A control method of a posture-orbit combined interceptor

Abstract

The application discloses a control method of a posture-orbit composite interceptor, and relates to the technical field of automatic control. According to an angular acceleration calculation formula of a line-of-sight angle and a dynamic equation of the interceptor, a strong coupling state space equation is obtained; according to the strong coupling state space equation, a high-order multi-dimensional sliding mode matrix is constructed, and according to the high-order multi-dimensional sliding mode matrix, a strong coupling control law of posture control engines and orbit control engines of the interceptor is designed; according to a stable condition and a fixed time convergence condition of a guidance control system of the posture-orbit composite interceptor, an interference term in the strong coupling control law is solved to obtain an estimated value of the interference term; and the estimated value of the interference term and state data of the interceptor are substituted into the strong coupling control law of the interceptor to obtain lateral thrusts generated by the posture control engines and the orbit control engines and perpendicular to the axis of the interceptor, so that the interceptor is controlled. The method can improve the control precision of the interceptor.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automatic control, in particular to a control method of a posture-orbit composite interceptor. BACKGROUND

[0002] At present, the posture-orbit composite control system is composed of a trajectory control engine installed near the center of mass of the interceptor and a posture control engine installed at a relatively long distance in front of or behind the center of mass, which can provide lateral thrust and moment for the interceptor to change the acceleration and attitude of the interceptor.

[0003] However, in practical applications, the posture-orbit composite control system has the following characteristics: (1) the direction of the direct lateral force is fixedly coupled with the body and is perpendicular to the body axis, and changes with the attitude of the interceptor; (2) the posture control engine simultaneously generates force and moment increments; (3) due to fuel consumption and installation errors, the deviation between the trajectory control engine thrust point and the center of mass is inevitable, which causes the trajectory control engine to also simultaneously generate force and moment increments. The above characteristics cause the interceptor guidance control system under the posture-orbit composite control to face serious control input coupling and trajectory / attitude coupling problems.

[0004] However, the control of the interceptor under the posture-orbit composite control in the prior art has the problem of low precision. SUMMARY

[0005] Therefore, it is necessary to provide a control method of an interceptor under posture-orbit composite control, which can improve the control precision of the interceptor.

[0006] The present application adopts the following technical solutions:

[0007] The present application provides a control method of an interceptor under posture-orbit composite control, comprising:

[0008] According to the relative kinematic equation of the interceptor and the target in the longitudinal plane, a formula for calculating the angular acceleration of the line-of-sight angle between the interceptor and the target is determined;

[0009] According to the formula for calculating the angular acceleration of the line-of-sight angle and the dynamic equation of the interceptor, a strongly coupled state space equation is constructed;

[0010] According to the strongly coupled state space equation, a high-order multi-dimensional sliding mode matrix is constructed, and according to the high-order multi-dimensional sliding mode matrix, a strongly coupled control law for the posture control engine and the trajectory control engine of the interceptor is designed; the high-order multi-dimensional sliding mode matrix satisfies the field-of-view constraint of the strapdown seeker of the interceptor;

[0011] According to the stability condition and the fixed-time convergence condition of the interceptor guidance control system under posture-orbit composite control, the disturbance term in the strongly coupled control law is solved to obtain an estimated value of the disturbance term;

[0012] The estimated value of the interference term and the state data of the interceptor are substituted into the strong coupling control law of the interceptor to obtain side thrusts generated by the attitude control engine and the orbit control engine and perpendicular to the axis of the missile body;

[0013] The interceptor is controlled by the side thrusts generated by the attitude control engine and the orbit control engine and perpendicular to the axis of the missile body.

[0014] Optionally, an angular acceleration calculation formula of the line-of-sight angle between the interceptor and the target is determined according to a relative kinematics equation of the interceptor and the target in a longitudinal plane, and the angular acceleration calculation formula of the line-of-sight angle comprises:

[0015] The trajectory inclination velocity calculation formula of the interceptor is substituted into the relative kinematics equation of the interceptor and the target in the longitudinal plane to obtain the angular acceleration calculation formula of the line-of-sight angle.

[0016] The relative kinematics equation of the interceptor and the target in the longitudinal plane is:

[0017] ;

[0018] wherein, and respectively represent a relative distance of the missile and the line-of-sight angle, denotes a first-order derivative with respect to time, denotes a first-order derivative with respect to time, and respectively represent trajectory inclinations of the interceptor and the target, and respectively represent velocities of the interceptor and the target;

[0019] The trajectory inclination velocity calculation formula of the interceptor is:

[0020] ;

[0021] wherein, denotes the trajectory inclination velocity of the interceptor, which is a first-order derivative of , represents a velocity vector of the interceptor, and respectively represent side thrusts generated by the attitude control engine and the orbit control engine and perpendicular to the axis of the missile body, represents an attack angle, denotes lift generated by air power, denotes a mass of the interceptor;

[0022] The angular acceleration calculation formula of the line-of-sight angle is:

[0023] ;

[0024] wherein, denotes the angular acceleration of the line-of-sight angle, denotes a first disturbance parameter, denotes the gravitational acceleration, denotes the acceleration of the target.

[0025] Optionally, the dynamics equation of the interceptor is:

[0026] ;

[0027] wherein, denotes the body line-of-sight angle of the interceptor, denotes a first derivative, denotes the pitch rate of the interceptor, denotes a second disturbance parameter, denotes a third disturbance parameter, denotes a first derivative with respect to time, denotes the dynamic pressure, denotes the maximum cross-sectional area of the interceptor, denotes the length of the interceptor, and denote the pitch moment coefficients with respect to and , respectively, and denote the distance from the thrust point of the attitude control engine and the orbit control engine to the warhead, respectively, denotes the distance from the center of mass to the warhead, denotes the moment of inertia of the interceptor about the z-axis of the body coordinate system.

