A method and system for intelligent control of aircraft attitude

Abstract

The invention relates to an intelligent control method and system for aircraft attitude, and relates to the field of aircraft attitude control. The method comprises the following steps: obtaining performance parameters and performance design targets of an aircraft attitude control system; determining the amplitude and phase of an initial state of the system according to the performance parameters and optimizing the amplitude and phase thereof, determining a current parameter solution population, a current individual historical optimal solution, and an optimal solution individual of the parameter solution population under a current iteration; judging whether the current parameter solution population meets the performance design target, and if so, retaining the optimal solution individual of the parameter solution population under the current iteration; if not, generating a next generation temporary parameter solution population, a next generation individual historical optimal solution, and an optimal solution individual of the next generation temporary parameter solution population; performing a transformation process on the next generation temporary parameter solution population to obtain a next generation parameter solution population; updating the current parameter solution population, the optimal solution individual of the parameter solution population under the current iteration, and the current individual historical optimal solution, and returning the judgment step, thereby improving the reliability of the attitude control system.

Description

Technical Field

[0001] The present invention relates to the field of aircraft attitude control, and in particular to an aircraft attitude intelligent control method and system. Background Art

[0002] The attitude control system is a system that controls the attitude of the aircraft. Its basic task is to ensure that the aircraft has good stability and operability during flight. In recent years, the aerospace technology of various countries has developed rapidly, and the rapid development of aircraft is also a hot topic among researchers. The parameters of the attitude control system determine whether the flight indicators of the aircraft meet the design requirements during flight. These parameters will affect whether the aircraft can operate stably after launch and accurately enter the predetermined orbit or destination. The parameter design of the aircraft attitude control system takes a lot of time in the process of aircraft development, and it is a difficult problem that must be overcome for the rapid development of aircraft.

[0003] The existing mainstream method is to manually set parameters, which requires a lot of time and manpower. Due to the large number of attitude control system parameters, the solution space generated by the combination of different values ​​of different parameters is extremely large. The search range of possible solutions that can be covered by the manual method is very limited, and it is impossible to guarantee that the set parameter combination can achieve the optimal aircraft control performance.

[0004] The existing PAM-based automatic generation algorithm for attitude control parameters performs well in some simple attitude control parameter generation types. However, when faced with complex optimization types such as "rigid body stability, elastic amplitude stability and elastic phase stability" and "rigid body stability, elastic amplitude stability and sway phase stability", it has not been able to successfully generate attitude control parameters that meet the design requirements. There is still much room for improvement in its reliability, global optimization capability and convergence speed.

[0005] The standard PAM algorithm is difficult to find the optimal solution when applied to complex optimization scenarios in the attitude control system of an aircraft, and is prone to falling into the local optimum. There are two main reasons for falling into the local optimum. The first is that the PAM algorithm will aggregate in the later stages of iteration, resulting in a decrease in population diversity. The second is that only individual experience and the guiding role of the optimal particles are considered in the algorithm iteration update formula, which will also lead to the problem of decreased diversity and premature maturity in the algorithm. Therefore, a method is needed to solve the three major defects of the PAM algorithm in the problem of automatic generation of attitude control system parameters: low reliability, easy to fall into the local optimum, and slow convergence speed. Summary of the invention

[0006] The purpose of the present invention is to provide an aircraft attitude intelligent control method and system to improve the reliability of the attitude control system.

[0007] To achieve the above object, the present invention provides the following solutions:

[0008] An intelligent control method for aircraft attitude, comprising:

[0009] Acquiring performance parameters and performance design targets of an aircraft attitude control system; the performance design targets include target amplitude, target phase, and target frequency; the performance parameters include a transfer function of the aircraft attitude control system, servo mechanism data, and zero-order holder sampling period data;

[0010] Determining the amplitude of the system initial state and the phase of the system initial state according to the performance parameter;

[0011] Optimizing using a correction network according to the amplitude of the initial state of the system and the phase of the initial state of the system to determine the current parameter solution population, the current individual historical optimal solution, and the optimal solution individual of the parameter solution population under the current iteration;

[0012] Determine whether the current parameter solution population meets the performance design target to obtain a first determination result; if the first determination result indicates yes, control the aircraft attitude control system according to the optimal solution individual of the parameter solution population under the current iteration; if the first determination result indicates no, generate the next generation temporary parameter solution population, the next generation individual historical optimal solution and the optimal solution individual of the next generation temporary parameter solution population according to the current parameter solution population and the current individual historical optimal solution by using particle velocity and position update;

[0013] The next generation temporary parameter solution population is sequentially mutated, crossovered and selected to obtain the next generation parameter solution population;

[0014] The current parameter solution population is updated using the next generation parameter solution population, the optimal solution individual of the next generation temporary parameter solution population is updated using the optimal solution individual of the parameter solution population under the current iteration, the current individual historical optimal solution is updated using the next generation individual historical optimal solution and the process returns to the step of "determining whether the current parameter solution population meets the performance design goal to obtain a first judgment result".

[0015] Optionally, the determining whether the current parameter solution population meets the performance design target to obtain a first determination result specifically includes:

[0016] Determine the system amplitude, system phase and system frequency according to the current parameter solution population;

[0017] It is determined whether the system amplitude, the system phase, and the system frequency all meet the performance design target to obtain a first determination result.

