CN105139328A - Travel time real-time prediction method facing license plate data identification and device - Google Patents
Travel time real-time prediction method facing license plate data identification and device Download PDFInfo
- Publication number
- CN105139328A CN105139328A CN201510518220.8A CN201510518220A CN105139328A CN 105139328 A CN105139328 A CN 105139328A CN 201510518220 A CN201510518220 A CN 201510518220A CN 105139328 A CN105139328 A CN 105139328A
- Authority
- CN
- China
- Prior art keywords
- hourage
- time interval
- time
- module
- rate
- 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
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a travel time real-time prediction method facing license plate data identification and a device. The travel time real-time prediction method comprises three steps of exploration of prior rules, practical measurement and calculation of travel time and prediction calculation of travel time. The invention overcomes the problem that the data scale and effectiveness are limited, the prediction calculation respond is delayed and the accuracy is not high. The method based on the license plate identification data and the device reduce the calculation respond time , improve the prediction accuracy, can be realized in an Apache Storm and Hadoop MapReduce cluster environment, finish travel time prediction in the real-time data environment, and can be used for the road state monitoring and travel service publication in the traffic field. The invention greatly improves the instantaneity and reliability of smart traffic application in the big data environment, provides the traffic big data information to the user for real-time inquiry and prediction, and improves the user experience.
Description
Technical field
The present invention relates to areas of information technology and intelligent transportation field, specifically, relate to and use described method in ApacheStorm and HadoopMapReduce cluster environment, complete the real-time estimate technology of predicting travel time, and a kind of method and apparatus of real-time estimate road trip time.
Background technology
Road trip time is the traffic circulation status information of extensive concern the most in field of traffic, compared with other such as place traffic parameter, can the unimpeded degree of evaluation path better, and the conevying efficiency of reflection road, embodies road traffic congestion state.Along with the expansion of city size, the traffic congestion in rush hour becomes normality, how to issue the trip service of road ahead traffic congestion, to the variation tendency of traveler prediction future time instance condition of road surface, being guided out passerby selects Optimization route to arrive destination, has become the problem that colleague both at home and abroad falls over each other to study.Road trip time is predicted, is traffic flow guidance system primary study content, is considered to prediction modern city congested problem one of effective means the most.So-called hourage, typically refer to road trip time all in system-wide net; And road trip time refers to that certain section is all through the mean value of vehicle by the time in preset time interval.For vehicle supervision department, by road conditions can be passed judgment on hourage and optimize line design.Participate in the common people for traffic, care is the time that current time experiences from destination, arrival front, present position, and this is also just the business implication contained hourage.
For traffic information, current techniques and method can make frequency acquisition between 30 seconds to 2 minutes, Information issued frequency was at 5 minutes to about 10 minutes, and the prediction of hourage as complicated aggregate operation, calculated rate is also many at several tens minutes, and this and growing transport information instantaneity demand create contradiction.Basic data according to predicting travel time is originated, and divides following two classes: (1) is based on the Forecasting Methodology of mobile model traffic information collection technology by present Research domestic and international hourage.Such as, towards the GPS location data that Floating Car gathers, can Kalman filter model be passed through, realize actual measurement hourage and predicting travel time computing.But the sample size due to Floating Car collection is limited and data are more single, causes predicting travel time precision not high.(2) based on the Forecasting Methodology of fixed traffic information collection technology.Using the parameter such as the magnitude of traffic flow, density, speed, time occupancy, space occupancy of traffic detector collection as the input variable of predicting travel time.Such as, towards the car plate data set of key crossing identification camera, can realize actual measurement and the prediction and calculation of hourage, have higher precision of prediction, real-time is better simultaneously.
Also the prediction that some technical schemes realize hourage is disclosed in prior art.As the patent documentation " a kind of method of predicting travel time " that application number is CN200910083285; Application number is the patent documentation " acquisition methods of hourage and device, prognoses system " of CN201310227307; Application number is the patent documentation " Forecasting Methodology of expressway travel time and device " of CN201310227309; Application number is the patent documentation " a kind of quick and precisely prediction vehicle is by the method for road trip time " of CN201410270534; The patent documentation " a kind of system and method predicting hourage needed for the future time instance of through street " etc. that application number is CN201410291290.The ins and outs analyzed wherein are known, and the research of the real-time estimate of hourage is still in developing stage, and technology is still immature.The problem of three aspects below main existence or defect:
First aspect, work at present is mostly towards limited history data set.Such as, as the monitor data of Floating Car in a couple of days, several months, and non real-time monitor data.It is not enough that this brings complicacy to consider to current huge traffic data, causes the low or inquiry for the treatment of effeciency under large data environment slowly, limited to the directive significance of recent predicting travel time.
Second aspect, work at present realizes prediction based on mobile model transport information mostly.Such as, for the vehicle GPS data of vehicle of travelling frequently from bus, highway, but not the fixed transport information that accuracy of identification is higher.This makes vehicle limited coverage area, and effective time is discontinuous, causes the prediction effect limited by practical under large data environment.
The third aspect, does not manyly consider the feature of vehicle as mobile object when current method calculates, just a kind of specific in line computation or off-line operation.In fact, no matter which kind of vehicle image data, necessarily has the information of Time and place two kinds of dimensions simultaneously; And on these two the separate attributes, there is law and stream often.Also namely, first there is the demand of instantaneity hourage, require that calculating can return results as early as possible, secondly, also there is the demand of reflection recent trend hourage, if this kind of rule in historical data can not be taken into account, certainly will the limited precision that predicts the outcome be caused.