[0028] Optionally, according to the angular acceleration calculation formula of the line-of-sight angle and the dynamics equation of the interceptor, a strongly coupled state space equation is constructed, including:

[0029] Based on the angular acceleration calculation formula of the line-of-sight angle and the dynamics equation of the interceptor, the following are defined: , , , and , and the self-coupled state space equation is determined as:

[0030] ;

[0031] wherein, denotes the first derivative of , denotes the first derivative of , denotes the first derivative of , , , , , , , , , , , , , , , and denote the disturbance terms.

[0032] Optionally, according to the high-order multi-dimensional sliding mode matrix, a strong coupling control law of the attitude control engine and the orbit control engine of the interceptor is designed, comprising:

[0033] According to the high-order multi-dimensional sliding mode matrix, a reaching law is designed;

[0034] The high-order multi-dimensional sliding mode matrix is substituted into the reaching law to obtain the strong coupling control law of the attitude control engine and the orbit control engine of the interceptor;

[0035] The high-order multi-dimensional sliding mode matrix is:

[0036] ;

[0037] The reaching law is:

[0038] ;

[0039] wherein, and respectively represent an estimated value and an estimation error of , , ; , , , , , and are all normal numbers;

[0040] The strong coupling control law of the attitude control engine and the orbit control engine of the interceptor is:

[0041] ;

[0042] wherein, , , , , .

[0043] Optionally, according to the stability condition and the fixed-time convergence condition of the interceptor guidance control system under the posture-orbit composite, the disturbance term in the strong coupling control law is solved to obtain an estimated value of the disturbance term, including:

[0044] According to the stability condition and the fixed-time convergence condition, a constraint condition of a learning rate in the neural network model is determined;

[0045] According to the constraint condition of the learning rate in the neural network model, the neural network model is trained;

[0046] The relative distance between the interceptor and the target, the line-of-sight angle change rate of the interceptor, the body line-of-sight angle and the pitch angle rate are input into the trained neural network model to obtain the estimated value of the disturbance term.

[0047] Optionally, according to the stability condition and the fixed-time convergence condition, a constraint condition of a learning rate in the neural network model is determined, including:

[0048] According to the stability condition, a first constraint condition of the learning rate in the neural network model is determined;

[0049] According to the fixed-time convergence condition, a second constraint condition of the learning rate in the neural network model is determined;

[0050] According to the first constraint condition and the second constraint condition, the constraint condition of the learning rate in the neural network model is determined.

[0051] Optionally, according to the stability condition, a first constraint condition of the learning rate in the neural network model is determined, including:

[0052] According to the stability condition, a Lyapunov function is designed;

[0053] A first derivative of the Lyapunov function with respect to time is calculated;

[0054] Under the condition that the first derivative of the Lyapunov function with respect to time satisfies a condition of being less than or equal to 0, the first constraint condition is determined.

[0055] Optionally, according to the fixed-time convergence condition, a second constraint condition of the learning rate in the neural network model is determined, including:

[0056] A derivative of a cost function in the neural network model with respect to time is calculated;

[0057] Under the condition that the derivative of the cost function in the neural network model with respect to time satisfies the fixed-time convergence condition, the second constraint condition is determined.

[0058] Optionally, the first constraint condition is: , , ;

[0059] The second constraint condition is: ;

[0060] The constraint condition of the learning rate is:

[0061] ;

[0062] Wherein, the disturbance term in the strong coupling control law includes a first disturbance term , a second disturbance term and a third disturbance term , The corresponding learning rate of the first disturbance term is denoted as The corresponding learning rate of the second disturbance term is denoted as The corresponding learning rate of the third disturbance term is denoted as , , All are normal numbers, , and are all greater than 0, , The output of the first neuron in the hidden layer in the neural network is denoted as .

[0063] The application provides a control device of an interceptor under the combined action of attitude and orbit, comprising:

[0064] A determination module is configured to determine an angular acceleration calculation formula of a line-of-sight angle between the interceptor and a target according to a relative kinematics equation of the interceptor and the target in a longitudinal plane;

[0065] A self-definition module is configured to construct a strong coupling state space equation according to the angular acceleration calculation formula of the line-of-sight angle and a dynamics equation of the interceptor;

[0066] A design module is configured to construct a high-order multi-dimensional sliding mode matrix according to the strong coupling state space equation, and design a strong coupling control law of the interceptor according to the high-order multi-dimensional sliding mode matrix; the high-order multi-dimensional sliding mode matrix satisfies a field-of-view constraint of a strapdown seeker of the interceptor;

[0067] A solving module is configured to solve a disturbance term in the strong coupling control law according to a stable condition and a fixed time convergence condition of a guidance control system of the interceptor under the combined action of attitude and orbit, and obtain an estimated value of the disturbance term;

[0068] A obtaining module is configured to obtain a side thrust perpendicular to a missile body axis generated by an attitude control engine and an orbit control engine by substituting the estimated value of the disturbance term and state data of the interceptor into the strong coupling control law of the interceptor.

[0069] A control module is configured to control the interceptor by means of the lateral thrusts generated by the attitude control engine and the orbit control engine and perpendicular to the axis of the projectile body.

