CN104318757B - Bus bus or train route section Forecasting Methodology working time on a kind of public transportation lane - Google Patents
Bus bus or train route section Forecasting Methodology working time on a kind of public transportation lane Download PDFInfo
- Publication number
- CN104318757B CN104318757B CN201410609740.5A CN201410609740A CN104318757B CN 104318757 B CN104318757 B CN 104318757B CN 201410609740 A CN201410609740 A CN 201410609740A CN 104318757 B CN104318757 B CN 104318757B
- Authority
- CN
- China
- Prior art keywords
- bus
- working time
- section
- vehicle
- public transportation
- 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.)
- Expired - Fee Related
Links
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/123—Traffic control systems for road vehicles indicating the position of vehicles, e.g. scheduled vehicles; Managing passenger vehicles circulating according to a fixed timetable, e.g. buses, trains, trams
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to bus bus or train route section Forecasting Methodology working time on a kind of public transportation lane, input vector is mapped in the space of a higher-dimension by kernel function by it, and the learning process of regression function is solved at this higher dimensional space, thus nonlinear problem is converted into linear problem, significantly improve the precision of the bus car Forecasting of Travel Time on public transportation lane; Compound kernel function is obtained additionally by by basic kernel function weighting summation, remain the characteristic of different IPs function, dissimilar data input can be processed better, this model is made to have stronger robustness, advance the construction process of intelligent transportation, all there is far-reaching influence to the Optimized Operation of the bus system on public transportation lane and the facility of passenger's trip.Therefore, the bus bus or train route section that the present invention can be widely used on bus special lane predicts field working time.
Description
Technical field
The present invention relates to a kind of time forecasting methods, particularly about bus bus or train route section Forecasting Methodology working time on a kind of public transportation lane.
Background technology
Public transportation lane is at road conditions license, congested in traffic Important Sections, marks one or several track by mark, graticule etc., round-the-clock or limit passing through of other vehicles at times, special for bus.APTS (AdvancedPublicTransportationSystem, advanced public transportation system) and ATIS (AdvancedTravelerInformationSystem, prefabricated box girder) be ITS (IntelligentTransportSystem, intelligent transportation system) two large kernel subsystems, the bus running time reasonably and accurately on prediction public transportation lane is for ATIS, effective trip information can be provided, thus allow the plan of travel of passenger's reasonable arrangement oneself, improve Bus Service quality, for APTS, network operator can be allowed to take adequate measures to adjust operation plan, effectively carry out operation management.Therefore, bus car accurately reasonably on prediction public transportation lane is ITS the most key ingredient working time, for the Bus Service level promoted on public transportation lane, strengthen bus trip attraction dynamics, alleviate urban traffic blocking and improve human settlements trip situation all significant.
Bus car on public transportation lane is subject to the impact of a lot of enchancement factor working time, such as: the change etc. of weather, traffic and passenger flow, there is very complicated relation between the operation conditions of bus and these factors, be therefore difficult to applied mathematical model and demarcate accurately and predict.The neural network algorithm of existing prediction bus bus or train route section working time on public transportation lane is determined, is crossed study and owe the problems such as study and local convergence due to structure, make the precision of neural network algorithm not high, and another kind of Forecasting Methodology---SVM (SupportVectorMachine, standard support vector machine) precision depend on the selection of kernel function unduly, and do not have ripe theory to instruct this at present, cause its robustness poor, usually can not get in actual applications predicting the outcome preferably.And multiple basic kernel function is obtained compound kernel function by the form of weighting summation by multi-kernel learning support vector function, avoid the blindness of Selection of kernel function, reduce the dependence to Selection of kernel function, improve the robustness of prediction.
Summary of the invention
For the problems referred to above, the present invention adopts bus bus or train route section Forecasting Methodology working time provided on a kind of public transportation lane.
For achieving the above object, the present invention takes following technical scheme: bus bus or train route section Forecasting Methodology working time on a kind of public transportation lane, and it comprises the following steps: 1) gather bus running line information, bus vehicle operating information and the bus GPS service data on bus special lane; 2) according to bus running line information, bus vehicle operating information and the bus GPS service data on the bus special lane that obtains, the working time of acquisition bus vehicle between each section, and working time is normalized; 3) choose some basic kernel functions and set up preliminary multi-kernel learning SVM prediction model, and calculate the weight shared by each kernel function; 4) adopt step 2) in the data obtained train preliminary multi-kernel learning SVM prediction model, obtain multi-kernel learning SVM prediction model; 5) test for multi-kernel learning SVM prediction model, adopt root-mean-square error RMSE evaluation test result, if dissatisfied, then return step 3) again choose basic kernel function, build preliminary multi-kernel learning SVM prediction model; If test result is satisfied, then enter next step; 6) for working time between bus vehicle website in a circuit any in roadnet, adopt multi-kernel learning SVM prediction model to predict, obtain predicted value corresponding working time.