For the problems referred to above in correlation technique, at present effective solution is not yet proposed up to now.
Summary of the invention
The object of the invention is to overcome above-mentioned technological deficiency, thus solve the real-time estimate efficiency and the not high problem of accuracy hourage towards license plate identification data.
The present invention is interval and section rule base by definition time, a kind of real-time predicting method hourage towards license plate identification data is proposed, the summary responses of calculating can be ensured on the one hand based on two benches spatial and time parallelism technology towards real-time flow data, on the other hand can towards the mass historical data of accumulation based on naive Bayesian timing mining rule, and in conjunction with the low delay of HadoopMapReduce on ApacheStorm, achieve actual measurement and prediction and calculation.
Hadoop is a kind of distributed system base frame, achieve a distributed file system, MapReduce is a kind of programming model be structured on Hadoop file system, for the concurrent operation of massive data sets, HadoopMapReduce has general unique implication in areas of information technology, is known in those skilled in the art.ApacheStorm is a large data handling system of being increased income by Twitter, and for the real-time calculating of the large data of streaming, it has general unique implication in areas of information technology, is known in those skilled in the art.
Specifically, one aspect of the present invention, provide a kind of real-time predicting method hourage towards license plate identification data, it is characterized in that, described method comprises:
(1) priori rules excavation step, for building priori rules storehouse;
(2) hourage Actual measurement step, for calculating the road trip time measured value in current time road network, and accumulation measured result collection;
(3) predicting travel time calculation procedure, for predicting the road trip time rate of change in subsequent time road network, and calculates corresponding predicting travel time value.
Preferably, wherein said priori rules excavation step, is excavate based on the priori rules of naive Bayesian theory, comprises:
(1) priori rules of section attribute is excavated, and measured result collection hourage of input magnanimity history, exports section attribute priori rules storehouse;
(2) priori rules of time interval attribute is excavated, measured result collection hourage of input magnanimity history, output time Range Attributes priori rules storehouse.
Preferably, towards real-time predicting method hourage of license plate identification data, priori rules excavation step wherein, for section attribute, excavates section attribute priori rules storehouse; This excavation section attribute priori rules storehouse process is towards measured result collection hourage, its section attribute is divided, by time interval sequentially, value hourage of certain time interval under each section, rate of change is calculated with the value of next time interval, put under the classification of corresponding rate of change, added up the distribution situation of rate of change classification hourage in each section; This excavation section attribute priori rules storehouse process, under given rate of change classification hourage, calculated rate, time interval unit, can be described as following two MapReduce job steps:
(1) first operation, with under given calculated rate and time interval unit, divides measured result collection according to section attribute, sorts out the time interval under respective stretch, and value hourage under this time interval;
(2) second operations, with under given calculated rate and time interval unit, excavate the Changing Pattern of section attribute lower hourage, according to rate of change classification hourage of specifying, add up each classification under each section hourage rate of change distribution.
Preferably, described real-time predicting method hourage towards license plate identification data, priori rules excavation step wherein, for time interval attribute, excavates time interval attribute priori rules storehouse; This excavation time interval attribute priori rules storehouse process is towards measured result collection hourage, time interval in one day is divided as major key section, value hourage of certain time interval under each section, rate of change is calculated with the value of next time interval, put under the classification of corresponding rate of change, added up the distribution situation of rate of change classification hourage in each section; This excavation time interval attribute priori rules storehouse process, under given rate of change classification hourage, calculated rate, time interval unit, can be described as following two MapReduce job steps:
(1) first operation, with under given calculated rate and time interval unit, divides measured result collection according to section attribute, sorts out the time interval under respective stretch, and value hourage under this time interval;
(2) second operations, with under given calculated rate and time interval unit, excavate the Changing Pattern of time interval lower hourage, according to specify hourage rate of change classification, add up each classification under each time interval hourage rate of change distribution.
Preferably, described real-time predicting method hourage towards license plate identification data, wherein said Actual measurement step hourage, for calculating the road trip time in current time road network, and builds up measured result collection; This process is divided into following two stages, and the output of previous stage is using data-stream form as the input of the latter half:
(1) bicycle calculation stages hourage: under preset time interval δ, to sort out, for each vehicle v by car plate with given frequency
i, calculate this car in all bicycle hourages through section
(2) road trip time calculation stages: under preset time interval δ, to sort out, for each section s by section with given frequency
j, gather all bicycle hourages through vehicle, calculate road trip time tras
j.
Preferably, predicting travel time calculation procedure, for predicting rate of change hourage of current road segment subsequent time, and calculates corresponding predicted value; This step performs after Actual measurement completes; When Actual measurement obtains section i at time interval δ
jmeasured value tra after, section i is at time interval δ
j+1predicting travel time value, be calculated by following two steps to obtain:
(1) the time interval attribute priori rules storehouse described in utilization, and described section attribute priori rules storehouse, predict the interval δ of the future time of same section i by Bayes's condition probability formula
j+1under the classification x at rate of change △ hourage place;
(2) according to specify hourage rate of change classification, to the rate of change △ ' that each class definition is predicted, calculate δ
j+1predicted value tra '=(1+ △ ') * tra of interval lower hourage.
In addition, on the other hand, present invention also offers a kind of real-time estimate device hourage towards license plate identification data, specifically, this device can be used for realizing real-time predicting method hourage as above, and this device comprises as lower component:
(1) data memory module, for storing measured result collection hourage, time interval attribute priori rules storehouse, section attribute priori rules storehouse and predicting travel time result set;
(2) online computing module, for according to given calculated rate, calculates measured value hourage in current time interval, and the predicting travel time value in future time interval and in internal memory buffer memory;
(3) calculated off-line module, for according to given calculated rate, is excavated by priori rules, upgrades and also exports section attribute priori rules storehouse to data memory module, and upgrades and to data memory module output time Range Attributes priori rules storehouse;
(4) human-computer interaction module, for realizing the interactive operation of user and this real-time estimate device.
