Adaptive and robust stability control method for strong static unstable launch vehicle

Abstract

The present application relates to a strong static unstable launch vehicle adaptive and robust stability control method, gap measurement is performed on the rated and deviation models and the nominal model at different times of the whole flight segment, the segmentation of the rated and deviation models in the time dimension is completed, and the robust controller in each segment is designed; the nominal model is tuned by using loop shaping method to obtain the model G s ; the model G s is obtained by order reduction of G sb , the uncertainty envelope between G sb and the rated and deviation model set is measured to obtain the controlled object G p containing multiplicative uncertainty; the control weight function W u and the error weight function W e are designed; G p , W u and W e are formed into a generalized robust control design object P, which is substituted into the robust control design architecture, and the robust controller is obtained by using the mu synthesis iterative method and order reduction; the adaptive adjustment law of the loop gain is designed and multiplied by the robust controller to stabilize the flight attitude of the launch vehicle in the current segment. The problem of attitude stability control of a strong static unstable launch vehicle is solved.

Description

Technical Field

[0001] This invention belongs to the fields of launch vehicle overall design, launch vehicle dynamics and attitude control technology, and relates to adaptive and robust stability control methods for strongly statically unstable launch vehicles. Background Technology

[0002] To enhance market competitiveness and simplify structural design for cost reduction and efficiency improvement, next-generation liquid-propellant launch vehicles often require larger fairing envelopes and simplified tail fin designs, exhibiting stronger aerodynamic and static instability compared to previous rocket configurations. Strongly statically unstable liquid-propellant launch vehicles are more significantly affected by airflow disturbances during atmospheric flight, necessitating a high dynamic response performance from their attitude control systems. However, considering the effects of elastic vibrations of the rocket body, propellant sloshing, and deviations in system characteristics, the controlled object exhibits strong low-to-mid-frequency coupling and high uncertainty, making the design of a controller that simultaneously balances dynamic response performance and robustness extremely challenging. Traditional rocket control technologies based on PD control and notch filter algorithms cannot meet design requirements; therefore, advanced control schemes that simultaneously guarantee high dynamic performance and resistance to model perturbations are needed. Summary of the Invention

[0003] The technical problem solved by this invention is: to solve the attitude stabilization control problem of a statically unstable launch vehicle, an adaptive and robust stabilization control method for a statically unstable launch vehicle is proposed. This method forms a robust controller design method for complex and highly uncertain dynamic models, and proposes an adaptive control architecture for online adjustment of control gain, thereby comprehensively improving the robustness of the system.

[0004] The solution to the technical problem of this invention is: to propose an adaptive and robust stability control method for a strongly statically unstable launch vehicle, comprising the following steps:

[0005] Step 1: Using the rated and deviation models as the evaluated models, select the nominal model, measure the gap between the rated and deviation models and the nominal model at different times throughout the flight segment, and complete the segmentation of the rated and deviation models in the time dimension. The rated and deviation model set in each segment uses the same robust controller for attitude control.

[0006] Design robust controllers for each segment, including steps two through six:

[0007] Step 2: The nominal model is tuned using the loop shaping method to obtain the tuned nominal model G. s ;

[0008] Step 3: Use the equilibrium truncation method based on canonical coprime decomposition to tune the nominal model G. s The order is reduced to obtain the reduced nominal tuning model G. sb Measure Gsb The uncertainty envelope between the set of nominal and deviation models yields the controlled object G containing multiplicative uncertainties. p ;

[0009] Step 4: Design the control weight function W in robust control u Sum of error weights W e ;

[0010] Step 5: The controlled object G containing multiplicative uncertainties... p Control weight function W u Sum of error weights W e A generalized robust control design object P is synthesized, which is then substituted into the standard robust control design architecture. The robust controller is then obtained using the μ synthesis iterative method, and the robust controller is reduced in order.

[0011] Step 6: In the case of unforeseen deviations between the actual flight system dynamics model and the ground-based dynamics model, an adaptive augmented control method is adopted to design an adaptive adjustment law for the loop gain, which is used to adaptively adjust the control loop gain.

[0012] Step 7: Multiply the robust controller from Step 5 and the adaptive loop gain control law from Step 6 to obtain the adaptive robust controller for stabilizing the flight attitude of the launch vehicle in the current segment.