Described step 1) in, the bus running line information on public transportation lane comprises bus running circuit mileage, bus stop position and website number and line condition; Weather conditions when bus vehicle operating information comprises bus departure interval, bus passenger flow peak period, bus vehicle operating and road traffic condition; Bus GPS service data comprises bus stop title, bus vehicle license plate numbering, bus vehicle order of classes or grades at school numbering, bus vehicle frequency, bus vehicle sail website into and sail out of the website moment.
Described step 2) in, according to bus GPS service data, obtain the moment of vehicle through two end points in any section, the moment of these two end points is subtracted each other, obtain the working time of bus vehicle on this section, and will be normalized working time, its normalization formula is:
Wherein,
for the bus vehicle hour value after normalization, S
ifor original bus vehicle hour numerical value, S
minfor the minimum value of bus vehicle hour in sample sequence, S
maxfor the maximal value of bus vehicle hour in sample sequence.
Described step 4) comprise following content: by step 2) in by the input vector x that working time, 5 variablees were formed in the working time on weather, time period, section, current road segment, next section
iwith bus section y working time of correspondence
ithe sample data of composition is divided into training sample and test sample book, and chooses 70% as training sample, all the other 30% test sample books being used for checking training result.
Described step 6) comprise following content: if 1. this car is first car, adopts and only comprise time period x
1, weather x
2and section
three variablees form three-dimensional input vector, predict the virtual up-to-date working time in each section with multi-kernel learning SVM prediction model; After vehicle crosses, upgrade the up-to-date working time in each section, as the input variable of next bus car; If 2. this car is not first car, then by time period x
1, weather x
2and section
the working time of current road segment
the working time that next section is up-to-date
these five variablees, as input variable, are input in multi-kernel learning SVM prediction model; To export between bus vehicle website predicted value y working time
k+1.
The present invention is owing to taking above technical scheme, it has the following advantages: 1, the present invention first obtains basic data by collecting the methods such as gps data, calculate the travelling speed of the bus bus or train route section on bus special lane, then basic kernel function weight is tried to achieve by simpleMKL method, basic for difference kernel function weighting summation is set up preliminary multi-kernel learning SVM prediction model, again by weather, time, section, current road segment travelling speed, next section travelling speed is used as variable, be input to support vector machine to carry out training and testing and obtain multi-kernel learning SVM prediction model, finally utilize this model in a circuit any in public transport roadnet between bus vehicle website working time predict.2, the present invention adopts the convex combination of basic kernel function to set up forecast model, both SVM had been remained by kernel function by the feature space of input data-mapping to higher dimensional space, and then linear proximity finds mapping function, be applicable to advantage that is complicated and nonlinear problem, adopt single kernel function unlike SVM again, and the forecast model caused depends on the problem of Selection of kernel function unduly.3, input vector is mapped in the space of a higher-dimension by kernel function by the present invention, and the learning process of regression function is solved at this higher dimensional space, thus nonlinear problem is converted into linear problem, significantly improve the precision of the bus car Forecasting of Travel Time on public transportation lane; Compound kernel function is obtained additionally by by basic kernel function weighting summation, remain the characteristic of different IPs function, dissimilar data input can be processed better, this model is made to have stronger robustness, advance the construction process of intelligent transportation, all there is far-reaching influence to the Optimized Operation of the bus system on public transportation lane and the facility of passenger's trip.In view of above reason, the bus bus or train route section that the present invention can be widely used on bus special lane predicts field working time.
Accompanying drawing explanation
Fig. 1 is process schematic of the present invention
Fig. 2 is the bus running circuit example schematic on public transportation lane
Fig. 3 is that the present invention and neural net prediction method to predict the outcome comparison diagram the working time under different weather, Different periods
Embodiment
Below in conjunction with the drawings and specific embodiments, illustrate the present invention further, these examples should be understood only be not used in for illustration of the present invention and limit the scope of the invention, after having read the present invention, the amendment of those skilled in the art to the various equivalent form of value of the present invention has all fallen within the application's claims limited range.
