Robust control method for hypersonic vehicle based on weighted recursive filtering algorithm

Abstract

This invention discloses a robust control method for hypersonic vehicles based on a weighted recursive filtering algorithm. To accurately estimate and correct the impact of the wind field on the reentry vehicle, environmental disturbances are divided into stable and unstable disturbances. Based on an improved cubature Kalman filter, a weighted recursive algorithm is introduced to optimize the nonlinear disturbance observer and controller. The weights of each sampling point in the filtering process are adjusted using the weighted recursive method. A table of relationships between the optimal weight coefficient, confidence interval, and measurement error is statistically integrated based on simulation results, providing a better weight coefficient scheme for error observation in the vehicle reentry control system. This invention improves the real-time performance of the filter, reduces the mean square value of the tracking error through the nonlinear disturbance observer and robust controller, and enhances the vehicle's reentry command tracking capability.

Description

Technical Field

[0001] The present invention belongs to the technical field of aircraft guidance and control, and in particular relates to a hypersonic aircraft robust control method based on a weighted recursive filtering algorithm. Background Art

[0002] During the spacecraft's re-entry process, after the guidance system optimizes the instructions, it transmits them to the control system. The control system design primarily involves sensors, comparators, actuators, the controlled object, and measurement feedback. The control system transmits the received attitude angle command to the attitude controller. The attitude controller compares the received control command with the actual aircraft attitude using a comparator, and the controller calculates the attitude angular rate control command. The attitude angular rate controller then compares the received control command with the aircraft state to generate a control torque command. This control torque is then distributed to the control surfaces or RCS actuators to control the aircraft's attitude. The sensors then transmit the aircraft's state after execution back to the controller for feedback control.

[0003] Most current control methods simply incorporate a certain amount of interference, without considering the impact of atmospheric wind on the control system. Wind field fluctuations can interfere with the aircraft's actuators, affecting the aircraft's attitude angular rate and, consequently, the attitude angle, leading to deviations from the intended control value. Therefore, addressing wind field interference requires not only addressing the guidance law design but also considering how to improve the control method to better adapt to wind field variations and enhance overall performance. Summary of the Invention

[0004] Purpose of the invention: In order to improve the robustness of the control system and the attitude angle control tracking performance, the present invention provides a hypersonic aircraft robust control method based on a weighted recursive filtering algorithm.

[0005] Technical solution: The present invention provides a hypersonic vehicle robust control method based on a weighted recursive filtering algorithm, comprising the following steps:

[0006] (1) Acquiring reentry targets based on hypersonic vehicles;

[0007] (2) Obtain guidance instructions that meet the target mission through the aircraft's own guidance algorithm;

[0008] (3) After the attitude angle controller receives the guidance command, it calculates through the internal controller, combines the real-time state of the current aircraft attitude angle and the interference obtained by the interference observer and filter, and compares and analyzes to obtain the latest attitude angle rate control signal that meets the attitude angle control command;

[0009] (4) After the attitude angular rate controller receives the attitude angular rate control signal from the attitude angle controller, it calculates through the internal controller, combines the real-time state quantity of the current aircraft's attitude angular rate and the interference obtained by the interference observer, and compares and analyzes to obtain the latest control torque that meets the attitude angular rate control instruction;

[0010] (5) After the control surface actuator receives the control torque signal, it distributes it to the aircraft control surface actuator according to the signal, and the aircraft control surface adjusts to control the aircraft attitude;

[0011] (6) After the wind field is disturbed, the real-time attitude angle and real-time attitude angular rate of the hypersonic aircraft will change. The feedback system transmits the attitude angle and attitude angular rate signals affected by the wind field disturbance to the disturbance observer. The disturbance observer compares the ideal control quantity with the actual state quantity to obtain the disturbance error and transmits it to the weighted recursive improvement filter.

[0012] (7) Based on the weighted recursive improved filtering algorithm, the control deviation caused by wind turbulence is estimated to obtain the accurately estimated attitude angular rate interference and attitude angle interference, and the wind field interference is transmitted to the attitude angle controller and the attitude angular rate controller to form a closed loop.

[0013] Furthermore, the reentry target task in step (1) is the initial conditions, terminal constraints, and process constraints of the spacecraft reentry.

