating component moves into position and impacts the buffer is set to a peak impact force of 3500N.

[0036] From the dynamic response of the simulated waveforms, it can be seen that at the moment of transient impact, traditional PI control, due to the inherent lag in the integral element, cannot respond to sudden torque changes in a timely manner, causing the gimbal to... Immediately afterwards, a significant "head-up" motion occurred, with the maximum dynamic position deviation reaching 0.025 rad (approximately). Such amplitude jitter will cause the baseline to deviate significantly from the target during precision operations. In contrast, the MPC-LADRC control strategy employed in this invention benefits from the high-bandwidth extended state observer configured in the inner-loop LADRC, which can identify high-frequency transient impacts as the system's "total disturbance" in real time and rapidly generate a reverse compensation torque within 0.001s through the feedforward channel. Experimental results show that the dynamic position deviation of this method is strictly limited to 0.0005 rad (approximately...). Within a certain range, compared to traditional PI control, the disturbance suppression capability is improved by approximately 98%, effectively achieving steady-state maintenance during transient impacts.

[0037] During the recovery phase following the mechanical impact, the outer-loop MPC controller utilizes a rolling optimization algorithm to plan a smooth regression trajectory while strictly adhering to virtual angular acceleration constraints. The system... The left and right movements completely eliminate oscillations and restore steady state, adjusting the time. Furthermore, although the system faces a large moment of inertia ( ), MPC's high-rigidity weight configuration ( This ensures that the final steady-state error strictly converges to zero, fully verifying the algorithm's high-precision positioning capability and robustness under heavy load and strong impact conditions.

[0038] Therefore, the present invention adopts the above-mentioned high-precision anti-disturbance control method for UAV gimbal with transient impact load, which shortens the adjustment time to 0.18s and reduces the maximum dynamic position error during firing to 0.0005 rad, thereby achieving accurate steady-state maintenance under high impact environment.

[0039] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A high-precision anti-disturbance control method for a UAV gimbal under transient impact load, characterized in that, Includes the following steps: Built on dq The mathematical model of the motor in the rotating coordinate system includes the stator voltage balance equation and flux linkage equation, the electromagnetic torque equation and control strategy, and the mechanical motion equations including multi-stage transient impact disturbances. Construct a piecewise mathematical model containing four typical physical stages The total disturbance of multi-stage transient impact is obtained. ; Based on dq A mathematical model of the motor in a rotating coordinate system is used to construct a predictive model, and an outer-loop predictive control trajectory is planned to obtain the outer-loop speed command. Based on the outer ring speed command, the total disturbance of multi-stage transient impacts is offset in real time, combined with... dq A mathematical model of the motor in a rotating coordinate system is used to design a linear active disturbance rejection controller, thereby achieving high-precision disturbance rejection control of the UAV gimbal.

2. The high-precision anti-disturbance control method for a UAV gimbal with transient impact load according to claim 1, characterized in that, Establish the stator voltage balance equation and flux linkage equation: （1）； in, , The stator voltage is respectively at , Components on the axis, , The stator currents are respectively , Components on the axis, For stator winding resistance, The angular velocity of the electric rotor, and They are respectively , Shaft inductance, for surface-mount motors , It is a permanent magnet flux chain. for Shaft stator flux linkage components, for Shaft stator flux linkage components, This is the stator equivalent inductance.

3. The high-precision anti-disturbance control method for a UAV gimbal with transient impact load according to claim 1, characterized in that, Establish electromagnetic torque equations and control strategies, and adopt... Vector control strategy, electromagnetic torque and The shaft current is proportional to: （2）； in, This represents the number of pole pairs of the motor.

4. The high-precision anti-disturbance control method for a UAV gimbal with transient impact load according to claim 1, characterized in that, The mechanical equations of motion, which include the total disturbance of multi-stage transient impacts, are expressed as: （3）； （4）； in, The mechanical angular velocity of the motor is represented by J, which is the total moment of inertia of the gimbal system, including the inertia of the motor rotor, load, and gimbal frame. B The coefficient of viscous friction, For normal load torque, This refers to the total disturbance torque caused by nonlinear multi-stage transient impacts under special loads. It is the first derivative of angular velocity with respect to time t.

5. A high-precision anti-disturbance control method for a UAV gimbal with transient impact load according to claim 1, characterized in that, Total disturbance of multi-stage transient shock: （5）； in, Let the lever arm be a given value, and introduce the force transmission coefficient. Then, the combined transient impact The piecewise mathematical model is as follows: （6）； In the formula, F1(t) is the transient thrust generated by the expansion of the high-pressure fluid in the driving cavity, F2(t) is the forward reaction force generated by the damping piston during the aerodynamic feedback period, F3(t) is the jet recoil force generated by the residual high-pressure fluid ejecting from the nozzle during the after-effect period, and F4(t) is the rigid impact force generated by the reciprocating motion component inside the end rigid impact period. t g The moment when the linear motion load passes the bypass pressure relief hole. t local Relative time, , For the firing moment, t k The moment when the linear motion load leaves the pipe opening. t end This is the time when the effect of the discharge ends. At the moment of impact, This is the impact pulse width coefficient.

