CN110704948A - Design method of intelligent controller of unmanned aerial vehicle - Google Patents
Design method of intelligent controller of unmanned aerial vehicle Download PDFInfo
- Publication number
- CN110704948A CN110704948A CN201910906869.5A CN201910906869A CN110704948A CN 110704948 A CN110704948 A CN 110704948A CN 201910906869 A CN201910906869 A CN 201910906869A CN 110704948 A CN110704948 A CN 110704948A
- Authority
- CN
- China
- Prior art keywords
- particle
- algorithm
- intelligent controller
- unmanned aerial
- design method
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
Abstract
The invention discloses a design method of an intelligent controller of an unmanned aerial vehicle, which adopts an improved particle swarm algorithm, namely an improved algorithm of inertia weight and learning factor in a self-adaptive transformation algorithm is designed aiming at a standard particle swarm algorithm.
Description
Technical Field
The invention belongs to the field of artificial intelligence control, and particularly relates to a design method of an intelligent controller of an unmanned aerial vehicle.
Background
With the development of control theory, more and more control theory is widely applied to the flight control of unmanned aerial vehicles, Parry and Muray-Smith, et al, firstly utilize the EA method to research the flight control system, Samblancat, et al, configure the characteristic structure method and H￥The method is combined, firstly, a characteristic structure configuration method is adopted to design the inner ring of the controller, so that the system is decoupled for the longitudinal channel, the transverse channel and the course of the aircraft, and then H is adopted for a single channel∞The method designs the ACAH system. Hess and Gorder apply Quantitative Feedback Theory (QFT) to longitudinal flight control of AH-64, and the like.
With the development of intelligent technology, many intelligent algorithms play an important role in the flight control of an aircraft, such as Genetic Algorithm (GA), particle swarm algorithm (PSO), ant colony Algorithm (ACO), neural network, and the like. H.Boubertakh and S.Labiod and the like design a flight controller by a method of combining a PSO algorithm and fuzzy PI control, and because a plurality of parameters do not have certain design standards in the design process of a flight control system, the flight controller can be optimized by using an intelligent algorithm, so that the workload of a designer can be reduced, and the stability of the designed system is more reliable.
In recent years, advanced modern control theory is applied to a flight control system, foreign research is more active, a plurality of good effects are obtained, the flight control research in China starts late, and the gap between the research and the foreign research is large in the aspect.
The particle swarm optimization is a swarm intelligent optimization algorithm except for ant colony algorithm and fish colony algorithm in the field of computers. This algorithm was first proposed by Kennedy and Eberhart in 1995. Many scientific advances of human beings can not leave the help of the nature, and the shadow of the nature can be seen from Newton's law of motion to the plane soaring the sky. The particle swarm algorithm is also adopted, researchers inspire the predation behavior of birds, research the predation strategy of the birds, find that the most effective strategy for the birds to find food is to search the area of the birds nearest to the food in the current state, and research and expansion are carried out on the area, so that the prototype of the particle swarm algorithm is formed. The particle swarm optimization algorithm has a simple structure, does not need gradient information, is easy to realize, can flexibly adjust global and local search capabilities by adjusting inertial weight and learning factors, and is one of the hot spots of random search at present. Although the PSO has a fast search speed in the early stage, the PSO is still easy to be in local optimum in the later stage, and the PSO is also one of the improved hotspots of the particle swarm optimization.
Each particle in the PSO algorithm is equivalent to a bird in a bird group, and represents a potential solution of the problem, and a fitness function, i.e., an objective function, in the PSO algorithm can calculate the fitness of the particle to evaluate the quality of the particle. In order to find the optimal solution in the population, each particle has a velocity vector, and the velocity vectors of the particles are continuously changed through continuous learning, so that the optimization of an individual in a feasible solution space is realized.
Disclosure of Invention
The invention aims to provide an improved particle swarm optimization algorithm for optimally designing an aircraft controller aiming at the defects of the existing aircraft controller.
The purpose of the invention is realized by the following technical scheme: a design method for an intelligent controller of an unmanned aerial vehicle comprises the following steps:
the selection of the inertia weight and the learning factor in the basic particle swarm algorithm is researched, the self-adaptive transformation inertia weight and the learning factor are improved, and the Schwefel function is used for verification;
in the design of an aircraft display model tracking control system, an integral matrix and a forward gain matrix in the display model tracking control system are optimized by adopting a particle swarm algorithm with the minimum tracking error as a target;
and (4) building a Simulnk simulation model, obtaining the value of the optimization item in the display model, and verifying the robustness of the model.