Preferably, data memory module is the storage realized based on distributed file system; This module stores measured result hourage collection, time interval attribute priori rules storehouse, section attribute priori rules storehouse and predicting travel time result set; This module is connected with calculated off-line module, for calculated off-line module is provided as the measured result collection and be stored as time interval attribute priori rules storehouse, the section attribute priori rules storehouse of output hourage of input; This module is connected with online computing module, for online computing module is provided as time interval attribute priori rules storehouse, the section attribute priori rules storehouse of basic data; This module is connected with human-computer interaction module, for human-computer interaction module provide hourage measured result collection and predicting travel time result set;
Preferably, online computing module input license plate identification data stream, according to given calculated rate, by the Actual measurement step of the above-mentioned hourage towards license plate identification data described in real-time predicting method, calculates measured value hourage in current time interval; This module is connected with data memory module, reads section attribute priori rules storehouse and time interval attribute priori rules storehouse, data maintenance based in internal memory; The measured result that this module exports, sends to calculated off-line module as historical data, forms measured result collection hourage; The measured result that this module exports, also can for secondary development as data stream; This module, according to given calculated rate, by the prediction and calculation step of the above-mentioned hourage towards license plate identification data in real-time predicting method, calculates the predicting travel time value in future time interval, and in internal memory buffer memory; Only need for a section predicted value safeguarding future time interval in internal memory; What this module exported predicts the outcome, and sends to data memory module as historical data, forms predicting travel time result set; This module is realized by the Topology of 1 in ApacheStorm software (topological structure), wherein, in each stage in two stages of the above-mentioned hourage towards license plate identification data in real-time predicting method, a Bolt (node) respectively as this Topology (topological structure) realizes; This module is connected with human-computer interaction module, obtains the parameter of user's input, the parameter of configuration Actual measurement step, and these parameters comprise the frequency of calculating, the time interval of calculating;
Preferably, what calculated off-line module inputted is measured result collection hourage, be connected with data memory module, according to given calculated rate, by the priori rules excavation step of the above-mentioned hourage towards license plate identification data described in real-time predicting method, upgrade and also export section attribute priori rules storehouse to data memory module, and to upgrade and to data memory module output time Range Attributes priori rules storehouse; This module is realized by 3 operations (Job) of HadoopMapReduce software systems, wherein, priori rules based on section attribute is excavated and is realized by 2 operations (Job), priori rules based on time interval attribute is excavated and is also realized by 2 operations (Job), because first operation (Job) of two mining processes is identical, therefore can be shared; This module is connected with human-computer interaction module, obtains the parameter of user's input, configures the parameter of the priori rules excavation step of the section attribute in each operation, comprise rate of change classification hourage, calculated rate, time interval unit.
Preferably, human-computer interaction module is connected with data memory module, for user provides interactive interface, supports user input query parameter, the section that described parameter comprises time interval, specifies, to present the measured result and predicting the outcome hourage of inquiry in Web page; This module is connected with online computing module, and for user provides interactive interface, support user's input parameter, described parameter comprises the frequency of calculating, the time interval of calculating, for being configured in the calculating parameter of line computation module; This module is connected with calculated off-line module, for user provides interactive interface, support that user inputs calculating parameter, described parameter comprises rate of change classification hourage of the priori rules excavation step of the section attribute in each operation, calculated rate, time interval unit.
Preferably, described priori rules is excavated, and for section attribute, excavates section attribute priori rules storehouse; This process is towards measured result collection hourage, its section attribute is divided, by time interval sequentially, value hourage of certain time interval under each section, rate of change is calculated with the value of next time interval, put under the classification of corresponding rate of change, added up the distribution situation of rate of change classification hourage in each section; This process, under given rate of change classification hourage, calculated rate, time interval unit, can be described as following two MapReduce job steps:
(1) first operation, with under given calculated rate and time interval unit, divides measured result collection according to section attribute, sorts out the time interval under respective stretch, and value hourage under this time interval;
(2) second operations, with under given calculated rate and time interval unit, excavate the Changing Pattern of section attribute lower hourage, according to rate of change classification hourage of specifying, add up each classification under each section hourage rate of change distribution.
Preferably, described priori rules is excavated, and for time interval attribute, excavates time interval attribute priori rules storehouse; This process is towards measured result collection hourage, time interval in one day is divided as major key section, value hourage of certain time interval under each section, rate of change is calculated with the value of next time interval, put under the classification of corresponding rate of change, added up the distribution situation of rate of change classification hourage in each section; This process, under given rate of change classification hourage, calculated rate, time interval unit, can be described as following two MapReduce job steps:
(1) first operation, with under given calculated rate and time interval unit, divides measured result collection according to section attribute, sorts out the time interval under respective stretch, and value hourage under this time interval;
(2) second operations, with under given calculated rate and time interval unit, excavate the Changing Pattern of time interval lower hourage, according to specify hourage rate of change classification, add up each classification under each time interval hourage rate of change distribution.
Two first operations in above-mentioned are identical, and two job steps, in the priori rules for time attribute and section attribute is excavated, can be shared completely, are also that calculated off-line module realizes indeed through 3 job steps.
