Automatic Optimization Design Method of Missile Stabilization Control Parameters Based on Genetic Algorithm

Abstract

The present invention discloses a method for automatically optimizing the design of missile stability control parameters based on genetic algorithms. The design of missile stability control parameters is a multi-parameter combination optimization problem. Aiming at the problems in the current design of missile stability control systems that rely on the engineering design experience of designers and may not be able to obtain the combined optimized controller parameters globally, a method for automatically optimizing the design of missile stability control system parameters based on genetic algorithms is proposed. The invention makes full use of the global search and optimization characteristics of genetic algorithms, introduces the performance indicators required by the system into the fitness function, and thus realizes the automatic optimization design of controller parameters. Applying the present invention to the automatic optimization design of the parameters of a certain type of missile stability control system, the simulation results show good control effects, verifying the correctness and feasibility of the present invention. The present invention realizes the automated design of missile stability control parameters and has strong practical value.

Description

Technical Field

[0001] The present invention relates to an automatic optimization design method for missile stability control parameters based on genetic algorithms, belonging to the field of aerospace. Background Art

[0002] The design of missile stability control parameters is mostly based on classical control theory. First, based on small perturbation linearization, the time-varying nonlinear missile body model is transformed into several linear systems with piecewise constant parameters at several characteristic points of the missile body change. Then, for these constant linear systems, the control system is designed while retaining sufficient stability margins and performance. The design of missile stability control parameters based on this conventional idea needs to be based on a certain "trial and error", and the whole design process includes multiple cycles of repeated design and verification, and the design results may not be able to obtain the combined optimized controller parameters globally. Summary of the Invention

[0003] Object of the Invention: In order to overcome the deficiencies in the prior art, the present invention provides an automatic optimization design method for missile stability control parameters based on genetic algorithms, which uses genetic algorithms to realize the self-optimization design of controller parameters at different characteristic points, which can free designers from multiple cycles of repeated design and verification and realize the automatic design of missile stability control parameters.

[0004] Technical Solution: To solve the above technical problems, an automatic optimization design method for missile stability control parameters based on genetic algorithms of the present invention includes the following steps:

[0005] (1) Determine the missile stability control loop structure according to missile design requirements;

[0006] (2) Determine the variables to be optimized among the control parameters;

[0007] (3) Design the fitness function and the optimization termination condition;

[0008] (4) Initialize the optimization parameters, including: the serial number n of the characteristic point = 1, the total number N of characteristic points, the number of generations of evolution Gen = 1, the maximum number of iterations, the population size, the optimization range of the parameters to be optimized, the coding method, the initialization population, and the genetic operation method;

[0009] (5) For the nth characteristic point, start the automatic optimization design of the parameters;

[0010] (6) For the population of the Gen-th generation, calculate the fitness values of the individuals in the population;

[0011] (7) For the population of the Gen-th generation, perform selection, crossover, and mutation operations in sequence according to the set genetic operation method;

[0012] (8) Determine whether the current evolution generation Gen is less than the maximum number of iterations. If it is less, set Gen = Gen + 1, and execute step (6); otherwise, execute step (9).

[0013] (9) Calculate the time-domain response performance under the optimal individual in step (8), and determine whether the time-domain response performance of the optimal individual meets the requirements of the time-domain performance index. If it does not meet the requirements, execute step (4); otherwise, execute step (10).

[0014] (10) Output and save the optimization result of the nth feature point as the design result of the parameters to be optimized at the nth feature point.

[0015] (11) Determine whether the feature point serial number n is less than the total number of feature points N. If it is less, set n = n + 1 and execute step (4); otherwise, end the automatic optimization design process.

[0016] Preferably, the missile stable control loop structure in step (1) is that the input signal enters the fourth parameter adjustment module, which is sequentially provided with a first comparator, a first parameter adjustment module, a second comparator, a second parameter adjustment module, a third comparator, and a third parameter adjustment module. The signal acts on the rudder system of the missile body, and the signals collected by the gyroscope on the missile body are simultaneously fed back to the second comparator and the third comparator, and the signals collected by the accelerometer on the missile body are fed back to the first comparator.

[0017] Preferably, in step (1), the parameter adjustment parameters of the first parameter adjustment module, the second parameter adjustment module, the third parameter adjustment module, and the fourth parameter adjustment module are K1, K2, K3, and K4 respectively.

[0018] Preferably, in step (2), the parameters to be optimized are the time constants t1, t2, and the damping coefficient ξ.

[0019] Preferably, the relationship model between the parameters to be optimized t1, t2, and ξ and the control loop parameter adjustment parameters K1, K2, K3, and K4 is as follows:

[0020]

[0021]

[0022]

[0023]

[0024]

[0025] Among them: a1, a2, a3, a4, and a5 are missile dynamic coefficients.

