CN107798428A - A kind of locomotive automatic Pilot control forecasting molding machine learning method - Google Patents
A kind of locomotive automatic Pilot control forecasting molding machine learning method Download PDFInfo
- Publication number
- CN107798428A CN107798428A CN201710990981.2A CN201710990981A CN107798428A CN 107798428 A CN107798428 A CN 107798428A CN 201710990981 A CN201710990981 A CN 201710990981A CN 107798428 A CN107798428 A CN 107798428A
- Authority
- CN
- China
- Prior art keywords
- locomotive
- model
- mrow
- population
- state
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- 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)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Business, Economics & Management (AREA)
- Biophysics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Mathematical Optimization (AREA)
- Human Resources & Organizations (AREA)
- Pure & Applied Mathematics (AREA)
- Strategic Management (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Economics (AREA)
- Operations Research (AREA)
- Mathematical Analysis (AREA)
- Computational Mathematics (AREA)
- Probability & Statistics with Applications (AREA)
- Tourism & Hospitality (AREA)
- Biomedical Technology (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Genetics & Genomics (AREA)
- Feedback Control In General (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- Game Theory and Decision Science (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Quality & Reliability (AREA)
Abstract
The present invention relates to a kind of locomotive automatic Pilot control forecasting molding machine learning method, it includes:Step S10, mechanism model recurrence learning is carried out for locomotive, obtains basic mechanism model;Step S20, according to the influence factor during locomotive operation, the study of model is iterated by Symbolic Regression mode, the iterative model after being restrained;Step S30, State Forecasting Model is collectively constituted by basic mechanism model and iterative model;Step S40, differentiates whether the error of State Forecasting Model meets to minimize, if satisfied, then using the currently available State Forecasting Model as preferable dbjective state forecast model;If not satisfied, then return to step S20.Bonding machine car mechanism model of the present invention, learn to be applied to embedded system and the high State Forecasting Model of precision of prediction by the way of machine learning, obtain the dbjective state forecast model of locomotive automatic Pilot control;By the dbjective state forecast model, make precision of prediction higher.
Description
Technical field
The present invention relates to railway traffic control field, especially one kind to be related to locomotive automatic Pilot control forecasting molding machine
Learning method.
Background technology
Locomotive automatic Pilot optimal control is one of Core Feature of locomotive automatic control system, and its main function is basis
Optimal velocity curvilinear path, corresponding control input amount is calculated based on actual motion state and with reference to model- following control algorithm,
And the control input amount is acted on to the actual mechanical process of locomotive completion speed control.
Locomotive automated driving system need to realize on schedule, the index such as comfortable and energy-saving run.The optimization of optimal velocity curve
Generation be locomotive automated driving system meet on schedule, the guarantee of the index such as comfortable and energy-conservation, and carry out locomotive driving control
Foundation.But in actual moving process, because the influence of various external factor, the actual motion track of locomotive is more difficult bent with ideal
Line overlaps, therefore the main target of optimal velocity curve model- following control algorithm is to reduce machine as far as possible in locomotive automated driving system
The error of car actual motion rate curve and optimal velocity curve, to ensure that locomotive can be completed to run according to optimal velocity curve
Task, thus locomotive automated driving system control algolithm be ensure locomotive realize on schedule, steady and energy-saving run key technology
One of.
Predictive control algorithm is one of conventional System design based on model algorithm, and existing predictive control algorithm need to rely on
The forecast model of process dynamics behavior is described.Current classical forecast model is mainly in terms of the PREDICTIVE CONTROL of processing linear system
There is good performance.But for the nonlinear system of general significance, when being predicted control using approximate linear model,
It is difficult to the target for reaching optimal control because deviation is big, then occurs some and studied for nonlinear prediction method.
Include the PREDICTIVE CONTROL based on mechanism model currently for nonlinear prediction method research, but mechanism model is present
Partial parameters, which need to test, to be determined.But some factors in locomotive actual moving process be present for the dry of forecast result of model
Disturb, easily cause the error increase of the locomotive actual motion rate curve of prediction, therefore existing rely solely on the pre- of mechanism model
Surveying control algolithm can not Accurate Prediction locomotive actual motion state.
The content of the invention
The purpose of the present invention is the problem of presence for prior art, there is provided a kind of locomotive automatic Pilot control forecasting model
Machine learning method, it being capable of Accurate Prediction locomotive virtual condition.
