CN105930565A - Method for calibrating traffic simulation model parameters based on cross entropy algorithm of linear strategy - Google Patents
Method for calibrating traffic simulation model parameters based on cross entropy algorithm of linear strategy Download PDFInfo
- Publication number
- CN105930565A CN105930565A CN201610231247.3A CN201610231247A CN105930565A CN 105930565 A CN105930565 A CN 105930565A CN 201610231247 A CN201610231247 A CN 201610231247A CN 105930565 A CN105930565 A CN 105930565A
- Authority
- CN
- China
- Prior art keywords
- sample
- speed
- model
- following
- iteration
- 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.)
- Granted
Links
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 41
- 238000000034 method Methods 0.000 title claims abstract description 23
- 238000004088 simulation Methods 0.000 title claims abstract description 19
- 238000011156 evaluation Methods 0.000 claims abstract description 10
- 230000001133 acceleration Effects 0.000 claims description 5
- 241000208340 Araliaceae Species 0.000 claims description 2
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 claims description 2
- 235000003140 Panax quinquefolius Nutrition 0.000 claims description 2
- 235000008434 ginseng Nutrition 0.000 claims description 2
- 238000002372 labelling Methods 0.000 claims description 2
- 238000004364 calculation method Methods 0.000 abstract description 3
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000004308 accommodation Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000000227 grinding Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000008929 regeneration Effects 0.000 description 1
- 238000011069 regeneration method Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000002922 simulated annealing Methods 0.000 description 1
- 238000005303 weighing Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/30—Circuit design
- G06F30/36—Circuit design at the analogue level
- G06F30/367—Design verification, e.g. using simulation, simulation program with integrated circuit emphasis [SPICE], direct methods or relaxation methods
Landscapes
- Engineering & Computer Science (AREA)
- Computer Hardware Design (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Microelectronics & Electronic Packaging (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Traffic Control Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a method for calibrating traffic simulation model parameters based on a cross entropy algorithm of a linear strategy. The method comprises following steps: a car-following model is used as an example, the space headway between two vehicles and the speed of the rear vehicle are used as evaluation indexes to determine the format of a target function and therefore to determine the objects of the parameter calibration; required traffic data is collected; parameters to be calibrated and a corresponding effective value range are determined according to the determined car-following model and the traffic data; parameter calibration calculation is performed on the car-following model by means of the cross entropy algorithm; in the calculation process, the linear strategy is adopted to determine the sample size: assuming that a negative linear relation exists between the generation quantity of the next iteration sample and the variance of the current iteration sample, that is the large the variance of the current sample is, the larger the generation quantity of the next iteration sample is, and the smaller the variance is, the smaller the sample generation quantity is; a sample mean value is finally output as the optimal value of the calibration until the sample variance is less than a preset threshold value of the cross entropy algorithm.
Description
Technical field
The present invention relates to the technical field of traffic simulation parameter calibration, especially relate to a kind of based on linear plan
The Traffic Flow Simulation Models parameter calibration method of Cross-Entropy Algorithm slightly, is the improvement of a kind of parameter calibration method.
Background technology
For the most serious traffic jam issue, Microscopic Traffic Simulation Mathematic Model is being analyzed and is being processed again
Hybridize understanding and considerate scape day by day to play a significant role.As a example by following-speed model, this model introduces more independence ginseng
Number carrys out drive simulating person's driving behavior under different traffic sights.But, the driving behavior of driver with
The difference of driver and different, additionally, in the behavior of the same driver in the place that the different time is different
Also can change.Therefore, before following-speed model is as emulation tool drive simulating behavior, it is necessary to profit
By measured data, following-speed model carried out parameter calibration, thus improve the precision of Microscopic Traffic Simulation Mathematic Model.