[0026] Preferably, in the step (3), the fitness function is in the form of:

[0027] J = σ1(M zn - M znzb ) 2 + σ2(P zn - P znzb ) 2 + σ3(M js - M jszb ) 2 + σ4(P js - P jszb ) 2

[0028] Or

[0029] J = σ1|M zn - M znzb | + σ2|P zn - P znzb | + σ3|M js - M jszb | + σ4|P js - P jszb |

[0030] Wherein, M zn and P zn are the open-loop amplitude margin and phase margin disconnected at K3, M js and P js are the open-loop amplitude margin and phase margin disconnected at K1, M znzb , P znzb , M jszb , P jszb are the requirements for the frequency-domain performance indexes of the control system, and σ1, σ2, σ3, σ4 are the weighting coefficients.

[0031] Preferably, in the step (3), the optimization termination condition is t r < t rzb and σ < σ zb , where t r and σ are respectively the rise time and overshoot of the time-domain response of the control loop, and t rzb and σ zb are respectively the requirements for the time-domain performance indexes of the control loop.

[0032] The present invention aims at each characteristic point of the missile stable control loop, designs a fitness function related to the frequency-domain performance indexes, and at the same time uses the time-domain performance indexes as the necessary conditions for optimization termination, and adopts a genetic algorithm to design a method capable of automatically optimizing the controller parameters, realizing the automatic design of the missile stable control parameters.

[0033] In the present invention, the fitness function is mainly composed of frequency-domain performance indexes. Considering the differences in the orders of magnitude of various frequency-domain performance indexes, a weighted coefficient is used to normalize each frequency-domain index. In addition, the time-domain performance index is used as a necessary condition for terminating the optimization. That is, if the optimization result does not meet the time-domain performance index, the iterative optimization is restarted until the optimization result meets the requirements of the time-domain performance index. The present invention optimizes the time constants t1, t2 and the damping coefficient ξ that vary within a small range, and then indirectly calculates K1, K2, K3 and K4 by using the relationship model between the above t1, t2, ξ and K1, K2, K3 and K4, which can effectively avoid the problem that it is impossible to directly find the approximate optimal solution due to the large variation range of K1, K2, K3 and K4.

[0034] Beneficial effects: The automatic optimization design method for missile stability control parameters based on the genetic algorithm of the present invention has the following advantages:

[0035] 1. The design method of the present invention liberates designers from repeated design and verification, realizes the automatic design of missile stability control parameters, and improves the design efficiency;

[0036] 2. The design method of the present invention has design results that meet both time-domain performance indexes and frequency-domain performance indexes, and the design method has strong practical value. Description of the drawings

[0037] Figure 1 is the structural diagram of the missile lateral stability control loop.

[0038] Figure 2 is the implementation flowchart of the design method.

[0039] Figure 3 is the frequency-domain characteristic curve of the characteristic point.

[0040] Figure 4 is the time-domain unit step overload response curve of the characteristic point.

[0041] Figure 5 is the six-degree-of-freedom overload response curve. Specific implementation manners

[0042] The present invention will be further described below with reference to the drawings.

[0043] As Figures 1 to 5 shown, determine the structural diagram of the missile stability control loop. The present invention takes the missile lateral stability control loop as an example, and its control loop structural diagram is shown in the appendix Figure 1 . It should be noted that the principle of the present invention is equally applicable to the automatic optimization design of control parameters of other forms of stability control loops and rolling stability loops. In this control loop:

[0044] uz For executing the control instruction of the execution system, n y For overload response, K1, K2, K3, and K4 are tuning parameters. The rudder system, gyroscope, and accelerometer model are based on the actual system model. Taking the missile body in the pitch plane as an example, the pitch motion model of the missile body is as follows:

[0045]

[0046] Among them, a1, a2, a3, a4, and a5 are missile dynamic coefficients. These five parameters are provided by the overall aerodynamic specialty at the beginning of the design and are known for the design of control parameters. a1 is the damping coefficient, a2 is the static stability coefficient, a3 is the rudder efficiency coefficient, a4 is the normal force coefficient, and a5 is the rudder lift coefficient, collectively referred to as missile dynamic coefficients. θ is the pitch angle, θ is the ballistic inclination angle, α is the angle of attack, Δδ is the rudder deflection angle, V d is the missile speed, and g is the acceleration due to gravity.

