CN105091889A

CN105091889A - Hotspot path determination method and hotspot path determination equipment

Info

Publication number: CN105091889A
Application number: CN201410167036.9A
Authority: CN
Inventors: 袁明轩; 曾嘉; 谭浩宇
Original assignee: Huawei Technologies Co Ltd
Current assignee: Huawei Technologies Co Ltd
Priority date: 2014-04-23
Filing date: 2014-04-23
Publication date: 2015-11-25
Anticipated expiration: 2034-04-23
Also published as: CN105091889B

Abstract

Embodiments of the invention disclose a hotspot path determination method and hotspot path determination equipment. The method comprises: carrying out hidden markov model-based fuzzy map matching calculation on each measurement path in acquired user track measurement data to obtain the optional path set of each measurement path; determining the vehicle flowrate data of all road sections in a road network by using the optional path set of each measurement path and the topology structure information of the road network; determining a candidate path set by using the topology structure information, and the starting position and the destination position input by the user; and determining the probability-based hotspot path according to the vehicle flowrate data of all road sections and the candidate path set. According to the present invention, the hidden markov model-based fuzzy map matching calculation is performed on the user track measurement data to determine the hotspot path so as to effectively improve the accuracy of the hotspot path.

Description

A kind of determination method and apparatus of hotspot path

Technical field

The present invention relates to the method for network path search, particularly relate to a kind of determination method and apparatus of hotspot path.

Background technology

In recent years, along with the explosive growth of mobile network and the widespread use of intelligent movable equipment, the track data of mobile subscriber becomes a kind of important large Data Source.The track data of user is otherwise known as user's spatial and temporal distributions data.Such as user opens GPS (English full name is: GlobalPositioningSystem, is abbreviated as: GPS) when serving, and this user is exactly the track data of this user in the information of space-time movement.When user uses mobile network, the track data also containing a large number of users in mobile broadband (English full name is: MobileBroadband, is abbreviated as: the MBB) data of base station record.The track data of user to be added up and the degree of depth is excavated and brought the application of much new business, such as shop addressing, service recommendation, traffic administration and map reparation etc.The application of these new business is one of focus of current industrial circle research.

For automobile navigation service, use hotspot path (English full name is: MostFrequentPath, is abbreviated as: MFP) algorithm can obtain the optimal path of the group wisdom reflecting trip colony to the track data of user.Compare the algorithm that shortest path first, the fastest path algorithm etc. only use road attribute, the common results of many factors can be embodied in the path that hotspot path algorithm obtains, have better Consumer's Experience, wherein, hotspot path refers to from origin to destination the most often by path that user uses.But, MBB data set use existing hotspot path algorithm be faced with problem, namely the customer position information utilizing MBB data to determine only can be accurate to about 100 meters, and existing map-matching algorithm be difficult to by this have accurately be mapped on road network compared with the data of high level error.Therefore hotspot path can be caused to calculate inaccurate problem.

Summary of the invention

Embodiments providing a kind of determination method and apparatus of hotspot path, calculating inaccurate problem for solving hotspot path of the prior art.

First aspect present invention provides a kind of defining method of hotspot path, comprising:

Fuzzy map matching primitives based on Hidden Markov Model (HMM) is carried out to each measuring route in the user trajectory measurement data got, obtain the feasible path set of each measuring route described, wherein, comprise normalization matching probability in all feasible paths of a measuring route in feasible path set described in and come the feasible path of front N, described user trajectory measurement data is the trajectory measurement data of the user of the road network travelling on preset regions in preset time period, and described N is positive integer;

The described feasible path set of each measuring route and the topology information of described road network is utilized to determine the vehicle flowrate data in all sections of described road network;

Utilize described topology information, user input reference position and destination locations determination alternative path set, the path in described alternative path set is all with described reference position for starting point, and described destination locations is the set in the path of terminal;

The hotspot path based on probability is determined according to the vehicle flowrate data in described all sections and described alternative path set.

In the first possible implementation of first aspect, the described vehicle flowrate data according to described all sections and described alternative path set also comprise before determining the hotspot path based on probability:

Whether the vehicle flowrate data judging described all sections are non-sparse data;

If the vehicle flowrate data in described all sections are non-sparse data, then determine that the vehicle flowrate data in described all sections are the vehicle flowrate data acquisition in the section of described road network;

If the vehicle flowrate data in described all sections are sparse data, then based on the common triage techniques of multiple linear regression, or fill up based on the common triage techniques of described multiple linear regression and the training mechanism of semi-supervised learning the vehicle flowrate data belonging to unavailable section in described all sections, the vehicle flowrate data in all sections after being optimized, using the vehicle flowrate data acquisition of the vehicle flowrate data in all sections after described optimization as the section of described road network, described unavailable section refers to that vehicle flowrate data are less than the section of the threshold value pre-set.

In conjunction with the first possible implementation of first aspect, in the implementation that first aspect the second is possible, the described common triage techniques based on multiple linear regression fills up the vehicle flowrate data belonging to unavailable section in described all sections, the vehicle flowrate data in all sections after being optimized, comprising:

Be matrix M by the vehicle flowrate data construct in described all sections, and matrix M is as follows:

Wherein, v ₁, v ₂..., v _nfor the beginning or end in described section, F ₁₂, F ₁₃, F _1n..., F _nnall represent the vehicle flowrate data in section;

Carrying out low-rank approximation decomposition to described matrix M, is X by described matrix M approximate representation ^tx, wherein said X representing matrix X, described X ^tthe transposed matrix of representing matrix X;

Solve objective function || M-X ^tx|| ₂minimum value when condition rank (X) <r, obtains matrix X ';

By the vehicle flowrate data X in section unavailable in described matrix M ^{' T}the vehicle flowrate data in unavailable section described in X ' replace, and the matrix M after being optimized, the matrix M after described optimization is the vehicle flowrate data in all sections after described optimization.