Preferably, the inertial weight of the improved particle swarm algorithm is proportional to the fitness of the algorithm.
Preferably, the fitness of the algorithm reflects the degree of goodness and badness of a feasible solution.
Preferably, the learning factor of the improved particle swarm optimization is c1And c2Wherein c is1Representing the ability of the particle to learn from its own experience, c2Representing the ability of the particle to learn from the population.
Preferably, the learning factor of the particle swarm optimization should be reduced appropriately when the fitness of a particle is not good enough for the minimum problem, and the pair c corresponding to the particle is not good enough1Increase its corresponding c2When the fitness of a particle is goodC corresponding to the particle should be increased appropriately1Decrease its corresponding c2。
Preferably, the integral matrix and the forward gain matrix in the explicit model tracking control system are optimized by using an improved particle swarm optimization algorithm.
Compared with the prior art, the invention has the following beneficial effects:
(1) the improved particle swarm algorithm is adopted, so that the defect that the standard particle swarm algorithm is easy to fall into a local minimum value is avoided;
(2) by adopting the improved particle swarm algorithm, the parameter matrix in the aircraft controller is optimized, the interference of human factors is avoided, and the design efficiency is greatly improved.
Drawings
FIG. 1 is a flow chart of a method for designing an intelligent controller for an unmanned aerial vehicle according to the present invention;
FIG. 2 is a graph of the iterative variation of a standard particle swarm algorithm;
FIG. 3 is a modified particle swarm algorithm iteration variation;
FIG. 4 is an unmanned aerial vehicle display model;
FIG. 5 is a Simulink simulation model of the unmanned your aircraft display model;
FIG. 6 is a graph of robust performance verification in an embodiment of the invention;
fig. 7 is a graph of robust performance verification in an embodiment of the invention.
Detailed Description
As shown in fig. 1 to 6, the invention discloses a design method of an intelligent controller of an unmanned aerial vehicle, which comprises the following steps:
1. a learning factor for improving the particle swarm optimization, c1And c2The learning ability of the particle to own experience and the learning ability of the particle to the population are represented respectively. For the minimum problem, when the fitness of a particle is not good enough, the pair c corresponding to the particle should be reduced appropriately1Increase its corresponding c2When the fitness of a particle is better, c corresponding to the particle should be increased appropriately1Decrease its corresponding c2. Thus, pair c1、c2The following modifications are carried out:
wherein f isminIs the minimum value of the fitness of the particle in a certain iteration, fitnessiIs the fitness of the ith particle.
2. Improving the size of the inertial weight in the particle swarm algorithm embodies the fact that the particles inherit the previous speed and capacity, larger weight is beneficial to global search, smaller weight is beneficial to local search, and in order to balance the global search capacity and the local search capacity in the particle swarm algorithm, the inertial weight is redefined aiming at the problem of the minimum value, so that the redefined inertial weight is in direct proportion to the fitness, and the reconstruction is as follows:
wherein: w is amaxMaximum inertial weight, w, is 0.9min0.4 is the minimum inertial weight, fmaxIs the maximum value of the fitness of the particles in a certain iteration process.
3. The Schwefel function is used for testing the algorithm, and the search space of the particles is-500 ≦ xi≤500，
The population size is 20, the iteration times are 300, and the test results are shown in figure 1 (standard particle swarm iteration change diagram) and figure 2 (improved particle swarm algorithm iteration change diagram).
4. And designing a display model of the unmanned aerial vehicle (figure 4). In the tracking control of the display model, an operation command is transmitted to the display model, the output of the display model is used as the input variable quantity, and the variable quantity is adjusted in real time according to a loop in the model, so that the control result required by the display model is achieved.
5. Integration constant matrix G in pair display model by using standard particle swarm algorithm4Optimally designing with a forward gain diagonal matrix R, and in the common display model control, performing G4The design with R is often artificial, which not only results in low design efficiencyAnd the designed system has no persuasion due to different matrixes designed by different designers, and the two matrixes are optimized by adopting the particle swarm optimization algorithm, so that the design of the whole display model system is more scientific and reasonable.
6、G4The array is an integral constant array, a diagonal array, G4The elements on the diagonal line represent the integral constants of 4 channels, which respectively correspond to a longitudinal channel integral constant, a transverse channel integral constant, a course channel integral constant and a total distance channel integral constant, and the 4 integral constants have the function of improving the tracking steady-state error in one beat time. When the values of the diagonal elements are different, the step response of each channel is also different, G4The form of the matrix is as follows:
7. r is a diagonal matrix, the elements on the diagonal correspond to the gain of each channel, when the diagonal elements take different values, the step response of each channel is also different, and the form of the R matrix is as follows:
8. the designed optimized fitness function is as follows:
fmin＝w1|q-C11|+w2|f-C22|+w3|r-C33|+w4|w-C44|
wherein q, f, r, w are the four outputs of a mechanical aircraft control System (MFCS), C11、C22、C33、C44Respectively corresponding sensitivity coefficients, w1、w2、w3、w4Taking w in the design herein as the weight occupied by the error of four outputs1＝w1＝w1＝w1The weights can be adjusted later as needed, 1.