[0014] Furthermore, the attitude angle controller calculation method in step (3) is:

[0015]

[0016] x c =[α c ,β c ,σ c ] T 、x s =[α,β,σ] T

[0017] Among them, ω c is the ideal control value of attitude angular rate obtained by the controller algorithm; g s is the attitude angle control loop system matrix; f s is a nonlinear function of the attitude angle control loop state vector; k s =diag{k si} n×n is the design parameter of the attitude angle loop, k si is a positive value, is the estimated term of the constant wind field disturbance in the attitude angle control loop; x c To control the attitude angle, x s is the real-time attitude angle; is the differential term for controlling the attitude angle, To control the attitude angle estimate, is the real-time attitude angle estimation value; α c ,β c ,σ c are the ideal control values ​​of angle of attack, sideslip angle and roll angle, and α, β, σ are the actual values ​​of angle of attack, sideslip angle and roll angle given by the feedback system.

[0018] Furthermore, it is characterized in that the calculation formula of the attitude angular rate controller in step (4) is:

[0019]

[0020] x c(f) =[p c ,q c ,r c ] T 、x f =[p,q,r] T

[0021] Among them, M c is the ideal control value of the torque obtained by the controller algorithm; g f represents the attitude angular rate control loop system matrix; f f represents the nonlinear function of the state vector of the attitude angular rate control loop; ω c is the control value of the attitude angular rate, is the differential term of the ideal control value of the attitude angular rate; k f =diag{k fi} n×n is the design parameter of the attitude angular rate loop, k fi is a positive value, is the estimated term of the constant wind field disturbance in the attitude angular rate control loop; x c(f) To control the attitude angular rate, x f is the real-time attitude angular rate; is the estimated value of the real-time attitude angular rate; p c ,q c ,r c are the ideal control values ​​of the roll, pitch, and yaw rates, and p, q, and r are the actual values ​​of the roll, pitch, and yaw rates given by the feedback system.

[0022] Furthermore, the real-time attitude angle and real-time attitude angular rate in step (6) are:

[0023]

[0024]

[0025] in, is the differential term of the real-time attitude angle, is the differential term of the real-time attitude angular rate; f s represents the nonlinear function of the attitude angle control loop state vector, g s represents the attitude angle control loop system matrix, f f represents the nonlinear function of the attitude angular rate control loop state vector, g f represents the attitude angular rate control loop system matrix, M C =[l c ,m c ,n c ] T represents the control torque vector; d s and d f They are the wind field interferences on the attitude angle control loop and the wind field interferences on the attitude angular rate control loop respectively.

[0026] Furthermore, the calculation method of the attitude angle and attitude angular rate interference observer in step (6) is:

[0027]

[0028] Furthermore, the implementation process of step (7) is as follows:

[0029] The weighted recursive algorithm of the weighted recursive filter is to distribute the sample weights of the recursive sample length H linearly with a slope of K. The specific expression of the weight a(ni) of the nith sample in the interval is:

[0030]

[0031] Multiply the sample value by the corresponding weight to get the new sample value Z i The expression is:

[0032] Z i =s i ×a (n-i)

[0033] in represents the wind field disturbance estimation value obtained by the attitude angle and attitude angular rate disturbance observer;

[0034] The filtering algorithm is optimized based on the volumetric Kalman filter:

[0035] The observation quantity before improvement is Z i , the improved observation quantity is Assuming that the interference within the confidence interval is linearly distributed, The calculation method is:

[0036]

[0037] In the formula Indicates Z k-H to Z k-1 The average value of all observations between is expressed as:

[0038]

[0039] Where △Z k-1|k-2 It represents the difference between the mean of Z at time k-1 and k-2 within the confidence interval; △Z k-1|k-2 The expression is:

[0040]

[0041] According to the previous moment Z k-1 Estimated value of and variance P k-1|k-1 Get the volume point:

[0042]

[0043] There are six volume points in total, and the volume point propagation formula is:

[0044] χ i,k|k-1 =F(χ i,k-1|k-1 )

[0045] Based on predictions and the prediction variance P k|k-1 , calculate the volume point:

[0046]

[0047] according to The calculation equation for volume point propagation is:

[0048] Z i,k|k-1 =H(χ i,k|k-1 ),

[0049] Calculate measurement prediction Z k|k-1 , measurement prediction error variance S k , state and measurement cross-covariance C k , gain K k ; The measured variable Z k According to probability distribution With Z k / H, the filter output and variance are:

[0050]

[0051] Obtained Instead of the controller algorithm and

[0052] Beneficial effects: Compared with the existing technology, the beneficial effects of the present invention are: on the basis of cubature Kalman filtering, a weighted recursive algorithm is further introduced to adjust the weights of each sampling point in the filtering process, and a nonlinear disturbance observer and controller are designed; the present invention integrates the relationship table between the optimal weight coefficient, confidence interval and measurement error, provides a better weight coefficient scheme for error observation of the aircraft re-entry control system, and improves the robustness of the control system and the attitude angle control tracking performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 is a flow chart of the present invention;

[0054] Figure 2 It is the distribution diagram of the optimal weight coefficient K;

[0055] Figure 3 This is a diagram showing the tracking effect of the attitude angular rate by the control system after adopting the present invention;

[0056] Figure 4 This is a diagram showing the tracking effect of the attitude angle by the control system after adopting the present invention;

[0057] Figure 5 is the mean square value of the angle of attack error obtained after 60 simulations using the present invention;

[0058] Figure 6 It is the mean square value of the roll angle error obtained after 60 groups of simulations are performed using the present invention. DETAILED DESCRIPTION

[0059] The present invention will be further described in detail below with reference to the accompanying drawings.

[0060] The present invention provides a hypersonic aircraft robust control method based on a weighted recursive filtering algorithm, such as Figure 1 As shown, the specific steps include:

[0061] Step 1: Obtain the reentry target mission based on the hypersonic aircraft, including mission indicators such as the aircraft's initial reentry conditions, terminal constraints, and process constraints.

[0062] Step 2: Obtain guidance instructions α that meet the target mission through the aircraft’s own guidance algorithm c , β c , σ c .

[0063] Step 3: The attitude angle controller receives the guidance instruction α c , βc , σ c After that, the internal controller calculates the current state of the aircraft's attitude angle α, β, σ and the interference obtained by the interference observer and filter. The comparative analysis yields the latest attitude angular rate control signal p that satisfies the attitude angle control instruction. c ,q c 、r c .

[0064] The attitude angle controller is calculated as follows:

[0065]

[0066] x c =[α c ,β c ,σ c ] T 、x s =[α,β,σ] T

[0067] Among them, ω c is the ideal control value of attitude angular rate obtained by the controller algorithm; g s is the attitude angle control loop system matrix; f s is a nonlinear function of the attitude angle control loop state vector; k s =diag{k si} n×n is the design parameter of the attitude angle loop, k si is a positive value, is the estimated term of the constant wind field disturbance in the attitude angle control loop; x c To control the attitude angle, x s is the real-time attitude angle; is the differential term for controlling the attitude angle, To control the attitude angle estimate, is the real-time attitude angle estimation value; α c ,β c ,σ c are the desired control values ​​for the angle of attack, sideslip angle, and bank angle, and α, β, and σ are the actual values ​​of the angle of attack, sideslip angle, and bank angle given by the feedback system. Variables preceded by a “·” are derivative terms, and variables preceded by a “^” are estimated terms.

[0068] Step 4: The attitude angular rate controller obtains the angular rate control instruction p given by the attitude angle controller c ,q c 、r c After that, the internal controller calculates and combines the real-time state quantities p, q, r of the current aircraft's attitude angular rate and the interference obtained by the interference observer. Comparative analysis of the latest control torque l that meets the attitude angular rate control instruction c 、m c 、n c .