As shown in Figure 1, the present invention includes following steps:
1) bus running line information, bus vehicle operating information and the bus GPS service data on public transportation lane is gathered;
Bus running line information on public transportation lane comprises bus running circuit mileage, bus stop position and website number and line condition.Weather conditions when bus vehicle operating information comprises h minute bus departure interval (h generally determines according to the situation of the volume of the flow of passengers on working line, line length and vehicle number, generally gets 1-5min), bus passenger flow peak period, bus vehicle operating and road traffic condition.Bus GPS service data comprise bus stop title, bus vehicle license plate numbering, bus vehicle order of classes or grades at school numbering k (k=1,2 ..., 30), bus vehicle frequency, bus vehicle sail website into and sail out of the website moment.
As shown in Figure 2,10 time points are set on this public transportation lane in this example, are divided into 9 sections by bus special lane, with these 9 sections for research object, predict the working time on these sections.
2) according to bus running line information, bus vehicle operating information and the bus GPS service data on the public transportation lane that obtains, the working time of acquisition bus vehicle between each section, and working time is normalized;
According to bus GPS service data, obtain the moment of vehicle through two end points in any section, the moment of these two end points is subtracted each other, obtain the working time of bus vehicle on this section, and will be normalized working time, make between data-mapping to 0 to 1 working time, thus eliminate the dimension impact between index, improve the prediction effect of multi-kernel learning SVM prediction model, its normalization formula is:
Wherein,
for the bus vehicle hour value after normalization, S
ifor original bus vehicle hour numerical value, S
minfor the minimum value of bus vehicle hour in sample sequence, S
maxfor the maximal value of bus vehicle hour in sample sequence.
3) choose some basic kernel functions and set up preliminary multi-kernel learning SVM prediction model, and calculate the weight shared by each kernel function;
Multiple basic kernel function is adopted to set up preliminary multi-kernel learning SVM prediction model
wherein, i (i=1,2,3 ..., l) represent the numbering of input vector; L represents the group number of input vector in sample data; x
irepresent the input vector x that working time, these 5 variablees were formed in the working time on the weather in i-th input vector (adopting step 2 in the present invention) the data obtained, time period, section, current road segment, next section
i); X is vector data to be predicted; B is intercept, is obtained by Training Support Vector Machines; K (x
i, x) be the convex combination of multiple basic kernel function, thus avoid and adopt single basic kernel function to bring the impact compared with big error to prediction, it embodies as follows:
Wherein, m=1,2 ..., M, M are total numbers of basic kernel function; d
mweight coefficient corresponding to each basic kernel function; K
mbe basic kernel function, this basic kernel function is corresponding in turn to gaussian kernel function, Polynomial kernel function, perceptron kernel function, Spline Kernel function, i.e. K
1(x
i, x) be gaussian kernel function, K
2(x
i, x) be Polynomial kernel function ..., the like.5 input variables are above mapped to the feature space of higher-dimension by nonlinear transformation, and carry out linear proximity in this space, find mapping function can well approach data-oriented sample, improve the precision of prediction.
for Lagrange multiplier,
obtain by solving following quadratic programming problem:
Wherein, α=(α
1, α
2..., α
l) and
be all Lagrange multiplier vector, ε is error threshold value, and C is penalty factor, and C>0; K (x
i, x
j) represent vector x
iand vector x
jhigher dimensional space is mapped to, y by kernel function K
irepresent the bus bus or train route section operation time of i-th group of sample.
Above-mentioned quadratic programming problem is changed into its dual problem, and substitutes into the convex combination K (x of multiple basic kernel function
i, x
j), problem can be expressed from the next:
Wherein, d
mrepresent kernel function K
m(x
i, x
j) corresponding weight coefficient, d=(d
1..., d
m..., d
m)
trepresent by d
m(m=1,2 ..., M) weight vectors that forms, computation process is as follows:
Order
So J (d) is to d
mlocal derviation be:
Make u be the sequence number of greatest member in weight vectors d, then:
Suppose D=(D
1, D
2..., D
m..., D
m)
tthe reduced gradient of J (d), wherein D
mbe calculated as follows
According to simpleMKL algorithm, vectorial D is utilized to solve the weight d of each basic kernel function
mthus by MKL-SVM (MultipleKernelLearning-SupportVectorMachine, multi-kernel learning support vector machine) be converted to standard SVM (SupportVectorMachine) quadratic programming problem, be convenient to solve with traditional sequential minimal optimization algorithm.