[0070] The application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to realize the control method of the interceptor under the attitude and orbit coupling.

[0071] The application provides a computer device, which comprises a memory, a processor and a computer program stored in the memory and capable of running on the processor, and the processor realizes the control method of the interceptor under the attitude and orbit coupling when executing the program.

[0072] The application adopts the above at least one technical scheme to achieve the following beneficial effects:

[0073] By constructing a high-order multi-dimensional sliding mode matrix, a strong coupling control law of the attitude control engine and the orbit control engine of the interceptor is designed, efficient utilization and compensation of the coupling effect of the attitude control engine and the orbit control engine are realized, the disturbance term in the strong coupling control law is solved in combination with the stability condition and the fixed time convergence condition of the guidance control system of the interceptor under the attitude and orbit coupling, an estimated value of the disturbance term is obtained, the fixed time convergence of the approximation error is realized on the premise of considering the stability of the control system, and therefore the control precision of the interceptor under complex disturbance is improved. BRIEF DESCRIPTION OF DRAWINGS

[0074] The accompanying drawings, which are included to provide a further understanding of the application and constitute a part of this application, illustrate certain illustrative embodiments of the application and together with the description serve to explain the application. In the drawings:

[0075] Figure 1 A control method flowchart of the interceptor under the attitude and orbit coupling is provided in the application;

[0076] Figure 2 A schematic diagram of the attitude and orbit coupling control effect in the initial state and the interception process is provided;

[0077] Figure 3 A response curve schematic diagram of the SCC and the GC in the body line angle change with time under different degrees of mass center drift is provided;

[0078] Figure 4 A response curve schematic diagram of the SCC and the GC in the pitch angle rate change with time under different degrees of mass center drift is provided;

[0079] Figure 5 A state response curve schematic diagram of the FTRBFNN and the RBFNN is provided in the application;

[0080] Figure 6 A schematic diagram of a miss distance result of a Monte Carlo shooting of a control method of a posture and orbit compound under interceptor provided by the present application is shown in the figure;

[0081] Figure 7 A schematic diagram of a control device of a posture and orbit compound under interceptor provided by the present application is shown in the figure;

[0082] Figure 8 A schematic diagram of a computer device for implementing the control method of the posture and orbit compound under interceptor provided by the present application is shown in the figure. DETAILED DESCRIPTION

[0083] In order to make the purpose, technical solutions and advantages of the present application clearer, the technical solutions of the present application will be described clearly and completely below in combination with specific embodiments of the present application and corresponding drawings. Obviously, the described embodiments are only some of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0084] Near space vehicles generally have the characteristics of fast flight speed, strong maneuvering ability and need for kinetic energy killing, which puts forward very high requirements on the guidance and control precision and response speed of the interceptor.

[0085] The near space target interceptor is affected by static factors and time-varying factors during flight, and faces strong uncertainty challenges caused by multi-source interference. On the one hand, as the flight altitude rises, the space atmospheric density, air temperature and dynamic pressure change significantly, thus forming unpredictable high-temperature gas effect, rarefied gas effect and flow transition phenomenon, causing uncertainty of the pressure distribution on the surface of the projectile, leading to dramatic changes in the aerodynamic moment of the interceptor projectile; on the other hand, the factors such as attitude and orbit compound control error, combination coordination error and time-varying mass characteristics caused by fuel consumption bring great internal uncertainty to the guidance and control. The traditional interference identification technology is proposed for a single interference source, and when facing multi-source interference, the strategy of independent identification of each interference is often adopted, which lacks consistency in the identification speed of each interference, and is not conducive to the unified regulation and control of the control system precision and convergence speed.

[0086] With the development of computing technology, for low-precision model of uncertain identification and control problem, researchers have proposed data-driven intelligent control methods, such as neural networks, fuzzy logic systems, support vector machines and Gaussian process regression. Neural networks are widely used to construct disturbance observers because of their ability to approximate nonlinear functions with arbitrary precision, low dependence on models and strong adaptive ability. In particular, radial basis function neural networks are widely used in aircraft control research because of their simple network structure and low computational pressure. However, the existing stability and convergence of neural network control system theory research is not perfect, and the related parameter setting greatly depends on the designer's experience.

[0087] However, the existing interceptor control method has the following defects: (1) for the guidance control system under strong coupling, there is a lack of integrated research on guidance control under the influence of complex coupling of attitude and orbit composite control; (2) there is a lack of theoretical research on the approximation performance of the neural network disturbance observer, and the approximation time of the neural network to the disturbance cannot be flexibly regulated.

[0088] Based on this, the present application provides a control method for an interceptor under attitude and orbit composite control for high-speed target interception problem, which utilizes favorable coupling and compensates for unfavorable coupling. At the same time, combined with the fixed-time neural network disturbance observer, the synchronous estimation and compensation of multi-source non-matching disturbance are realized, and the overall performance of the interceptor guidance control system is improved.

[0089] The technical solutions provided by the embodiments of the present application will be described in detail below with reference to the drawings.

[0090] Figure 1 The flowchart of the control method for an interceptor under attitude and orbit composite control in the present application specifically includes the following steps:

[0091] S101, according to the relative kinematics equation of the interceptor and the target in the longitudinal plane, determine the angular acceleration calculation formula of the line-of-sight angle between the interceptor and the target.