Preferably, hourage, Actual measurement step, for calculating the road trip time in current time road network, and built up measured result collection, and this process is divided into following two stages, and the output of previous stage is using data-stream form as the input of the latter half:
(1) bicycle calculation stages hourage: under preset time interval δ, to sort out, for each vehicle v by car plate with given frequency
i, calculate this car in all bicycle hourages through section
(2) road trip time calculation stages: under preset time interval δ, to sort out, for each section s by section with given frequency
j, gather all bicycle hourages through vehicle, calculate road trip time tras
j.
Preferably, predicting travel time calculation procedure, for predicting rate of change hourage of current road segment subsequent time, and calculates corresponding predicted value, when Actual measurement obtains section i at time interval δ
jmeasured value tra after, section i is at time interval δ
j+1predicting travel time value, be calculated by following two steps to obtain:
(1) utilize described time interval attribute priori rules storehouse and described section attribute priori rules storehouse, predict the interval δ of the future time of same section i by Bayes's condition probability formula
j+1under the classification x at rate of change △ hourage place;
(2) according to specify hourage rate of change classification, to the rate of change △ ' that each class definition is predicted, calculate δ
j+1predicted value tra '=(1+ △ ') * tra of interval lower hourage.
Be not difficult to find out by technique scheme, the present invention has following beneficial effect:
1, utilize license plate identification data, as vehicle location property and timeliness hourage record according to one of, have cover extensively, position accurately and the feature of Time Continuous, effectively can improve the accuracy of predicting travel time;
2, by time forecasting methods hourage of efficiently and accurately, substantially increase real-time estimate efficiency hourage, can respond accurately in real time, relevant information service is become in field of traffic can wide popularization and application, improve practicality, solve existing time prediction technology is difficult to practical application problem due to problems such as response time delayed and accuracy rate are not high.
Accompanying drawing explanation
The description that the present invention can carry out with reference to following Figure and being better understood, and in all of the figs, employ identical or similar Reference numeral to identify.Described accompanying drawing comprises in this manual together with detailed description below and forms the part of this instructions, and is used for illustrating preferred embodiment of the present invention further and explaining principle and advantage of the present invention.
Fig. 1 be of the present invention towards license plate identification data hourage real-time estimate device Organization Chart;
Fig. 2 is the computation process that the priori rules of section attribute of the present invention is excavated;
Fig. 3 is the computation process that the priori rules of time interval attribute of the present invention is excavated;
Fig. 4 is Actual measurement step and predicting travel time calculation procedure hourage of the present invention.
Embodiment
For making the technical problem to be solved in the present invention, technical scheme and advantage clearly, be described in detail below in conjunction with the accompanying drawings and the specific embodiments.Those skilled in the art should know, following specific embodiment or embodiment, to be the present invention be explains the set-up mode of the series of optimum that concrete summary of the invention is enumerated further, and all can be combined with each other or interrelated use between those set-up modes, cannot carry out associating with other embodiment or embodiment and arrange unless clearly proposed wherein some or a certain specific embodiment or embodiment in the present invention or jointly use.Meanwhile, following specific embodiment or embodiment only as optimized set-up mode, and not as limiting the understanding of protection scope of the present invention.
Embodiment 1
The invention provides a kind of real-time estimate device hourage towards license plate identification data, mainly comprise four parts: data memory module, online computing module, calculated off-line module, human-computer interaction module.Below with reference to the accompanying drawings 1 modules is described in detail.
Data memory module: this module is the storage realized based on distributed file system; This module stores measured result hourage collection, time interval attribute priori rules storehouse, section attribute priori rules storehouse and predicting travel time result set; This module is connected with calculated off-line module, for calculated off-line module is provided as the measured result collection and be stored as time interval attribute priori rules storehouse, the section attribute priori rules storehouse of output hourage of input; This module is connected with online computing module, for online computing module is provided as time interval attribute priori rules storehouse, the section attribute priori rules storehouse of basic data; This module is connected with human-computer interaction module, for human-computer interaction module provide hourage measured result collection and predicting travel time result set;
Online computing module: this module input license plate identification data stream, according to given calculated rate, by Actual measurement step, calculates measured value hourage in current time interval; This module is connected with data memory module, reads section attribute priori rules storehouse and time interval attribute priori rules storehouse, data maintenance based in internal memory; The measured result that this module exports, sends to calculated off-line module as historical data, forms measured result collection hourage; The measured result that this module exports, also can for secondary development as data stream; This module, according to given calculated rate, by prediction and calculation step, calculates the predicting travel time value in future time interval, and in internal memory buffer memory; Only need for a section predicted value safeguarding future time interval in internal memory; What this module exported predicts the outcome, and sends to data memory module as historical data, forms predicting travel time result set; This module is realized by 1 Topology (topological structure) of ApacheStorm software, wherein, hourage Actual measurement step each stage in two stages, a Bolt (node) respectively as this Topology (topological structure) realizes; This module is connected with human-computer interaction module, obtains the parameter of user's input, configures the parameter of described Actual measurement step, comprise the frequency of calculating, the time interval of calculating;
Above-mentioned Actual measurement step hourage, for calculating the road trip time in current time road network, and builds up measured result collection, and this process is divided into following two stages, and the output of previous stage is using data-stream form as the input of the latter half:
(1) bicycle calculation stages hourage: under preset time interval δ, to sort out, for each vehicle v by car plate with given frequency
i, calculate this car in all bicycle hourages through section
(2) road trip time calculation stages: under preset time interval δ, to sort out, for each section s by section with given frequency
j, gather all bicycle hourages through vehicle, calculate road trip time tras
j.