In the conventional research about parameter calibration algorithm, mainly based on stochastic search methods, such as, lose
Propagation algorithm, simulated annealing, Cross-Entropy Algorithm etc..The main thought of these algorithms is all in iteration mistake
Cheng Zhong, uses respective strategy, the degree of accuracy of the model in raising next iteration.Wherein cross entropy is calculated
Method, with its succinct and complete theoretical basis, is widely used day by day.This algorithm is based on Mondicaro
Thought, constantly approaches the probability sensitivity function of parameter of optimum during each iteration, thus quickly
The parameter combination that location is optimum.But, traditional Cross-Entropy Algorithm the most all generates identical number
Amount sample, particularly when sample variance is less, the sample that quantity is identical can't improve search optimal value
Precision, the simulation times of repetition can be increased on the contrary, this process can take a substantial amount of time.
Summary of the invention
It is an object of the invention to overcome deficiency present in existing cross entropy technology, it is provided that a kind of based on linearly
The Traffic Flow Simulation Models parameter calibration method of the Cross-Entropy Algorithm of strategy, is that a kind of Linear Strategy determines often
The sample of an iteration generates quantity, reduces the simulation times repeated, and improves the calculating effect of Cross-Entropy Algorithm
Rate, has easy to operate, it is achieved the high efficiency of parameter calibration, it is ensured that the result accuracy of parameter calibration,
The feature of wide accommodation, and successfully this algorithm is applied in the parameter calibration of Traffic Flow Simulation Models.
To achieve these goals, the technical scheme is that
A kind of Traffic Flow Simulation Models parameter calibration method of Cross-Entropy Algorithm based on Linear Strategy, concrete
Step is as follows:
A, choose the space headway in two workshops and the speed of rear car as evaluation index, determine target letter
The form of number, so that it is determined that the target of parameter calibration.
The target of traditional parameter calibration is that the traffic flow data minimizing observation obtains with phantom
Difference between traffic flow data.As it has been described above, in following-speed model with the space headway of two cars and after
Vehicle speed is as evaluation index, and therefore in object function, the object of traffic flow is exactly space headway and speed,
The form of concrete calculating difference is:
In formula, S (x) represents the value of object function, and x represents the parameter set of following-speed model;T represents total imitating
True time is long, and t represents that certain emulates the moment;Represent the space headway calculated at t following-speed model,Represent and observe, in t, the space headway obtained;WithIt is illustrated respectively in t model to calculate
Go out and observe the velocity amplitude obtained;λ is weighted value, and λ is the highest, represents space headway in object function
Proportion is higher, otherwise, the proportion of velocity amplitude is higher.
Traffic data required for b, collection, and determine the form of following-speed model;According to the friendship set up
Logical model and field survey data, determine effective span of parameter to be calibrated and correspondence.
Specifically, the data demarcating following-speed model needs include: (track refers to front truck wheelpath
Certain car is at speed sometime, absolute position, acceleration etc.), with the original state etc. of car of speeding.According to
The measured data obtained, it may be determined that the calibrating parameters of following-speed model and effective span of correspondence.
In intelligent driving model (one type of following-speed model), the parameter of demarcation includes peak acceleration, phase
Hope deceleration, desired speed, expected time spacing, minimum space headway.
C, utilize Cross-Entropy Algorithm that above-mentioned following-speed model is carried out the computing of parameter calibration.Before the computation,
Use and linearly determine sample size strategy, i.e. assume the quantity of sample and the variance of sample exist one negative
The sample variance of linear relationship, i.e. current iteration is big, is accomplished by substantial amounts of sample when next iteration and receives
Collect more excellent sample;The sample variance of current iteration is few, just uses a small amount of sample when next iteration, subtracts
Few number of times repeating emulation.Until the threshold values that the variance of sample is preset less than Cross-Entropy Algorithm, finally export
The optimal value demarcated.
Compared with prior art, the invention have the advantage that the Cross-Entropy Algorithm of improvement, in every single-step iteration
Before, it is assured that the sample next time needing to generate many small number by simple linear relationship, from
And ensure that Cross-Entropy Algorithm has bigger hunting zone when initial, there is higher computational efficiency before the end.