[0092] Ideally, the effects of attitude control and orbit control are independent of each other, but in fact, attitude / orbit control is severely coupled, and the coupled disturbance increases with the increase of the center of mass drift. The attitude and orbit composite control effect in the initial state and the interception process is as shown in Figure 2 , wherein Figure 2 , wherein Figure 2 , wherein , and represent the lateral thrusts perpendicular to the axis of the missile body generated by the attitude control engine and the orbit control engine, respectively. , and L represents the distance from the point of the attitude control engine and the point of the orbit control engine to the center of mass of the interceptor, L represents the distance from the point of the attitude control engine and the point of the orbit control engine to the center of mass of the interceptor, L represents the distance from the point of the attitude control engine and the point of the orbit control engine to the center of mass of the interceptor, L represents the distance from the point of the attitude control engine and the point of the orbit control engine to the center of mass of the interceptor, L represents the distance from the point of the attitude control engine and the point of the orbit control engine to the center of mass of the interceptor, L represents the distance from the point of the attitude control engine and the point of the orbit control engine to the center of mass of the interceptor,

[0093] The forces and torques acting on the interceptor in the attitude-orbit control system can be written as follows:

[0094] (1);

[0095] wherein, and are the sum of the force increment and the torque increment of the interceptor, and represent the lift and the pitching moment generated by the aerodynamic force. L represents the distance from the point of the attitude control engine and the point of the orbit control engine to the center of mass of the interceptor, and are abbreviated as and hereinafter.

[0096] The relative kinematic equation of the interceptor and the target in the longitudinal plane is:

[0097] (2);

[0098] wherein, and represent the relative distance and the line-of-sight angle of the interceptor and the target, denotes the first derivative with respect to time, denotes the first derivative with respect to time, and represent the trajectory inclination angles of the interceptor and the target, and represent the velocities of the interceptor and the target.

[0099] The trajectory inclination velocity of the interceptor is calculated by the formula:

[0100] (3);

[0101] wherein, denotes the trajectory inclination velocity of the interceptor, which is the first derivative of , represents the velocity vector of the interceptor, and respectively represent the lateral thrusts generated by the attitude control engine and the orbit control engine, respectively, and perpendicular to the axis of the missile body, represent the angle of attack, represent the lift generated by the air force, represent the mass of the interceptor.

[0102] Therefore, according to the relative kinematic equation of the interceptor and the target in the longitudinal plane, the angular acceleration calculation formula of the line-of-sight angle between the interceptor and the target is determined, including: substituting the ballistic inclination angle velocity calculation formula of the interceptor into the relative kinematic equation of the interceptor and the target in the longitudinal plane to obtain the angular acceleration calculation formula of the line-of-sight angle. That is, substituting formula (3) into formula (2) to obtain the angular acceleration calculation formula of the line-of-sight angle:

[0103] (4);

[0104] wherein, represent the angular acceleration of the line-of-sight angle, represent the first interference parameter, represent the gravitational acceleration, represent the acceleration of the target.

[0105] S102, according to the angular acceleration calculation formula of the line-of-sight angle and the dynamic equation of the interceptor, a strong coupling state space equation is constructed.

[0106] The strapdown seeker is widely used in precise guided interceptors due to its small size, light weight, low cost, high reliability and insensitivity to missile body overload. The field of view of the strapdown seeker is fixed with the missile body, and its measurement information is the body line-of-sight angle , in the longitudinal plane , wherein , and respectively represent the pitch angle and the pitch angle rate of the interceptor. On this basis, the dynamic equation of the interceptor is:

[0107] (5);

[0108] wherein, represent the body line-of-sight angle of the interceptor, represent the first derivative of , represent the pitch angle rate of the interceptor, represent the second interference parameter, represent the third interference parameter, represent the first derivative of with respect to time, represent the dynamic pressure, represent the maximum cross-sectional area of the interceptor, represent the length of the interceptor, and respectively represent the pitch moment coefficients of and respectively represent the pitch moment coefficients of and respectively represent the distance from the thrust point of the attitude control engine and the orbit control engine to the warhead, represents the distance from the center of mass to the warhead, represents the rotational inertia of the interceptor around the z-axis of the body coordinate system.

[0109] It should be noted that the first disturbance parameter, the second disturbance parameter and the third disturbance parameter are composed of control input error, aerodynamic parameter perturbation, measurement error, unknown dynamic characteristics, external disturbance, etc.

[0110] In one embodiment, according to the angular acceleration calculation formula of the line-of-sight angle and the dynamic equation of the interceptor, a strongly coupled state space equation is constructed, including: based on the angular acceleration calculation formula of the line-of-sight angle and the dynamic equation of the interceptor, defining , , , and , the strongly coupled state space equation is determined as:

[0111] (6);

[0112] wherein, represents the first derivative of , represents the first derivative of , represents the first derivative of , , , , , , , , , , , , , , , and represent disturbance terms.

[0113] That is, based on formula (4) and formula (5), the parameters are self-defined to obtain the strongly coupled state space equation.

[0114] S103, constructing a high-order multi-dimensional sliding mode matrix according to the strong coupling state space equation, and designing the strong coupling control law of the attitude control engine and the orbit control engine of the interceptor according to the high-order multi-dimensional sliding mode matrix; the high-order multi-dimensional sliding mode matrix satisfies the field of view constraint of the strapdown seeker of the interceptor.

[0115] According to the requirement of the interceptor to accurately hit the target while satisfying the field of view constraint of the strapdown seeker of the interceptor, a high-order multi-dimensional sliding mode matrix is constructed S As follows:

[0116] (7);

[0117] wherein, , S =0 represents a sliding surface, represents that the interceptor can hit the target, represents that the target is located in the center of the field of view of the seeker.