Above-mentioned predicting travel time calculation procedure, for predicting rate of change hourage of current road segment subsequent time, and calculates corresponding predicted value, when Actual measurement obtains section i at time interval δ
jmeasured value tra after, section i is at time interval δ
j+1predicting travel time value, be calculated by following two steps to obtain:
(1) utilize described time interval attribute priori rules storehouse and described section attribute priori rules storehouse, predict the interval δ of the future time of same section i by Bayes's condition probability formula
j+1under the classification x at rate of change △ hourage place;
(2) according to specify hourage rate of change classification, to the rate of change △ ' that each class definition is predicted, calculate δ
j+1predicted value tra '=(1+ △ ') * tra of interval lower hourage.
Calculated off-line module: what this module inputted is measured result collection hourage, be connected with data memory module, according to given calculated rate, by priori rules excavation step, upgrade and also export section attribute priori rules storehouse to data memory module, and to upgrade and to data memory module output time Range Attributes priori rules storehouse; This module is realized by 3 Job (operation) of HadoopMapReduce, wherein, priori rules based on section attribute is excavated and is realized by 2 Job (operation), priori rules based on time interval attribute is excavated and is also realized by 2 Job (operation), because first Job (operation) of two mining processes is identical, therefore can be shared; This module is connected with human-computer interaction module, obtains the parameter of user's input, configures the parameter of the priori rules excavation step of the section attribute in each operation, comprise rate of change classification hourage, calculated rate, time interval unit.
In a concrete embodiment, this priori rules excavation step, is excavate based on the priori rules of naive Bayesian theory, comprises:
(1) priori rules of section attribute is excavated, and measured result collection hourage of input magnanimity history, exports section attribute priori rules storehouse;
(2) priori rules of time interval attribute is excavated, measured result collection hourage of input magnanimity history, output time Range Attributes priori rules storehouse.
In a concrete embodiment, priori rules excavation step wherein, for section attribute, excavates section attribute priori rules storehouse; This process is towards measured result collection hourage, its section attribute is divided, by time interval sequentially, value hourage of certain time interval under each section, rate of change is calculated with the value of next time interval, put under the classification of corresponding rate of change, added up the distribution situation of rate of change classification hourage in each section; This process, under given rate of change classification hourage, calculated rate, time interval unit, can be described as following two MapReduce job steps:
(1) first operation, with under given calculated rate and time interval unit, divides measured result collection according to section attribute, sorts out the time interval under respective stretch, and value hourage under this time interval;
(2) second operations, with under given calculated rate and time interval unit, excavate the Changing Pattern of section attribute lower hourage, according to rate of change classification hourage of specifying, add up each classification under each section hourage rate of change distribution.
In a concrete embodiment, priori rules excavation step wherein, for time interval attribute, excavates time interval attribute priori rules storehouse; This process is towards measured result collection hourage, time interval in one day is divided as major key section, value hourage of certain time interval under each section, rate of change is calculated with the value of next time interval, put under the classification of corresponding rate of change, added up the distribution situation of rate of change classification hourage in each section; This process, under given rate of change classification hourage, calculated rate, time interval unit, can be described as following two MapReduce job steps:
(1) first operation, with under given calculated rate and time interval unit, divides measured result collection according to section attribute, sorts out the time interval under respective stretch, and value hourage under this time interval;
(2) second operations, with under given calculated rate and time interval unit, excavate the Changing Pattern of time interval lower hourage, according to specify hourage rate of change classification, add up each classification under each time interval hourage rate of change distribution
In a concrete embodiment, this human-computer interaction module: this module is connected with data memory module, for user provides interactive interface, support user input query parameter, the section that described parameter comprises time interval, specifies, to present the measured result and predicting the outcome hourage of inquiry in Web page; This module is connected with online computing module, and for user provides interactive interface, support user's input parameter, described parameter comprises the frequency of calculating, the time interval of calculating, for being configured in the calculating parameter of line computation module; This module is connected with calculated off-line module, for user provides interactive interface, support that user inputs calculating parameter, described parameter comprises rate of change classification hourage of the priori rules excavation step of the section attribute in each operation, calculated rate, time interval unit.
Embodiment 2
Composition graphs 2 basic procedure is described the computation process that the priori rules of section attribute is excavated.In a concrete embodiment, the priori rules mining process of section attribute, can be described as following two MapReduce job steps:
(1) first operation, with under given calculated rate and time interval unit, divides measured result collection according to section attribute, sorts out the time interval under respective stretch, and value hourage under this time interval; Wherein, the Map stage (i.e. mapping phase) is loaded into measured result collection hourage, and divide measured result collection hourage according to section attribute, obtaining with section is major key, measured value set hourage of sorting in chronological order; The Reduce stage (i.e. merging phase) is integrated according to time interval sequencing, and obtaining with section is major key, measured value set hourage of temporally interval sequence;
(2) second operations, be loaded into the result of first operation, with under given calculated rate and time interval unit, excavate the Changing Pattern of section attribute lower hourage, according to rate of change classification hourage of specifying, add up each classification under each section hourage rate of change distribution; Wherein, the Map stage (i.e. mapping phase) is loaded into the result of first operation, calculates adjacent time rate of change interval hourage, and this value is brought in the classification of rate of change affiliated hourage; The Reduce stage (i.e. merging phase), for each rate of change classification hourage, counts each distribution of classifying under each section.
