CN112037289A - Off-line parameter calibration method based on genetic algorithm - Google Patents
Off-line parameter calibration method based on genetic algorithm Download PDFInfo
- Publication number
- CN112037289A CN112037289A CN202010972259.8A CN202010972259A CN112037289A CN 112037289 A CN112037289 A CN 112037289A CN 202010972259 A CN202010972259 A CN 202010972259A CN 112037289 A CN112037289 A CN 112037289A
- Authority
- CN
- China
- Prior art keywords
- variable
- input
- genetic algorithm
- output
- parameter
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/80—Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S17/00—Systems using the reflection or reradiation of electromagnetic waves other than radio waves, e.g. lidar systems
- G01S17/88—Lidar systems specially adapted for specific applications
- G01S17/89—Lidar systems specially adapted for specific applications for mapping or imaging
-
- 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]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Data Mining & Analysis (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Remote Sensing (AREA)
- Radar, Positioning & Navigation (AREA)
- Electromagnetism (AREA)
- Computer Networks & Wireless Communication (AREA)
- Physiology (AREA)
- Genetics & Genomics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Complex Calculations (AREA)
- Feedback Control In General (AREA)
Abstract
The invention discloses an off-line parameter calibration method based on a genetic algorithm, which belongs to the technical field of instant positioning and map construction and comprises the following steps of determining calibration parameters, selecting proper input and output variables, and establishing a mathematical model of the parameters to be calibrated; acquiring N groups of input and output variable data, and establishing a minimized residual function for representing the difference between the output variable space mapped to the mathematical model and the actual output vector space; and selecting the optimal individual in the input and output variables by using a genetic algorithm, and substituting the optimal individual into the minimized residual error function to solve the parameter to be calibrated. The method is convenient, efficient, high in adaptability and strong in global optimization capability.
Description
Technical Field
The invention belongs to the technical field of instant positioning and map construction, and particularly relates to an offline parameter calibration method based on a genetic algorithm.
Background
Parameter calibration work in the field of SLAM (instant positioning and mapping) is crucial, such as: the accuracy and precision of calibration results of geometric model parameters of camera imaging in the visual SLAM, a relative pose matrix of a laser radar and a chassis center in the laser SLAM and the like directly influence the accuracy and stability of an SLAM algorithm. On the basis of establishing a parameter calibration mathematical model, mathematical methods for solving parameters include a least square method, a gauss-newton method and the like, and the methods can converge to a global optimal solution given a good initial iteration value, but are easy to fall into a local optimal solution given an improper initial value.
The Genetic Algorithm (GA) is a parallel Algorithm and has strong global convergence capability, so the invention provides an off-line parameter calibration method based on the Genetic Algorithm (GA), and the method is particularly suitable for the condition that a calibration parameter mathematical model is complex.
Disclosure of Invention
The invention aims to provide an off-line parameter calibration method based on a genetic algorithm in order to solve the problem that the parameter calibration needs to depend on the setting of an initial value, otherwise, the calibrated parameter variable is easy to fall into a local optimal solution.
The invention realizes the aim through the following technical scheme, and an off-line parameter calibration method based on a genetic algorithm comprises the following steps:
defining calibration parameters, selecting proper input and output variables, and establishing a mathematical model of the parameters to be calibrated;
acquiring N groups of input and output variable data, and establishing a minimized residual function for representing the difference between the output variable space mapped to the mathematical model and the actual output vector space;
selecting the optimal individual in the input and output variables by using a genetic algorithm, and substituting the optimal individual into the minimized residual error function to solve the parameter to be calibrated, wherein the genetic algorithm comprises the following steps:
a. designing an individual coding mode which can be identified by a processor to represent N groups of input and output variable data, and forming an individual code representing each group of variable data into a coding string;
b. taking the coding string as an initial generation population, performing the minimum residual function operation on each parameter variable in the initial generation population, and selecting an optimal individual;
c. and (4) iterating the optimal individuals in the initial generation population, and outputting the optimal individuals after the maximum iteration times is reached.
Preferably, the mathematical model of the parameter to be calibrated is as follows:
[y1,y2,…]＝f(a1x1+a2x2+…)or Y＝f(P·X)
wherein, f represents the functional relationship between the input variable and the output variable, i.e. the mathematical model to be established,' · represents the vector inner product, and X represents the input vector: x ═ X1,x2,…]，xi(i ═ 1,2, …) represents the ith input variable; y ═ Y1,y2,…]，yi(i ═ 1,2, …) denotes the ith output variable, P denotes the scaled parameter vector: p ═ a1,a2,…]，ai(i ═ 1,2, …) denotes the ith calibration parameter.