[0118] Optionally, the strong coupling control law of the attitude control engine and the orbit control engine of the interceptor is designed according to the high-order multi-dimensional sliding mode matrix, including: designing a reaching law according to the high-order multi-dimensional sliding mode matrix; and substituting the high-order multi-dimensional sliding mode matrix into the reaching law to obtain the strong coupling control law of the attitude control engine and the orbit control engine of the interceptor.

[0119] Specifically, to control the high-order multi-dimensional sliding mode matrix S to converge to zero, the reaching law is designed as:

[0120] (8);

[0121] wherein, and represent an estimated value and an estimated error of , , ; , , , , , and are all normal numbers.

[0122] Substituting the derivative of formula (7) with respect to time into formula (8) to obtain the strong coupling control law of the attitude control engine and the orbit control engine of the interceptor is:

[0123] (9);

[0124] wherein, , , , , .

[0125] S104, according to the stability condition and the fixed time convergence condition of the interceptor guidance control system under the posture and orbit composite, the disturbance term in the strong coupling control law is solved to obtain the estimated value of the disturbance term.

[0126] The neural network is a radial basis neural network, which is composed of an input layer, a hidden layer and an output layer, wherein the nonlinear activation function of the hidden layer is as follows:

[0127] (10);

[0128] wherein, , , , , , The number of hidden layer nodes is represented. The output of the first neuron in the hidden layer is represented. The Euclidean distance between and is represented, is a normal number representing the width of the Gaussian basis function, is a center vector.

[0129] The definitions of and , The network weight of the first neuron is represented, and the disturbance in the system can be represented as:

[0130] (11);

[0131] wherein, represents the approximation error, represents the ideal weight. It is assumed that , , and are all normal numbers.

[0132] The radial basis neural network estimates by adjusting the actual weight , so:

[0133] (12).

[0134] The error between the actual weight and the ideal weight is defined, and (11) is substituted into (12) to obtain:

[0135] (13).

[0136] The smaller the value, the better the neural network performs. The smaller the approximation error, the more cost function is constructed as follows:

[0137] (14);

[0138] in, .

[0139] Design using gradient descent method The adaptive change law makes Gradient descent, i.e. ,in, The learning rate of the neural network. Substituting (14) into it, we get:

[0140] (15).

[0141] The learning rate in a neural network needs to meet the following two requirements: (1) ensure the stability of the control system based on the neural network disturbance observer; (2) the neural network disturbance observer needs to synchronously approximate multiple unmatched disturbances within a fixed time.

[0142] Optionally, based on the stability conditions and fixed-time convergence conditions of the interceptor guidance and control system under attitude-orbit composite conditions, the disturbance term in the strongly coupled control law is solved to obtain an estimate of the disturbance term. This includes: determining the learning rate constraint in the neural network model based on the stability conditions and fixed-time convergence conditions; training the neural network model based on the learning rate constraint; and inputting the relative distance between the interceptor and the target, the rate of change of the interceptor's line-of-sight angle, the volume line-of-sight angle, and the pitch rate into the trained neural network model to obtain an estimate of the disturbance term.

[0143] In one embodiment, determining the constraints on the learning rate in a neural network model based on a stability condition and a fixed-time convergence condition includes: determining a first constraint on the learning rate in the neural network model based on the stability condition; determining a second constraint on the learning rate in the neural network model based on the fixed-time convergence condition; and determining the constraints on the learning rate in the neural network model based on the first constraint and the second constraint.

[0144] Optionally, based on the stability condition, the first constraint condition for the learning rate in the neural network model is determined, including: designing a Lyapunov function based on the stability condition, calculating the first derivative of the Lyapunov function with respect to time; and determining the first constraint condition under the condition that the first derivative of the Lyapunov function with respect to time is less than or equal to 0.

[0145] The Lyapunov function is as follows:

[0146] (16);

[0147] wherein, , , are all normal numbers.

[0148] The derivative of formula (8), (13) and (15) with respect to time is substituted into (16) to obtain:

[0149] (17).

[0150] Let when , , , , , is always true, and at this time the control system based on the neural network disturbance observer is stable, that is, the first constraint condition is: , , .

[0151] Alternatively, according to the fixed time convergence condition, a second constraint condition of the learning rate in the neural network model is determined, including: calculating the derivative of the cost function in the neural network model with respect to time; under the condition that the derivative of the cost function in the neural network model with respect to time satisfies the fixed time convergence condition, the second constraint condition is determined.

[0152] Based on formula (13) and (15), the derivative of the cost function with respect to time can be written as follows:

[0153] (18).

[0154] Let satisfy the following inequality

[0155] (19);

[0156] wherein, , and are all greater than 0, .

[0157] Substituting (19) into (18) obtains:

[0158] (20).

[0159] The formula (20) is true, which indicates It is stable over a fixed period of time. Convergence time for:

[0160] (twenty one).

[0161] Unaffected by real-time status, can be determined by parameters , and Regulation. In By selecting the same parameters, it is possible to achieve synchronous approximation of multiple unmatched disturbances.

[0162] Combining formulas (17) and (19), the learning rate The constraints are:

[0163] (twenty two);

[0164] The disturbance term in the strongly coupled control law includes the first disturbance term. Second interference item and the third interference term , Indicates the first interference term The corresponding learning rate, Indicates the second interference term The corresponding learning rate, Indicates the third interference term The corresponding learning rate, , , All are positive numbers. , and All are greater than 0. , This represents the th hidden layer in a neural network. The output of each neuron.