[0018] Optionally, the step of updating the generation of the next generation temporary parameter solution population, the next generation individual historical optimal solution and the optimal solution individual of the next generation temporary parameter solution population by using particle speed and position according to the current parameter solution population and the current individual historical optimal solution specifically includes:

[0019] Generate a next generation of temporary parameter solution population by updating particle speed and position according to the current parameter solution population, the current individual historical optimal solution and the optimal solution individual of the parameter solution population under the current iteration;

[0020] The next generation individual historical optimal solution and the optimal solution individual of the next generation temporary parameter solution population are generated according to the next generation temporary parameter solution population and the optimal solution individual of the parameter solution population in the current iteration.

[0021] Optionally, the individual expressions in the next generation temporary parameter solution population are:

[0022]

[0023] in, The i-th individual in the population is the temporary parameter solution for the next generation, is the j-th dimension position of the ith particle at the t-th iteration, is the velocity of the jth dimension of the i-th particle in the next generation of temporary parameter solution population, t is the number of iterations, i is the particle number, and j is the dimension;

[0024] Among them, the expression of the j-dimensional position of the ith particle in the t-th iteration is:

[0025]

[0026] Where w is the inertia weight factor, is the optimal parameter solution of the individual history of the i-th particle, Refers to the optimal solution of the parameter solution population, c1 is the first learning factor, c2 is the second learning factor, r1 is the first random number, r2 is the first random number, is the j-th dimension velocity of the ith particle at the t-th iteration.

[0027] Optionally, the expression of the next generation individual historical optimal solution is:

[0028]

[0029] Among them, pbest i The optimal solution for the next generation of individuals’ history. is the position of the ith particle at the tth iteration, is the fitness of the ith particle in the tth iteration, The fitness of the i-th individual in the population is the temporary parameter solution for the next generation, Solve the i-th individual in the population for the temporary parameters of the next generation;

[0030] The expression of the optimal solution individual of the next generation temporary parameter solution population is:

[0031]

[0032] Among them, gbest is the optimal solution individual of the next generation temporary parameter solution population, and f(gebst) is the fitness of the optimal solution individual of the next generation temporary parameter solution population.

[0033] Optionally, the step of sequentially performing mutation, crossover and selection on the next generation temporary parameter solution population to obtain the next generation parameter solution population specifically includes:

[0034] Performing a mutation operation on the next generation temporary parameter solution population according to the scaling factor to obtain a mutated next generation temporary parameter solution population;

[0035] Performing crossover operation on the next generation temporary parameter solution population and the mutated next generation temporary parameter solution population to obtain the next generation temporary parameter solution population after crossover;

[0036] The next generation parameter solution population is obtained by selecting the next generation parameter solution population based on the next generation temporary parameter solution population and the next generation temporary parameter solution population after the crossover using a fitness function.

[0037] An aircraft attitude intelligent control system, comprising:

[0038] An acquisition module is used to acquire performance parameters and performance design targets of an aircraft attitude control system; the performance design targets include target amplitude, target phase and target frequency; the performance parameters include a transfer function of the aircraft attitude control system, servo mechanism data and zero-order holder sampling period data;

[0039] An initial state determination module, used to determine the amplitude of the system initial state and the phase of the system initial state according to the performance parameter;

[0040] An optimization module, used to optimize using a correction network according to the amplitude of the initial state of the system and the phase of the initial state of the system, and determine the current parameter solution population, the current individual historical optimal solution, and the optimal solution individual of the parameter solution population under the current iteration;

[0041] A judgment module, used to judge whether the current parameter solution population meets the performance design target, and obtain a first judgment result;

[0042] A control module, configured to control the aircraft attitude control system according to the optimal solution individual of the parameter solution population under the current iteration if the first judgment result indicates yes;

[0043] An updating module, configured to update the next generation temporary parameter solution population, the next generation individual historical optimal solution and the optimal solution individual of the next generation temporary parameter solution population according to the current parameter solution population and the current individual historical optimal solution by using particle speed and position if the first judgment result indicates no;

[0044] A mutation, crossover and selection module, used to sequentially perform mutation, crossover and selection on the next generation temporary parameter solution population to obtain the next generation parameter solution population;

[0045] A return module is used to update the current parameter solution population using the next generation parameter solution population, update the optimal solution individual of the parameter solution population under the current iteration using the optimal solution individual of the next generation temporary parameter solution population, update the current individual historical optimal solution using the next generation individual historical optimal solution and return to the step of "determining whether the current parameter solution population meets the performance design goal to obtain a first judgment result".

[0046] Optionally, the judging module specifically includes:

[0047] A system amplitude, system phase and system frequency determination unit, used to determine the system amplitude, system phase and system frequency according to the current parameter solution population;

[0048] The judgment unit is used to judge whether the system amplitude, the system phase and the system frequency all meet the performance design target, and obtain a first judgment result.

[0049] Optionally, the update module specifically includes:

[0050] An updating unit, configured to generate a next generation temporary parameter solution population by updating the particle speed and position according to the current parameter solution population, the current individual historical optimal solution and the optimal solution individual of the parameter solution population under the current iteration;

[0051] A generation module is used to generate the next generation individual historical optimal solution and the optimal solution individual of the next generation temporary parameter solution population according to the next generation temporary parameter solution population and the optimal solution individual of the parameter solution population under the current iteration.