In conjunction with the first possible implementation of first aspect, in the third possible implementation of first aspect, the training mechanism of the described common triage techniques based on described multiple linear regression and semi-supervised learning fills up the vehicle flowrate data belonging to unavailable section in described all sections, the vehicle flowrate data in all sections after being optimized, comprising:

Be matrix M by the vehicle flowrate data construct in described all sections ₀, and matrix M ₀as follows:

Wherein, v ₁, v ₂..., v _nfor the beginning or end in described section, F ₁₂, F ₁₃, F _1n..., F _nnall represent the vehicle flowrate data in section; The initial value of i is 0, performs following steps:

To described matrix M _icarry out low-rank approximation decomposition, by described matrix M _iapproximate representation X _i ^tx _i, wherein said X _irepresenting matrix X _i, described X _i ^tmatrix X _itransposed matrix;

Solve objective function || M _i-X _i ^tx _i|| at condition rank (X _i) <r time minimum value, obtain matrix X _i', wherein r is the numerical value pre-set;

If described matrix M _iin the quantity in unavailable section be greater than z, then by described matrix M _iin the vehicle flowrate data in any Z unavailable section be used in X _i ^{' T}x _i' the vehicle flowrate data in the unavailable section of middle correspondence replace, and obtain matrix M _i+1; Make i=i+1, return execution described to described matrix M _icarry out the step of low-rank approximation decomposition;

If described matrix M _iin the quantity in unavailable section be less than or equal to z, then by described matrix M _iin the vehicle flowrate data in unavailable section be used in X _i ^{' T}x _i' described in unavailable section vehicle flowrate data replace, obtain matrix M _i+1, by described matrix M _i+1as the vehicle flowrate data in all sections after optimization.

In conjunction with first aspect or the first possible implementation of first aspect, in first aspect the 4th kind of possible implementation, each measuring route in the described user trajectory measurement data to getting carries out the fuzzy map matching primitives based on Hidden Markov Model (HMM), obtain the feasible path set of each measuring route described, comprising:

Calculate the candidate matches limit collection of each node in each measuring route in the trajectory measurement data of the user got according to described topology information, described candidate matches limit integrates the set in the section that may be positioned at as node;

Utilize the candidate matches limit collection of each node in each measuring route described to build the Hidden Markov Model of each measuring route described, obtain the feasible path set of each measuring route described.

In conjunction with first aspect the 4th kind of possible implementation, in first aspect the 5th kind of possible implementation, the described topology information according to described road network calculates the candidate matches limit collection of each node in each measuring route described, comprising:

Determine the candidate matches limit collection of each node of measuring route in the following manner:

If the trajectory measurement data of the node i of measuring route A are (X _i, Y _i, T _i), then the measuring position of node i is (X _i, Y _i), suppose the physical location of described node i be (X ' _i, Y ' _i), then calculate the section meeting matching formula in described topology information, the described set meeting the section of matching formula is the candidate matches limit collection of described node i, described candidate matches limit concentrate comprise section that described node i may be positioned at and described may at the probability in section;

Described matching formula is:

P{ (X _i, Y _i) | (X ' _i, Y ' _i) on e > θ

Wherein, X _irepresent the longitude of described node i, Y _irepresent the latitude of described node i, e represents section e, and P represents probability, and θ is the numerical value pre-set.

In conjunction with first aspect the 5th kind of possible implementation, in first aspect the 6th kind of possible implementation, the candidate matches limit collection of each node in each measuring route described in described utilization builds the Hidden Markov Model of each measuring route described, and the feasible path set obtaining each measuring route described comprises:

Determine the feasible path set of each measuring route in the following manner:

The candidate matches limit collection of each node on described measuring route A is utilized to build Hidden Markov Model (HMM), determine all feasible paths of described measuring route A, the probability in the section on described feasible path is the probability of candidate matches limit centralized node on described section;

Based on the candidate matches limit collection of the topology information of road network and each node of measuring route A, transition probability when adjacent two nodes get different candidate matches limits in computation and measurement path A;

According to the probability in the section on each feasible path of described measuring route A and the matching probability of described transition probability calculating each feasible path described, described matching probability equals the product between the transition probability between the probability in all sections of feasible path and the connected section of described feasible path;

From described all feasible paths, select matching probability to come the feasible path set of feasible path as described measuring route A of front N, and the matching probability of each feasible path in described feasible path set is normalized the normalization matching probability obtaining each feasible path described.

In conjunction with first aspect the 6th kind of possible implementation, in first aspect the 7th kind of possible implementation, the vehicle flowrate data in all sections of described road network are determined in the feasible path set of each measuring route described in described utilization, comprising:

Determine the set of the number of times S that each section in described road network occurs in all feasible path set of described all measuring route and the section probability that S time occurs, described S is positive integer;

The set of the section probability occurred according to the number of times S of the appearance in each section described and S time calculates the vehicle flowrate data in each section, and comprise the set that section occurrence number is the probability of k in described vehicle flowrate data, the value of described k is 1 to S.

In conjunction with first aspect the 6th kind of possible implementation, in first aspect the 8th kind of possible implementation, describedly determine to comprise based on the hotspot path of probability according to described vehicle flowrate data and described alternative path set:

The vehicle flowrate data in each section of each path candidate in described alternative path set are obtained from described vehicle flowrate data acquisition;

The family of functions that pre-sets is utilized to calculate flow distribution value corresponding to the vehicle flowrate data in each section of the path candidate in described alternative path set;

The flow distribution value corresponding according to the vehicle flowrate data in each section of each paths described determines the hotspot path based on probability.

In conjunction with first aspect the 8th kind of possible implementation, in first aspect the 9th kind of possible implementation, the flow distribution value that the vehicle flowrate data in each section of each paths described in described basis are corresponding is determined to comprise based on the hotspot path of probability:

Flow distribution value corresponding for all sections of each path candidate in described alternative path set is carried out the path temperature sorted as each path candidate described according to order from small to large;

From described alternative path set, select two path candidates, respectively as path candidate a and path candidate b, wherein, the path temperature of described path candidate a is Freqa, and Freqa=<F _a1, F _a2..., F _aLathe flow distribution value in La section is comprised, the path temperature Freq of path candidate b in >, Freqa _b, and Freq _b=<F _b1, F _b2..., F _bLb>, Freq _bin comprise the flow distribution value in Lb section;

Judge that the path temperature of described path candidate a is whether higher than the path temperature of described path candidate b;