[0165] S105, substitute the estimated value of the disturbance term and the state data of the interceptor into the strongly coupled control law of the interceptor to obtain the lateral thrust perpendicular to the missile axis generated by the attitude control engine and the orbit control engine. The interceptor is controlled by the lateral thrust perpendicular to the missile axis generated by the attitude control engine and the orbit control engine.

[0166] Substituting the estimated value of the disturbance term and the state data of the interceptor into the strongly coupled control law of the interceptor, we obtain... and ,Right now and .

[0167] The application faces high-speed target interceptor under the posture and orbit composite control, and proposes a control method of the interceptor under the posture and orbit composite control, realizes efficient utilization and compensation of coupling through construction of a high-order multi-dimensional sliding mode control law, saves engine fuel while improving control performance, proposes a control system stability condition and a fixed time convergence condition for a radial basis neural network weight updating law in combination of the sliding mode control law and the fixed time theory, realizes fixed time convergence of approximation error on the premise of considering control system stability, and improves response speed and control precision of the interceptor under complex interference.

[0168] The effects of the application are verified through the following three parts: 1) strong coupling control method verification; 2) fixed time neural network disturbance observer performance verification; and 3) strong coupling control method verification based on the fixed time neural network disturbance observer.

[0169] 1) Strong coupling control method verification

[0170] Under different degrees of coupling (reflected by different degrees of center of mass drift), the control effects of the control method (SCC) of the interceptor under the posture and orbit composite proposed in the application and the traditional guidance-control loop separation method (guidance and control, GC) are compared and tested. As shown in Figure 3 、 Figure 4 and Table 1, wherein Figure 3 is a response curve diagram of the angle of the line of sight of SCC and GC with time under different degrees of center of mass drift, Figure 3 , wherein (a) is a response curve of the angle of the line of sight of the method SCC with time under different degrees of center of mass drift, Figure 3 , wherein (b) is a response curve of the angle of the line of sight of the traditional method GC with time under different degrees of center of mass drift; Figure 4 is a response curve diagram of the pitch angle rate of SCC and GC with time under different degrees of center of mass drift, Figure 4 , wherein (a) is a response curve of the pitch angle rate of the method SCC with time under different degrees of center of mass drift, Figure 4 , wherein (b) is a response curve of the pitch angle rate of the traditional method GC with time under different degrees of center of mass drift; and Table 1 is a miss distance and consumed impulse under SCC and GC control under different degrees of center of mass drift.

[0171] Table 1

[0172]

[0173] From Figure 3 and Figure 4It can be seen that, with the aggravation of coupling, the state response curves under SCC are almost coincident and finally converge to zero, while the control performance of GC significantly decreases with the increase of coupling; it can be known from the data in Table 1 that, under different couplings, SCC successfully intercepts the target, and the consumption of attitude and orbit control thrust impulse is not much different, while the impulse consumed by GC increases with the aggravation of coupling, and GC cannot successfully intercept the target when the coupling is serious. Based on the simulation results, it can be known that the robustness of SCC to unknown coupling is significantly stronger than that of GC, and SCC can realize high-precision control under attitude and orbit combination.

[0174] Performance verification of fixed-time convergence neural network disturbance observer: The disturbance approximation performance of the disturbance observer based on the fixed-time convergence radial basis function neural network (FTRBFNN) proposed in the application and the disturbance observer based on the asymptotic convergence radial basis function neural network (RBFNN) are compared and tested.

[0175] As shown in Figure 5 , FIGS. (a), (b) and (c) in Figure 5 are respectively response curves of the line-of-sight angular rate, the body line-of-sight angle and the pitch angle rate with time in the FTRBFNN and the RBFNN, and it can be seen from Figure 5 that, compared with the RBFNN, the state response curve converges faster and the tracking error is smaller under the action of the FTRBFNN, which indicates that the FTRBFNN can realize fast estimation and compensation of the disturbance, and thus improve the control precision and response speed of the interceptor. Figure 5 Performance verification of strong coupling control method based on fixed-time neural network disturbance observer: The performance of the strong coupling control method based on the fixed-time neural network disturbance observer proposed in the application is tested through 300 times of Monte Carlo shooting tests, and the test results are shown in

[0176] , which is a schematic diagram of the miss distance results of the control method of the interceptor under attitude and orbit combination provided by the application through Monte Carlo shooting, and it can be seen from Figure 6 that the average miss distance of 300 times of Monte Carlo tests is 0.88 m, and only once interception fails, which proves that the strong coupling control method based on the fixed-time neural network disturbance observer has strong robustness and can realize accurate interception of high-speed targets under complex disturbance. Figure 6 Figure 6

[0177] ​​The application faces a strong coupling system under the posture and orbit composite control, and proposes a control method of the interceptor under the posture and orbit composite control, improves the control performance by using the coupling effect between the posture control and the orbit control, and saves the engine fuel, and proposes a stable condition and a fixed time convergence condition for the weight updating law of the neural network, forms a strong coupling control method based on the fixed time neural network disturbance observer, realizes the rapid and synchronous approximation of the non-matching disturbance, and effectively improves the control precision and the response speed of the interceptor.