[0047] 2. Determine the parameters to be optimized as the time constants t1, t2, and the damping coefficient ξ. The optimization range of the parameters to be optimized is 0.05 ≤ t1 ≤ 0.1, 0.06 ≤ t2 ≤ 0.15, 0.7 ≤ ξ ≤ 0.9. At the same time, establish the relationship model between the parameters to be optimized t1, t2, and ξ and the tuning parameters K1, K2, K3, and K4 of the control loop as follows:

[0048]

[0049]

[0050]

[0051]

[0052]

[0053] 3. Design the fitness function as:

[0054] J = σ1(M zn - M znzb ) 2 + σ2(P zn - P znzb ) 2 + σ3(M js - M jszb ) 2 + σ4(P js - P jszb ) 2

[0055] Or

[0056] J = σ1|M zn - Mznzb | + σ2 | P zn - P znzb | + σ3 | M js - M jszb | + σ4 | P js - P jszb |

[0057] Among them, M zn and P zn are Figure 1 the open-loop amplitude margin and phase margin that are disconnected at K3 in js and P js are Figure 1 the open-loop amplitude margin and phase margin that are disconnected at K1 in znzb , P znzb , M jszb , P jszb are the requirements for the frequency-domain performance indicators of the control system. σ1, σ2, σ3, and σ4 are weighting coefficients.

[0058] 4. Design the optimization termination conditions, taking the maximum number of iterations and the time-domain performance indicators as the necessary conditions for optimization termination:

[0059] t r < t rzb and σ < σ zb

[0060] Among them, t r and σ are the rise time and overshoot of the time-domain response of the control loop respectively, and t rzb and σ zb are the requirements for the time-domain performance indicators of the control loop respectively;

[0061] 5. Initialize the optimization parameters, including: the serial number of the characteristic point n = 1, the total number of characteristic points N (the characteristic points are representative characteristic aerodynamic points in the entire flight airspace of the missile, such as the points with the maximum and minimum dynamic pressure, the point with the maximum aerodynamic time constant, the highest point of the elastic vibration frequency, etc., which are selected according to the data provided by the aerodynamic specialty at the beginning of the control parameter design. Since the selection of the characteristic points does not involve the design of the control parameters, the selection of the characteristic points is not given in the text and is considered known for the control parameter design), the generation number Gen = 1, the maximum number of iterations, the population size, the optimization range of the parameters to be optimized, the coding method, the initialization population, and the genetic operation method, etc.;

[0062] 6. For the nth characteristic point, start the automatic parameter optimization design;

[0063] 7. For the Gen-th generation population, calculate the fitness values of the individuals in the population;

[0064] 8. For the Gen-th generation population, perform selection, crossover, and mutation operations in sequence according to the set genetic operation method;

[0065] 9. Determine whether the current evolutionary generation Gen is less than the maximum number of iterations. If it is less, set Gen = Gen + 1 and execute step 7; otherwise, execute step 10.

[0066] 10. Calculate the time-domain response performance under the optimal individual in step 9. Determine whether the time-domain response performance of the optimal individual meets the requirements of the time-domain performance index. Calculate the corresponding time-domain response (rise time t r , overshoot σ) of the optimal individual, and directly compare it with the requirements of the time-domain performance index t rzb and σ zb . If t r <t rzb and σ < σ zb , it is considered to meet the requirements; otherwise, it does not meet the requirements. If it does not meet the requirements, execute step 5 (when executing step 5, only some conditions need to be re-initialized. Specifically, only initialize n = 1, evolutionary generation Gen = 1, and initialize the population in step 5); otherwise, execute step 11.

[0067] 11. Output and save the optimization result of the nth feature point as the design result of the parameter to be optimized at the nth feature point.

[0068] 12. Determine whether the feature point serial number n is less than the total number of feature points N. If it is less, set n = n + 1 and execute step 5; otherwise, end the automatic optimization design process. By using the method of the present invention, the optimal control strategy at each feature point can be quickly found, reducing the workload of designers' repeated design and verification in multiple cycles, and realizing the automatic design of missile stability control parameters.

[0069] In the present invention, the genetic algorithm is used for parameter optimization. The main difficulty lies in designing an objective function that matches the actual problem. Commonly used performance indicators for the objective function include integral of error (IE), integral of absolute error (IAE), integral of squared error (ISE), integral of the product of time and absolute error (ITAE), etc. However, these conventional performance indicator functions only consider errors and do not combine with the time-domain and frequency-domain performance index requirements in actual engineering, resulting in the optimization result having no direct relationship with the time-frequency domain performance. Simulation results show that if the time-domain and frequency-domain performances are simultaneously considered in the performance indicator function, a satisfactory optimization result cannot be obtained. The objective function designed in this patent separately considers the time-domain performance and the frequency-domain performance, takes the frequency-domain performance as the main factor of the objective function, takes the time-domain performance as the terminable condition, and considers the normalization problem, obtaining a satisfactory optimization result and improving the optimization efficiency.