If there is a k, and the span of k be 1,2 ...., min (La, Lb) }, make F _ak<F _bk; And to all i being less than k, meet F _ai=F _bi, or when La<Lb for all i being less than or equal to La, meet F _ai=F _biwherein, k and i is positive integer, then determine that the path temperature of path candidate b is higher than path candidate a, if and comprise not by the path candidate selected in alternative path set, then make path candidate a=path candidate c, return and perform the described path temperature judging described path candidate a whether higher than the path temperature of described path candidate b, described path candidate c is not by the path candidate selected in described alternative path set; If do not comprise in alternative path set not by the path candidate selected, then determine that described path candidate b is hotspot path;

If there is a k, and the span of k be 1,2 ...., min (La, Lb) }, make F _ak>F _bk; And to all i being less than k, meet F _ai=F _bi, or for all i being less than or equal to Lb, meet F at La>Lb _ai=F _biwherein, k and i is positive integer, then determine that the path temperature of path candidate a is higher than path candidate b, if and comprise not by the path candidate selected in alternative path set, make path candidate b=path candidate c, return and perform the described path temperature judging described path candidate a whether higher than the path temperature of described path candidate b, described path candidate c is not by the path candidate selected in described alternative path set, if do not comprise in alternative path set not by the path candidate selected, then determine that described path candidate a is hotspot path;

If La=Lb, and the span of k be 1,2 ...., during La or Lb}, F _ai=F _bithen determine that path candidate a is identical with the path temperature of path candidate b, if and comprise not by the path candidate selected in alternative path set, then make path candidate a=path candidate c, return and perform the described path temperature judging described path candidate a whether higher than the path temperature of described path candidate b, or, make path candidate b=path candidate c, return and perform the described path temperature judging described path candidate a whether higher than the path temperature of described path candidate b, described path candidate c is not by the path candidate selected in described alternative path set; If do not comprise in alternative path set not by the path candidate selected, then determine that described path candidate a or path candidate b is hotspot path.

Second aspect present invention provides a kind of hotspot path locking equipment really, comprising:

First computing unit, for carrying out the fuzzy map matching primitives based on Hidden Markov Model (HMM) to each measuring route in the user trajectory measurement data got, obtain the feasible path set of each measuring route described, wherein, comprise normalization matching probability in all feasible paths of a measuring route in feasible path set described in and come the feasible path of front N, described user trajectory measurement data is the trajectory measurement data of the user of the road network travelling on preset regions in preset time period, and described N is positive integer;

First determining unit, for obtain each measuring route described at described first computing unit feasible path set after, utilize the described feasible path set of each measuring route and the topology information of described road network to determine the vehicle flowrate data in all sections of described road network;

Second determining unit, after determining the vehicle flowrate data in all sections of described road network in described first determining unit, utilize described topology information, user input reference position and destination locations determination alternative path set, path in described alternative path set is all with described reference position for starting point, and described destination locations is the set in the path of terminal;

3rd determining unit, after obtaining described alternative path set in described second determining unit, determines the hotspot path based on probability according to the vehicle flowrate data in described all sections and described alternative path set.

In the first possible implementation of second aspect, described equipment also comprises:

Judging unit, after obtaining described alternative path set in described second determining unit, judges whether the vehicle flowrate data in described all sections are non-sparse data;

4th determining unit, if determine that the vehicle flowrate data in described all sections are non-sparse data for described judging unit, then determines that the vehicle flowrate data in described all sections are the vehicle flowrate data acquisition in the section of described road network;

5th determining unit, if determine that the vehicle flowrate data in described all sections are sparse data for described judging unit, then based on the common triage techniques of multiple linear regression, or fill up based on the common triage techniques of described multiple linear regression and the training mechanism of semi-supervised learning the vehicle flowrate data belonging to unavailable section in described all sections, the vehicle flowrate data in all sections after being optimized, using the vehicle flowrate data acquisition of the vehicle flowrate data in all sections after described optimization as the section of described road network, described unavailable section refers to that vehicle flowrate data are less than the section of the threshold value pre-set.

In conjunction with the first possible implementation of second aspect, in the implementation that second aspect the second is possible, described 5th determining unit specifically for:

If described judging unit determines that the vehicle flowrate data in described all sections are sparse data, be matrix M, and matrix M is as follows by the vehicle flowrate data construct in described all sections:

In conjunction with the first possible implementation of second aspect, in the third possible implementation of second aspect, described 5th determining unit specifically for:

If described judging unit determines that the vehicle flowrate data in described all sections are sparse data, is matrix M by the vehicle flowrate data construct in described all sections ₀, and matrix M ₀as follows:

In conjunction with second aspect or the first possible implementation of second aspect, in second aspect the 4th kind of possible implementation, described first computing unit comprises:

Limit collection computing unit, for calculating the candidate matches limit collection of each node in each measuring route in the trajectory measurement data of the user got according to described topology information, described candidate matches limit integrates the set in the section that may be positioned at as node;

Feasible path computing unit, after obtaining described candidate matches limit collection at described limit collection computing unit, utilize the candidate matches limit collection of each node in each measuring route described to build the Hidden Markov Model of each measuring route described, obtain the feasible path set of each measuring route described.

In conjunction with second aspect the 4th kind of possible implementation, in second aspect the 5th kind of possible implementation, described limit collection computing unit specifically for:

Described matching formula is:

P{ (X _i, Y _i) | (X ' _i, Y ' _i) on e > θ

In conjunction with second aspect the 5th kind of possible implementation, in second aspect the 6th kind of possible implementation, described feasible path computing unit specifically for:

In conjunction with second aspect the 6th kind of possible implementation, in second aspect the 7th kind of possible implementation, described first determining unit comprises:

6th determining unit, for obtain each measuring route described at described first computing unit feasible path set after, determine the set of the number of times S that each section in described road network occurs in all feasible path set of described all measuring route and the section probability that S time occurs, described S is positive integer;

Second computing unit, after set for the section probability of the occurrence number S and S appearance that determine described each article of section in described 6th determining unit, the set of the section probability occurred according to the number of times S of the appearance in each section described and S time calculates the vehicle flowrate data in each section, comprise the set that section occurrence number is the probability of k in described vehicle flowrate data, the value of described k is 1 to S.