[0178] The control performance of the interceptor has very high requirements for the high-speed target interception task, the existing posture and orbit composite control research regards the posture control and the orbit control as two relatively independent links, lacks the in-depth analysis and effective use of the coupling effect caused by the posture control and the orbit control, and thus leads to low control precision and poor robustness of the control system. The application proposes a control method of the interceptor under the posture and orbit composite control, regards the posture control and the orbit control as a whole, introduces the coupling effect in the control law, and effectively improves the control performance of the interceptor under the posture and orbit composite control. Moreover, due to the complex flight environment in the near space and the unknown coupling caused by the posture and orbit composite control, it is difficult to accurately model the high-speed target interceptor, the existing disturbance identification method based on the data driving and facing the low-precision model lacks the research on the speed regulation technology of the identification, and it is difficult to realize the rapid and accurate estimation of the non-matching disturbance in the strong coupling system. The application proposes a fixed time neural network disturbance observer, combines the fixed time neural network disturbance observer with the strong coupling control, realizes the regulation of the disturbance identification speed, provides a technical approach for the rapid and synchronous estimation / compensation of the multi-source and non-matching disturbance, and improves the control precision and the response speed of the high-speed target interceptor under the posture and orbit composite control.

[0179] When the control method of the interceptor under the posture and orbit composite control provided by the application is applied, the order of the steps shown in Figure 1 may be executed, and the execution order of the specific steps can be determined according to the needs, and the application does not limit this.

[0180] The control method of the interceptor under the posture and orbit composite control provided by one or more embodiments of the application is based on the same idea, and the application also provides a corresponding control device of the interceptor under the posture and orbit composite control, as shown in Figure 7 .

[0181] Figure 7 A control device of the interceptor under the posture and orbit composite control provided by the application is shown in the figure, and the device 700 comprises:

[0182] A determination module 701 is configured to determine the angular acceleration calculation formula of the line-of-sight angle between the interceptor and the target according to the relative kinematic equation of the interceptor and the target in the longitudinal plane.

[0183] A self-definition module 702 is configured to construct a strong coupling state space equation according to the angular acceleration calculation formula of the line-of-sight angle and the dynamic equation of the interceptor.

[0184] The design module 703 is configured to construct a high-order multi-dimensional sliding mode matrix according to the strong coupling state space equation, and design a strong coupling control law of the interceptor according to the high-order multi-dimensional sliding mode matrix, and the high-order multi-dimensional sliding mode matrix satisfies the field of view constraint of the strapdown seeker of the interceptor.

[0185] The solving module 704 is configured to solve the interference term in the strong coupling control law according to the stability condition and the fixed time convergence condition of the interceptor guidance control system under the attitude-orbit combination, to obtain an estimated value of the interference term.

[0186] The obtaining module 705 is configured to substitute the estimated value of the interference term and the state data of the interceptor into the strong coupling control law of the interceptor, to obtain the lateral thrusts perpendicular to the axis of the missile body generated by the attitude control engine and the orbit control engine.

[0187] The control module 706 is configured to control the interceptor through the lateral thrusts perpendicular to the axis of the missile body generated by the attitude control engine and the orbit control engine.

[0188] The specific limitation of the control device of the interceptor under the attitude-orbit combination can be referred to the limitation of the control method of the interceptor under the attitude-orbit combination, which will not be repeated here. The modules in the control device of the interceptor under the attitude-orbit combination can be realized by software, hardware and combinations thereof. The modules can be embedded in or independent of the processor in the computer device in the form of hardware, or stored in the memory in the computer device in the form of software, so as to be called and executed by the processor.

[0189] The application further provides a computer readable storage medium, which stores a computer program, and the computer program can be used to execute the control method of the interceptor under the attitude-orbit combination. Figure 1 The control method of the interceptor under the attitude-orbit combination is provided.

[0190] The application further provides a computer device, which is configured to execute the control method of the interceptor under the attitude-orbit combination. Figure 8 The structure diagram of the computer device is shown in the accompanying drawings, and the computer device includes a processor, an internal bus, a network interface, a memory and a non-volatile memory. Figure 8 The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs, so as to execute the control method of the interceptor under the attitude-orbit combination. Figure 1 The control method of the interceptor under the attitude-orbit combination is provided.

[0191] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when executed, can include the processes of the above-mentioned embodiments of the methods. In the embodiments of the present application, any reference to memory, storage, database or other medium can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0192] The technical features of the above embodiments can be combined in any way. In order to make the description simple, all possible combinations of the technical features in the above embodiments are not described, but as long as the combination of the technical features does not exist, it should be considered as the scope of the present application.