[0069]

[0070] x c(f) =[p c ,q c ,r c ] T 、x f =[p,q,r] T

[0071] Among them, M c is the ideal control value of the torque obtained by the controller algorithm; g f represents the attitude angular rate control loop system matrix; f f represents the nonlinear function of the state vector of the attitude angular rate control loop; ω c is the control value of the attitude angular rate, is the differential term of the ideal control value of the attitude angular rate; k f =diag{k fi} n×n is the design parameter of the attitude angular rate loop, k fi is a positive value, is the estimated term of the constant wind field disturbance in the attitude angular rate control loop; x c(f) To control the attitude angular rate, x f is the real-time attitude angular rate; is the estimated value of the real-time attitude angular rate; p c ,q c ,r c are the ideal control values ​​of the roll, pitch, and yaw rates, and p, q, and r are the actual values ​​of the roll, pitch, and yaw rates given by the feedback system.

[0072] Step 5: After the control surface actuator receives the control torque signal, it is distributed to the aircraft control surface actuator according to the signal, and the aircraft control surface adjusts to control the aircraft attitude.

[0073] Step 6: After the wind field is disturbed, the real-time attitude angle and real-time attitude angular rate of the hypersonic aircraft will change. The feedback system transmits the attitude angle and attitude angular rate signals affected by the wind field disturbance to the disturbance observer. The disturbance observer obtains the disturbance error S by comparing the ideal control quantity with the actual state quantity. s With S f , and passed to the weighted recursive improvement filter.

[0074] The real-time attitude angle and real-time attitude angular rate are:

[0075]

[0076]

[0077] in, is the differential term of the real-time attitude angle, is the differential term of the real-time attitude angular rate; f s represents the nonlinear function of the attitude angle control loop state vector, g s represents the attitude angle control loop system matrix, f f represents the nonlinear function of the attitude angular rate control loop state vector, g f represents the attitude angular rate control loop system matrix, M C =[l c ,m c ,n c ] T represents the control torque vector; d s and d f They are the wind field interferences on the attitude angle control loop and the wind field interferences on the attitude angular rate control loop respectively.

[0078] The calculation method of attitude angle and attitude angular rate disturbance observer is:

[0079] s s =x s -x c(s) s f =x f -x c(f)

[0080]

[0081]

[0082] Step 7: Estimate the control deviation caused by wind turbulence based on the weighted recursive improved filtering algorithm to obtain accurately estimated attitude angular rate interference and attitude angle interference, and transmit the wind field interference to the attitude angle controller and attitude angular rate controller to form a closed loop.

[0083] The weighted recursive algorithm of the weighted recursive filter is to distribute the sample weights of the recursive sample length H linearly with a slope of K. The specific expression of the weight a(ni) of the nith sample in the interval is:

[0084]

[0085] Multiply the sample value by the corresponding weight to get the new sample value Z iThe expression is:

[0086] Z i =s i ×a (n-i)

[0087] in represents the wind field disturbance estimation value obtained by the attitude angle and attitude angular rate disturbance observer;

[0088] The filtering algorithm is optimized based on the volumetric Kalman filter:

[0089] The observation quantity before improvement is Z i , the improved observation quantity is Assuming that the interference within the confidence interval is linearly distributed, The calculation method is:

[0090]

[0091] In the formula Indicates Z k-H to Z k-1 The average value of all observations between is expressed as:

[0092]

[0093] Where △Z k-1|k-2 It represents the difference between the mean of Z at time k-1 and k-2 within the confidence interval. k-1|k-2 The expression is:

[0094]

[0095] According to the previous moment Z k-1 Estimated value of and variance P k-1|k-1 Get the volume point:

[0096]

[0097] According to step (6) s and d f Algorithm, since the state of the system is a 3D state equation, and the number of volume points is twice the dimension, there are 6 volume points in total, and the volume point propagation formula is:

[0098] χ i,k|k-1 =F(χ i,k-1|k-1 ),i=1,2,3,4,5,6

[0099] Compute the forecast and forecast variance:

[0100]

[0101] According to the current and variance P k|k-1 Calculate volume point:

[0102]

[0103] according to The calculation equation for volume point propagation is:

[0104] Z i,k|k-1 =H(χ i,k|k-1 ),i=1,2,3,4,5,6

[0105] Measure and predict Z k|k-1 , measurement forecast error variance (innovation variance) S k , state and measurement cross-covariance C k , gain K k The expression is:

[0106]

[0107] The measured variable Z k According to probability distribution With Z k / H. The filtered output and variance are:

[0108]