[0013] Furthermore, the clearance measurement of the rated and deviation models and the nominal model at different times throughout the flight segment is as follows:

[0014] Based on the launch vehicle dynamics model, the dynamic parameters are fixed at a fixed second point and the deviation is linearized to obtain the corresponding rocket body transport characteristics G. a The model being evaluated; a selected flight time period, during which multiple rocket body transfer characteristics exist at different second points under rated and deviation conditions. The transfer characteristics of the entire rocket body during the flight period are used as a design model set. Used for subsequent construction of controlled objects containing multiplicative uncertainties;

[0015] A transfer characteristic is selected from the set of design models as the nominal model G, and the gap metric δ between the nominal model G and the evaluated model is used. ν Expressed as:

[0016]

[0017] In the formula The supremum of the function values ​​within the frequency ω range is ψ(G(jω),G a (jω) is used to calculate the amplitude difference between two closed-loop systems at different frequencies, when δ ν(G,G a When the value is less than the constant a, it indicates that the two models are close, where 0 < a < 1, and a is a constant close to 0.

[0018] Furthermore, the nominal model G is selected as the rated operating condition model with the median time within the flight period.

[0019] Furthermore, the tuning of the nominal model is as follows:

[0020] First, the nominal model G is preprocessed, i.e., the high-frequency elasticity information of the notch filter and low-pass filter attenuation is designed. The designed notch filter and low-pass filter are denoted as F. The preprocessed model GF is tuned using the loop shaping method to generate the tuning nominal model for robust control.

[0021] G s (s)=G(s)·F(s)·W1(s)

[0022] Wherein, model GF represents the model obtained by series notch filter and low-pass filter F of nominal model G, and W1 is the pre-weighting function, which is obtained by loop shaping method and used to improve the open-loop characteristics of mathematical model;

[0023] Before using the loop forming method, it is necessary to select the optimized target model G. d :

[0024]

[0025] Where ω c For the target cutoff frequency, G d The input is fed into the loop shaping method, and the existing loop shaping algorithm tools are used to solve it, resulting in the pre-weighting function W1.

[0026] Furthermore, the nominal model G is tuned... s The order is reduced as follows:

[0027] The launch vehicle dynamics model is a non-minimum phase model. After canonical coprime decomposition, two minimum phase models are obtained. The two models obtained by canonical coprime decomposition are then truncated to reduce the order of the non-minimum phase model, resulting in the reduced-order tuning nominal model G. sb ;

[0028] The gap metric method is used to evaluate the reduced-order tuning nominal model G. sb Compared with the nominal model G before order reduction s gap δ ν1 And compare the model before and after order reduction with the loop shaping controller K. s After closing the loop, the system performance is comparable to G. s gap δ ν1A model with a step response overshoot of less than 0.005 and a closed-loop step response overshoot and transition time difference of less than 5% is used as the evaluation index for order reduction accuracy. The order-reduced model that meets the evaluation index is used as the nominal tuning model G after order reduction. sb .

[0029] Furthermore, the controlled object G containing multiplicative uncertainty p as follows:

[0030]

[0031] According to the design model set M G The model within the model, and the frequency response of the notch filter and low-pass filter F concatenated with the pre-weighting function W1, are used to construct the multiplicative uncertainty envelope transfer function W. I (s), Δ I (s) is an arbitrary stable transfer function whose amplitude is less than or equal to 1 at any frequency.

[0032] Furthermore, the control weight function W u Sum of error weights W e as follows:

[0033]

[0034] Among them, M s Let ω be the closed-loop amplitude-frequency peak value of the sensitivity function. b Let M be the desired closed-loop bandwidth, ε be the steady-state error of the closed-loop system tracking the step response, and k ≥ 1. u Take the desired amplitude-frequency peak value of the controller input to the control command, ω bc ε1 represents the constrained controller bandwidth and the desired attenuation amplitude of the controller output at high frequencies.

[0035] Furthermore, the robust controller is reduced in order as follows:

[0036] If the robust controller is a minimum phase model, it is directly reduced using the balanced truncation method; if the robust controller is a non-minimum phase model, it is first reduced using the balanced truncation method based on canonical coprime decomposition, and then the order reduction accuracy of the robust controller is evaluated using the order reduction accuracy evaluation index, finally obtaining the minimum order order reduction robust controller K that meets the stability requirements.