[0052] Optionally, the individual expressions in the next generation temporary parameter solution population are:

[0053]

[0054] in, The i-th individual in the population is the temporary parameter solution for the next generation, is the j-th dimension position of the ith particle at the t-th iteration, is the velocity of the jth dimension of the i-th particle in the next generation of temporary parameter solution population, t is the number of iterations, i is the particle number, and j is the dimension;

[0055] Among them, the expression of the j-dimensional position of the ith particle in the t-th iteration is:

[0056]

[0057] Where w is the inertia weight factor, is the optimal parameter solution of the individual history of the i-th particle, Refers to the optimal solution of the parameter solution population, c1 is the first learning factor, c2 is the second learning factor, r1 is the first random number, r2 is the first random number, is the j-th dimension velocity of the ith particle at the t-th iteration.

[0058] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0059] The present invention obtains performance parameters and performance design targets of an aircraft attitude control system; determines the amplitude of an initial state of the system and the phase of an initial state of the system according to the performance parameters; optimizes using a correction network according to the amplitude of the initial state of the system and the phase of the initial state of the system to determine the current parameter solution population, the current individual historical optimal solution and the optimal solution individual of the parameter solution population under the current iteration; determines whether the current parameter solution population meets the performance design target, and if so, controls the aircraft attitude control system according to the optimal solution individual of the parameter solution population under the current iteration; if not, uses the particle velocity and position according to the current parameter solution population and the current individual historical optimal solution Update and generate the next generation temporary parameter solution population, the next generation individual historical optimal solution and the optimal solution individual of the next generation temporary parameter solution population; perform mutation, crossover and selection on the next generation temporary parameter solution population in turn to obtain the next generation parameter solution population; use the next generation parameter solution population to update the current parameter solution population, use the optimal solution individual of the next generation temporary parameter solution population to update the optimal solution individual of the parameter solution population under the current iteration, use the next generation individual historical optimal solution to update the current individual historical optimal solution and return to the step "determine whether the current parameter solution population meets the performance design target and obtain the first judgment result", thereby improving the reliability of the attitude control system. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0061] Figure 1 A flow chart of the aircraft attitude intelligent control method provided by the present invention;

[0062] Figure 2 A schematic diagram of the flow chart of the intelligent control method for aircraft attitude provided by the present invention;

[0063] Figure 3 Success / failure details graph for PAM and RAM;

[0064] Figure 4 The best optimization ratio details for PAM and RAM;

[0065] Figure 5 The figure shows the convergence speed details of PAM and RAM. DETAILED DESCRIPTION

[0066] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0067] The purpose of the present invention is to provide an aircraft attitude intelligent control method and system to improve the reliability of the attitude control system.

[0068] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0069] like Figure 1 As shown, the present invention provides an aircraft attitude intelligent control method, comprising:

[0070] Step 101: Acquire performance parameters and performance design targets of an aircraft attitude control system; the performance design targets include target amplitude, target phase, and target frequency; the performance parameters include a transfer function of the aircraft attitude control system, servo mechanism data, and zero-order holder sampling period data.

[0071] Step 102: Determine the amplitude of the system initial state and the phase of the system initial state according to the performance parameters.

[0072] Step 103: Optimize using a correction network according to the amplitude of the system initial state and the phase of the system initial state to determine the current parameter solution population, the current individual historical optimal solution, and the optimal solution individual of the parameter solution population under the current iteration.

[0073] Step 104: Determine whether the current parameter solution population meets the performance design target, and obtain a first determination result; if the first determination result indicates yes, execute step 105; if the first determination result indicates no, execute step 106.

[0074] Step 105: Control the aircraft attitude control system according to the optimal solution individual of the parameter solution population in the current iteration.

[0075] Step 106: Generate the next generation temporary parameter solution population, the next generation individual historical optimal solution and the optimal solution individual of the next generation temporary parameter solution population by using particle speed and position update according to the current parameter solution population and the current individual historical optimal solution.

[0076] Step 107: performing mutation, crossover and selection on the next generation temporary parameter solution population in sequence to obtain the next generation parameter solution population.

[0077] Step 108: Use the next generation parameter solution population to update the current parameter solution population, use the optimal solution individual of the next generation temporary parameter solution population to update the optimal solution individual of the parameter solution population under the current iteration, use the next generation individual historical optimal solution to update the current individual historical optimal solution and return to step 104.

[0078] As an optional implementation, step 104 specifically includes:

[0079] Determine the system amplitude, system phase and system frequency according to the current parameter solution population; judge whether the system amplitude, the system phase and the system frequency all meet the performance design target, and obtain a first judgment result.

[0080] As an optional implementation, step 106 specifically includes:

[0081] The next generation of temporary parameter solution population is generated by updating the particle speed and position according to the current parameter solution population, the current individual historical optimal solution and the optimal solution individual of the parameter solution population under the current iteration.

[0082] The next generation individual historical optimal solution and the optimal solution individual of the next generation temporary parameter solution population are generated according to the next generation temporary parameter solution population and the optimal solution individual of the parameter solution population in the current iteration.

[0083] As an optional implementation, the individual expressions in the next generation temporary parameter solution population are:

[0084]

[0085] in, The i-th individual in the population is the temporary parameter solution for the next generation, is the j-th dimension position of the ith particle at the t-th iteration, is the velocity of the jth dimension of the i-th particle in the next generation of temporary parameter solution population, t is the number of iterations, i is the particle number, and j is the dimension.