6. A high-precision anti-disturbance control method for a UAV gimbal with transient impact load according to claim 5, characterized in that, Comprehensive transient impact The piecewise mathematical model specifically includes: High voltage working period The high-pressure power-on period simulates the thrust generated by the transient release of high-pressure fluid within the drive cavity, and the pressure rise process within the cavity is described using a sinusoidal half-wave function: （7）； in, S For the area of ​​force application, For the maximum peak driving pressure, This is the moment of peak pressure. Bypass feedback period During the bypass feedback period, after the simulated linear motion load passes the bypass pressure relief hole, some high-pressure fluid enters the feedback pipeline, pushing the damping piston to generate a forward reaction force, which offsets part of the transient impact. （8）； in, The area of ​​the piston subjected to feedback damping is the force area. The fluid pressure attenuation coefficient, The initial discharge pressure at the pipe opening. The constant representing the pressure attenuation of the fluid inside the pipe; Post-leakage effect period After the discharge effect period, simulating the linear motion load leaving the pipe opening, the recoil force generated by the residual high-pressure fluid ejecting at high speed from the pipe opening exhibits an exponential decay characteristic: （9）； in, The decay time constant; End-of-life rigid impact period After the fluid finishes performing work, the internal reciprocating motion component enters the inertial free sliding phase. Subsequently, when the component moves into position, the impact at the end of the buffer generates a rigid impact, which is modeled as a Gaussian impulse function: （10）； in, This represents the peak impact force. This is the pulse width coefficient.

7. A high-precision anti-disturbance control method for a UAV gimbal with transient impact load according to claim 1, characterized in that, Based on dq The mathematical model of the motor in a rotating coordinate system is used to construct a predictive model and plan the outer-loop predictive control trajectory, including: Based on the mathematical model of the motor, at the current moment status Starting from this point, we can deduce the future. System output at each moment: （11）； in, To predict the output sequence, The control input sequence to be solved is... For the current moment The rotor angular position, For the current moment The rotor angular velocity, where T is the matrix transpose symbol. This is the system state prediction matrix. To control the input prediction matrix, For prediction in the time domain; Design the cost function, and construct the following quadratic cost function. : （12）； In the formula, For the future The reference angle position at that moment. Based on the current moment Predicted future Angular position at time, For Let be the squared Euclidean norm of the weight matrix. For Let be the squared Euclidean norm of the weight matrix. To control the time domain (and satisfy) (i) represents the prediction step size index, j represents the control step size index, and k represents the current discrete time. For the future The increment of the angular velocity control command at time t, where R and Q are both weight matrices; This indicates a penalty for position tracking error. This indicates an increase in the amount of punishment controlled; By introducing physical constraints and transforming the quadratic cost function into a standard quadratic programming problem, the optimal control sequence is obtained. According to the rolling optimization principle, only the first element of the optimal control sequence is applied to the system, and the outer loop output at the current moment is: （13）； In the formula, For the current moment The optimal angular velocity control command increment obtained by solving the problem. This is the outer loop output at the current moment.

8. A high-precision anti-disturbance control method for a UAV gimbal with transient impact load according to claim 7, characterized in that, Physical constraints include: Speed ​​constraints: ; Acceleration constraints: .

9. A high-precision anti-disturbance control method for a UAV gimbal with transient impact load according to claim 1, characterized in that, Based on the outer ring speed command, the total disturbance of multi-stage transient impacts is offset in real time, combined with... dq A mathematical model of the motor in a rotating coordinate system is established, and a linear active disturbance rejection controller is designed, specifically including: According to the speed command given by the outer ring To counteract the total disturbance caused by multi-stage transient impacts in real time, a second-order linear active disturbance rejection controller is designed based on the mechanical motion equations of the permanent magnet synchronous motor. The total disturbance of a system, including stage transient impacts, frictional forces, and parameter perturbations, is defined as the extended state variable. , For the total disturbance moment, the second-order LESO is established as follows: （14）； in, This is an estimated value for angular velocity. This is the estimated total system disturbance. This is an estimated value for angular acceleration. Let be the rate of change of the total system disturbance, and let the observer bandwidth be . ,but , , For system control variables, For actual measured speed , For proportional gain, For integral gain; To offset the estimated total system disturbance The following control law is designed: （15）； in, This is an estimated value for the system control gain. This is the disturbance compensation component; To control the input of the unperturbed integral cascade model, proportional control is adopted as follows: （16）； in, For controller bandwidth Determined proportional gain, .

Images