[0037] Furthermore, the adaptive adjustment law for the gain of the designed loop is as follows:

[0038] k T =k0+k a

[0039] a = a - k T

[0040] In the formula, k T Here, k is the open-loop adjustment coefficient, k0 is the initial value of the adaptive gain, and k a The adaptive term for gain, where 'a' is the current control cycle loop gain. - This is the loop gain from the previous control cycle;

[0041] Adaptive term k a The regulation law is designed as follows:

[0042]

[0043] In the formula, For k a The first derivative, γ is the adaptive error gain, α is the spectral damper gain, β is the overflow gain, k max For k a The amplitude limit, (k max -k a ) / k max For logical damping terms; e r The tracking error of the attitude angle relative to the program angle. This is the error term, used to address rigid body error problems; increasing the gain improves the system's dynamic performance. -αk a y s The output of the spectrum damper is used to address the control-structure coupling problem, reduce gain, and improve the steady-state performance of the system; -β(k T -1) is an overflow term to prevent the gain from changing too quickly; the final adaptive robust controller is aK.

[0044] A computer-readable storage medium is proposed, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it implements the steps of the adaptive and robust stability control method for a strongly statically unstable launch vehicle.

[0045] The advantages of this invention compared to the prior art are:

[0046] This invention addresses the attitude stabilization control problem of strongly statically unstable launch vehicles, proposing an advanced control solution based on robust control. This solution encompasses key technologies such as controller segmentation using quantization model gaps, high-order model reduction and system uncertainty description, and robust controller design, overcoming the limitations of manual parameter tuning in traditional PD control architectures. Furthermore, the invention proposes a gain adaptive adjustment strategy, further enhancing the system's robustness to unforeseen deviations and disturbances. Attached Figure Description

[0047] Figure 1 Flowchart for loop forming design;

[0048] Figure 2 Design control structure diagram for loop shaping;

[0049] Figure 3 A flowchart for order reduction of a non-minimum phase model based on canonical coprime decomposition;

[0050] Figure 4 A robust control structure diagram considering uncertainties;

[0051] Figure 5 For multiplicative uncertainty envelope;

[0052] Figure 6 For a robust control model based on performance constraints for control commands and errors;

[0053] Figure 7 For robust control of generalized objects;

[0054] Figure 8 Flowchart for designing a robust controller based on a nominal model;

[0055] Figure 9 A schematic diagram of the full-loop adaptive robust control. Detailed Implementation

[0056] To address the attitude stabilization control problem of a statically unstable launch vehicle, this invention provides an adaptive and robust stabilization control method, comprising the following steps:

[0057] Step 1: Using the rated and deviation models as the evaluated models, select the nominal model, measure the gap between the rated and deviation models and the nominal model at different times throughout the flight segment, and complete the segmentation of the rated and deviation models in the time dimension. The rated and deviation model set in each segment uses the same robust controller for attitude control.

[0058] Design robust controllers for each segment, including steps two through six:

[0059] Step 2: The nominal model is tuned using the loop shaping method to obtain the tuned nominal model G. s ;

[0060] Step 3: Use the equilibrium truncation method based on canonical coprime decomposition to tune the nominal model G. s The order is reduced to obtain the reduced nominal tuning model G. sb Measure G sb The uncertainty envelope between the set of nominal and deviation models yields the controlled object G containing multiplicative uncertainties. p ;

[0061] Step 4: Design the control weight function W in robust control uSum of error weights W e ;

[0062] Step 5: The controlled object G containing multiplicative uncertainties... p Control weight function W u Sum of error weights W e A generalized robust control design object P is synthesized, which is then substituted into the standard robust control design architecture. The robust controller is then obtained using the μ synthesis iterative method, and the robust controller is reduced in order.

[0063] Step 6: In the case of unforeseen deviations between the actual flight system dynamics model and the ground-based dynamics model, an adaptive augmented control method is adopted to design an adaptive adjustment law for the loop gain, which is used to adaptively adjust the control loop gain.