Preferably, the minimized residual function is:
and S represents that the input variables are mapped to the difference value between the theoretically corresponding output variable space and the actual output vector space through the established mathematical model, and N groups of input and output values are subjected to accumulation operation to calculate the minimum difference value.
Preferably, the genetic algorithm needs to be initialized and the iteration number and population size are set before running.
Preferably, the population iteration includes performing selection operation, crossover operation and variation operation on the optimal individuals in the initial generation population to generate diversified population individuals.
Preferably, the individual coding mode is a binary coding mode, the ith variable ai in the parameter vector P to be calibrated is represented by an M-bit binary number, each variable can be represented by an M-bit binary length, the range of the variable is limited to [ a, b ], and then an M-bit binary number group M is randomly generated, and the value actually representing the variable is
Where D (M) represents the conversion of an 8-bit binary number to a decimal value.
Compared with the prior art, the invention has the beneficial effects that:
the method is convenient, efficient, high in adaptability and strong in global optimization capability.
Drawings
FIG. 1 is a flow chart of a parameter calibration method of the present invention.
FIG. 2 is a flow chart of the genetic algorithm of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
As shown in fig. 1, an off-line parameter calibration method based on genetic algorithm includes the following steps:
s101, defining calibration parameters, selecting proper input and output variables, and establishing a mathematical model of the parameters to be calibrated; the mathematical model of the parameter to be calibrated is as follows:
[y1,y2,…]＝f(a1x1+a2x2+…)or Y＝f(P·X)
wherein, f represents the functional relationship between the input variable and the output variable, i.e. the mathematical model to be established,' · represents the vector inner product, and X represents the input vector: x ═ X1,x2,…]，xi(i ═ 1,2, …) represents the ith input variable; y ═ Y1,y2,…]，yi(i ═ 1,2, …) denotes the ith output variable, P denotes the scaled parameter vector: p ═ a1,a2,…]，ai(i ═ 1,2, …) denotes the ith calibration parameter.
Step S102, collecting N groups of input and output variable data, and establishing a minimized residual error function for representing the difference value between the output variable space mapped to the theoretical corresponding mathematical model and the actual output vector space; the minimized residual function is:
and S represents that the input variables are mapped to the difference value between the theoretically corresponding output variable space and the actual output vector space through the established mathematical model, and N groups of input and output values are subjected to accumulation operation to calculate the minimum difference value.
Step S103, selecting the optimal individual in the input and output variables by using a genetic algorithm, substituting the optimal individual into the minimized residual error function, and solving the parameter to be calibrated, wherein the genetic algorithm comprises the following steps:
a. designing an individual coding mode which can be identified by a processor to represent N groups of input and output variable data, and forming an individual code representing each group of variable data into a coding string;
b. taking the coding string as an initial generation population, performing the minimum residual function operation on each parameter variable in the initial generation population, and selecting an optimal individual;
c. and (4) iterating the optimal individuals in the initial generation population, and outputting the optimal individuals after the maximum iteration times is reached.
The individual coding mode is a binary coding mode, the ith variable ai in the parameter vector P to be calibrated is represented by M-bit binary numbers, each variable can be represented by M-bit binary length, the range of the variable is limited to [ a, b ], and then an M-bit binary number group M is randomly generated, and the value actually representing the variable is
Where D (M) represents the conversion of an 8-bit binary number to a decimal value.
The data acquisition of the input variable and the output variable adopts an online acquisition mode, the optimal solution of the calibration parameters is selected by using an offline genetic algorithm after the data acquisition is finished, the genetic algorithm needs to be initialized and iteration times and population specification modulus are set before operation, and population iteration comprises selection operation, cross operation and variation operation on the optimal individuals in the initial population to generate diversified population individuals. As shown in fig. 2, binary strings of the parameter variables are sequentially and randomly generated through binary coding, and the binary strings of the parameter variables are sequentially combined into a binary string, namely a population individual in the genetic algorithm, so that the population size and the number of individuals are randomly generated and set; decoding the population individuals to obtain values of parameter variables in the parameter vector, and substituting the values into a minimized residual error function in the step S102, wherein the formula can calculate a minimized residual error and also can be called a fitness function to represent the superiority and inferiority of the individuals, and the smaller the value obtained by substituting the decoded values of the population individuals into the formula in the step S102 is, the higher the fitness of the individuals is represented, otherwise, the smaller the fitness is;
step S1, the fitness value of each individual is calculated, the wheel roulette principle is used for selecting population individuals, and good individuals are inherited to the next generation with high probability, and the step belongs to selection operation in genetic algorithm;
step S2, performing cross operation on the populations generated in the selection operation in pairs in sequence to generate new populations, aiming at inheriting excellent individuals to next generation and diversified population individuals with high probability to improve the global optimization searching capability of the algorithm, wherein the step belongs to the cross operation in the genetic algorithm;
s3, carrying out individual variation on population individuals generated in the cross operation with a certain probability in sequence, aiming at diversifying the population individuals and improving the global optimization capability of the algorithm, wherein the step belongs to variation operation in a genetic algorithm;
and step S4, substituting the parameter variables in the population calculated in the variation operation into a minimized residual error function, calculating individual fitness values, selecting the optimal individual, adding 1 to the population iteration frequency, judging whether the maximum iteration frequency is reached, if so, outputting the optimal individual, otherwise, skipping to the step S1 and executing in sequence.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference sign in a claim should not be construed as limiting the claim concerned.