4) selecting step 2) weather in the data obtained, the time period, section, working time on current road segment, next section the input vector x that working time, these 5 variablees were formed
i, choose corresponding bus bus or train route section working time as y
i, and (x
i, y
i) form sample, this sample is divided into training sample and test sample book, for obtaining good training result, chooses 70% as training sample, all the other 30% as test sample book be used for check training result.Carry out Training Support Vector Machines with training sample set, obtain the parameter Ge Lang multiplier of forecast model
and intercept b, thus the multi-kernel learning SVM prediction model that is optimized.
Real data is utilized to obtain Ge Lang multiplier according to the above-mentioned QUADRATIC PROGRAMMING METHOD FOR that solves
(i=1,2 ..., l), and the computing method of intercept b are as follows:
Wherein, N
nSVfor standard support vector number, SV is support vector, is the input vector near plane in higher dimensional space, is automatically obtained by Training Support Vector Machines.Substitute into Lagrange's multiplier
intercept b and step 3) the weight d of each basic kernel function that obtains
m, obtain forecast model
5) test support vector machine by test sample book, evaluate test result RMSE (root-mean-squareerror, root-mean-square error), RMSE is less, then represent that predicated error is less, effect is unreasonable to be thought, concrete formula is as follows:
Wherein, H is number working time needing prediction,
for the operation time of prediction, the test sample book value that x (k) is working time.
If test result is unsatisfied with, in this example, RMSE > 3min is considered as being unsatisfied with, then returns step 3) again choose basic kernel function, build preliminary multi-kernel learning SVM prediction model; If test result is satisfied, then completes the foundation of forecast model, enter next step;
6) for working time between bus vehicle website in a circuit any in roadnet, adopt multi-kernel learning SVM prediction model to predict, obtain predicted value corresponding working time, its process is as follows:
If 1. this car is first car, due to the bus vehicle process of equal not this circuit in all sections, the up-to-date working time in section cannot be obtained, therefore adopt and only comprise time period x
1, weather x
2and section
three variablees form three-dimensional input vector, predict the virtual up-to-date working time in each section on public transportation lane by multi-kernel learning support vector machine; After vehicle crosses, upgrade the up-to-date working time in each section, as the input variable of next bus car;
If 2. this car is not first car, then by time period x
1, weather x
2and section
the working time of current road segment
the working time that next section is up-to-date
five variablees form five dimension input vectors, are input in multi-kernel learning SVM prediction model; To export between bus vehicle website predicted value y working time
k+1.
For verifying validity of the present invention further, by the present invention and Neural Network Prediction at fine day peak period (SP class), the fine day flat peak period (SO class), compare in peak period rainy day (RP class) and flat peak period rainy day (RO class) four kinds of situations, result is as shown in Figure 3: as seen from the figure, when in the situation of fine day peak, only have the result of section 7 neural network prediction algorithm less than multi-kernel learning SVM prediction model about 5 seconds, other 8 section the present invention predict the outcome little than neural network prediction resultant error, especially section 2 error is little nearly 15 seconds.In other three kinds of situations, multi-kernel learning SVM prediction model result RMSE is also less than neural network.Visible the present invention has higher precision, has stronger robustness, demonstrates the feasibility in the bus running time prediction of multi-kernel learning support vector machine on bus special lane.
The foregoing is only preferred embodiment of the present invention, not in order to limit the present invention, within the spirit and principles in the present invention all, any amendment done, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.
Claims (6)
1. bus bus or train route section Forecasting Methodology working time on public transportation lane, it comprises the following steps:
1) bus running line information, bus vehicle operating information and the bus GPS service data on bus special lane is gathered;
2) according to bus running line information, bus vehicle operating information and the bus GPS service data on the bus special lane that obtains, the working time of acquisition bus vehicle between each section, and working time is normalized;
3) choose some basic kernel functions and set up preliminary multi-kernel learning SVM prediction model, and calculate the weight shared by each kernel function;
4) adopt step 2) in the data obtained train preliminary multi-kernel learning SVM prediction model, obtain multi-kernel learning SVM prediction model;
5) test for multi-kernel learning SVM prediction model, adopt root-mean-square error RMSE evaluation test result, if dissatisfied, then return step 3) again choose basic kernel function, build preliminary multi-kernel learning SVM prediction model; If test result is satisfied, then enter next step;
6) for working time between bus vehicle website in a circuit any in roadnet, adopt multi-kernel learning SVM prediction model to predict, obtain predicted value corresponding working time.