[0070] Using genetic algorithms to optimize the controller parameters can, on the one hand, solve the existing problem of repeated calculations, and on the other hand, make the design results optimal and unified, and can quantitatively measure the design results. When using conventional design methods, the design results of different designers may all meet the requirements of time-frequency domain performance indicators, but the design results are inevitably affected by the experience of the designers and cannot be quantitatively compared and judged.

[0071] The above are only the preferred embodiments of the present invention. It should be pointed out that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention.

Claims

1. A method for automatically optimizing the design of missile stability control parameters based on genetic algorithm, characterized in that It includes the following steps: (1) Determine the missile stable control loop structure according to the missile control design requirements; (2) Determine the variables to be optimized among the control parameters; (3) Design the fitness function and the optimization termination condition; (4) Initialize the optimization parameters, including: the serial number n of the characteristic point = 1, the total number N of characteristic points, the generation number Gen = 1, the maximum number of iterations, the population size, the optimization range of the parameters to be optimized, the coding method, the initialization population, and the genetic operation method; (5) For the nth characteristic point, start the automatic parameter optimization design; (6) For the population of the Gen-th generation, calculate the fitness values of the individuals in the population; (7) For the population of the Gen-th generation, perform selection, crossover, and mutation operations in sequence according to the set genetic operation method; (8) Judge whether the current generation number Gen is less than the maximum number of iterations. If it is less, set Gen = Gen + 1 and execute step (6). Otherwise, execute step (9); (9) Calculate the time-domain response performance under the optimal individual in step (8), and judge whether the time-domain response performance of the optimal individual meets the requirements of the time-domain performance index. If it does not meet, execute step (4). Otherwise, execute step (10); (10) Output and save the optimization result of the nth characteristic point as the design result of the parameter to be optimized at the nth characteristic point; (11) Judge whether the serial number n of the characteristic point is less than the total number N of characteristic points. If it is less than N, set n = n + 1 and execute step (4). Otherwise, end the automatic optimization design process and output the final design result; The missile stable control loop structure in step (1) is that the input signal enters the fourth parameter adjustment module. The fourth parameter adjustment module is successively provided with a first comparator, a first parameter adjustment module, a second comparator, a second parameter adjustment module, a third comparator, and a third parameter adjustment module. The signal acts on the rudder system of the missile body. The signals collected by the gyroscope on the missile body are simultaneously fed back to the second comparator and the third comparator, and the signals collected by the accelerometer on the missile body are fed back to the first comparator; In step (1), the parameter adjustment parameters of the first parameter adjustment module, the second parameter adjustment module, the third parameter adjustment module, and the fourth parameter adjustment module are K1, K2, K3, and K4 respectively; In step (3), the form of the fitness function is: J = σ1(M zn - M znzb ) 2 + σ2(P zn - P znzb ) 2 + σ3(M js - M jszb ) 2 + σ4(P js - P jszb ) 2 or J = σ1|M zn -M znzb | + σ2|P zn -P znzb | + σ3|M js -M jszb | + σ4|P js -P jszb | Among them, M zn and P zn are the open-loop amplitude margin and phase margin disconnected at K3, M js and P js are the open-loop amplitude margin and phase margin disconnected at K1, M znzb , P znzb , M jszb , P jszb are the requirements for the frequency-domain performance indicators of the control system, and σ1, σ2, σ3, σ4 are the weighting coefficients.

2. The automatic optimization design method for missile stability control parameters based on genetic algorithm according to claim 1, characterized in that In step (2), the parameters to be optimized are the time constants t1, t2, and the damping coefficient ξ, and the optimization range of the parameters to be optimized is 0.05 ≤ t1 ≤ 0.1, 0.06 ≤ t2 ≤ 0.15, 0.7 ≤ ξ ≤ 0.

9.

3. The automatic optimization design method for missile stability control parameters based on the genetic algorithm according to claim 2, characterized in that The relationship model between the parameters to be optimized t1, t2, and ξ and the parameter adjustment parameters K1, K2, K3, and K4 of the control loop is: Where: a1, a2, a3, a4, and a5 are missile dynamic coefficients, and V d is the missile velocity.

4. The automatic optimization design method for missile stability control parameters based on genetic algorithm according to claim 1, characterized in that In the said step (3), the optimization termination condition is t r <t rzb and σ < σ zb , where t r and σ are respectively the rise time and overshoot of the time-domain response of the control loop, and t rzb and σ zb are respectively the requirements for the time-domain performance indexes of the control loop.

Images