In conjunction with second aspect the 7th kind of possible implementation, in second aspect the 8th kind of possible implementation, described 3rd determining unit comprises:

Acquiring unit, for after described 4th determining unit or described 5th determining unit, obtains the vehicle flowrate data in each section of each path candidate in described alternative path set from described vehicle flowrate data acquisition;

3rd computing unit, for obtain each section of each path candidate in described alternative path set at described acquiring unit vehicle flowrate data after, utilize the family of functions pre-set to calculate flow distribution value corresponding to the vehicle flowrate data in each section of the path candidate in described alternative path set;

6th determining unit, for after the flow distribution value that the vehicle flowrate data obtaining described each article of section at described 3rd computing unit are corresponding, the flow distribution value corresponding according to the vehicle flowrate data in each section of each paths described determines the hotspot path based on probability.

In conjunction with second aspect the 8th kind of possible implementation, in second aspect the 9th kind of possible implementation, described 6th determining unit specifically for:

As can be seen from the above technical solutions, the embodiment of the present invention has the following advantages:

Fuzzy map matching primitives based on Hidden Markov Model (HMM) is carried out to each measuring route in the user trajectory measurement data got, obtain the feasible path set of each measuring route, contain normalization matching probability in all feasible paths of a measuring route in a feasible path set and come the feasible path of front N, and the topology information of the feasible path set and road network that utilize each measuring route is determined the vehicle flowrate data in all sections of this road network, utilize this topology information, the reference position of user's input and destination locations determination alternative path set, the hotspot path based on probability is determined according to the vehicle flowrate data in this all section and alternative path set, by carrying out the fuzzy map matching primitives based on Hidden Markov Model (HMM) to user trajectory measurement data, obtain the feasible path set of each measuring route, remain a part of uncertain information in user trajectory measurement data, obtain the path profile of likely mating, effectively can improve the accuracy of hotspot path.

Term " first ", " second ", " the 3rd " " 4th " etc. (if existence) in instructions of the present invention and claims and above-mentioned accompanying drawing are for distinguishing similar object, and need not be used for describing specific order or precedence.Should be appreciated that the data used like this can be exchanged in the appropriate case, so as embodiments of the invention described herein such as can with except here diagram or describe those except order implement.In addition, term " comprises " and " having " and their any distortion, intention is to cover not exclusive comprising, such as, contain those steps or unit that the process of series of steps or unit, method, system, product or equipment is not necessarily limited to clearly list, but can comprise clearly do not list or for intrinsic other step of these processes, method, product or equipment or unit.

Refer to Fig. 1, be a kind of in embodiment of the present invention embodiment of defining method of hotspot path, comprise:

101, the fuzzy map matching primitives based on Hidden Markov Model (HMM) is carried out to each measuring route in the user trajectory measurement data got, obtain the feasible path set of each measuring route;

In embodiments of the present invention, hotspot path really locking equipment (being called for short: equipment) can get user trajectory measurement data from server, wherein, equipment is the trajectory measurement data through the user of the road network of preset regions in preset time period from the user trajectory measurement data that server gets, the trajectory measurement data of user refer to user in the driving process of road network the positional information and temporal information etc. in section of process, such as: if user vehicle A drives to be in urban district, Shenzhen, on the automobile of then this user vehicle A for determining that the equipment of hotspot path can obtain the trajectory measurement data of the user of the road network travelling on urban district, Shenzhen within half an hour from server.

In embodiments of the present invention, equipment division carries out the fuzzy map matching primitives based on Hidden Markov Model (HMM) to each measuring route in the user trajectory measurement data got, obtain the feasible path set of each measuring route, wherein, comprise normalization matching probability in all feasible paths of a measuring route in one feasible path set and come the feasible path of front N, and this N is positive integer.

The part uncertainty of user trajectory measurement data can be retained in the feasible path set obtained by carrying out fuzzy map matching primitives based on Hidden Markov Model (HMM), obtain the path profile of likely mating, the hotspot path that the feasible path set utilizing measuring route is obtained can be more accurate and effective.

102, the vehicle flowrate data in all sections of the feasible path set of each measuring route and the topology information determination road network of road network are utilized;

In embodiments of the present invention, equipment is carrying out fuzzy map matching primitives based on Hidden Markov Model (HMM), after obtaining the feasible path set of each measuring route, by the vehicle flowrate data in all sections of the topology information determination road network of the feasible path set and road network that utilize each measuring route.

103, utilize topology information, user input reference position and destination locations determination alternative path set, the path in alternative path set is all take reference position as starting point, and destination locations is the set in the path of terminal;

In embodiments of the present invention, equipment will utilize the topology information of this road network, the reference position of user's input and destination locations determination alternative path set, and the path in this alternative path set is all take reference position as starting point, destination locations is the set in the path of terminal.Such as: user is when using this equipment to search route, can input reference position is GuoMao Building, destination locations is Window on the World, then this equipment can utilize the topology information of network road to determine, from GuoMao Building to the optional path of Window on the World, namely to determine alternative path set.

104, the hotspot path based on probability is determined according to the vehicle flowrate data in all sections and alternative path set.

In embodiments of the present invention, equipment after the alternative path set determining the reference position that user inputs and destination locations, the hotspot path can determining based on probability according to the vehicle flowrate data in all sections and alternative path set.

In embodiments of the present invention, equipment carries out the fuzzy map matching primitives based on Hidden Markov Model (HMM) to each measuring route in the user trajectory measurement data got, obtain the feasible path set of each measuring route, and the topology information of the feasible path set and road network that utilize each measuring route is determined the vehicle flowrate data in all sections of this road network, utilize this topology information, the reference position of user's input and destination locations determination alternative path set, the hotspot path based on probability is determined according to the vehicle flowrate data in this all section and alternative path set, by carrying out the fuzzy map matching primitives based on Hidden Markov Model (HMM) to user trajectory measurement data, obtain the feasible path set of each measuring route, remain a part of uncertain information in user trajectory measurement data, effectively can improve the accuracy of hotspot path.

In order to understand the technical scheme of the defining method of hotspot path in the embodiment of the present invention better, referring to Fig. 2, being a kind of in embodiment of the present invention embodiment of defining method of hotspot path, comprising:

201, calculate the candidate matches limit collection of each node in each measuring route in the user trajectory measurement data got according to topology information, candidate matches limit integrates the set in the section that may be positioned at as node;

In embodiments of the present invention, measuring route in user trajectory measurement data is made up of multiple node, this node is actual node, equipment is after getting user trajectory measurement data, by the candidate matches limit collection according to each node in each measuring route in the topology information calculating user trajectory measurement data of road network, wherein, candidate matches limit integrates the set in the section that may be positioned at as node.