Compared with traditional Cross-Entropy Algorithm, the algorithm after improvement is adapted to increasingly complex model calibration problem,
And the linear relationship assumed is easy to operate, applied widely.
Accompanying drawing explanation
Fig. 1 is the flow chart demarcating following-speed model.
Fig. 2 is for improving Cross-Entropy Algorithm calculation flow chart.
Detailed description of the invention
The present invention is further illustrated with example below in conjunction with the accompanying drawings.
As it is shown in figure 1, before following-speed model puts into microscopic simulation drive simulating person's driving behavior, need
Through the calibration process of parameter, the present embodiment proposes the traffic of a kind of Cross-Entropy Algorithm based on Linear Strategy
The step of simulation parameters scaling method includes:
A, choose evaluation index and determine object function:
In the example of the present invention, the standard of evaluation is the traffic flow data that draws of following-speed model and observe
Difference between the traffic flow data arrived is minimum.The size weighing difference is exactly object function, and traffic flow
Flow data is general in following-speed model uses rear car in information such as the speed in all moment and absolute positions.?
In example, the concrete form of object function is as follows:
In formula, S (x) represents the value of object function, and x represents the parameter set of following-speed model;T represents total imitating
True time is long, and t represents that certain emulates the moment;Represent the space headway calculated at t following-speed model,Representing and observe, in t, the space headway obtained, space headway is that the absolute position by front truck deducts
The vehicle commander of front truck deducts what the absolute position of rear car obtained again;WithIt is illustrated respectively in t model meter
Calculate and draw and observe the velocity amplitude obtained;λ is weighted value, and λ is the highest, represents in object function between headstock
Away from proportion higher, otherwise, the proportion of velocity amplitude is higher.
It should be noted that the form of object function is different, the calibration result obtained may be different;Target
Evaluation index in function is different, and the calibration result obtained is likely to differ.Therefore, in this example,
Have employed the weighted number of two evaluation indexes as object function, and have selected generated data and demarcate.
Contrasted with calibration result by default optimal value, be selected to maximize the target letter reappearing driving behavior
Array closes (target function value i.e. obtaining minimum), improves the descriptive power of following-speed model.
B, collection are demarcated the traffic data needed and determine the form of following-speed model:
In this example, demarcating the traffic data source needed is that the project from NGSIM obtains.This project
It is that United States highways federation organizes and implements, is upstairs set up photographic head record California I-80 by building high
The information of speed highway.The technical finesse by video image of this tissue, obtains each car on this highway
Wheelpath, including each car in certain moment speed, acceleration, absolute position etc. and open
The traffic data information obtained.According to the data message gathered, this example have selected intelligent driving model,
This model introduces 5 variables model with the driving behavior of car of speeding, and because its outstanding descriptive power is grinding
It is widely used in studying carefully.Demarcate parameter include peak acceleration, expectation deceleration, desired speed,
Expected time spacing, minimum space headway.
C, the pretreatment of data:
The track data of NGSIM is drawn by video image processing technology, wherein contains bigger mistake
Difference, may have bigger difference with truth.In order to consistent with reality, need in demarcation
The front data to NGSIM carry out pretreatment.In this example, the method using filtering is abnormal some
Irrational data point is directly removed, and the method re-using linear interpolation adds reasonably value again.