[0109] The filtering method obtained Instead of the controller algorithm and

[0110] The main purpose of this invention is to study the design of the attitude control system of a hypersonic reentry vehicle under wind interference. Therefore, observing and tracking the attitude angle error caused by wind field interference is a prerequisite for achieving optimization. The present invention uses the mean square value of the tracking error of the state quantity and the control quantity to evaluate the quality of the tracking performance. Taking into account the randomness of the wind field, the different sampling periods of different aircraft, and the different confidence intervals, repeated experiments are carried out to take the weight coefficient K corresponding to the minimum mean square value of the error as the optimal value. The optimal weight coefficient K values ​​obtained by simulation under different confidence intervals and sampling errors are shown in Table 1:

[0111] Table 1 Optimal weight distribution table

[0112]

[0113]

[0114] like Figure 2As shown in Figure 1, after the Monte Carlo simulation test, the optimal coefficient K setting diagram corresponding to Table 1 is more intuitively displayed. It can be seen that under different confidence intervals and sampling errors, the value of the corresponding weight coefficient K when the minimum tracking error is taken.

[0115] like Figure 3 、 Figure 4 The figure shows a comparison between the control values ​​of the attitude angular rate and attitude angle during reentry obtained from a single experiment using the present invention and the actual values ​​under wind field interference using a conventional control method. It can be seen that the tracking error after the improvement is smaller than before the improvement, and the attitude angular rate and attitude angle tracking performance is better.

[0116] like Figure 5 、 Figure 6 Figure 2 shows the mean squared tracking error of the attitude angular rate and attitude angle during reentry, obtained from experiments using the improved method after 60 Monte Carlo target simulations. It can be seen that the mean squared tracking error of the attitude angular rate and attitude angle after the improvement is stable within 0.05, and the mean squared tracking error of the attitude angle after the improvement is stable within 0.3.

[0117] In addition, it should be understood that although this specification is described according to embodiments, not every embodiment contains only one independent technical solution. This narrative method of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation methods that can be understood by those skilled in the art.

Claims

1. A hypersonic vehicle robust control method based on a weighted recursive filtering algorithm, characterized in that: The following steps are involved: (1) Acquiring reentry targets based on hypersonic vehicles; (2) Obtain guidance instructions that meet the target mission through the aircraft's own guidance algorithm; (3) After the attitude angle controller receives the guidance command, it calculates through the internal controller, combines the real-time state of the current aircraft attitude angle and the interference obtained by the interference observer and filter, and compares and analyzes to obtain the latest attitude angle rate control signal that meets the attitude angle control command; (4) After the attitude angular rate controller obtains the attitude angular rate control signal given by the attitude angle controller, it performs calculations through the internal controller, combining the real-time state quantity of the current aircraft's attitude angular rate and the interference obtained by the interference observer; Comparative analysis yields the latest control torque that satisfies the attitude angular rate control command; (5) After the control surface actuator receives the control torque signal, it distributes it to the aircraft control surface actuator according to the signal, and the aircraft control surface adjusts to control the aircraft attitude; (6) After the wind field is disturbed, the real-time attitude angle and real-time attitude angular rate of the hypersonic aircraft will change. The feedback system transmits the attitude angle and attitude angular rate signals affected by the wind field disturbance to the disturbance observer. The disturbance observer compares the ideal control quantity with the actual state quantity to obtain the disturbance error and transmits it to the weighted recursive improvement filter. (7) Based on the weighted recursive improved filtering algorithm, the control deviation caused by wind turbulence is estimated to obtain the accurately estimated attitude angular rate interference and attitude angle interference, and the wind field interference is transmitted to the attitude angle controller and the attitude angular rate controller to form a closed loop.

2. The hypersonic vehicle robust control method based on weighted recursive filtering algorithm according to claim 1, characterized in that: The reentry target task in step (1) is the initial conditions, terminal constraints, and process constraints of the spacecraft reentry.