[0086] Among them, the expression of the j-dimensional position of the ith particle in the t-th iteration is:

[0087]

[0088] Where w is the inertia weight factor, is the optimal parameter solution of the individual history of the i-th particle, Refers to the optimal solution of the parameter solution population, c1 is the first learning factor, c2 is the second learning factor, r1 is the first random number, r2 is the first random number, is the j-th dimension velocity of the ith particle at the t-th iteration.

[0089] As an optional implementation, the expression of the next generation individual historical optimal solution is:

[0090]

[0091] Among them, pbest i The optimal solution for the next generation of individuals’ history. is the position of the ith particle at the tth iteration, is the fitness of the ith particle in the tth iteration, The fitness of the i-th individual in the population is the temporary parameter solution for the next generation, Solve the i-th individual in the population for the temporary parameters of the next generation.

[0092] The expression of the optimal solution individual of the next generation temporary parameter solution population is:

[0093]

[0094] Among them, gbest is the optimal solution individual of the next generation temporary parameter solution population, and f(gebst) is the fitness of the optimal solution individual of the next generation temporary parameter solution population.

[0095] As an optional implementation, step 107 specifically includes:

[0096] A mutation operation is performed on the next generation temporary parameter solution population according to the scaling factor to obtain a mutated next generation temporary parameter solution population.

[0097] A crossover operator is used to perform a crossover on the next generation temporary parameter solution population and the mutated next generation temporary parameter solution population to obtain a next generation temporary parameter solution population after the crossover.

[0098] The next generation parameter solution population is obtained by selecting the next generation parameter solution population based on the next generation temporary parameter solution population and the next generation temporary parameter solution population after the crossover using a fitness function.

[0099] The present invention also provides an aircraft attitude intelligent control system, comprising:

[0100] The acquisition module is used to acquire the performance parameters and performance design targets of the aircraft attitude control system; the performance design targets include target amplitude, target phase and target frequency; the performance parameters include the aircraft attitude control system transfer function, servo mechanism data and zero-order holder sampling period data.

[0101] The initial state determination module is used to determine the amplitude of the system initial state and the phase of the system initial state according to the performance parameters.

[0102] The optimization module is used to optimize the current parameter solution population, the current individual historical optimal solution and the optimal solution individual of the parameter solution population under the current iteration by using the correction network according to the amplitude of the initial state of the system and the phase of the initial state of the system.

[0103] The judgment module is used to judge whether the current parameter solution population meets the performance design target and obtain a first judgment result.

[0104] A control module is used to control the aircraft attitude control system according to the optimal solution individual of the parameter solution population under the current iteration if the first judgment result indicates yes.

[0105] An updating module is used to update the next generation temporary parameter solution population, the next generation individual historical optimal solution and the optimal solution individual of the next generation temporary parameter solution population based on the current parameter solution population and the current individual historical optimal solution by using particle speed and position if the first judgment result indicates no.

[0106] The mutation, crossover and selection module is used to sequentially perform mutation, crossover and selection on the next generation temporary parameter solution population to obtain the next generation parameter solution population.

[0107] A return module is used to update the current parameter solution population using the next generation parameter solution population, update the optimal solution individual of the parameter solution population under the current iteration using the optimal solution individual of the next generation temporary parameter solution population, update the current individual historical optimal solution using the next generation individual historical optimal solution and return to the step of "determining whether the current parameter solution population meets the performance design goal to obtain a first judgment result".

[0108] As an optional implementation manner, the judgment module specifically includes:

[0109] A system amplitude, system phase and system frequency determination unit is used to determine the system amplitude, system phase and system frequency according to the current parameter solution population; a judgment unit is used to judge whether the system amplitude, the system phase and the system frequency all meet the performance design target to obtain a first judgment result.

[0110] As an optional implementation, the update module specifically includes:

[0111] An updating unit is used to generate the next generation temporary parameter solution population based on the current parameter solution population, the current individual historical optimal solution and the optimal solution individual of the parameter solution population under the current iteration using particle speed and position update; a generating module is used to generate the next generation individual historical optimal solution and the optimal solution individual of the next generation temporary parameter solution population based on the next generation temporary parameter solution population and the optimal solution individual of the parameter solution population under the current iteration.

[0112] As an optional implementation, the individual expressions in the next generation temporary parameter solution population are:

[0113]

[0114] in, The i-th individual in the population is the temporary parameter solution for the next generation, is the j-th dimension position of the ith particle at the t-th iteration, is the velocity of the jth dimension of the i-th particle in the next generation of temporary parameter solution population, t is the number of iterations, i is the particle number, and j is the dimension;

[0115] Among them, the expression of the j-dimensional position of the ith particle in the t-th iteration is:

[0116]

[0117] Where w is the inertia weight factor, is the optimal parameter solution of the individual history of the i-th particle, Refers to the optimal solution of the parameter solution population, c1 is the first learning factor, c2 is the second learning factor, r1 is the first random number, r2 is the first random number, is the j-th dimension velocity of the ith particle at the t-th iteration.

[0118] In order to solve the three major defects of the PAM algorithm in the automatic generation of attitude control system parameters, namely low reliability, easy to fall into local optimum and slow convergence speed, the present invention proposes a high-reliability aircraft attitude control system parameter intelligent generation method based on swarm intelligence algorithm - RAM, and then designs the aircraft control system according to the generated system parameters to realize aircraft attitude control.