[0064] Step 7: Multiply the robust controller from Step 5 and the adaptive loop gain control law from Step 6 to obtain the adaptive robust controller for stabilizing the flight attitude of the launch vehicle in the current segment.

[0065] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0066] Example 1

[0067] Step 1: Using the rated and deviation models as the evaluated models, select the nominal model, measure the gap between the rated and deviation models and the nominal model at different times throughout the flight segment, and complete the segmentation of the rated and deviation models in the time dimension. The rated and deviation model set in each segment uses the same robust controller for attitude control.

[0068] The clearance measurement of the rated and deviation models and the nominal model at different times throughout the flight segment is as follows:

[0069] Based on the fixed-second dynamic parameters of the launch vehicle dynamics model and the small-deviation linearization, the corresponding rocket body transport characteristics G can be obtained. a That is, the evaluated model G a A selected flight time period is defined, during which multiple rocket body transmission characteristics exist at different second points under rated and deviation operating conditions. The transfer characteristics of the entire rocket body during the flight period are used as a design model set. Used for subsequent construction of robust control objects with uncertainties.

[0070] In this design model set, a transfer characteristic is selected as the nominal model G. Generally, the nominal model G can be selected as the rated operating condition model with the median time within the flight period. Nominal model G and the model being evaluated G a The gap measurement value δ between ν Expressed as:

[0071]

[0072] In the formula The supremum of the function values ​​within the frequency ω range is ψ(G(jω),G a (jω) is used to calculate the amplitude difference between two closed-loop systems at different frequencies, when δ ν (G,G a When the value of δ is less than the constant a (0 < a < 1, and a is a constant close to 0), it indicates that the two models are similar. In this case, the controller of one stable model will also stabilize the other controller. Based on engineering experience, δ ν When the value is less than 0.3, a single set of control parameters can be used. Based on this principle, the controller segmentation, i.e., the parameter switching time, can be determined. For each segment, the design model set M... G The following steps are used to design a robust controller for this time period.

[0073] Design robust controllers for each segment, including steps two through six:

[0074] Step 2: The nominal model is tuned using the loop shaping method to obtain the tuned nominal model G. s .

[0075] The tuning of the nominal model is as follows:

[0076] according to Figure 1 The model tuning process first preprocesses the nominal model G, i.e., designs the high-frequency elasticity information for attenuation of the notch filter and low-pass filter; denoted as F, the designed notch filter and low-pass filter are then tuned using a loop shaping method to obtain the preprocessed model GF (nominal model in series with notch filter and low-pass filter), generating the tuned nominal model for robust control. The tuned nominal model is denoted as:

[0077] G s (s)=G(s)·F(s)·W1(s)

[0078] Wherein, W1 is the pre-weighting function, obtained by the loop shaping method, used to improve the open-loop characteristics of the mathematical model and tune it to an ideal open-loop shape;

[0079] Before using the loop forming method, it is necessary to select the optimized target model G. d G d The design goal is generally high low-frequency gain and fast high-frequency attenuation. The optimized target model G d It can be designed as follows:

[0080]

[0081] Where ω cThe target cutoff frequency is G. d The input is fed into the loop shaping method. At this point, the loop shaping design can be solved using mature loop shaping algorithm tools to obtain the pre-weighting function W1.

[0082] Figure 1 and Figure 2 K in s This is the loop shaping control function, another output of the loop shaping method, which satisfies the stability condition. The optimization model and its closed-loop control structure diagram using loop shaping are shown below. Figure 2 .

[0083] Step 3: Use the equilibrium truncation method based on canonical coprime decomposition to tune the nominal model G. s The order is reduced to obtain the reduced nominal tuning model G. sb Measure G sb With the rated and deviation model set M G The uncertainty envelope between them yields the controlled object G containing multiplicative uncertainty. p .