Embodiment 3
Composition graphs 3 basic procedure is described the computation process that the priori rules of time interval attribute is excavated.In a concrete embodiment, the priori rules mining process of time interval attribute, can be described as following two MapReduce job steps:
(1) first operation, with under given calculated rate and time interval unit, divides measured result collection according to section attribute, sorts out the time interval under respective stretch, and value hourage under this time interval; Wherein, the Map stage (i.e. mapping phase) is loaded into measured result collection hourage, and divide measured result collection hourage according to section attribute, obtaining with section is major key, measured value set hourage of sorting in chronological order; The Reduce stage (i.e. merging phase) is integrated according to time interval sequencing, and obtaining with section is major key, measured value set hourage of temporally interval sequence;
(2) second operations, with under given calculated rate and time interval unit, excavate the Changing Pattern of time interval lower hourage, according to specify hourage rate of change classification, add up each classification under each time interval hourage rate of change distribution; Wherein, the Map stage (i.e. mapping phase) calculates rate of change hourage in adjacent time interval, and this value is included in the classification of rate of change affiliated hourage, to time interval Further Division; The Reduce stage (i.e. merging phase), for each of rate of change classification, counts distribution of each classification under each time interval hourage.
Embodiment 4
Composition graphs 4 basic procedure to hourage Actual measurement step and predicting travel time calculation procedure be described.In a concrete embodiment, hourage, Actual measurement step, for calculating the road trip time in current time road network, and built up measured result collection; This process is divided into following two stages, and the output of previous stage is using data-stream form as the input of the latter half:
(1) bicycle calculation stages hourage: with given frequency, sort out by car plate, for each vehicle, calculate this car in all bicycle hourages through section under preset time interval;
(2) road trip time calculation stages: with given frequency, sort out by section, for each section, gather all bicycle hourages through vehicle under preset time interval, calculate road trip time.
Predicting travel time calculation procedure afterwards, for predicting rate of change hourage of current road segment subsequent time, and calculates corresponding predicted value; This step performs after Actual measurement according to claim 5 completes; When Actual measurement obtains section i at time interval δ
jmeasured value tra after, section i is at time interval δ
j+1predicting travel time value, be calculated by following two steps to obtain:
(1) utilize the time interval attribute priori rules storehouse described in claim 3, and utilize the section attribute priori rules storehouse described in claim 4, predict the interval δ of the future time of same section i by Bayes's condition probability formula
j+1under the classification x at rate of change △ hourage place;
(2) according to specify hourage rate of change classification, to the rate of change △ ' that each class definition is predicted, calculate δ
j+1predicted value tra '=(1+ △ ') * tra of interval lower hourage.
Be not difficult to find out by above-mentioned specific embodiment and embodiment, beneficial effect of the present invention is:
1, utilize car plate data, as one of the location property of vehicle and the foundation of timeliness record hourage, and these time datas are all obtained by field survey, effectively improve the accuracy of the time data of predicting travel time institute foundation;
2, by correct time prediction algorithm, substantially increase real-time estimate efficiency hourage, make the data existing hourage based on Real-time Obtaining, just can carry out the prediction of vehicle journeys time in real time, time prediction is become in vehicle drive can the method for wide popularization and application, improve practicality, solve existing time prediction technology is difficult to practical application problem due to problems such as arithmetic speeds.
The above is the preferred embodiment of the present invention; it should be pointed out that for those skilled in the art, under the prerequisite not departing from principle of the present invention; can also make some improvements and modifications, these improvements and modifications also should be considered as protection scope of the present invention.
Claims (11)
1. towards real-time predicting method hourage for license plate identification data, it is characterized in that, described method comprises:
(1) priori rules excavation step, for building priori rules storehouse;
(2) hourage Actual measurement step, for calculating the road trip time measured value in current time road network, and accumulation measured result collection;
(3) predicting travel time calculation procedure, for predicting the road trip time rate of change in subsequent time road network, and calculates corresponding predicting travel time value.
2. real-time predicting method hourage towards license plate identification data according to claim 1, is characterized in that, wherein said priori rules excavation step, is to excavate based on the priori rules of naive Bayesian theory, comprising:
(1) priori rules of section attribute is excavated, and measured result collection hourage of input history, exports section attribute priori rules storehouse;
(2) priori rules of time interval attribute is excavated, measured result collection hourage of input history, output time Range Attributes priori rules storehouse.
3. real-time predicting method hourage towards license plate identification data according to claim 2, is characterized in that, described priori rules excavation step, for section attribute, excavates section attribute priori rules storehouse; This excavation section attribute priori rules storehouse process is towards measured result collection hourage, its section attribute is divided, by time interval sequentially, value hourage of certain time interval under each section, rate of change is calculated with the value of next time interval, put under the classification of corresponding rate of change, added up the distribution situation of rate of change classification hourage in each section; This excavation section attribute priori rules storehouse process, under given rate of change classification hourage, calculated rate, time interval unit, can be described as following two MapReduce job steps:
(1) first operation, with under given calculated rate and time interval unit, divides measured result collection according to section attribute, sorts out the time interval under respective stretch, and value hourage under this time interval;
(2) second operations, with under given calculated rate and time interval unit, excavate the Changing Pattern of section attribute lower hourage, according to rate of change classification hourage of specifying, add up each classification under each section hourage rate of change distribution.