The purpose of the present invention is achieved through the following technical solutions:
The present invention provides a kind of locomotive automatic Pilot control forecasting molding machine learning method, and it includes:
Step S10, mechanism model recurrence learning is carried out for locomotive, obtains basic mechanism model;
Step S20, according to influence factor in locomotive actual moving process, model is iterated by way of Symbolic Regression
Study, the iterative model after being restrained;
Step S30, State Forecasting Model is collectively constituted by basic mechanism model and iterative model;
Step S40, function is selected to differentiate whether the error of State Forecasting Model meets to minimize by equation, if satisfied,
Then perform step S50, will the currently available State Forecasting Model as preferable dbjective state forecast model;If it is unsatisfactory for
Condition, then continue return to step S20.
It is highly preferred that the step S10 includes:
Step 1, using locomotive as research object, by the kinetic model of locomotive, determine locomotive mechanism model;It is described
Locomotive mechanism model, represent as follows:
In above-mentioned formula, C represents that locomotive is made a concerted effort in track direction of advance;M represents the weight of locomotive;V works as locomotive
The preceding speed of service;T is the current run time of locomotive;The acceleration of as current locomotive;S is that the position of locomotive is public affairs
In mark;F (s) represents the tractive force of current locomotive position;W is locomotive operation resistance;B is right caused by locomotive brake device
Locomotive plays the brake drag acted on backward;
Step 2, according to identified locomotive mechanism model, the recurrence learning of mechanism model is carried out, obtains basic mechanism mould
Type.
It is highly preferred that the step 2 includes:
Using polynomial regression method, mechanism model recurrence learning is carried out to locomotive mechanism model, obtains error sum of squares
Optimization object function when minimum, and by mechanism model based on the optimization object function.
It is highly preferred that the step S20 includes:
Step S201, function variable is determined according to locomotive state information, and associative function collection determines the initial of iterative model
Population;
Step S202, calculate the fitness value of population;
State abstraction is carried out according to the input of train information, driver information and line information, kind is calculated using equation below
The fitness (degree of approximation of population approaching to reality solution) of group:
In formula, i represents i-th of population, f (xj) j-th of individual calculated value in i-th of population is represented, g (x) is represented
The desired value of j-th of individual in i-th of population;FiRepresent the fitness value of j-th of individual in i-th of population;
Step S203, judge colony fitness value whether match state jump condition, if satisfied, then performing step
S204;If not satisfied, then performing step S206, i.e., variation is performed with the probability of setting on the basis of current population, inserts, move
Move, genetic recombination operator, produce new population, jump to step S202;
Step S204, terminate computing, export corresponding expression formula, obtain scheduling actions optimal under current state, then
It is transferred to step S205;
Step S205, selection control action is performed according to optimal scheduling actions under current state, and by the machine after change
Car status information feeds back to iterative model with function variables such as selection control actions.
The present invention has the following technical effect that it can be seen from the technical scheme of the invention described above:
Bonding machine car mechanism model of the present invention, learn to be applied to embedded system and prediction essence by the way of machine learning
High State Forecasting Model is spent, obtains the dbjective state forecast model of locomotive automatic Pilot control;Predicted by the dbjective state
Model so that the precision of prediction of prediction algorithm is higher.
Brief description of the drawings
Fig. 1 is the implementation Organization Chart of the present invention;
Fig. 2 is the implementing procedure figure of the present invention.
Embodiment
Technical scheme is described in further details below with reference to accompanying drawing.
Embodiment one
The present invention provides a kind of locomotive automatic Pilot control forecasting molding machine learning method, and it is based on mechanism model, knot
Experimental data is closed, the method that symbolization returns is predicted the study of model.It is implemented as depicted in figs. 1 and 2, including as follows
Step:
Step S10, mechanism model recurrence learning is carried out for locomotive, obtains basic mechanism model.
It is one of mode of reduction model error to carry out mechanism model recurrence learning for locomotive, therefore first in the present embodiment
First mechanism model recurrence learning is carried out for research object obtain basic mechanism model.
Step S10 implementation procedure is specific as follows:
First, locomotive mechanism model is determined.
Using a certain locomotive as research object, by the physical principle and calculation formula of the kinetic model for analyzing locomotive,
Locomotive mechanism model is determined, it is specific as follows:
Locomotive in the process of moving can be by different directions and the active force of size, and stressing conditions are complicated, but in usual feelings
Active force of the locomotive along track direction of advance is only considered under condition, thus may determine that locomotive mechanism model is according to locomotive
Stressing conditions and the motor sport equation that determines, represent as follows:
In above-mentioned formula, C represents that locomotive is made a concerted effort in track direction of advance;M represents the weight of locomotive;V works as locomotive
The preceding speed of service;T is the current run time of locomotive;The acceleration of as current locomotive;S is that the position of locomotive is public affairs
In mark;F (s) represents the tractive force of current locomotive position;W is locomotive operation resistance;B is right caused by locomotive brake device
Locomotive plays the brake drag acted on backward.