[0084] The nominal model G is tuned. s The order is reduced as follows:

[0085] Since the dynamic model of a launch vehicle is generally a non-minimum phase model, two minimum phase models can be obtained after canonical coprime decomposition. Balancing and truncation of these two models can reduce the order of the non-minimum phase model, resulting in the reduced-order tuning nominal model G. sb The reduced-order tuning nominal model G is evaluated using the same gap metric method as in step one. sb Compared with the nominal model G before order reduction s gap δ ν1 And compare the model before and after order reduction with the loop shaping controller K. s After closing the loop, the system performance is comparable to G. s gap δ ν1 Models with a step response overshoot of less than 0.005 and a closed-loop step response overshoot and transition time difference of less than 5% are used as evaluation indicators for order reduction accuracy. The order-reduced model that meets the evaluation indicators is used as the nominal tuning model G after order reduction. sb The order reduction process for the non-minimum phase model based on canonical coprime decomposition is shown below. Figure 3 .

[0086] Furthermore, measuring G sb The uncertainty envelope between the set of nominal and deviation models yields the controlled object G containing multiplicative uncertainties. p The details are as follows:

[0087] Construct robust control structures with multiplicative input uncertainties, such as Figure 4 As shown, G sb This is a reduced-order model with balanced truncation after loop shaping optimization. It is assumed that there is a multiplicative model uncertainty W between the nominal model and the actual model. I Construct a controlled object G containing multiplicative uncertainty. p as follows:

[0088]

[0089] Based on the design model set M determined in step one G The model within the model, and the frequency response of the notch filter and low-pass filter F concatenated with the pre-weighting function W1, are used to construct the multiplicative uncertainty envelope transfer function W. I (s), Δ I (s) is an arbitrary stable transfer function whose amplitude is less than or equal to 1 at any frequency. An example of the multiplicative uncertainty envelope constructed using a booster stage model of a certain type of rocket is shown below. Figure 5 .

[0090] Step 4: Design the control weight function W in robust control u Sum of error weights W e .

[0091] Construct a robust control architecture that meets performance requirements for tracking error and control signals, such as Figure 6 As shown, where G p For a controlled object with multiplicative uncertainty, W u W is the control weight function. e Design the control weight function W as the error weight function, according to the following weight function form. u and W e :

[0092]

[0093] Among them, M s Let ω be the closed-loop amplitude-frequency peak value of the sensitivity function. b Let M be the desired closed-loop bandwidth, ε be the steady-state error of the closed-loop system tracking the step response, and k≥1 make the sensitivity function achieve a steeper transition at high and low frequencies. u Take the desired amplitude-frequency peak value of the controller input to the control command, ω bc ε1 represents the constrained controller bandwidth and the desired attenuation amplitude of the controller output at high frequencies.

[0094] Step 5: The controlled object G containing multiplicative uncertainties... p Control weight function W u Sum of error weights W eA generalized robust control design object P is synthesized, which is then substituted into the standard robust control design architecture. Finally, the robust controller K0 is obtained by using the μ synthesis iterative method, and the robust controller is reduced in order.

[0095] like Figure 8 To obtain the controlled object G considering multiplicative uncertainties p After applying the weighting function, the robust control design object P in the augmented sense is obtained and substituted into the following... Figure 7 The standard robust control design framework is presented. The robust controller K0 is obtained using the μ-synthesis iterative method. If the order of the obtained robust controller K0 is too high, it can be reduced in order as follows:

[0096] If the robust controller K0 is a minimum phase model, it is directly reduced using the balanced truncation method. If the robust controller K0 is a non-minimum phase model, it is first reduced using the same balanced truncation method based on canonical coprime decomposition as in step three. Then, the same evaluation method as in step three, based on the gap metric and the difference in the system closed-loop response characteristics, is used to evaluate the order reduction accuracy of the robust controller. Finally, the minimum order reduced robust controller K that meets the stability requirements is obtained.

[0097] Step Six: When there is an unforeseen deviation between the actual flight system dynamics model and the ground-based dynamics model, an adaptive augmented control method is adopted to design an adaptive adjustment law for the loop gain, which is used to adaptively adjust the control loop gain.