[0119] The main idea of ​​RAM is to integrate the mutation, crossover and selection operations in the evolutionary algorithm into the algorithm's iteration process, where the mutation operator and crossover operator can increase the diversity of the population. The adaptive mutation operator used in the present invention can generate new individuals when the algorithm is stuck in a local optimum for a long time, thereby enhancing the algorithm's ability to jump out of the local optimum. At the same time, the mutation operator can automatically adjust the scaling factor according to the iteration process, better cope with complex optimization scenarios, and improve the reliability of the algorithm in the problem of automatic generation of attitude control parameters; the selection operator simulates the "survival of the fittest" process in nature. Since the new individuals generated by mutation and crossover may not perform better than the original population individuals, it is necessary to eliminate some individuals with poor performance to ensure that each round of iteration retains individuals with better performance to enter the next round of iteration. This "greedy" selection operation can greatly accelerate the convergence speed of the algorithm. The entire RAM design scheme is as follows: Figure 2 shown.

[0120] (1) The input module input is the file name of the aircraft attitude control system data, such as tf1.trn, tf2.trn, sf.cuv, etc., and the output is the data of the aircraft attitude control system. The system performance parameters include the transfer function G of the aircraft attitude control system 传函 (s), servo mechanism data Sf, zero-order holder sampling period data T, system design requirements including amplitude design requirements C 幅值 , phase design requirements C 相位 , frequency design requirements C 频率 . Among them, the frequency is directly obtained.

[0121] (2) The input of the system initial state amplitude and phase calculation module is the data of the aircraft attitude control system, including G 传函 (s), Sf and T, the output is the amplitude L(ω) and phase of the aircraft attitude control system The amplitude and phase of the system initial state are shown in equations (1) and (2) respectively.

[0122] L(ω)=L 传函 (ω)+L零阶 (ω)+L 伺服 (ω), (1)

[0123] Where L 传函 (ω) is the amplitude of the transfer function, L 零阶 (ω) is the amplitude of the zero-order holder, L 伺服 (ω) is the amplitude of the servo mechanism and L(ω) is the system handling state.

[0124]

[0125] in is the phase of the transfer function, is the phase of the zero-order holder, is the phase of the servo mechanism.

[0126] (3) The initial solution generation module of the correction network does not require input and is a separate module. The output is the parameter solution population. The individual historical optimal solution pbest and the optimal solution individual gbest of the parameter solution population. Assuming the population size is n and the dimension of each parameter solution is D, then the parameter solution population X 0 is an n*D matrix, the parameter solution for the i-th individual in the population is a row vector, and each individual represents a correction network parameter combination. Since this is the first generation population, the individual historical optimal solution pbest is X 0 , the optimal solution individual gbest of the parameter solution population is the individual with the best fitness in the entire population. The parameters in the correction network are the parameters that need to be designed for the aircraft attitude control system. There are two strategies for generating the initial solution: a) If the initial value is not specified, the entire parameter solution population is randomly generated in the solution space, and each dimension satisfies the uniform distribution; b) If the initial value is specified, the entire parameter solution population is generated near the initial value.

[0127] (4) Whether the design requirements are met. The input is the system design requirements C 幅值 , C 相位 , C 频率 and the t-th generation parameter solution population X t , the output is whether the parameter solution meets the design requirements.

[0128] (5) The input of the module generating new parameter solutions is the t-th generation parameter solution population X t , the output is the population of temporary parameter solutions of the t+1th generation

[0129] The module for generating new parameter solutions is mainly responsible for generating new parameter solutions in each iteration of the algorithm, so that the algorithm moves towards the position of the global optimal solution. The main idea is to use the individual historical best position and the overall best parameter solution to determine the position of the new parameter solution. Generate the next generation of temporary parameter solution population X temp The i-th individual Each dimension The method is shown in equations (3) and (4). i Each individual in X corresponds to a correction network. temp It is by X i The temporary population generated is not necessarily the next generation population X i+1 .

[0130]

[0131]

[0132] Where: w is the inertia weight factor, refers to the j-th dimension velocity of the ith particle at the t-th iteration, refers to the j-th dimension position of the ith particle at the t-th iteration, refers to the jth dimension of the historical optimal parameter solution of the i-th particle individual in the t-th iteration, Refers to the j-th dimension of the optimal solution of the population of parameter solutions for the t-th iteration; c1 and c2 are learning factors, respectively, and r1 and r2 are random numbers in [0, 1].

[0133] pbest and gbest are updated according to equations (5) and (6), where f is the fitness function.

[0134]

[0135]

[0136] (6) The input of the parameter solution enhancement module is the next generation temporary parameter solution population X temp , the output is the next generation parameter solution population X t+1 The parameter solution enhancement module is mainly responsible for enhancing the generated new parameter solutions, which includes three sub-modules: mutation, crossover, and selection.

[0137] 1. Mutation module. The input of this module is the next generation of temporary parameter solution population X temp , the output is the next generation of temporary parameter solution population

[0138] In the tth iteration, for individual Generate a temporary variant individual according to formula (7) Solving the population X from temporary parameters tempThree individuals are randomly selected from the population And p1≠p2≠p3≠i. The generated mutation vector is:

[0139]

[0140] in is the difference vector; F is the scaling factor used to control the influence of the difference vector.