9. Constructing simulink simulation model (figure 5) to obtainG4And the value of R. simulink is a conventional technology, and is a common technology in laboratories of colleges and universities, and is not described herein.
10. And (3) robustness verification:
in order to verify the robustness of the designed system, the robustness of the system is analyzed in two cases:
① when the system parameter floats-10%, the state matrix A in the state space model1Control array B0.9A1＝0.9B
② when the system parameter is floating by + 10%, i.e. the state matrix A in the state space model1Control array B1.1A1＝1.1B
As shown in fig. 6, when A1＝0.9A，B1When equal to 0.9B, the four channels output Δ We＝1，ΔWYao (a)＝1，ΔWr＝1，ΔWcRobust performance verification when 1.
When A is shown in FIG. 71＝1.1A，B1When 1.1B, the four channels output Δ We＝1，ΔWYao (a)＝1，ΔWr＝1，ΔWcRobust performance verification when 1.
The robustness is mainly to detect the stability of the system when the external environment is disturbed, A1＝0.9A B10.9B and A1＝1.1A,B1The two cases 1.1B represent disturbance caused by the external environment to the system, and the disturbance of the external environment does not have a great influence on the system according to the output graph, so that the system has better robustness, that is, the system is stable.
Claims (6)
1. A design method for an intelligent controller of an unmanned aerial vehicle is characterized by comprising the following steps:
selecting inertia weight and learning factor in a basic particle swarm algorithm to be researched, improving self-adaptive transformation inertia weight and learning factor, and verifying by using a Schwefel function;
in the design of an aircraft display model tracking control system, an integral matrix and a forward gain matrix in the display model tracking control system are optimized by adopting a particle swarm algorithm with the minimum tracking error as a target;
and (4) building a Simulnk simulation model, obtaining the value of the optimization item in the display model, and verifying the robustness of the model.
2. The design method of the intelligent controller of the unmanned aerial vehicle as claimed in claim 1, wherein the inertial weight of the improved particle swarm algorithm is proportional to the fitness of the algorithm.
3. The design method of the intelligent controller of the unmanned aerial vehicle as claimed in claim 1, wherein the fitness of the improved particle swarm algorithm is a degree reflecting the goodness and badness of a feasible solution.
4. The design method of an UAV intelligent controller according to claim 1, wherein the learning factor of the improved PSO algorithm is c1And c2Wherein c is1Representing the ability of the particle to learn from its own experience, c2Representing the ability of the particle to learn from the population.
5. The design method of unmanned aerial vehicle intelligent controller according to claim 4, wherein the learning factor of the improved particle swarm optimization algorithm is to appropriately reduce the corresponding pair c of a certain particle when the fitness of the particle is not good enough for the problem of minimum value1Increase its corresponding c2(ii) a When the fitness of a particle is better, c corresponding to the particle should be increased appropriately1Decrease its corresponding c2。
6. The design method of the intelligent controller of the unmanned aerial vehicle as claimed in claim 1, wherein the integration matrix and the forward gain matrix in the explicit model tracking control system are optimized by using an improved particle swarm optimization.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910906869.5A CN110704948A (en) | 2019-09-24 | 2019-09-24 | Design method of intelligent controller of unmanned aerial vehicle |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910906869.5A CN110704948A (en) | 2019-09-24 | 2019-09-24 | Design method of intelligent controller of unmanned aerial vehicle |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110704948A true CN110704948A (en) | 2020-01-17 |
Family
ID=69194811
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910906869.5A Pending CN110704948A (en) | 2019-09-24 | 2019-09-24 | Design method of intelligent controller of unmanned aerial vehicle |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110704948A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111711396A (en) * | 2020-04-13 | 2020-09-25 | 山东科技大学 | Method for setting control parameters of speed ring of permanent magnet synchronous motor based on fractional order sliding mode controller |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102507131A (en) * | 2011-09-26 | 2012-06-20 | 中国航空工业第六一八研究所 | Method for acquiring unknown pneumatic parameters in UAV (Unmanned Aerial Vehicle) mathematical model |