In embodiments of the present invention, equipment by the candidate matches limit collection of each node in each measuring route of determining in the following manner in user trajectory measurement data, the node i for measuring route A:

1) if the trajectory measurement data of the node i of measuring route A are (X _i, Y _i, T _i), then the measuring position of node i is (X _i, Y _i), Measuring Time is T _i, suppose the physical location of node i be (X ' _i, Y ' _i), then calculate the section meeting matching formula in topology information, the set meeting the section of matching formula is the candidate matches limit collection of node i, and candidate matches limit is concentrated and comprised section that node i may be positioned at and the probability in the section that may be positioned at;

Wherein, matching formula is:

P{ (X _i, Y _i) | (X ' _i, Y ' _i) on e > θ

Wherein, X _irepresent the longitude of node i, Y _irepresent the latitude of node i, e represents section e, and P represents probability, and θ is the numerical value pre-set.

By above-mentioned matching formula, equipment can calculate the candidate matches limit collection of each node in each measuring route.

202, utilize the candidate matches limit collection of each node in each measuring route to build the Hidden Markov Model of each measuring route, carry out fuzzy map matching primitives based on Hidden Markov Model (HMM), obtain the feasible path set of each measuring route;

In embodiments of the present invention, after equipment obtains the candidate matches limit collection of all nodes difference correspondences in trajectory measurement data, the candidate matches limit collection of each node utilized in each measuring route is built the Hidden Markov Model (HMM) of this measuring route, to carry out fuzzy map matching primitives, carry out fuzzy map matching primitives based on Hidden Markov Model (HMM), obtain the feasible path set of each measuring route.

In embodiments of the present invention, equipment can determine the feasible path set of each measuring route in the following manner, for measuring route A:

1) the candidate matches limit collection of each node on equipment utilization measuring route A builds Hidden Markov Model (HMM), determine all feasible paths of measuring route A, the probability in the section on feasible path is the probability of candidate matches limit centralized node on section, therefore, equipment can obtain the probability in all feasible paths of measuring route A and each section of each feasible path.

2) based on the candidate matches limit collection of the topology information of road network and each node of measuring route A, transition probability when adjacent two nodes get different candidate matches limits in computation and measurement path A;

In embodiments of the present invention, equipment is by the candidate matches limit collection based on the topology information of road network and each node of measuring route A, adjacent two nodes of computation and measurement path A get transition probability during different candidate matches limits, such as: such as: if the node-locus of measuring route A is a1->a2->a3, and the candidate matches limit collection of node a1 comprises candidate matches limit e1 and e2, the candidate matches limit collection of node a2 comprises candidate matches limit e3, e4 and e5, the candidate matches limit collection of node a3 is for comprising candidate matches limit e6.The topology information based on road network is then needed to calculate the transition probability of e1->e3, e1->e4, e1->e5, e2->e3, e2->e4, e2->e5, e3->e6, e4->e6, e5->e6 respectively.

Wherein, candidate matches limit collection is node in corresponding original measuring route, and the transition probability between two candidate limits is by Distance geometry Time dependent.Such as, for feasible path e1 to the e3 of node a1 to a2, projected to by a1 on e1 and obtain a b1, namely b1 is point nearest from a1 on the e1 of limit; Similarly, a2 is projected on e3 and obtain a b2.For this original measurement section of a1 to a2, the transition probability of candidate matches limit e1 to e3 is by the shortest path length L of b1 to b2, and time Δ t used determines, wherein, shortest path length can be determined according to the particular location of b1 and b2 in the topology information of road network.Particularly, it has been generally acknowledged that car speed meets normal distribution N (V, dv), wherein V, dv be respectively the average velocity of all vehicles and variance (this average velocity and variance be calculated as prior art, do not repeat herein), then the speed from b1 to b2 is V '=L/ Δ t, and therefore the transition probability of e1 to e3 is that V ' is at normal distribution N (V, dv) probability under, i.e. (1/dv) * ((V '-V)/dv), wherein for the probability density function of normal distribution.

In the manner described above, two adjacent nodes that can obtain on measuring route A get transition probability during different candidate matches limits.

3) calculate the matching probability of each feasible path according to the probability in the section on each feasible path of measuring route A and transition probability, matching probability equals the product between the transition probability between the probability in all sections of feasible path and the connected section of feasible path; Such as: path candidate is: e ₃→ e ₄→ e ₄→ e ₃, the matching probability of this path candidate is: P{e ₃→ e ₄→ e ₄→ e ₃}=P{ (x ₁, y ₁) | (x ₁', y ₁') at e ₃on × P ₃₄(t ₂-t ₁) × P{ (x ₂, y ₂) | (x ₂', y ₂') at e ₄on × P ₄₄(t ₃-t ₂) × P{ (x ₃, y ₃) | (x ₃', y ₃') at e ₄on × P ₄₃(t ₄-t ₃) × P{ (x ₄, y ₄) | (x ₄', y ₄') at e ₃on, wherein P{ (x ₁, y ₁) | (x ₁', y ₁') at e ₃on represent t ₁the probability of measurement point on time measurement path on e3, P ₃₄(t ₂-t ₁) represent between e3 and e4 at time interval t ₂-t ₁the transition probability of time period, other implication can be by that analogy.

4) from all feasible paths, select matching probability to come the feasible path set of feasible path as measuring route A of front N, and the matching probability of each feasible path in feasible path set is normalized the normalization matching probability obtaining each feasible path.Such as: the feasible path set expression of measuring route A is H={ (path1, p1), (path2, p2),, (pathn, pn) }, wherein, p1+p2+ ... + pn=1, wherein, path represents feasible path, and p represents the matching probability of feasible path.

Refer to 2, the fuzzy map-matching algorithm based on Hidden Markov Model (HMM) for a measuring route in the embodiment of the present invention obtains the exemplary plot of path candidate, wherein, and t ₁, t ₂, t ₃, t ₄four nodes that Measuring Time respectively in expression measuring route is different, dotted line and solid line form all set of paths in measuring route, wherein, solid line forms the feasible path set of measuring route, and in fig. 2, the feasible path set of measuring route comprises six feasible paths, is respectively: e3 → e1 → e1 → e1, e3 → e1 → e1 → e1, e3 → e1 → e3 → e1, e3 → e1 → e3 → e3, e3 → e4 → e3 → e1, e3 → e4 → e3 → e3, e3 → e4 → e4 → e3.