[0141] The present invention adopts an adaptive mutation operator, and the scaling factor F is shown in formula (8):

[0142]

[0143] Where T m is the maximum number of iterations, t is the current number of iterations, and F0 is the set parameter.

[0144] (ii) Crossover module. The input of this module is the next generation of temporary parameter solution population X temp and X H , the output is the next generation of temporary parameter solution population

[0145] The introduction of the crossover operator can enhance the diversity of the population. In the tth iteration, for each individual and the intermediate vector it generates Crossover, specifically, is to select temporary variant individuals for each allele (each dimension of the individual) according to a certain probability (Otherwise it is the original temporary individual ) to generate temporary crossover individuals Each dimension of is calculated according to formula (9):

[0146]

[0147] Among them, P cr is the crossover rate.

[0148] (III) Selection module. The input of this module is the next generation of temporary parameter solution population X temp and X M , the output is the next generation of temporary parameter solution population

[0149] This process selects temporary crossover individuals according to the value of the fitness function. and temporary individuals The next generation will be selected from the group with higher fitness. and The n individuals with the highest fitness are selected from a total of 2n individuals, but for each individual of this generation and Each of them is judged once, and the one with the best fitness is retained as the individual at the same position in the next generation, as shown in formula (10), where f is the fitness function.

[0150]

[0151] (1) Output module. The input of this module is the best parameter solution gbest, and the output is the parameter solution in the form of a file named net.trn. gbest represents the best correction network parameter combination in the entire population. (Aircraft attitude control system parameters = correction network parameters = an individual in the population = each row in the matrix, gbest is the individual with the best performance in the population)

[0152] The present invention selects the transmission function data of three typical optimization design types in the automatic optimization design of aircraft attitude control parameters, namely "rigid body stability and elastic amplitude stability", "rigid body stability, elastic amplitude stability and elastic phase stability", and "rigid body stability, elastic amplitude stability and sway phase stability" for experimental verification, and compares the performance of standard PAM and RAM. To simplify the description, let Opt1 represent "rigid body stability and elastic amplitude stability", Opt2 represent "rigid body stability, elastic amplitude stability and elastic phase stability", and Opt2 represent "rigid body stability, elastic amplitude stability and sway phase stability".

[0153] When comparing the performance of PAM and RAM in the problem of automatic generation of attitude control parameters, the present invention uses PAM and RAM to automatically generate 100 experiments for the three attitude control parameter optimization types Opt1, Opt2, and Opt3, respectively. The algorithm iterates 1000 times in each experiment. Finally, the algorithm is evaluated from three aspects: algorithm reliability (the proportion of the number of attitude control parameters generated that meet the design requirements to the total number of experiments), optimal optimization ratio (the multiple of the optimization result to the optimization target), and convergence speed.

[0154] In terms of reliability, the statistical information is shown in Table 1. In the simpler Opt1, both PAM and RAM can successfully generate attitude control parameters that meet the design requirements with a probability greater than 95%. In Opt2, the success rate of PAM is only 3%, while the success rate of RAM is still as high as 97%. In Opt3, PAM did not successfully generate attitude control parameters that meet the design requirements once, while the success rate of RAM is still as high as 96%. Figure 3 As shown, Figure 3 (a) is the reliability comparison diagram of PAM and RAM under the elastic amplitude optimization type. Figure 3 (b) is the reliability comparison diagram of PAM and RAM under the elastic phase optimization type. Figure 3 (c) is a reliability comparison chart of PAM and RAM under the sway phase optimization type.

[0155] Table 1 Comparison of PAM and RAM reliability

[0156]

[0157] The statistical information of the best optimization ratio is shown in Table 2. Except for the standard deviation, which retains one decimal place, the rest of the maximum value, minimum value, mean, and median retain the integer part. The best optimization ratio for each experiment is shown in Table 2. Figure 4 As shown, Figure 4 (a) is the comparison chart of the best optimization ratios of PAM and RAM under the elastic amplitude optimization type. Figure 4 (b) is the comparison chart of the best optimization ratios of PAM and RAM under the elastic phase optimization type. Figure 4 (c) is a comparison chart of the best optimization ratios of PAM and RAM under the sway phase optimization type. In the simpler Opt1, the average value of the best optimization ratio of RAM is 46.5% higher than that of PAM. Although both PAM and RAM can generate attitude control parameters that are much larger than the optimization target at a higher ratio, RAM is better than PAM in terms of maximum, minimum, average, and median. In terms of the standard deviation of the best optimization ratio stability, RAM is also smaller than PAM. This shows that RAM can achieve better optimization results for some simple optimization types based on PAM. In the more complex Opt2, the situation is the same as Opt1, and RAM is better than PAM in all aspects. In Opt3, since the success rate of PAM is 0, all statistical information is 0, but RAM still maintains a high optimization ratio at this time.