Above-mentioned locomotive operation resistance W includes two parts:A part is the datum drag of locomotive, is mainly being run by locomotive
Resistance caused by the friction of journey middle (center) bearing, wheel rolling etc.;When another part is that locomotive is run in circuit slope section, curve, tunnel etc.
Caused additional drag.Therefore locomotive operation resistance W can be calculated by equation below:
W=Rb(v)+Rl(s) (2)
Wherein,
In above-mentioned formula, v is the current speed of service of locomotive;Rb(v) it is the datum drag of locomotive;S is the position of locomotive
Put i.e. kilometer post;Rl(s) additional drag caused by being run for current locomotive in position.
For the datum drag R of locomotiveb(v), according to urban track traffic locomotive traction specification, carried out by below equation
Calculate:
Rb(v)=m (a1+a2v+a3v2) (4)
In above formula, v is the speed of service of locomotive;M represents the weight of locomotive;a1, a2, a3For constant coefficient, it is necessary to according to reality
Determination is tested, it is different with locomotive type of vehicle.
Additional drag Rl(s) mainly caused as the line condition where it, including the additional resistance of additional resistance due to grade, curve
Power and tunnel additional drag etc..
Additional resistance due to grade wgFor:
wg=i (N/kN) (5)
I is the value of slope where locomotive in formula, and unit is thousand indexing, and the value is if just, then it represents that goes up a slope;If bearing, then
Represent descending.
Additional resistance due to curve wcFor:
R is the radius of curvature of curve where locomotive in formula, and unit is rice;a4Constant coefficient is represented, it is necessary to true according to experiment
Determine, it is necessary to be determined according to experiment.
The additional drag w in tunneltFor:
wt=a5Ls (N/kN) (7)
L in formulasFor the length in tunnel, unit is rice;a5Experiment coefficient is represented, it is necessary to be determined according to experiment.
Therefore locomotive additional drag is three's sum, i.e.,:
Rl(s)=wg+wc+wt (8)
2nd, the locomotive mechanism model determined by, carries out the recurrence learning of mechanism model, obtains basic mechanism model.
Detailed process is as follows:
When locomotive mechanism model according to represented by above formula (1) carries out mechanism model recurrence learning, using multinomial
Homing method, function when optimization object function is error sum of squares minimum.The explanation of polynomial regression analysis is given below:
If variable y and x relation are p order polynomials, and in xiRandom error ε of the place to yiNormal Distribution, then exist
Nonlinear multinomial model:
In formula (9), β0、β1.。。βpAs correspond to the coefficient of item number, εiFor random error.
OrderThen above-mentioned nonlinear multinomial model is converted into polynary
Linear model, i.e.,:
yi=β0+β1xi1+β2xi2+...+βpxip+εi(i=1,2 ..., n) 10)
β in formula (10)0、β1.。。βpAs correspond to the coefficient of item number, εiFor random error.
Optimization object function is obtained using the method for multiple linear regression analysis afterwards.It is specific as follows:
The coefficient matrices As of above-mentioned formula (10), structure matrix X, constant matrices B are extracted, is respectively:
In above-mentioned formula (11), (12) and (13), p is polynomial number, and n is variable y and x quantity.
The least-squares estimation of regression equation coefficient is:
B=A ' B=(XTX)-1XY (14)
Function during the error sum of squares minimum then finally given is optimization object function, as basic mechanism model.
Step S20, according to influence factor in locomotive actual moving process, model is iterated by way of Symbolic Regression
Study, the iterative model after being restrained.
In locomotive actual moving process, having several factors can have an impact to basic mechanism model, these factors such as circuit
Information, locomotive information and operation information etc..In view of various factors is to the accuracy of basic mechanism model in actual moving process
Influence, therefore the study of model is iterated by way of Symbolic Regression.
The process of Symbolic Regression is the process of function modelling, in the process, according to given one group of argument value and one group
Functional value (is referred to as training data), finds out the functional relation of fitting training data.Symbolic Regression is done using genetic programming (GP)
Method, specific implementation procedure are as follows:
Step S201, function variable and collection of functions are determined, and according to the initial population of its determination iterative model.