By above-mentioned steps 1) to 4), the fuzzy map matching primitives based on Hidden Markov Model (HMM) can be realized.

203, determine the set of the number of times S that each section in road network occurs in all feasible path set of all measuring route and the section probability that S time occurs, S is positive integer;

In embodiments of the present invention, equipment after the feasible path set obtaining each measuring route in user trajectory measurement data, by the number of times S occurred in all feasible path set of all measuring route based on each section in the topology information determination road network of road network and the set occurring section probability for S time.

The set of the section probability 204, occurred according to the number of times S of the appearance in each section and S time calculates the vehicle flowrate data in each section, and comprise the set that section occurrence number is the probability of k in vehicle flowrate data, the value of k is 1 to S;

In embodiments of the present invention, the set of the section probability that the number of times S occurred according to each section in the network topology structure of road network and S time occur by equipment calculates the vehicle flowrate data in each section, and will the set that this section occurrence number is the probability of K be comprised in vehicle flowrate data, wherein, K is 1 to S, such as: if the number of times S that section occurs is 5 times, then will calculate this section respectively and occur the probability of 1 time, there is the probability of 2 times, there is the probability of 3 times, there is the probability of 4 times and occur the probability of 5 times, and forming the vehicle flowrate data in this section.Such as: section e _ijstarting point and terminal be respectively v _iand v _j, then F can be used _ij=(f ₀, f ₁..., f _s) represent section e _ijvehicle flowrate data, wherein, f _krepresent section e _ijoccurrence number (actual flow) probability that is k, and vehicle flowrate data for a better understanding of the present invention in embodiment, to calculate section e _ijvehicle flowrate data instance the account form of vehicle flowrate data is described: if section e _ijthe number of times occurred is 3 times, and the set of the section probability of 3 times is (p1, p2, p3), then f ₀=(1-p1) (1-p2) (1-p3), f ₁=p1 (1-p2) (1-p3)+p2 (1-p1) (1-p3)+p3 (1-p1) (1-p2), f ₂=p1*p2 (1-p3)+p1* (1-p2) * p3+ (1-p1) * p2*p3, f ₃=p1*p2*p3.By utilizing the account form of above-mentioned vehicle flowrate data, the object of the flow describing section with probabilistic manner can be realized.

Whether the vehicle flowrate data 205, judging all sections are non-sparse data, if so, then perform step 206, if not, then perform step 207;

206, determine that the vehicle flowrate data in all sections are the vehicle flowrate data acquisition in the section of road network, continue to perform step 208;

207, based on the common triage techniques of multiple linear regression, or fill up based on the common triage techniques of multiple linear regression and the training mechanism of semi-supervised learning the vehicle flowrate data belonging to unavailable section in all sections, the vehicle flowrate data in all sections after being optimized, using the vehicle flowrate data acquisition of the vehicle flowrate data in all sections after optimization as the section of road network, continue to perform step 208;

In embodiments of the present invention, when the vehicle flowrate data in all sections are sparse data, to the possibility cannot determining hotspot path be there is in the vehicle flowrate data in this all section of equipment utilization, therefore, the problem cannot determining hotspot path caused in order to avoid factor data is sparse, whether equipment will be that non-sparse data judges to the vehicle flowrate data in all sections, if and the vehicle flowrate data in all sections are non-sparse data, then determine that the vehicle flowrate data in this all section are the vehicle flowrate data acquisition in the section of road network.

If the vehicle flowrate data in all sections are sparse data, then equipment is by the common triage techniques based on multiple linear regression, or adjust the vehicle flowrate data of mending and belonging to unavailable section in this all section based on the common triage techniques of multiple linear regression and the training mechanism of semi-supervised learning, the vehicle flowrate data in all sections after being optimized, and using the vehicle flowrate data acquisition of the vehicle flowrate data in all sections after optimization as road network, wherein, unavailable section refers to that vehicle flowrate data are less than the section of the threshold value pre-set, such as: the number of times that section occurs in all alternative path set of all measuring route is less than or equal to 10 times.

In embodiments of the present invention, equipment is after determining that the vehicle flowrate data in all sections are sparse data, the vehicle flowrate data belonging to unavailable section in this all section can be filled up based on the common triage techniques of multiple linear regression, concrete, can in the following way:

The vehicle flowrate data construct in all sections is matrix M by equipment, and matrix M is as follows:

Wherein, v ₁, v ₂..., v _nfor the beginning or end in section, F ₁₂, F ₁₃, F _1n..., F _nnall represent the vehicle flowrate data in section;

Carrying out low-rank approximation decomposition to matrix M, is X by matrix M approximate representation ^tx, wherein X representing matrix X, X ^tthe transposed matrix of representing matrix X;

By the vehicle flowrate data X in section unavailable in matrix M ^{' T}in X ', the vehicle flowrate data in unavailable section replace, the matrix M after being optimized, and the matrix M after optimization is the vehicle flowrate data in all sections after optimizing.

In embodiments of the present invention, equipment is after determining that the vehicle flowrate data in all sections are sparse data, the vehicle flowrate data belonging to unavailable section in all sections can be filled up based on the training mechanism of the common triage techniques of multiple linear regression and semi-supervised learning, the vehicle flowrate data in all sections after being optimized, comprising:

The vehicle flowrate data construct in all sections is matrix M by equipment ₀, and matrix M ₀as follows:

Wherein, v ₁, v ₂..., v _nfor the beginning or end in section, F ₁₂, F ₁₃, F _1n..., F _nnall represent the vehicle flowrate data in section; The initial value of i is 0, performs following steps:

Equipment is to matrix M _icarry out low-rank approximation decomposition, by matrix M _iapproximate representation X _i ^tx _i, wherein X _irepresenting matrix X _i, X _i ^tmatrix X _itransposed matrix;

Then, equipment will solve objective function || M _i-X _i ^tx _i|| at condition rank (X _i) <r time minimum value, obtain matrix X _i', wherein r is the numerical value pre-set;

If matrix M _iin the quantity in unavailable section be greater than z, then by matrix M _iin the vehicle flowrate data in any z unavailable section be used in X _i ^{' T}x _i' the vehicle flowrate data in the unavailable section of middle correspondence replace, and obtain matrix M _i+1; Make i=i+1, return and perform matrix M _icarry out the step of low-rank approximation decomposition;

If matrix M _iin the quantity in unavailable section be less than or equal to z, then by matrix M _iin the vehicle flowrate data in unavailable section be used in X _i ^{' T}x _i' in this unavailable section vehicle flowrate data replace, obtain matrix M _i+1, by matrix M _i+1as the vehicle flowrate data in all sections after optimization.Such as: if M _iin F ₁₂for the data in unavailable section, then use X _i ^{' T}x _i' middle F ₁₂replace M _iin F ₁₂.