Claims

1. A control method of a posture-orbit combined interceptor, characterized by, The method comprises the following steps: determining an angular acceleration calculation formula of a line-of-sight angle between the interceptor and the target according to a relative kinematics equation of the interceptor and the target in a longitudinal plane; constructing a strong coupling state space equation according to the angular acceleration calculation formula of the line-of-sight angle and a dynamics equation of the interceptor; constructing a high-order multi-dimensional sliding mode matrix according to the strong coupling state space equation, and designing a strong coupling control law of an attitude control engine and a trajectory control engine of the interceptor according to the high-order multi-dimensional sliding mode matrix; the high-order multi-dimensional sliding mode matrix meets a field of view constraint of a strapdown seeker of the interceptor; solving an interference term in the strong coupling control law according to a stable condition and a fixed time convergence condition of a guidance control system of the interceptor under the attitude and trajectory combination, to obtain an estimated value of the interference term; substituting the estimated value of the interference term and state data of the interceptor into the strong coupling control law of the interceptor, to obtain lateral thrusts generated by the attitude control engine and the trajectory control engine and perpendicular to an axis of the interceptor; controlling the interceptor through the lateral thrusts generated by the attitude control engine and the trajectory control engine and perpendicular to the axis of the interceptor; the step of solving the interference term in the strong coupling control law according to the stable condition and the fixed time convergence condition of the guidance control system of the interceptor under the attitude and trajectory combination, to obtain the estimated value of the interference term, comprises the following steps: determining a constraint condition of a learning rate in a neural network model according to the stable condition and the fixed time convergence condition; 2. The method of claim 1, wherein, training the neural network model according to the constraint condition of the learning rate in the neural network model, and inputting a relative distance between the interceptor and the target, a line-of-sight angle change rate of the interceptor, a body line-of-sight angle and a pitch angle rate into the trained neural network model, to obtain the estimated value of the interference term. the step of determining the angular acceleration calculation formula of the line-of-sight angle between the interceptor and the target according to the relative kinematics equation of the interceptor and the target in the longitudinal plane, comprises the following steps: substituting a ballistic inclination angle velocity calculation formula of the interceptor into the relative kinematics equation of the interceptor and the target in the longitudinal plane, to obtain the angular acceleration calculation formula of the line-of-sight angle; ； wherein, and respectively represent the relative distance and the line-of-sight angle of the interceptor to the target, denotes the first order derivative with respect to time, denotes the rate of change of the line-of-sight angle, is the first order derivative with respect to time, and respectively represent the ballistic angle of the interceptor and the target, and respectively represent the velocity of the interceptor and the target; the relative kinematics equation of the interceptor and the target in the longitudinal plane is: ； where represents the ballistic angle velocity of the interceptor, is the first derivative of , represents the velocity vector of the interceptor, and represent the lateral thrusts perpendicular to the axis of the projectile produced by the attitude control engine and the orbit control engine, respectively, represents the angle of attack, represents the lift force generated by the air force, represents the mass of the interceptor; the ballistic inclination angle velocity calculation formula of the interceptor is: ； wherein denotes the angular acceleration of the line of sight angle, denotes a first disturbance parameter, denotes the gravitational acceleration, denotes the acceleration of the target.

3. The method of claim 2, wherein, the angular acceleration calculation formula of the line-of-sight angle is: ； wherein, represents the boresight angle of the interceptor, represents the first derivative of represents the pitch rate of the interceptor, represents the second interference parameter, represents the third interference parameter, represents the first derivative with respect to time, represents the dynamic pressure, represents the maximum cross-sectional area of the interceptor, represents the length of the interceptor, and respectively represent the pitch moment coefficients with respect to and respectively represent the pitch moment coefficients with respect to and respectively represent the distance from the thrust point of the attitude control engine and the orbit control engine to the warhead, represents the distance from the center of mass to the warhead, represents the moment of inertia of the interceptor about the z-axis of the body coordinate system.

4. The method of claim 3, wherein, the dynamics equation of the interceptor is: Based on the angular acceleration calculation formula of the line-of-sight angle and the dynamic equation of the interceptor, the strong coupling state space equation is defined as: , , , and . ； wherein denotes the first derivative of , denotes the first derivative of , denotes the first derivative of , , , , , , , , , , , , , , , and denote interference terms.

5. The method of claim 4, wherein, the step of constructing the strong coupling state space equation according to the angular acceleration calculation formula of the line-of-sight angle and the dynamics equation of the interceptor, comprises the following steps: the step of designing the strong coupling control law of the attitude control engine and the trajectory control engine of the interceptor according to the high-order multi-dimensional sliding mode matrix, comprises the following steps: designing a reaching law according to the high-order multi-dimensional sliding mode matrix; The high-order multi-dimensional sliding mode matrix is: ； The approach law is: is: ； wherein and represent the estimate and the estimation error, respectively, , , ; , , , , , and are normal numbers; Strong coupling control law of interceptors' attitude control engine and orbit control engine Is: ； wherein , , , , .

6. The method of claim 1, wherein, substituting the high-order multi-dimensional sliding mode matrix into the reaching law, to obtain the strong coupling control law of the attitude control engine and the trajectory control engine of the interceptor; the step of determining the constraint condition of the learning rate in the neural network model according to the stable condition and the fixed time convergence condition, comprises the following steps: determining a first constraint condition of the learning rate in the neural network model according to the stable condition; determining a second constraint condition of the learning rate in the neural network model according to the fixed time convergence condition; determining the constraint condition of the learning rate in the neural network model according to the first constraint condition and the second constraint condition.

7. The method of claim 6, wherein, The first constraint condition of the learning rate in the neural network model is determined according to a stability condition, and the first constraint condition comprises: designing a Lyapunov function according to the stability condition; calculating a first derivative of the Lyapunov function with respect to time; determining the first constraint condition when the first derivative of the Lyapunov function with respect to time satisfies a condition of being less than or equal to 0.

8. The method of claim 6, wherein, The second constraint condition of the learning rate in the neural network model is determined according to a fixed-time convergence condition, and the second constraint condition comprises: calculating a derivative of a cost function in the neural network model with respect to time; determining the second constraint condition when the derivative of the cost function in the neural network model with respect to time satisfies the fixed-time convergence condition.

9. The method of claim 6, wherein, The first constraint condition is: , , ; The second constraint condition is: ; The constraint condition of the learning rate is: ； The disturbance term in the strong coupling control law includes a first disturbance term , a second disturbance term , and a third disturbance term , The first disturbance term corresponds to a learning rate The second disturbance term corresponds to a learning rate The third disturbance term corresponds to a learning rate , , are all positive constants, , and are all greater than 0, , represents the output of the th neuron in the hidden layer of the neural network.

Images