2. bus bus or train route section Forecasting Methodology working time on a kind of public transportation lane as claimed in claim 1, it is characterized in that: described step 1) in, the bus running line information on public transportation lane comprises bus running circuit mileage, bus stop position and website number and line condition; Weather conditions when bus vehicle operating information comprises bus departure interval, bus passenger flow peak period, bus vehicle operating and road traffic condition; Bus GPS service data comprises bus stop title, bus vehicle license plate numbering, bus vehicle order of classes or grades at school numbering, bus vehicle frequency, bus vehicle sail website into and sail out of the website moment.
3. bus bus or train route section Forecasting Methodology working time on a kind of public transportation lane as claimed in claim 1, it is characterized in that: described step 2) in, according to bus GPS service data, obtain the moment of vehicle through two end points in any section, the moment of these two end points is subtracted each other, obtain the working time of bus vehicle on this section, and will be normalized working time, its normalization formula is:
Wherein,
for the bus vehicle hour value after normalization, S
ifor original bus vehicle hour numerical value, S
minfor the minimum value of bus vehicle hour in sample sequence, S
maxfor the maximal value of bus vehicle hour in sample sequence.
4. bus bus or train route section Forecasting Methodology working time on a kind of public transportation lane as claimed in claim 2, it is characterized in that: described step 2) in, according to bus GPS service data, obtain the moment of vehicle through two end points in any section, the moment of these two end points is subtracted each other, obtain the working time of bus vehicle on this section, and will be normalized working time, its normalization formula is:
Wherein,
for the bus vehicle hour value after normalization, S
ifor original bus vehicle hour numerical value, S
minfor the minimum value of bus vehicle hour in sample sequence, S
maxfor the maximal value of bus vehicle hour in sample sequence.
5. bus bus or train route section Forecasting Methodology working time on a kind of public transportation lane as claimed in claim 1 or 2 or 3 or 4, is characterized in that: described step 4) comprise following content: by step 2) in by the input vector x that working time, 5 variablees were formed in the working time on weather, time period, section, current road segment, next section
iwith bus section y working time of correspondence
ithe sample data of composition is divided into training sample and test sample book, and chooses 70% as training sample, all the other 30% test sample books being used for checking training result.
6. bus bus or train route section Forecasting Methodology working time on a kind of public transportation lane as claimed in claim 5, is characterized in that: described step 6) comprise following content:
If 1. this car is first car, adopts and only comprise time period x
1, weather x
2and section
three variablees form three-dimensional input vector, predict the virtual up-to-date working time in each section with multi-kernel learning SVM prediction model; After vehicle crosses, upgrade the up-to-date working time in each section, as the input variable of next bus car;
If 2. this car is not first car, then by time period x
1, weather x
2and section
the working time of current road segment
the working time that next section is up-to-date
these five variablees, as input vector, are input in multi-kernel learning SVM prediction model; To export between bus vehicle website predicted value y working time
k+1.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410609740.5A CN104318757B (en) | 2014-11-03 | 2014-11-03 | Bus bus or train route section Forecasting Methodology working time on a kind of public transportation lane |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410609740.5A CN104318757B (en) | 2014-11-03 | 2014-11-03 | Bus bus or train route section Forecasting Methodology working time on a kind of public transportation lane |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104318757A CN104318757A (en) | 2015-01-28 |
CN104318757B true CN104318757B (en) | 2016-04-20 |
Family
ID=52373981
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410609740.5A Expired - Fee Related CN104318757B (en) | 2014-11-03 | 2014-11-03 | Bus bus or train route section Forecasting Methodology working time on a kind of public transportation lane |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104318757B (en) |
Families Citing this family (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106845768B (en) * | 2016-12-16 | 2019-12-10 | 东南大学 | Bus travel time model construction method based on survival analysis parameter distribution |
CN107609698A (en) * | 2017-09-06 | 2018-01-19 | 上海享骑电动车服务有限公司 | Virtual stake |