4. real-time predicting method hourage towards license plate identification data according to claim 2, is characterized in that, described priori rules excavation step, for time interval attribute, excavates time interval attribute priori rules storehouse; This excavation time interval attribute priori rules storehouse process is towards measured result collection hourage, time interval in one day is divided as major key section, value hourage of certain time interval under each section, rate of change is calculated with the value of next time interval, put under the classification of corresponding rate of change, added up the distribution situation of rate of change classification hourage in each section; This excavation time interval attribute priori rules storehouse process, under given rate of change classification hourage, calculated rate, time interval unit, can be described as following two MapReduce job steps:
(1) first operation, with under given calculated rate and time interval unit, divides measured result collection according to section attribute, sorts out the time interval under respective stretch, and value hourage under this time interval;
(2) second operations, with under given calculated rate and time interval unit, excavate the Changing Pattern of time interval lower hourage, according to specify hourage rate of change classification, add up each classification under each time interval hourage rate of change distribution.
5. real-time predicting method hourage towards license plate identification data according to claim 1, it is characterized in that, wherein said Actual measurement step hourage, for calculating the road trip time in current time road network, and build up measured result collection, this computation process is divided into following two stages, and the output of previous stage is using data-stream form as the input of the latter half:
(1) bicycle calculation stages hourage: under preset time interval δ, to sort out, for each vehicle v by car plate with given frequency
i, calculate this car in all bicycle hourages through section
(2) road trip time calculation stages: under preset time interval δ, to sort out, for each section s by section with given frequency
j, gather all bicycle hourages through vehicle, calculate road trip time tras
j.
6. real-time predicting method hourage towards license plate identification data according to claim 1, it is characterized in that, predicting travel time calculation procedure, for predicting rate of change hourage of current road segment subsequent time, and calculate corresponding predicted value, when Actual measurement obtains section i at time interval δ
jmeasured value tra after, section i is at time interval δ
j+1predicting travel time value, be calculated by following two steps to obtain:
(1) utilize described time interval attribute priori rules storehouse and described section attribute priori rules storehouse, predict the interval δ of the future time of same section i by Bayes's condition probability formula
j+1under the classification x at rate of change △ hourage place;
(2) according to specify hourage rate of change classification, to the rate of change △ ' that each class definition is predicted, calculate δ
j+1predicted value tra '=(1+ △ ') * tra of interval lower hourage.
7. towards real-time estimate device hourage for license plate identification data, it is characterized in that, described device, for realizing the method for one of claim 1 ~ 6, comprises as lower component:
(1) data memory module, for storing measured result collection hourage, time interval attribute priori rules storehouse, section attribute priori rules storehouse and predicting travel time result set;
(2) online computing module, for according to given calculated rate, calculates measured value hourage in current time interval, and the predicting travel time value in future time interval and in internal memory buffer memory;
(3) calculated off-line module, for according to given calculated rate, is excavated by priori rules, upgrades and also exports section attribute priori rules storehouse to data memory module, and upgrades and to data memory module output time Range Attributes priori rules storehouse;
(4) human-computer interaction module, for realizing the interactive operation of user and this real-time estimate device.
8. real-time estimate device hourage towards license plate identification data according to claim 7, is characterized in that, described data memory module, is the storage realized based on distributed file system; This module is connected with calculated off-line module, for calculated off-line module is provided as the measured result collection and be stored as time interval attribute priori rules storehouse, the section attribute priori rules storehouse of output hourage of input; This module is connected with online computing module, for online computing module is provided as time interval attribute priori rules storehouse, the section attribute priori rules storehouse of basic data; This module is connected with human-computer interaction module, for human-computer interaction module provide hourage measured result collection and predicting travel time result set.
9. real-time estimate device hourage towards license plate identification data according to claim 7, is characterized in that, described online computing module input license plate identification data stream; This module is connected with data memory module, reads section attribute priori rules storehouse and time interval attribute priori rules storehouse, data maintenance based in internal memory; The measured result that this module exports, sends to calculated off-line module as historical data, forms measured result collection hourage; The measured result that this module exports, also can for secondary development as data stream; Only need for a section predicted value safeguarding future time interval in described internal memory; What this module exported predicts the outcome, and sends to data memory module as historical data, forms predicting travel time result set; This module is connected with human-computer interaction module, obtains the parameter of user's input, and the parameter of configuration Actual measurement step, the parameter of described Actual measurement step comprises the frequency of calculating, the time interval of calculating.
10. real-time estimate device hourage towards license plate identification data according to claim 7, is characterized in that, what described calculated off-line module inputted is measured result collection hourage, is connected with data memory module; This module is realized by 3 operations, and wherein, the priori rules based on section attribute is excavated and realized by 2 operations, and the priori rules based on time interval attribute is excavated and realized by 2 operations, and first of two mining processes operation is identical, can be shared; This module is connected with human-computer interaction module, obtains the parameter of user's input, and the parameter of the priori rules excavation step of configuration section attribute, the parameter of described priori rules excavation step comprises rate of change classification hourage, calculated rate, time interval unit.