3. The hypersonic vehicle robust control method based on weighted recursive filtering algorithm according to claim 1, characterized in that: The calculation method of the attitude angle controller in step (3) is: x c =[a c ,b c ,s c ] T ,x s =[a,b,c] T Among them, ω c is the ideal control value of attitude angular rate obtained by the controller algorithm; g s is the attitude angle control loop system matrix; f s is a nonlinear function of the attitude angle control loop state vector; k s =diag{k si } n×n is the design parameter of the attitude angle loop, k si is a positive value, is the estimated term of the constant wind field disturbance in the attitude angle control loop; x c To control the attitude angle, x s is the real-time attitude angle; is the differential term for controlling the attitude angle, To control the attitude angle estimate, is the real-time attitude angle estimation value; α c ,β c ,σ c are the ideal control values ​​of angle of attack, sideslip angle and roll angle, and α, β, σ are the actual values ​​of angle of attack, sideslip angle and roll angle given by the feedback system.

4. The hypersonic vehicle robust control method based on weighted recursive filtering algorithm according to claim 3, characterized in that: The calculation formula of the attitude angular rate controller in step (4) is: x c(f) =[p c ,q c ,r c ] T 、x f =[p,q,r] T Among them, M c is the ideal control value of the torque obtained by the controller algorithm; g f represents the attitude angular rate control loop system matrix; f f represents the nonlinear function of the state vector of the attitude angular rate control loop; ω c is the control value of the attitude angular rate, is the differential term of the ideal control value of the attitude angular rate; k f =diag{k fi } n×n is the design parameter of the attitude angular rate loop, k fi is a positive value, is the estimated term of the constant wind field disturbance in the attitude angular rate control loop; x c(f) To control the attitude angular rate, x f is the real-time attitude angular rate; is the estimated value of the real-time attitude angular rate; p c ,q c ,r c are the ideal control values ​​of the roll, pitch, and yaw rates, and p, q, and r are the actual values ​​of the roll, pitch, and yaw rates given by the feedback system.

5. The hypersonic vehicle robust control method based on weighted recursive filtering algorithm according to claim 4, characterized in that: The real-time attitude angle and real-time attitude angular rate in step (6) are: in, is the differential term of the real-time attitude angle, is the differential term of the real-time attitude angular rate; f s represents the nonlinear function of the attitude angle control loop state vector, g s represents the attitude angle control loop system matrix, f f represents the nonlinear function of the attitude angular rate control loop state vector, g f represents the attitude angular rate control loop system matrix, M C =[l c ,m c ,n c ] T represents the control torque vector; d s and d f They are the wind field interferences suffered by the attitude angle control loop and the wind field interferences suffered by the attitude angular rate control loop respectively; The calculation method of the attitude angle and attitude angular rate interference observer is as follows:

6. The hypersonic vehicle robust control method based on weighted recursive filtering algorithm according to claim 5, characterized in that: The implementation process of step (7) is as follows: The weighted recursive algorithm of the weighted recursive filter is to distribute the sample weights of the recursive sample length H linearly with a slope of K. The specific expression of the weight a(ni) of the nith sample in the interval is: Multiply the sample value by the corresponding weight to get the new sample value Z i The expression is: Z i =s i ×a (n-i) in represents the wind field disturbance estimation value obtained by the attitude angle and attitude angular rate disturbance observer; The filtering algorithm is optimized based on the cubature Kalman filter: The observation quantity before improvement is Z i , the improved observation quantity is Assuming that the interference within the confidence interval is linearly distributed, The calculation method is: In the formula Indicates Z k-H to Z k-1 The average value of all observations between is expressed as: Where ΔZ k-1|k-2 It represents the difference between the mean of Z at time k-1 and k-2 within the confidence interval; ΔZ k-1|k-2 The expression is: According to the previous moment Z k-1 Estimated value of and variance P k-1|k-1 Get the volume point: The volumetric point propagation formula is: x i,k|k-1 =F(x i,k-1|k-1 ) Based on predictions and the prediction variance P k|k-1 , calculate the volume point: according to The calculation equation for volume point propagation is: Z i,k|k-1 =H(χ i,k|k-1 ), Calculate measurement prediction Z k|k-1 , measurement prediction error variance S k , state and measurement cross-covariance C k , gain K k ; The measured variable Z k According to probability distribution With Z k / H, the filter output and variance are: Obtained Instead of the controller algorithm and 7. The hypersonic vehicle robust control method based on weighted recursive filtering algorithm according to claim 6, characterized in that: There are six volume points.

Images