[0158] Table 2 Comparison of the best optimization ratios of PAM and RAM

[0159]

[0160]

[0161] Due to the existence of the "premature" phenomenon, it is meaningless to simply consider the convergence speed. In terms of the convergence speed, the present invention will analyze whether the algorithm falls into the local optimum (optimization failure means falling into the local optimum) in each experiment. The statistical information of the convergence speed is shown in Table 3. Except for the standard deviation and the median, which retain one decimal place, the rest of the statistical information retains the integer part. The convergence speed of each experiment is shown in Table 3. Figure 5 As shown, Figure 5 (a) is a comparison of the convergence speed of PAM and RAM under the elastic amplitude optimization type. Figure 5 (b) is a comparison of the convergence speed of PAM and RAM under the elastic phase optimization type. Figure 5(c) is a comparison of the convergence speed of PAM and RAM under the sway phase optimization type. In the simpler Opt1, since the success rates of PAM and RAM are both greater than 95%, only the convergence speed is considered. The convergence speed of RAM 225 times is 48.9% faster than that of PAM 440 times, and Figure 5 It can be clearly seen in (a) that the convergence speed of RAM fluctuates less, which shows that when the success rates of both are high, RAM can significantly shorten the time for attitude control parameter design. Although the average convergence speed of RAM is better than PAM in Opt2 and Opt3, it is obvious that PAM has fallen into the local optimum in the early stage of most experiments, and the "premature" phenomenon has occurred, and the average convergence speed of RAM belongs to the middle and early stages in 1000 iterations, which is within the acceptable range. Even in the three experiments in Opt2 where PAM did not fall into the local optimum, the convergence speed of RAM is obviously slower than that of RAM, which also shows that RAM sacrifices a certain amount of time in exchange for higher reliability (success rate).

[0162] Table 3 Comparison of PAM and RAM convergence speed

[0163]

[0164]

[0165] Compared with PAM, RAM can better meet the characteristics of multiple optimization types in the problem of automatic generation of aircraft attitude control parameters; compared with PAM, RAM has higher reliability in various optimization types of aircraft attitude control parameters, and the success rate of automatically generating parameters that meet design requirements is greater than 95%; compared with PAM, RAM has a stronger ability to jump out of local optimality and perform global optimization in the problem of automatic generation of aircraft attitude control parameters, and converges faster without being "premature".

[0166] like Figure 3 As shown, the process of the implementation scheme of the present invention is as follows:

[0167] 1. Input the transfer function data, servo interpolation data and zero-order holder data required by the model.

[0168] 2. Set the design requirements of the system, including three major categories: amplitude design requirements, phase design requirements, and frequency design requirements. Each type has several design requirements.

[0169] 3. Calculate the amplitude and phase of the system's initial state (without correction network) based on the constructed attitude control system model.

[0170] 4. Generate the population of initial parameter solutions (correction network), which can be generated in two ways: randomly in the solution space and around a given initial value.

[0171] 5. Determine whether the parameter solution meets the design requirements. If not, continue with step 6; if yes, jump to step 12.

[0172] 6. Enter the "Generate New Solution" module and generate new parameter solutions based on formulas 1 and 2.

[0173] 7. Update individual historical best position and global optimal parameter solution.

[0174] 8. Enter the "Parameter Solution Enhancement" module and perform mutation operations on the new parameter solutions generated according to formulas 3 and 4.

[0175] 9. Perform crossover operation on the parameter solution according to formula 5.

[0176] 10. Select the parameter solution according to formula 6.

[0177] 11. Return to step 5.

[0178] 12. Save the best parameter solution (correct the network).

[0179] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0180] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. An intelligent control method for aircraft attitude, characterized in that: include: Obtain the performance parameters and performance design objectives of the aircraft attitude control system; The performance design targets include target amplitude, target phase and target frequency; The performance parameters include the aircraft attitude control system transfer function, servo mechanism data and zero-order holder sampling period data; Determining the amplitude of the system initial state and the phase of the system initial state according to the performance parameter; Optimizing using a correction network according to the amplitude of the initial state of the system and the phase of the initial state of the system to determine the current parameter solution population, the current individual historical optimal solution, and the optimal solution individual of the parameter solution population under the current iteration; Determine whether the current parameter solution population meets the performance design target to obtain a first determination result; if the first determination result indicates yes, control the aircraft attitude control system according to the optimal solution individual of the parameter solution population under the current iteration; if the first determination result indicates no, generate the next generation temporary parameter solution population, the next generation individual historical optimal solution and the optimal solution individual of the next generation temporary parameter solution population according to the current parameter solution population and the current individual historical optimal solution by using particle velocity and position update; The next generation temporary parameter solution population is sequentially mutated, crossovered and selected to obtain the next generation parameter solution population; Using the next generation parameter solution population to update the current parameter solution population, using the optimal solution individual of the next generation temporary parameter solution population to update the optimal solution individual of the parameter solution population under the current iteration, using the next generation individual historical optimal solution to update the current individual historical optimal solution and returning to step "determining whether the current parameter solution population meets the performance design goal to obtain a first determination result"; The step of updating the next generation of temporary parameter solution population, the next generation of individual historical optimal solution and the optimal solution individual of the next generation of temporary parameter solution population by using particle speed and position according to the current parameter solution population and the current individual historical optimal solution specifically includes: Generate a next generation of temporary parameter solution population by updating particle speed and position according to the current parameter solution population, the current individual historical optimal solution and the optimal solution individual of the parameter solution population under the current iteration; Generate the next generation individual historical optimal solution and the optimal solution individual of the next generation temporary parameter solution population according to the next generation temporary parameter solution population and the optimal solution individual of the parameter solution population under the current iteration; The individual expressions in the next generation temporary parameter solution population are: in, The i-th individual in the population is the temporary parameter solution for the next generation, is the j-th dimension position of the ith particle at the t-th iteration, is the velocity of the jth dimension of the i-th particle in the next generation of temporary parameter solution population, t is the number of iterations, i is the particle number, and j is the dimension; in, Where w is the inertia weight factor, is the optimal parameter solution of the individual history of the i-th particle, Refers to the optimal solution of the parameter solution population, c1 is the first learning factor, c2 is the second learning factor, r1 is the first random number, r2 is the first random number, is the j-th dimension velocity of the ith particle at the t-th iteration; The expression of the next generation individual historical optimal solution is: Among them, pbest i The optimal solution for the next generation of individuals’ history. is the position of the ith particle at the tth iteration, is the fitness of the ith particle in the tth iteration, The fitness of the i-th individual in the population is the temporary parameter solution for the next generation, Solve the i-th individual in the population for the temporary parameters of the next generation; The expression of the optimal solution individual of the next generation temporary parameter solution population is: Among them, gbest is the optimal solution individual of the next generation temporary parameter solution population, and f(gebst) is the fitness of the optimal solution individual of the next generation temporary parameter solution population.