[0098] The designed adaptive loop gain adjustment law is as follows:

[0099] k T =k0+k a

[0100] a = a - k T

[0101] In the formula, k T Here, k is the open-loop adjustment coefficient, k0 is the initial value of the adaptive gain, and k a The adaptive term for gain, where 'a' is the current control cycle loop gain. - This is the loop gain from the previous control cycle;

[0102] Adaptive term k a The regulation law is designed as follows:

[0103]

[0104] In the formula, For k a The first derivative, γ is the adaptive error gain, α is the spectral damper gain, β is the overflow gain, kmax For k a The amplitude limit, (k max -k a ) / k max This is a logic damping term; e r The tracking error of the attitude angle relative to the program angle. This is the error term, used to address rigid body error problems, increase the gain, and improve the system's dynamic performance; -αk a y s The output of the spectrum damper is used to address the control-structure coupling problem, reduce gain, and improve the steady-state performance of the system; -β(k T -1) is an overflow term to prevent the gain from changing too quickly. The initial value of a is set to a(t=0)=1. The adaptive augmented control (AAC) method is used for the adaptive adjustment of the loop gain. Based on the low-frequency tracking error and high-frequency energy information, the size of the system loop a is adaptively adjusted to achieve the overall upward and downward shift of the system amplitude-phase-frequency curve. The final adaptive robust controller is aK.

[0105] Step 7: Multiply the robust controller from Step 5 and the adaptive loop gain control law from Step 6 to obtain the adaptive robust controller for stabilizing the flight attitude of the launch vehicle in the current segment.

[0106] In summary, steps one through seven of the above steps achieved the design of a robust controller for the ground attitude control system and an adaptive adjustment strategy for the real-time flight control gain. The full-loop control architecture is shown below. Figure 9 This improves the robustness of the launch vehicle dynamics model to predict design uncertainties and flight anomalies.

[0107] This application provides a computer-readable storage medium storing computer instructions that, when executed on a computer, cause the computer to perform... Figure 1 The method described.

[0108] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0109] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0110] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0111] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0112] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

[0113] The contents not described in detail in this specification are common knowledge to those skilled in the art.

Claims

1. An adaptive and robust stability control method for a strongly statically unstable launch vehicle, characterized in that, Includes the following steps: Step 1: Using the rated and deviation models as the evaluated models, select the nominal model, measure the gap between the rated and deviation models and the nominal model at different times throughout the flight segment, and complete the segmentation of the rated and deviation models in the time dimension. The rated and deviation model set in each segment uses the same robust controller for attitude control. Design robust controllers for each segment, including steps two through six: Step 2: Tune the nominal model using the loop forming method to obtain the tuned nominal model. ; Step 3: Use the equilibrium truncation method based on canonical coprime decomposition to tune the nominal model. The order is reduced to obtain the reduced nominal tuning model. ,measure The uncertainty envelope between the set of nominal and deviation models yields the controlled object containing multiplicative uncertainties. ; Among them, the nominal model of the tuning The order is reduced as follows: The launch vehicle dynamics model is a non-minimum phase model. After canonical coprime decomposition, two minimum phase models are obtained. The two models obtained by canonical coprime decomposition are then truncated to reduce the order of the non-minimum phase model, resulting in the reduced-order nominal tuning model. ; The gap metric method is used to evaluate the reduced-order tuning nominal model. Compared with the nominal model before order reduction gap And compare the model before and after order reduction with the loop shaping controller. System performance after closed-loop control, and its relationship with gap A model with a step response overshoot of less than 0.005 and a closed-loop step response overshoot and transition time difference of less than 5% is used as the evaluation index for order reduction accuracy. The order-reduced model that meets the evaluation index is used as the nominal tuning model after order reduction. ; Step 4: Design the control weight function in robust control Sum of error weights ; Step 5: The controlled object containing multiplicative uncertainty... Control weight function Sum of error weights This synthesizes a generalized robust control design object. Substitute the standard robust control design architecture, and then utilize A robust controller is obtained by synthesizing iterative methods, and the robust controller is then reduced in order. Step 6: In the case of unforeseen deviations between the actual flight system dynamics model and the ground-based dynamics model, an adaptive augmented control method is adopted to design an adaptive adjustment law for the loop gain, which is used to adaptively adjust the control loop gain. Step 7: Multiply the robust controller from Step 5 and the adaptive loop gain control law from Step 6 to obtain the adaptive robust controller for stabilizing the flight attitude of the launch vehicle in the current segment.