D, the cross-entropy method improved is used to carry out the parameter calibration of intelligent driving model:
Determine the form of object function, specify that the parameter in the following-speed model of demarcation, have collected demarcation
The all data needed, next step is exactly the parameter calibration work using optimized algorithm to carry out model.At this
In example, use the Cross-Entropy Algorithm improved to carry out the staking-out work of parameter, idiographic flow as in figure 2 it is shown,
Step includes:
D1, starting before Cross-Entropy Algorithm, it is assumed that the quantity of random sample (parameter) and the side of sample
The sample variance that difference exists simple negative correlativing relation, i.e. current iteration is the biggest, the sample number of following iteration
Measuring the biggest, the sample variance of current iteration is the least, and the sample size of next iteration is the least;
D2, the probability density function of initialization sample.This example use normal distribution generate random sample,
Initial parameter includes average and the variance of normal distribution.Bigger variance is needed, it is ensured that generate time initial
Random sample there is bigger hunting zone;
D3, Cross-Entropy Algorithm iteration start to determine exactly the quantity of sample according to linear relationship.Such as step d1
Described, variance bigger time initial can obtain bigger sample size, thus ensures the data sample generated
Sufficient and contain all possible value;
D4, according to normal distribution generate a number of random sample;
D5, each sample is inputted intelligent driving model respectively, the driving situation of drive simulating person, and receive
Collect drive speed and the absolute position of each moment driver.Drawn by comparative simulation and observation is obtained
Speed and absolute position, obtain evaluating the target function value of each sample;
D6, the target function value obtained is arranged in order from small to large, producing minimum target functional value
The sample labeling of front 5% is elite sample;I.e. elite sample is exactly in each iteration, it is possible to produce minimum
The parameter set of target function value;
D7, judge whether current elite sample meets stop condition.Essence during stop condition in this example
The variance of English sample is less than certain threshold values.When being unsatisfactory for stop condition, by the letter according to elite sample
Breath updates average and the variance of normal distribution, and returns d3 step, proceeds iteration.Until elite
Sample is gathered near optimum sample, exits circulation, the optimum sample of record, and terminates Cross-Entropy Algorithm.
E, judge whether optimum sample meets current requirement.By the threshold values of goal-selling function in this example
Judge whether optimum sample meets requirement.If being unsatisfactory for, then re-start data prediction work, then
Carry out parameter calibration, until target function value meets requirement.
The Cross-Entropy Algorithm improved is exactly that sample size is along with variance with the difference of conventional cross entropy algorithm maximum
Change and change.Conventional cross entropy algorithm the most all generates the random sample of equal number,
This is to ensure that the global search of algorithm at the iteration initial stage, but reducing along with sample variance, regeneration
Become the sample of equal number, wherein will there is more identical sample, the target of double counting same sample
Functional value will reduce the efficiency of algorithm.The Cross-Entropy Algorithm improved is by assuming that sample size and sample variance
Negative correlativing relation, it is ensured that iteration initially has the substantial amounts of all possible value of quantity Covering samples, repeatedly
Reduce identical sample size for the later stage, on the premise of not affecting algorithmic statement, reduce simulation times, improve
Efficiency of algorithm.The result of example shows, the Cross-Entropy Algorithm of improvement is faster looked for than traditional Cross-Entropy Algorithm
To optimal value, and optimal value show that target function value is the most accurate, the traffic flow data obtained and observation
Data between difference less.
Claims (4)
1. a Traffic Flow Simulation Models parameter calibration method for Cross-Entropy Algorithm based on Linear Strategy, its
Being characterised by using linear relationship to generate sample size, concrete step is as follows:
A, select following-speed model as Traffic Flow Simulation Models, choose the space headway between two cars and after
The speed of car, as evaluation index, determines the form of object function, so that it is determined that the target of parameter calibration;
Traffic data required for b, collection, and determine the form of following-speed model;Determined by according to
Speed model and field survey data, determine effective span of parameter to be calibrated and correspondence thereof;
C, utilize Cross-Entropy Algorithm that above-mentioned following-speed model carries out parameter calibration calculating, and use linear plan
Slightly determine sample size.