For Symbolic Regression, it is defeated that function variable includes the locomotive state informations such as train information, driver information and line information
Enter variable, and the stochastic variable such as scheduling actions and selection control action.Collection of functions includes the required algorithm used.Root
Initial population is determined according to the function variables such as above-mentioned train information, driver information and line information and collection of functions.
Step S202, calculate the fitness value of population.
State abstraction is carried out according to the input of train information, driver information and line information, kind is calculated using equation below
The fitness (degree of approximation of population approaching to reality solution) of group:
In formula, i represents i-th of population, f (xj) j-th of individual calculated value in i-th of population is represented, g (x) is represented
The desired value of j-th of individual in i-th of population;FiRepresent the fitness value of j-th of individual in i-th of population.
Step S203, judge colony fitness value whether match state jump condition (the state jump condition can be
Continuous n does not change for fitness value, or the fitness threshold value of setting), if satisfied, then performing step S204;It is if discontented
Foot, then step S206 is performed, i.e., the heredity such as variation, insertion, migration, restructuring are performed with the probability of setting on the basis of current population
Operator so that population is evolved, and new population is produced according to Evolution of Population result, is jumped to step S202 execution and is calculated the novel species
The fitness value of group.
Step S204, that is, terminate computing, exports corresponding expression formula, obtains scheduling actions optimal under current state, so
After be transferred to step S205.
Step S205, selection control action is performed according to optimal scheduling actions under current state, and by the machine after change
Car status information feeds back to iterative model with function variables such as selection control actions.
Training and study are so constantly made iteratively, finally make it that iterative model is restrained.
Step S30, State Forecasting Model is collectively constituted by basic mechanism model and iterative model.
Step S40, function is selected to differentiate whether the error of State Forecasting Model meets to minimize by equation, if satisfied,
Then perform step S50, will the currently available State Forecasting Model as preferable dbjective state forecast model;If it is unsatisfactory for
Condition, then continue return to step S20.
Target error has been preset in equation selection function, using the target error as standard, has selected function to differentiate by equation
Whether the error of State Forecasting Model, which meets, minimizes.
From the above, it is seen that bonding machine car mechanism model of the present invention, is learnt by the way of machine learning suitable for embedding
Embedded system and the high State Forecasting Model of precision of prediction, obtain the dbjective state forecast model of locomotive automatic Pilot control;It is logical
Cross the dbjective state forecast model of locomotive automatic Pilot control so that the precision of prediction of prediction algorithm is higher.
Although the present invention is disclosed as above with preferred embodiment, embodiment does not limit the present invention.This hair is not being departed from
Any equivalence changes done in bright spirit and scope or retouching, also belong to the protection domain of the present invention.Therefore the present invention
Protection domain should be using the content that claims hereof is defined as standard.
Claims (4)
- A kind of 1. locomotive automatic Pilot control forecasting molding machine learning method, it is characterised in that the locomotive automatic Pilot control Forecast model machine learning method processed includes:Step S10, mechanism model recurrence learning is carried out for locomotive, obtains basic mechanism model;Step S20, according to influence factor in locomotive actual moving process, model is iterated by way of Symbolic Regression Practise, the iterative model after being restrained;Step S30, State Forecasting Model is collectively constituted by basic mechanism model and iterative model;Step S40, function is selected to differentiate whether the error of State Forecasting Model meets to minimize by equation, if satisfied, then holding Row step S50, will the currently available State Forecasting Model as preferable dbjective state forecast model;If it is unsatisfactory for bar Part, then continue return to step S20.
- A kind of 2. locomotive automatic Pilot control forecasting molding machine learning method according to claim 1, it is characterised in that The step S10 includes:Step 1, using locomotive as research object, by the kinetic model of locomotive, determine locomotive mechanism model;The locomotive Mechanism model, represent as follows:<mrow> <mi>C</mi> <mo>=</mo> <mi>m</mi> <mfrac> <mrow> <mi>d</mi> <mi>v</mi> </mrow> <mrow> <mi>d</mi> <mi>t</mi> </mrow> </mfrac> <mo>=</mo> <mi>f</mi> <mrow> <mo>(</mo> <mi>s</mi> <mo>)</mo> </mrow> <mo>-</mo> <mi>W</mi> <mo>-</mo> <mi>B</mi> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>1</mn> <mo>)</mo> </mrow> </mrow>In above-mentioned formula, C represents that locomotive is made a concerted effort in track direction of advance;M represents the weight of locomotive;V is that locomotive is current The speed of service;T is the current run time of locomotive;The acceleration of as current locomotive;S is that the position of locomotive is kilometer post; F (s) represents the tractive force of current locomotive position;W is locomotive operation resistance;B is to locomotive caused by locomotive brake device Play the brake drag acted on backward;Step 2, according to identified locomotive mechanism model, the recurrence learning of mechanism model is carried out, obtains basic mechanism model.