Furthermore, it should be understood that although the present description refers to embodiments, not every embodiment may contain only a single embodiment, and such description is for clarity only, and those skilled in the art should integrate the description, and the embodiments may be combined as appropriate to form other embodiments understood by those skilled in the art.
Claims (6)
1. An off-line parameter calibration method based on a genetic algorithm is characterized by comprising the following steps:
defining calibration parameters, selecting proper input and output variables, and establishing a mathematical model of the parameters to be calibrated;
acquiring N groups of input and output variable data, and establishing a minimized residual function for representing the difference between the output variable space mapped to the mathematical model and the actual output vector space;
selecting the optimal individual in the input and output variables by using a genetic algorithm, and substituting the optimal individual into the minimized residual error function to solve the parameter to be calibrated, wherein the genetic algorithm comprises the following steps:
a. designing an individual coding mode which can be identified by a processor to represent N groups of input and output variable data, and forming an individual code representing each group of variable data into a coding string;
b. taking the coding string as an initial generation population, performing the minimum residual function operation on each parameter variable in the initial generation population, and selecting an optimal individual;
c. and (4) iterating the optimal individuals in the initial generation population, and outputting the optimal individuals after the maximum iteration times is reached.
2. The off-line parameter calibration method based on the genetic algorithm as claimed in claim 1, wherein the mathematical model of the parameter to be calibrated is:
[y1,y2,…]＝f(a1x1+a2x2+…)or Y＝f(P·X)
wherein, f represents the functional relationship between the input variable and the output variable, i.e. the mathematical model to be established,' · represents the vector inner product, and X represents the input vector: x ═ X1,x2,…]，xi(i ═ 1,2, …) represents the ith input variable; y ═ Y1,y2,…]，yi(i ═ 1,2, …) denotes the ith output variable, P denotes the scaled parameter vector: p ═ a1,a2,…]，ai(i ═ 1,2, …) denotes the ith calibration parameter.
3. The off-line parameter calibration method based on genetic algorithm as claimed in claim 1, wherein the minimized residual function is:
and S represents that the input variables are mapped to the difference value between the theoretically corresponding output variable space and the actual output vector space through the established mathematical model, and N groups of input and output values are subjected to accumulation operation to calculate the minimum difference value.
4. The off-line parameter calibration method based on the genetic algorithm as claimed in claim 1, wherein the genetic algorithm needs to be initialized and the number of iterations and population size are set before running.
5. The method of claim 4, wherein the population iteration comprises selection, crossover and mutation operations on optimal individuals in the initial population to generate diversified population individuals.