Method the most according to claim 1, it is characterised in that choose two workshops in step a
The speed of space headway and rear car, as evaluation index, determines the form of object function, particularly as follows:
Using the space headway in two workshops and the speed of rear car as evaluation index in following-speed model, then mesh
In scalar functions, calculating is exactly the space headway in two workshops of model emulation and the speed of rear car and observation
Difference between the data obtained, concrete form is:
In formula, S (x) represents the value of object function, and x represents the parameter set of following-speed model;T represents total imitating
True time is long, and t represents that certain emulates the moment;Represent the space headway calculated at t following-speed model,Represent and observe, in t, the space headway obtained;WithIt is illustrated respectively in t model to calculate
The velocity amplitude that go out and observation obtains;λ is weighted value, and λ is the highest, represents space headway in object function
Proportion the highest, otherwise, the proportion of velocity amplitude is the highest.
Method the most according to claim 1, it is characterised in that described step b gathers required
Traffic data include front truck wheelpath and the original state with car of speeding;Following-speed model determined by according to
And field survey data, determine parameter to be calibrated;In intelligent driving model, parameter to be calibrated
Including peak acceleration, expectation deceleration, desired speed, expected time spacing, minimum space headway.
Method the most according to claim 1, it is characterised in that described step c utilizes cross entropy to calculate
Method carries out parameter calibration calculating to above-mentioned following-speed model, and uses Linear Strategy to determine sample size, specifically
The process of realization is:
C1, the variance of the sample setting quantity and the current iteration of the random sample of an iteration exist simple
Negative correlativing relation, i.e. current sample variance is the biggest, and the sample size that next iteration generates is the biggest, when
Front sample variance is the least, and the sample size that next iteration generates is the least;
C2, the probability density function of initialization sample, sample herein refers to the ginseng to be calibrated of following-speed model
Number;
C3, iteration start, and are first according to the quantity of the sample that linear relationship determines that first iteration is to be generated;
C4, assume sample be distributed as independent normal distribution, according to this distribution generate a number of at random
Sample;
C5, each sample is inputted following-speed model respectively, the driving situation of drive simulating person, and collect every
The drive speed of one moment rear car driver and absolute position;Drawn by contrast simulation and observation is obtained
The drive speed of rear car and absolute position, calculating target function formula, obtain evaluating the mesh of each sample
Offer of tender numerical value;
C6, the target function value obtained is arranged in order from small to large, producing minimum target functional value
The sample labeling of front 5% is elite sample;I.e. elite sample is exactly in each iteration, it is possible to produce minimum
The parameter set of target function value;
C7, judge that whether current elite sample meets the iteration stopping condition set;When being unsatisfactory for stopping
During condition, by average and the variance of the information updating normal distribution according to elite sample, and return step
C3, proceeds iteration;Until elite sample is gathered near optimum sample, exit circulation, record
Excellent sample, and terminate Cross-Entropy Algorithm.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610231247.3A CN105930565B (en) | 2016-04-13 | 2016-04-13 | The Traffic Flow Simulation Models parameter calibration method of Cross-Entropy Algorithm based on Linear Strategy |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610231247.3A CN105930565B (en) | 2016-04-13 | 2016-04-13 | The Traffic Flow Simulation Models parameter calibration method of Cross-Entropy Algorithm based on Linear Strategy |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105930565A true CN105930565A (en) | 2016-09-07 |