- A kind of 3. locomotive automatic Pilot control forecasting molding machine learning method according to claim 2, it is characterised in that The step 2 includes:Using polynomial regression method, mechanism model recurrence learning is carried out to locomotive mechanism model, obtains error sum of squares minimum When optimization object function, and by mechanism model based on the optimization object function.
- A kind of 4. locomotive automatic Pilot control forecasting molding machine learning method according to claim 1, it is characterised in that The step S20 includes:Step S201, function variable is determined according to locomotive state information, and associative function collection determines the initial population of iterative model;Step S202, calculate the fitness value of population;State abstraction is carried out according to the input of train information, driver information and line information, population is calculated using equation below Fitness (degree of approximation of population approaching to reality solution):<mrow> <msub> <mi>F</mi> <mi>i</mi> </msub> <mo>=</mo> <msubsup> <mi>&Sigma;</mi> <mrow> <mi>j</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>n</mi> </msubsup> <mo>|</mo> <mi>f</mi> <mrow> <mo>(</mo> <msub> <mi>x</mi> <mi>j</mi> </msub> <mo>)</mo> </mrow> <mo>-</mo> <mi>g</mi> <mrow> <mo>(</mo> <msub> <mi>x</mi> <mi>j</mi> </msub> <mo>)</mo> </mrow> <mo>|</mo> </mrow>In formula, i represents i-th of population, f (xj) j-th of individual calculated value in i-th of population is represented, g (x) is represented i-th The desired value of j-th of individual in population;FiRepresent the fitness value of j-th of individual in i-th of population;Step S203, judge colony fitness value whether match state jump condition, if satisfied, then performing step S204;If It is unsatisfactory for, then performs step S206, i.e., variation, insertion, migration, restructuring something lost is performed with the probability of setting on the basis of current population Operator is passed, new population is produced, jumps to step S202;Step S204, terminate computing, export corresponding expression formula, obtain scheduling actions optimal under current state, be then transferred to Step S205;Step S205, selection control action is performed according to optimal scheduling actions under current state, and by the locomotive shape after change State information feeds back to iterative model with function variables such as selection control actions.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710990981.2A CN107798428A (en) | 2017-10-23 | 2017-10-23 | A kind of locomotive automatic Pilot control forecasting molding machine learning method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710990981.2A CN107798428A (en) | 2017-10-23 | 2017-10-23 | A kind of locomotive automatic Pilot control forecasting molding machine learning method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107798428A true CN107798428A (en) | 2018-03-13 |
Family
ID=61533395
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710990981.2A Pending CN107798428A (en) | 2017-10-23 | 2017-10-23 | A kind of locomotive automatic Pilot control forecasting molding machine learning method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107798428A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109857098A (en) * | 2018-12-11 | 2019-06-07 | 东南大学 | A kind of automatic Pilot agricultural machinery Trajectory Tracking System and method based on MPC |
CN111830825A (en) * | 2019-04-17 | 2020-10-27 | 株式会社日立制作所 | Automatic driving control method and system for machine |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102442323A (en) * | 2011-10-25 | 2012-05-09 | 中国电子科技集团公司第三十二研究所 | Method for achieve automatic driving curve generation between stations during operation by automatic train driving system |
CN102514602A (en) * | 2011-12-29 | 2012-06-27 | 浙江网新中控创新技术研究开发有限公司 | Method and system for planning and controlling train travelling speed |
CN106529023A (en) * | 2016-11-09 | 2017-03-22 | 南京工程学院 | Iterative learning-based subway train automatic running speed control method |
CN106777809A (en) * | 2017-01-13 | 2017-05-31 | 清华大学 | A kind of locomotive traction computation model calibration method based on actual operating data |
CN106844947A (en) * | 2017-01-18 | 2017-06-13 | 清华大学 | A kind of locomotive energy saving optimizing automatic Pilot method based on high-order relational learning |
CN106842901A (en) * | 2017-01-19 | 2017-06-13 | 重庆交通大学 | It is the method for train automated driving system formation speed control command |