6. The off-line parameter calibration method based on genetic algorithm as claimed in claim 1, wherein the individual coding mode is binary coding mode, the ith variable ai in the parameter vector P to be calibrated is represented by M-bit binary number, each variable can be represented by M-bit binary length, the range of the variable is limited to [ a, b ], then an M-bit binary number group M is randomly generated, and the value actually representing the variable is
Where D (M) represents the conversion of an 8-bit binary number to a decimal value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010972259.8A CN112037289A (en) | 2020-09-16 | 2020-09-16 | Off-line parameter calibration method based on genetic algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010972259.8A CN112037289A (en) | 2020-09-16 | 2020-09-16 | Off-line parameter calibration method based on genetic algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112037289A true CN112037289A (en) | 2020-12-04 |
Family
ID=73590360
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010972259.8A Pending CN112037289A (en) | 2020-09-16 | 2020-09-16 | Off-line parameter calibration method based on genetic algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112037289A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114966581A (en) * | 2022-07-08 | 2022-08-30 | 南京慧尔视软件科技有限公司 | Radar calibration method, device, equipment and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102337979A (en) * | 2011-08-11 | 2012-02-01 | 浙江大学 | Automatic calibration parameter optimization method of engine based on genetic algorithm |
| CN103871068A (en) * | 2014-03-31 | 2014-06-18 | 河海大学常州校区 | High-precision calibration method based on genetic algorithm |
| CN105628425A (en) * | 2016-01-01 | 2016-06-01 | 三峡大学 | Rotation machinery early stage fault diagnosis method based on heredity annealing optimization multi-core support vector machine |
| CN106679648A (en) * | 2016-12-08 | 2017-05-17 | 东南大学 | Vision-inertia integrated SLAM (Simultaneous Localization and Mapping) method based on genetic algorithm |
-
2020
- 2020-09-16 CN CN202010972259.8A patent/CN112037289A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102337979A (en) * | 2011-08-11 | 2012-02-01 | 浙江大学 | Automatic calibration parameter optimization method of engine based on genetic algorithm |
| CN103871068A (en) * | 2014-03-31 | 2014-06-18 | 河海大学常州校区 | High-precision calibration method based on genetic algorithm |
| CN105628425A (en) * | 2016-01-01 | 2016-06-01 | 三峡大学 | Rotation machinery early stage fault diagnosis method based on heredity annealing optimization multi-core support vector machine |
| CN106679648A (en) * | 2016-12-08 | 2017-05-17 | 东南大学 | Vision-inertia integrated SLAM (Simultaneous Localization and Mapping) method based on genetic algorithm |
Non-Patent Citations (2)
| Title |
|---|
| MOOZ BOURHOOICHA ET AL: "A NON-LINEAR CAMERA CALIBRATION WITH GENETIC ALGORITHMS", 《SEVENTH INTERNATIONAL SYMPOSIUM ON SIGNAL PROCESSING AND ITS APPLICATIONS, 2003. PROCEEDINGS.》 * |
| 刘芬: "《数据挖掘中的核心技术研究》", 31 March 2019 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114966581A (en) * | 2022-07-08 | 2022-08-30 | 南京慧尔视软件科技有限公司 | Radar calibration method, device, equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10931027B2 (en) | Method for array elements arrangement of l-shaped array antenna based on inheritance of acquired character | |
| CN109840154B (en) | Task dependency-based computing migration method in mobile cloud environment | |
| CN110175168B (en) | Time sequence data filling method and system based on generation of countermeasure network | |
| Fang et al. | Deep reinforcement learning for scenario-based robust economic dispatch strategy in internet of energy | |
| CN112700060B (en) | Station terminal load prediction method and prediction device | |
| CN112015719A (en) | Regularization and adaptive genetic algorithm-based hydrological prediction model construction method | |
| CN109991950A (en) | The balance ameliorative way of cooperation robotic asssembly production line based on genetic algorithm | |
| CN112037289A (en) | Off-line parameter calibration method based on genetic algorithm | |
| CN115345297A (en) | Platform area sample generation method and system based on generation countermeasure network | |
| CN110222816B (en) | Deep learning model establishing method, image processing method and device | |
| CN112163808A (en) | Method for solving logistics center addressing problem by self-adaptive whale algorithm based on opponent learning | |
| CN111612160A (en) | Incremental Bayesian network learning method based on particle swarm optimization algorithm | |
| CN116992779A (en) | Simulation method and system of photovoltaic energy storage system based on digital twin model | |
| CN107944045B (en) | Image search method and system based on t distribution Hash | |
| CN116562113A (en) | Power consumption assessment method, system and medium based on clustering algorithm | |
| CN112531725B (en) | Method and system for identifying parameters of static var generator | |
| Wang et al. | Research on the prediction model of greenhouse temperature based on fuzzy neural network optimized by genetic algorithm | |
| CN112183843B (en) | Load optimization distribution method for thermal power plant based on hybrid intelligent algorithm | |
| CN114839940A (en) | Cross-domain collaborative workshop dynamic scheduling method based on balance index adaptive evolution | |
| CN114202063A (en) | Fuzzy neural network greenhouse temperature prediction method based on genetic algorithm optimization | |
| CN114336793B (en) | Method for determining flexibility of alternating current/direct current hybrid power distribution network | |
| CN110930172A (en) | Neural network-based cross-country market joint marginal price prediction method and system | |
| CN115423211A (en) | Method, device, equipment and medium for predicting block random power | |
| CN116345495B (en) | Power plant unit frequency modulation optimization method based on data analysis and modeling | |
| CN114841472B (en) | GWO optimization Elman power load prediction method based on DNA hairpin variation |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20201204 |