CN105930565B CN105930565B (en) | 2019-11-19 |
Family
ID=56839124
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610231247.3A Expired - Fee Related CN105930565B (en) | 2016-04-13 | 2016-04-13 | The Traffic Flow Simulation Models parameter calibration method of Cross-Entropy Algorithm based on Linear Strategy |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105930565B (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106529076A (en) * | 2016-11-28 | 2017-03-22 | 东南大学 | Two-stage parameter calibration method for highway traffic safety simulation analysis |
CN110069887A (en) * | 2019-05-05 | 2019-07-30 | 腾讯科技(深圳)有限公司 | A kind of driving simulation method, apparatus, equipment and storage medium |
CN111125862A (en) * | 2019-09-27 | 2020-05-08 | 长安大学 | Following model emission measurement and calculation method based on genetic algorithm and specific power |
CN111680889A (en) * | 2020-05-20 | 2020-09-18 | 中国地质大学(武汉) | Offshore oil leakage source positioning method and device based on cross entropy |
CN112509064A (en) * | 2020-12-22 | 2021-03-16 | 北京百度网讯科技有限公司 | Method, device, equipment and storage medium for displaying camera calibration progress |
CN112926025A (en) * | 2021-03-09 | 2021-06-08 | 江苏海洋大学 | Method for solving weapon target allocation problem based on cross entropy genetic algorithm |
CN113094875A (en) * | 2021-03-16 | 2021-07-09 | 东南大学 | Method and device for calibrating micro traffic simulation system in urban expressway intersection area |
CN113268855A (en) * | 2021-04-19 | 2021-08-17 | 东南大学 | Calibration method of microscopic traffic simulation model of ring intersection |
CN115618655A (en) * | 2022-12-15 | 2023-01-17 | 北京华录高诚科技有限公司 | Single-point intersection microscopic traffic simulation model parameter calibration method and device |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102521438A (en) * | 2011-12-02 | 2012-06-27 | 东南大学 | Traffic conflict simulation two-stage parameter calibrating method |
CN103077275A (en) * | 2013-01-06 | 2013-05-01 | 东南大学 | Parameter calibration method of highway ramp simulation model |
CN103942398A (en) * | 2014-05-05 | 2014-07-23 | 江苏物联网研究发展中心 | Traffic simulation correction method based on genetic algorithm and generalized recurrent nerve network |
-
2016
- 2016-04-13 CN CN201610231247.3A patent/CN105930565B/en not_active Expired - Fee Related
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102521438A (en) * | 2011-12-02 | 2012-06-27 | 东南大学 | Traffic conflict simulation two-stage parameter calibrating method |
CN103077275A (en) * | 2013-01-06 | 2013-05-01 | 东南大学 | Parameter calibration method of highway ramp simulation model |
CN103942398A (en) * | 2014-05-05 | 2014-07-23 | 江苏物联网研究发展中心 | Traffic simulation correction method based on genetic algorithm and generalized recurrent nerve network |
Non-Patent Citations (2)
Title |
---|
R.X. ZHONG等: "A cross-entropy method and probabilistic sensitivity analysis framework for calibrating microscopic traffic models", 《TRANSPORTATION RESEARCH PART C》 * |
曹金亮等: "基于城市主干路交通流数据的跟驰模型标定", 《交通信息与安全》 * |
Cited By (17)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106529076A (en) * | 2016-11-28 | 2017-03-22 | 东南大学 | Two-stage parameter calibration method for highway traffic safety simulation analysis |
CN106529076B (en) * | 2016-11-28 | 2019-07-30 | 东南大学 | The two stages parameter calibration method of expressway traffic safety simulation analysis |
CN110069887A (en) * | 2019-05-05 | 2019-07-30 | 腾讯科技(深圳)有限公司 | A kind of driving simulation method, apparatus, equipment and storage medium |
WO2020224434A1 (en) * | 2019-05-05 | 2020-11-12 | 腾讯科技(深圳)有限公司 | Driving simulation method and apparatus, electronic device and computer storage medium |
CN111125862A (en) * | 2019-09-27 | 2020-05-08 | 长安大学 | Following model emission measurement and calculation method based on genetic algorithm and specific power |
CN111125862B (en) * | 2019-09-27 | 2023-12-26 | 长安大学 | Following model emission measuring and calculating method based on genetic algorithm and specific power |
CN111680889B (en) * | 2020-05-20 | 2023-08-18 | 中国地质大学(武汉) | Cross entropy-based offshore oil leakage source positioning method and device |