11. real-time estimate devices hourage towards license plate identification data according to claim 7, it is characterized in that, described human-computer interaction module is connected with data memory module, support user input query parameter, the section that described parameter comprises time interval, specifies, and the measured result and predicting the outcome hourage of inquiry is presented in Web page; This module is connected with online computing module, and for user provides interactive interface, support user's input parameter, described parameter comprises the frequency of calculating, the time interval of calculating, for being configured in the calculating parameter of line computation module; This module is connected with calculated off-line module, for user provides interactive interface, support user input calculating parameter, described parameter comprise section attribute priori rules excavate rate of change classification hourage, calculated rate, time interval unit.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510518220.8A CN105139328B (en) | 2015-08-21 | 2015-08-21 | Hourage real-time predicting method and device towards license plate identification data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510518220.8A CN105139328B (en) | 2015-08-21 | 2015-08-21 | Hourage real-time predicting method and device towards license plate identification data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105139328A true CN105139328A (en) | 2015-12-09 |
| CN105139328B CN105139328B (en) | 2019-03-19 |
Family
ID=54724663
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510518220.8A Active CN105139328B (en) | 2015-08-21 | 2015-08-21 | Hourage real-time predicting method and device towards license plate identification data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105139328B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106981198A (en) * | 2017-05-24 | 2017-07-25 | 北京航空航天大学 | Deep learning network model and its method for building up for predicting travel time |
| CN107588779A (en) * | 2016-07-06 | 2018-01-16 | 中国第汽车股份有限公司 | A kind of intelligent navigation method based on journey time between any two node |
| CN108415944A (en) * | 2018-01-30 | 2018-08-17 | 长安大学 | Real time computation system and its implementation based on micro services under a kind of traffic environment |
| CN113808401A (en) * | 2021-09-18 | 2021-12-17 | 平安普惠企业管理有限公司 | Traffic congestion prediction method, device, equipment and storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103258430A (en) * | 2013-04-26 | 2013-08-21 | 青岛海信网络科技股份有限公司 | Road traveling time calculating and traffic road condition judging method and road traveling time calculating and traffic road condition judging device |
| CN104021674A (en) * | 2014-06-17 | 2014-09-03 | 武汉烽火众智数字技术有限责任公司 | Method for rapidly and accurately forecasting travel time of vehicles for passing through road sections |
-
2015
- 2015-08-21 CN CN201510518220.8A patent/CN105139328B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103258430A (en) * | 2013-04-26 | 2013-08-21 | 青岛海信网络科技股份有限公司 | Road traveling time calculating and traffic road condition judging method and road traveling time calculating and traffic road condition judging device |
| CN104021674A (en) * | 2014-06-17 | 2014-09-03 | 武汉烽火众智数字技术有限责任公司 | Method for rapidly and accurately forecasting travel time of vehicles for passing through road sections |
Non-Patent Citations (1)
| Title |
|---|
| 张帅: "基于海量车牌识别数据的旅行时间计算问题研究", 《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107588779A (en) * | 2016-07-06 | 2018-01-16 | 中国第汽车股份有限公司 | A kind of intelligent navigation method based on journey time between any two node |
| CN107588779B (en) * | 2016-07-06 | 2021-03-05 | 中国第一汽车股份有限公司 | Intelligent vehicle navigation method based on travel time between any two nodes |
| CN106981198A (en) * | 2017-05-24 | 2017-07-25 | 北京航空航天大学 | Deep learning network model and its method for building up for predicting travel time |
| CN106981198B (en) * | 2017-05-24 | 2020-11-03 | 北京航空航天大学 | Deep learning network model for travel time prediction and establishing method thereof |
| CN108415944A (en) * | 2018-01-30 | 2018-08-17 | 长安大学 | Real time computation system and its implementation based on micro services under a kind of traffic environment |
| CN108415944B (en) * | 2018-01-30 | 2019-03-22 | 长安大学 | Real time computation system and its implementation based on micro services under a kind of traffic environment |
| CN113808401A (en) * | 2021-09-18 | 2021-12-17 | 平安普惠企业管理有限公司 | Traffic congestion prediction method, device, equipment and storage medium |
| CN113808401B (en) * | 2021-09-18 | 2023-02-10 | 平安普惠企业管理有限公司 | Traffic congestion prediction method, device, equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105139328B (en) | 2019-03-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhao et al. | Truck traffic speed prediction under non-recurrent congestion: Based on optimized deep learning algorithms and GPS data | |
| Chen et al. | Finding reliable shortest paths in road networks under uncertainty | |
| Ashqar et al. | Modeling bike availability in a bike-sharing system using machine learning | |
| CN101325004B (en) | Method for compensating real time traffic information data | |
| WO2017166474A1 (en) | Method and system for intersection group-based traffic control | |
| CN102646332B (en) | Traffic state estimation device and method based on data fusion | |
| Yang et al. | Fundamental diagrams of airport surface traffic: Models and applications | |
| CN104464344B (en) | A kind of vehicle running path Forecasting Methodology and system | |
| Liu et al. | Bus arrival time prediction based on the k-nearest neighbor method | |
| Biem et al. | Real-Time Traffic Information Management using Stream Computing. | |
| CN105139328A (en) | Travel time real-time prediction method facing license plate data identification and device | |
| CN114730522A (en) | Traffic reasoning machine | |
| Nie et al. | Optimizing actuated traffic signal control using license plate recognition data: Methods for modeling and algorithm development | |
| Xia et al. | Quadtree-based domain decomposition for parallel map-matching on GPS data | |
| US11537767B2 (en) | Automated control through a traffic model | |
| Ghandeharioun et al. | Link travel time estimation for arterial networks based on sparse GPS data and considering progressive correlations | |
| Li et al. | Traffic control optimization strategy based on license plate recognition data | |
| Tilg et al. | Simulation-based design of urban bi-modal transport systems | |
| Kumar et al. | Analytical approach to identify the optimum inputs for a bus travel time prediction method | |
| Dai et al. | PARP: A parallel traffic condition driven route planning model on dynamic road networks | |
| Merrad et al. | A survey on smart traffic network control and optimization | |
| Zygouras et al. | HTTE: A hybrid technique for travel time estimation in sparse data environments | |
| Conlan et al. | Collective shortest paths for minimizing congestion on temporal load-aware road networks | |
| Lin et al. | Service-oriented dynamic data driven application systems to traffic signal control | |
| Xia et al. | A parallel fusion method for heterogeneous multi-sensor transportation data |
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 |