208, utilize topology information, user input reference position and destination locations determination alternative path set, the path in alternative path set is all take reference position as starting point, and destination locations is the set in the path of terminal;

In embodiments of the present invention, equipment is after the vehicle flowrate data acquisition obtaining road network, topology information will be utilized, the reference position of user's input and destination locations determination alternative path set, path in alternative path set is all be starting point with reference position, destination locations is the set in the path of terminal, makes equipment that the vehicle flowrate data acquisition of road network can be utilized from alternative path set to select hotspot path.

209, from vehicle flowrate data acquisition, obtain the vehicle flowrate data in each section of each path candidate in alternative path set;

In embodiments of the present invention, equipment, after the vehicle flowrate data acquisition obtaining road network and alternative path set, will obtain the vehicle flowrate data in each section of each path candidate in alternative path set from vehicle flowrate data acquisition.

210, the hotspot path based on probability is determined according to the vehicle flowrate data in the section of the path candidate in alternative path set and the family of functions that pre-sets.

In embodiments of the present invention, family of functions coma-α is for comparing probability size between two paths, and wherein, the scope of parameter alpha is [0,1].

Wherein, determine to be specifically as follows based on the hotspot path of probability according to the vehicle flowrate data in the section of the path candidate in alternative path set and the family of functions that pre-sets:

1) the vehicle flowrate data in all sections of the path candidate of each in alternative path set compare according to the family of functions pre-set by equipment, sort according to vehicle flowrate data order from small to large, as the path temperature of each path candidate;

Concrete, with section e ₁with section e ₂carry out likelihood ratio comparatively example be described, section e ₁with section e ₂vehicle flowrate data be respectively F ₁and F ₂, utilize F ₁and F ₂calculate P ₁and P ₂, wherein, P ₁represent section e ₁vehicle flowrate data F ₁be greater than section e ₂vehicle flowrate data F ₂probability, can be expressed as: P ₁=P{ActualFlow (e ₁) <ActualFlow (e ₂), P ₂represent section e ₂vehicle flowrate data F ₂be greater than section e ₁vehicle flowrate data F ₁probability, can be expressed as: P ₁=P{ActualFlow (e ₁) <ActualFlow (e ₂).

Concrete: make F ₁=(f ₀, f ₁..., f _s), F ₂=(g ₀, g ₁..., g _t)

Then

Wherein, 0<=i<=s, 0<=j<=t.

Wherein, race function comp-α is utilized to compare section e ₁with section e ₂vehicle flowrate data F ₁and F ₂size, be specially:

If | P ₁-P ₂| > α, and P ₁>P ₂time, Comp-α (F ₁, F ₂) value be-1, i.e. F ₁<F ₂; If | P ₁-P ₂| > α, and P ₁<P ₂time, Comp-α (F ₁, F ₂) value be 1, i.e. F ₁>F ₂; In other situations, Comp-α (F ₁, F ₂) value be 0, i.e. F ₁=F ₂.

Likelihood ratio can be carried out comparatively by the way to the vehicle flowrate data in all sections of path candidate, sort according to vehicle flowrate data order from small to large, such as: path candidate comprises section a, b, c, d, e, then can first by the vehicle flowrate data of section a respectively with section b, c, d, the vehicle flowrate data of e compare, by the vehicle flowrate data of section b respectively with section c, d, the vehicle flowrate data of e compare, by the vehicle flowrate data of section c respectively with section d, the vehicle flowrate data of section e compare, the vehicle flowrate data of section d and the vehicle flowrate data of section e are compared, in conjunction with all comparative results, section a can be determined, b, c, d, the vehicle flowrate data sequence from small to large of e, by the path temperature in this ranking results alternatively path.

2) from alternative path set, select two path candidates, respectively as path candidate a and path candidate b, wherein, the path temperature of path candidate a is Freqa, and Freqa=<F _a1, F _a2..., F _aLathe vehicle flowrate data in La section are comprised, the path temperature Freq of path candidate b in >, Freqa _b, and Freq _b=<F _b1, F _b2..., F _bLb>, Freq _bin comprise the vehicle flowrate data in Lb section;

Judge that the path temperature of path candidate a is whether higher than the path temperature of path candidate b;

If there is a k, and the span of k be 1,2 ...., min (La, Lb) }, make F _ak<F _bk; And to all i being less than k, meet F _ai=F _bi, or when La<Lb for all i being less than or equal to La, meet F _ai=F _biwherein, k and i is positive integer, then determine that the path temperature of path candidate b is higher than path candidate a, if and comprise not by the path candidate selected in alternative path set, then make path candidate a=path candidate c, return the path temperature that performs and judge path candidate a whether higher than the step of the path temperature of path candidate b, path candidate c is not by the path candidate selected in alternative path set; If do not comprise in alternative path set not by the path candidate selected, then determine that path candidate b is hotspot path;

If there is a k, and the span of k be 1,2 ...., min (La, Lb) }, make F _ak>F _bk; And to all i being less than k, meet F _ai=F _bi, or for all i being less than or equal to Lb, meet F at La>Lb _ai=F _biwherein, k and i is positive integer, then determine that the path temperature of path candidate a is higher than path candidate b, if and comprise not by the path candidate selected in alternative path set, make path candidate b=path candidate c, return the path temperature that performs and judge path candidate a whether higher than the step of the path temperature of path candidate b, path candidate c is not by the path candidate selected in alternative path set, if do not comprise in alternative path set not by the path candidate selected, then determine that path candidate a is hotspot path;

If La=Lb, and the span of k be 1,2 ...., during La or Lb}, F _ai=F _bithen determine that path candidate a is identical with the path temperature of path candidate b, if and comprise not by the path candidate selected in alternative path set, then make path candidate a=path candidate c, return the path temperature that performs and judge path candidate a whether higher than the step of the path temperature of path candidate b, or, make path candidate b=path candidate c, return the path temperature that performs and judge path candidate a whether higher than the step of the path temperature of path candidate b, path candidate c is not by the path candidate selected in alternative path set; If do not comprise in alternative path set not by the path candidate selected, then determine that path candidate a or path candidate b is hotspot path.