CN107239628A (en) * | 2017-06-15 | 2017-10-10 | 清华大学 | A kind of uncertain locomotive simulation model system construction method based on dynamic time sequence figure |
-
2017
- 2017-10-23 CN CN201710990981.2A patent/CN107798428A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102442323A (en) * | 2011-10-25 | 2012-05-09 | 中国电子科技集团公司第三十二研究所 | Method for achieve automatic driving curve generation between stations during operation by automatic train driving system |
CN102514602A (en) * | 2011-12-29 | 2012-06-27 | 浙江网新中控创新技术研究开发有限公司 | Method and system for planning and controlling train travelling speed |
CN106529023A (en) * | 2016-11-09 | 2017-03-22 | 南京工程学院 | Iterative learning-based subway train automatic running speed control method |
CN106777809A (en) * | 2017-01-13 | 2017-05-31 | 清华大学 | A kind of locomotive traction computation model calibration method based on actual operating data |
CN106844947A (en) * | 2017-01-18 | 2017-06-13 | 清华大学 | A kind of locomotive energy saving optimizing automatic Pilot method based on high-order relational learning |
CN106842901A (en) * | 2017-01-19 | 2017-06-13 | 重庆交通大学 | It is the method for train automated driving system formation speed control command |
CN107239628A (en) * | 2017-06-15 | 2017-10-10 | 清华大学 | A kind of uncertain locomotive simulation model system construction method based on dynamic time sequence figure |
Non-Patent Citations (3)
Title |
---|
JIN HUANG 等: "An Energy-Efficient Train Control Framework for Smart Railway Transporation", 《IEEE TRANSACTIONS ON COMPUTERS》 * |
胡仲康: "非线性回归在预测中的应用", 《南方冶金学院学报》 * |
金金宝 等: "遗传编程在符号回归中的应用", 《计算机与数字工程》 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109857098A (en) * | 2018-12-11 | 2019-06-07 | 东南大学 | A kind of automatic Pilot agricultural machinery Trajectory Tracking System and method based on MPC |
CN111830825A (en) * | 2019-04-17 | 2020-10-27 | 株式会社日立制作所 | Automatic driving control method and system for machine |
US11619929B2 (en) | 2019-04-17 | 2023-04-04 | Hitachi, Ltd. | Automatic operation control method and system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107399333B (en) | A kind of accurate parking method of train towards train automatic Pilot | |
CN103246943B (en) | Vehicle operating condition multi-scale predicting method based on Markov chain | |
CN103847749B (en) | Sequence generating method handled by a kind of double-deck multiple-objection optimization railway locomotive | |
CN106427589A (en) | Electric car driving range estimation method based on prediction of working condition and fuzzy energy consumption | |
CN110458336B (en) | Online appointment vehicle supply and demand prediction method based on deep learning | |
CN107169567A (en) | The generation method and device of a kind of decision networks model for Vehicular automatic driving | |
CN109783843A (en) | Analogue system vehicle dynamic model speed predicting method based on LSTM | |
CN106458049A (en) | Prediction of battery power requirements for electric vehicles | |
CN103745110B (en) | Method of estimating operational driving range of all-electric buses | |
CN105868787A (en) | Electric car driving range evaluation method based on working condition identification and fuzzy energy consumption | |
CN103745111A (en) | Method of predicting driving range of all-electric passenger vehicles | |
US11661067B2 (en) | Method for ascertaining driving profiles | |
CN110509957A (en) | A kind of train speed prediction technique and device | |
CN106956680A (en) | A kind of electric automobile driving behavior identifying and analyzing method | |
CN104102776A (en) | Automatic identification method for urban railway trains | |
CN107798428A (en) | A kind of locomotive automatic Pilot control forecasting molding machine learning method | |
CN108162953A (en) | Vehicle congestion region control method and device | |
Skarka | Reducing the energy consumption of electric vehicles | |
CN110427690A (en) | A kind of method and device generating ATO rate curve based on global particle swarm algorithm | |
CN113415288A (en) | Sectional type longitudinal vehicle speed planning method, device, equipment and storage medium | |
CN114013284B (en) | Calculation method and system for vehicle endurance mileage | |
CN115422747A (en) | Method and device for calculating discharge amount of pollutants in tail gas of motor vehicle | |
Alateef et al. | Energy consumption estimation for electric vehicles using routing API data | |
CN110610326A (en) | Driving management system based on driving data | |
Liu et al. | Development of Driver-Behavior Model Based onWOA-RBM Deep Learning Network |
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: 20180313 |