| CN102663514A (en) * | 2012-03-22 | 2012-09-12 | 长沙理工大学 | Constrained optimization evolution algorithm based on feasible equilibrium mechanism |
| CN106597851A (en) * | 2016-12-15 | 2017-04-26 | 南京航空航天大学 | Robust fault-tolerant control method for small unmanned aerial vehicle flight control system |
| CN109947124A (en) * | 2019-04-25 | 2019-06-28 | 南京航空航天大学 | Improve particle swarm algorithm Optimization of Fuzzy PID unmanned helicopter attitude control method |
-
2019
- 2019-09-24 CN CN201910906869.5A patent/CN110704948A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102507131A (en) * | 2011-09-26 | 2012-06-20 | 中国航空工业第六一八研究所 | Method for acquiring unknown pneumatic parameters in UAV (Unmanned Aerial Vehicle) mathematical model |
| CN102663514A (en) * | 2012-03-22 | 2012-09-12 | 长沙理工大学 | Constrained optimization evolution algorithm based on feasible equilibrium mechanism |
| CN106597851A (en) * | 2016-12-15 | 2017-04-26 | 南京航空航天大学 | Robust fault-tolerant control method for small unmanned aerial vehicle flight control system |
| CN109947124A (en) * | 2019-04-25 | 2019-06-28 | 南京航空航天大学 | Improve particle swarm algorithm Optimization of Fuzzy PID unmanned helicopter attitude control method |
Non-Patent Citations (1)
| Title |
|---|
| 荣鹤: "基于粒子群算法的直升机控制器设计", 《中国优秀博硕士学位论文全文数据库(硕士) 工程科技Ⅱ辑》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111711396A (en) * | 2020-04-13 | 2020-09-25 | 山东科技大学 | Method for setting control parameters of speed ring of permanent magnet synchronous motor based on fractional order sliding mode controller |
| CN111711396B (en) * | 2020-04-13 | 2022-04-01 | 山东科技大学 | Method for setting motor speed loop control parameters based on fractional order sliding mode controller |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Song et al. | Coordinated optimization on energy capture and torque fluctuation of wind turbines via variable weight NMPC with fuzzy regulator | |
| CN111580564B (en) | Parallel multi-target search cooperative operation method for three-dimensional group unmanned aerial vehicle | |
| CN110110380B (en) | Piezoelectric actuator hysteresis nonlinear modeling method and application | |
| CN110264012A (en) | Renewable energy power combination prediction technique and system based on empirical mode decomposition | |
| Yang et al. | An improved evolution fruit fly optimization algorithm and its application | |
| Gao et al. | Consensus evaluation method of multi-ground-target threat for unmanned aerial vehicle swarm based on heterogeneous group decision making | |
| Liu et al. | Task offloading optimization of cruising UAV with fixed trajectory | |
| Xu et al. | Competitive search algorithm: a new method for stochastic optimization | |
| CN110704948A (en) | Design method of intelligent controller of unmanned aerial vehicle | |
| Lu et al. | Radial basis function neural network based on an improved exponential decreasing inertia weight-particle swarm optimization algorithm for AQI prediction | |
| Gou et al. | A multi-strategy improved particle swarm optimization algorithm and its application to identifying uncorrelated multi-source load in the frequency domain | |
| CN113435595A (en) | Two-stage optimization method for extreme learning machine network parameters based on natural evolution strategy | |
| Yang et al. | Accuracy analysis of sports performance prediction based on BP neural network intelligent algorithm | |
| CN115758866A (en) | Transmission performance optimization method for wireless power transmission system of electric vehicle | |
| CN111552317B (en) | Multi-spacecraft four-dimensional collaborative track determination method | |
| CN114995106A (en) | PID self-tuning method, device and equipment based on improved wavelet neural network | |
| CN110365020B (en) | Reactive power optimization method based on ensemble learning | |
| CN114372418A (en) | Wind power space-time situation description model establishing method | |
| CN112698666A (en) | Aircraft route optimization method based on meteorological grid | |
| CN114021847A (en) | Short-term power load prediction method and system based on ISGU (inverse transformation and aggregation) hybrid model | |
| Li et al. | Sparse broad learning system via a novel competitive swarm optimizer | |
| CN112329995A (en) | Optimal scheduling method and device for distributed energy storage cluster and computer equipment | |
| Zhu et al. | Temperature Control Optimization for Heat Pipe Based on Particle Swarm Optimization | |
| Liu et al. | Adaptive multi-group fruit fly optimization algorithm | |
| Ying et al. | An enhanced hybrid quadratic particle swarm optimization |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20200117 |
|
| WD01 | Invention patent application deemed withdrawn after publication |