2. The adaptive and robust stability control method for a strongly statically unstable launch vehicle according to claim 1, characterized in that, The clearance measurement of the rated and deviation models and the nominal model at different times throughout the flight segment is as follows: Based on the launch vehicle dynamics model, the dynamic parameters are fixed at a fixed second point and the deviation is linearized to obtain the corresponding rocket body transport characteristics. The model being evaluated; a selected flight time period, during which multiple rocket body transfer characteristics exist at different second points under rated and deviation conditions. The transmission characteristics of all rocket bodies during the flight period are used as a design model set. This is used to subsequently construct controlled objects containing multiplicative uncertainties; Select a transfer characteristic from the set of design models as the nominal model. nominal model The gap metric between the evaluated model and the evaluated model Expressed as: In the formula Characterizing frequency The supremum of function values ​​within the range, Used to calculate the amplitude difference between two closed-loop systems at different frequencies, when The value is less than a constant. When, it indicates that the two models are close, where ,and It is a constant that is close to 0.

3. The adaptive and robust stability control method for a strongly statically unstable launch vehicle according to claim 2, characterized in that, The nominal model Select the rated operating condition model based on the median time within this flight period.

4. The adaptive and robust stability control method for a strongly statically unstable launch vehicle according to claim 2, characterized in that, The tuning of the nominal model is as follows: First, regarding the nominal model Preprocessing is performed, namely, designing notch filters and low-pass filters to attenuate high-frequency elastic information; the designed notch filters and low-pass filters are denoted as... The pre-processed model was tuned using a loop forming method. Generate the nominal tuning model for robust control: Among them, the model Nominal model Series notch and low-pass filter The obtained model, The pre-weighting function, obtained by the loop shaping method, is used to improve the open-loop characteristics of the mathematical model. Before using the loop forming method, an optimized target model needs to be selected. : in For the target cutoff frequency, The input is fed into the loop shaping method, and the existing loop shaping algorithm tools are used to solve it, resulting in the pre-weighting function. .

5. The adaptive and robust stability control method for a strongly statically unstable launch vehicle according to claim 4, characterized in that, The controlled object containing multiplicative uncertainty as follows: According to the design model set The model within, as well as notch and low-pass filters. With pre-weighted function The frequency response after cascading is used to construct the multiplicative uncertainty envelope transfer function. , It is an arbitrary stable transfer function, and its amplitude is less than or equal to 1 at any frequency.

6. The adaptive and robust stability control method for a strongly statically unstable launch vehicle according to claim 1, characterized in that, The control weight function Sum of error weights as follows: in, The closed-loop amplitude-frequency peak value of the sensitivity function. For the desired closed-loop bandwidth, The steady-state error of the closed-loop system tracking the step response is a constant. , Take the desired amplitude-frequency peak value of the control command input to the controller. For the constrained controller bandwidth, The desired attenuation amplitude of the controller output at high frequencies.

7. The adaptive and robust stability control method for a strongly statically unstable launch vehicle according to claim 1, characterized in that, The reduction of the robust controller order is specifically as follows: If the robust controller is a minimum-phase model, it is directly reduced using the balanced truncation method. If the robust controller is a non-minimum-phase model, it is first reduced using the balanced truncation method based on canonical coprime decomposition, and then the order reduction accuracy of the robust controller is evaluated using the aforementioned order reduction accuracy evaluation index. Finally, a minimum-order order reduction robust controller that meets the stability requirements is obtained. .

8. The adaptive and robust stability control method for a strongly statically unstable launch vehicle according to claim 7, characterized in that, The adaptive adjustment law for the gain of the designed loop is as follows: In the formula, This is the open-loop adjustment coefficient. This is the initial value for the adaptive gain. For the adaptive term of gain, This represents the current control cycle loop gain. This is the loop gain from the previous control cycle; Adaptive terms The regulation law is designed as follows: In the formula, for The first derivative, For adaptive error gain, It is the gain of the spectrum damper. It is the overflow gain. for The amplitude limit, For logical damping terms; The tracking error of the attitude angle relative to the program angle. This is the error term, used to address rigid body error problems; increasing the gain improves the system's dynamic performance. The output of the spectrum damper is used to solve the control-structure coupling problem, reduce the gain and improve the steady-state performance of the system; To prevent excessively rapid gain changes, an overflow term is used; the final adaptive robust controller is... .

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 8.

Images