CN107563566B (en) * | 2017-09-18 | 2020-08-11 | 东南大学 | Inter-bus-station operation time interval prediction method based on support vector machine |
JP6804792B2 (en) * | 2017-11-23 | 2020-12-23 | ベイジン ディディ インフィニティ テクノロジー アンド ディベロップメント カンパニー リミティッド | Systems and methods for estimating arrival time |
CN107945560A (en) * | 2017-12-21 | 2018-04-20 | 大连海事大学 | A kind of public transport smart electronics stop sign information display control method and system |
CN110361019B (en) * | 2018-04-11 | 2022-01-11 | 北京搜狗科技发展有限公司 | Method, device, electronic equipment and readable medium for predicting navigation time |
CN112309109B (en) * | 2019-08-01 | 2022-02-18 | 中移(苏州)软件技术有限公司 | Road traffic time prediction method, device and storage medium |
CN111161536B (en) * | 2019-12-25 | 2021-04-02 | 南京行者易智能交通科技有限公司 | Time interval and road section selection method, device and system for bus lane |
Family Cites Families (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP3205239B2 (en) * | 1995-11-28 | 2001-09-04 | 株式会社日立製作所 | Traffic providing information propagation monitoring method and device |
JP2006154962A (en) * | 2004-11-25 | 2006-06-15 | Sumitomo Electric Ind Ltd | Method for predicting travel time using status space expression method |
CN101388143B (en) * | 2007-09-14 | 2011-04-13 | 同济大学 | Bus arriving time prediction method based on floating data of the bus |
CN102074124B (en) * | 2011-01-27 | 2013-05-08 | 山东大学 | Dynamic bus arrival time prediction method based on support vector machine (SVM) and H-infinity filtering |
CN102663224A (en) * | 2012-03-07 | 2012-09-12 | 吉首大学 | Comentropy-based integrated prediction model of traffic flow |
CN103295414B (en) * | 2013-05-31 | 2016-01-20 | 北京建筑大学 | A kind of bus arrival time Forecasting Methodology based on magnanimity history GPS track data |
CN103678917B (en) * | 2013-12-13 | 2016-11-23 | 杭州易和网络有限公司 | A kind of real-time arrival time Forecasting Methodology of public transport based on simulated annealing |
-
2014
- 2014-11-03 CN CN201410609740.5A patent/CN104318757B/en not_active Expired - Fee Related
Also Published As
Publication number | Publication date |
---|---|
CN104318757A (en) | 2015-01-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104318757B (en) | Bus bus or train route section Forecasting Methodology working time on a kind of public transportation lane | |
CN105046956B (en) | Traffic flow simulating and predicting method based on turning probability | |
CN104157139B (en) | A kind of traffic congestion Forecasting Methodology and method for visualizing | |
Zhao et al. | Truck traffic speed prediction under non-recurrent congestion: Based on optimized deep learning algorithms and GPS data | |
CN105118294B (en) | A kind of Short-time Traffic Flow Forecasting Methods based on state model | |
CN109670277A (en) | A kind of travel time prediction method based on multi-modal data fusion and multiple-model integration | |
CN106710215B (en) | Bottleneck upstream lane grade traffic status prediction system and implementation method | |
CN109523786A (en) | A kind of traffic accidents genetic analysis method | |
CN106384509A (en) | Urban road driving time distribution estimation method considering taxi operation states | |
CN104821080A (en) | Intelligent vehicle traveling speed and time predication method based on macro city traffic flow | |
CN106228808A (en) | City expressway travel time prediction method based on Floating Car space-time grid data | |
CN105469602B (en) | A kind of Forecasting Methodology of the bus passenger waiting time scope based on IC-card data | |
CN104809112A (en) | Method for comprehensively evaluating urban public transportation development level based on multiple data | |
CN105513337A (en) | Passenger flow volume prediction method and device | |
CN105405293A (en) | Short-term prediction method of road travel time and system | |
CN106529608A (en) | Comprehensive stationing system for exhaust gas remote-measuring equipment of motor vehicle | |
CN102890862B (en) | Traffic condition analyzing device and method based on vector mode | |
CN105206051B (en) | A kind of city road network Rationality Assessment method and device | |
CN110188797A (en) | A kind of intelligent automobile method for rapidly testing based on Bayes's optimization | |
Yu et al. | Bus travel-time prediction based on bus speed | |
CN109191849A (en) | A kind of traffic congestion Duration Prediction method based on multi-source data feature extraction | |
CN101930670A (en) | Method for predicting social vehicle running time on bus travel road section | |
Shah et al. | Relationship between road traffic features and accidents: An application of two-stage decision-making approach for transportation engineers | |
CN104900057A (en) | City expressway main and auxiliary road floating vehicle map matching method | |
CN107220724A (en) | Passenger flow forecast method and device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20160420 Termination date: 20161103 |