2. The method for intelligent control of aircraft attitude according to claim 1, characterized in that: The determining whether the current parameter solution population meets the performance design target to obtain a first determination result specifically includes: Determine the system amplitude, system phase and system frequency according to the current parameter solution population; It is determined whether the system amplitude, the system phase, and the system frequency all meet the performance design target to obtain a first determination result.

3. The method for intelligent control of aircraft attitude according to claim 1, characterized in that: The step of sequentially performing mutation, crossover and selection on the next generation temporary parameter solution population to obtain the next generation parameter solution population specifically includes: Performing a mutation operation on the next generation temporary parameter solution population according to the scaling factor to obtain a mutated next generation temporary parameter solution population; Performing crossover operation on the next generation temporary parameter solution population and the mutated next generation temporary parameter solution population to obtain the next generation temporary parameter solution population after crossover; The next generation parameter solution population is obtained by selecting the next generation parameter solution population based on the next generation temporary parameter solution population and the next generation temporary parameter solution population after the crossover using a fitness function.

4. An intelligent control system for aircraft attitude, characterized in that: include: An acquisition module is used to obtain the performance parameters and performance design objectives of the aircraft attitude control system; The performance design targets include target amplitude, target phase and target frequency; The performance parameters include the aircraft attitude control system transfer function, servo mechanism data and zero-order holder sampling period data; An initial state determination module, used to determine the amplitude of the system initial state and the phase of the system initial state according to the performance parameter; An optimization module, used to optimize using a correction network according to the amplitude of the initial state of the system and the phase of the initial state of the system, and determine the current parameter solution population, the current individual historical optimal solution, and the optimal solution individual of the parameter solution population under the current iteration; A judgment module, used to judge whether the current parameter solution population meets the performance design target, and obtain a first judgment result; A control module, configured to control the aircraft attitude control system according to the optimal solution individual of the parameter solution population under the current iteration if the first judgment result indicates yes; An updating module, configured to update the next generation temporary parameter solution population, the next generation individual historical optimal solution and the optimal solution individual of the next generation temporary parameter solution population according to the current parameter solution population and the current individual historical optimal solution by using particle speed and position if the first judgment result indicates no; A mutation, crossover and selection module, used to sequentially perform mutation, crossover and selection on the next generation temporary parameter solution population to obtain the next generation parameter solution population; A return module is used to update the current parameter solution population using the next generation parameter solution population, update the optimal solution individual of the parameter solution population under the current iteration using the optimal solution individual of the next generation temporary parameter solution population, update the current individual historical optimal solution using the next generation individual historical optimal solution and return to the step of "determining whether the current parameter solution population meets the performance design target to obtain a first determination result"; The update module specifically includes: An updating unit, configured to generate a next generation temporary parameter solution population by updating the particle speed and position according to the current parameter solution population, the current individual historical optimal solution and the optimal solution individual of the parameter solution population under the current iteration; A generating module, used for generating the next generation individual historical optimal solution and the optimal solution individual of the next generation temporary parameter solution population according to the next generation temporary parameter solution population and the optimal solution individual of the parameter solution population under the current iteration; The individual expressions in the next generation temporary parameter solution population are: in, The i-th individual in the population is the temporary parameter solution for the next generation, is the j-th dimension position of the ith particle at the t-th iteration, is the velocity of the jth dimension of the i-th particle in the next generation of temporary parameter solution population, t is the number of iterations, i is the particle number, and j is the dimension; in, Where w is the inertia weight factor, is the optimal parameter solution of the individual history of the i-th particle, Refers to the optimal solution of the parameter solution population, c1 is the first learning factor, c2 is the second learning factor, r1 is the first random number, r 2 is the first random number, is the j-th dimension velocity of the ith particle at the t-th iteration.

5. The aircraft attitude intelligent control system according to claim 4, characterized in that: The judgment module specifically includes: A system amplitude, system phase and system frequency determination unit, used to determine the system amplitude, system phase and system frequency according to the current parameter solution population; The judgment unit is used to judge whether the system amplitude, the system phase and the system frequency all meet the performance design target, and obtain a first judgment result.

Images