CN111680889A (en) * | 2020-05-20 | 2020-09-18 | 中国地质大学(武汉) | Offshore oil leakage source positioning method and device based on cross entropy |
CN112509064A (en) * | 2020-12-22 | 2021-03-16 | 北京百度网讯科技有限公司 | Method, device, equipment and storage medium for displaying camera calibration progress |
CN112509064B (en) * | 2020-12-22 | 2024-06-04 | 阿波罗智联(北京)科技有限公司 | Method, device, equipment and storage medium for displaying camera calibration progress |
CN112926025A (en) * | 2021-03-09 | 2021-06-08 | 江苏海洋大学 | Method for solving weapon target allocation problem based on cross entropy genetic algorithm |
CN112926025B (en) * | 2021-03-09 | 2023-07-25 | 江苏海洋大学 | Method for solving weapon target distribution problem based on cross entropy genetic algorithm |
CN113094875A (en) * | 2021-03-16 | 2021-07-09 | 东南大学 | Method and device for calibrating micro traffic simulation system in urban expressway intersection area |
CN113094875B (en) * | 2021-03-16 | 2024-04-09 | 东南大学 | Method and device for calibrating microscopic traffic simulation system in urban expressway interweaving area |
CN113268855A (en) * | 2021-04-19 | 2021-08-17 | 东南大学 | Calibration method of microscopic traffic simulation model of ring intersection |
CN113268855B (en) * | 2021-04-19 | 2024-01-30 | 东南大学 | Calibration method of microscopic traffic simulation model of annular intersection |
CN115618655A (en) * | 2022-12-15 | 2023-01-17 | 北京华录高诚科技有限公司 | Single-point intersection microscopic traffic simulation model parameter calibration method and device |
Also Published As
Publication number | Publication date |
---|---|
CN105930565B (en) | 2019-11-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105930565A (en) | Method for calibrating traffic simulation model parameters based on cross entropy algorithm of linear strategy | |
CN109709956B (en) | Multi-objective optimized following algorithm for controlling speed of automatic driving vehicle | |
CN107169567B (en) | Method and device for generating decision network model for automatic vehicle driving | |
Dörr et al. | Online driving style recognition using fuzzy logic | |
US8645052B2 (en) | Motor vehicle acceleration determining system | |
CN109242251A (en) | Vehicular behavior safety detecting method, device, equipment and storage medium | |
CN103150677B (en) | Radical driving condition recognition methods and system | |
CN103077275B (en) | The parameter calibration method of highway ramp simulation model | |
CN112201069A (en) | Deep reinforcement learning-based method for constructing longitudinal following behavior model of driver | |
CN105930614A (en) | Cell transmission model parameter calibration and verification method specific to variable speed limit control | |
CN114140903B (en) | Road type recognition vehicle-mounted device based on decision tree generation rule | |
CN107953888A (en) | A kind of road type recognition methods and system | |
CN116702096B (en) | Method and device for measuring and calculating road sliding resistance of vehicle plateau environment | |
CN114169444B (en) | Driving style classification method considering risk potential field distribution under vehicle following working condition | |
Derbel | Driving style assessment based on the GPS data and fuzzy inference systems | |
CN115422747A (en) | Method and device for calculating discharge amount of pollutants in tail gas of motor vehicle | |
CN110119528B (en) | Bridge random traffic flow simulation system based on intelligent cellular automata | |
Akçelik | The relationship between capacity and driver behaviour | |
CN111125862A (en) | Following model emission measurement and calculation method based on genetic algorithm and specific power | |
CN113743466B (en) | Road type identification method and system based on decision tree | |
Wu et al. | Particle filter based traffic data assimilation with sensor informed proposal distribution | |
CN115841712A (en) | Driving data processing method, device and equipment based on V2X technology | |
WO2021258847A1 (en) | Driving decision-making method, device, and chip | |
Li et al. | Modeling and verification of heavy-duty truck drivers’ car-following characteristics | |
CN108268678A (en) | Driving behavior analysis method, apparatus and system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20191119 |
|
CF01 | Termination of patent right due to non-payment of annual fee |