In embodiments of the present invention, equipment calculates the candidate matches limit collection of each node in each measuring route in the user trajectory measurement data got according to topology information, and utilize the candidate matches limit collection of each node in each measuring route to build the Hidden Markov Model (HMM) of each measuring route, fuzzy map matching primitives is carried out based on Hidden Markov Model (HMM), obtain the feasible path set of each measuring route, determine the set of the number of times S that each section in road network occurs in all feasible path set of all measuring route and the section probability that S time occurs, the set of the section probability occurred according to the occurrence number S in each section and S time calculates the vehicle flowrate data in each section, if the vehicle flowrate data in all sections are non-sparse data, then determine that the vehicle flowrate data in this all section are the vehicle flowrate data acquisition in the section of road network, if the vehicle flowrate data in all sections are sparse data, then based on the common triage techniques of multiple linear regression, or fill up based on the common triage techniques of multiple linear regression and the training mechanism of semi-supervised learning the vehicle flowrate data belonging to unavailable section in all sections, using the vehicle flowrate data acquisition of the vehicle flowrate data in all sections after optimization as the section of road network, and from this vehicle flowrate data acquisition, obtain the vehicle flowrate data in each section of each path candidate in fixed alternative path set, utilize the flow distribution value that the vehicle flowrate data in each section of the path candidate in the family of functions's calculated candidate set of paths pre-set are corresponding, the hotspot path based on probability is determined according to the flow distribution value in each section of each paths, by the fuzzy map matching primitives based on Hidden Markov Model (HMM), the a part of uncertain information in measurement data can be retained in the feasible path set of the measuring route obtained, obtain the path profile of likely mating, the accuracy that effective raising hotspot path calculates, simultaneously, when determining hotspot path, based on the common triage techniques of multiple linear regression, or, the vehicle flowrate data belonging to unavailable section in all sections are filled up based on the common triage techniques of multiple linear regression and the training mechanism of semi-supervised learning, the problem cannot determining hotspot path avoiding sparse data to bring, improve the accuracy that hotspot path is determined.

It should be noted that, the defining method of the hotspot path described in the embodiment of the present invention is generally used for user vehicle determination traffic route, user can install hotspot path locking equipment really on automobile, or the software that can perform the defining method of above-mentioned hotspot path is installed in the existing control system of automobile, determines traffic route to help user

Refer to Fig. 3, be the structure of the determining device of hotspot path in the embodiment of the present invention, comprise:

First computing unit 301, for carrying out the fuzzy map matching primitives based on Hidden Markov Model (HMM) to each measuring route in the user trajectory measurement data got, obtain the feasible path set of each measuring route described, comprise normalization matching probability in all feasible paths of a measuring route in feasible path set described in and come the feasible path of front N, described user trajectory measurement data is the trajectory measurement data of the user of the road network travelling on preset regions in preset time period, and described N is positive integer;

First determining unit 302, for obtain each measuring route described at described first computing unit 301 feasible path set after, utilize the described feasible path set of each measuring route and the topology information of described road network to determine the vehicle flowrate data in all sections of described road network;

Second determining unit 303, after determining the vehicle flowrate data in all sections of described road network in described first determining unit 302, utilize described topology information, user input reference position and destination locations determination alternative path set, path in described alternative path set is all with described reference position for starting point, and described destination locations is the set in the path of terminal;

3rd determining unit 304, after obtaining described alternative path set in described second determining unit 303, determines the hotspot path based on probability according to the vehicle flowrate data in described all sections and described alternative path set.

In embodiments of the present invention, first computing unit 301 of hotspot path really in locking equipment carries out the fuzzy map matching primitives based on Hidden Markov Model (HMM) to each measuring route in the user trajectory measurement data got, obtain the feasible path set of each measuring route described, wherein, comprise normalization matching probability in all feasible paths of a measuring route in feasible path set described in and come the feasible path of front N, described user trajectory measurement data is the trajectory measurement data of the user of the road network travelling on preset regions in preset time period, described N is positive integer, then the first determining unit 302 utilizes the described feasible path set of each measuring route and the topology information of described road network to determine the vehicle flowrate data in all sections of described road network, and by the second determining unit 303 utilize described topology information, user input reference position and destination locations determination alternative path set, path in described alternative path set is all with described reference position for starting point, and described destination locations is the set in the path of terminal, finally, the 3rd determining unit 304, determines the hotspot path based on probability according to the vehicle flowrate data in described all sections and described alternative path set.

The determining device of the hotspot path for a better understanding of the present invention in embodiment, refer to Fig. 4, comprise: the first computing unit 301, first determining unit 302, second determining unit 303 and the 3rd determining unit 304 described in embodiment as shown in Figure 3, and similar to the content of description embodiment illustrated in fig. 3, do not repeat herein.

In embodiments of the present invention, equipment also comprises:

Judging unit 401, after obtaining described alternative path set in described second determining unit 302, judges whether the vehicle flowrate data in described all sections are non-sparse data;

4th determining unit 402, if determine that the vehicle flowrate data in described all sections are non-sparse data for described judging unit 401, then determines that the vehicle flowrate data in described all sections are the vehicle flowrate data acquisition in the section of described road network;

5th determining unit 403, if determine that the vehicle flowrate data in described all sections are sparse data for described judging unit 401, then based on the common triage techniques of multiple linear regression, or fill up based on the common triage techniques of described multiple linear regression and the training mechanism of semi-supervised learning the vehicle flowrate data belonging to unavailable section in described all sections, the vehicle flowrate data in all sections after being optimized, using the vehicle flowrate data acquisition of the vehicle flowrate data in all sections after described optimization as the section of described road network, described unavailable section refers to that vehicle flowrate data are less than the section of the threshold value pre-set.

In embodiments of the present invention, the 3rd determining unit 302, specifically for after described 4th determining unit 402 or described 5th determining unit 403, determines the hotspot path based on probability according to described vehicle flowrate data acquisition and described alternative path set.

Wherein, the 5th determining unit 403 specifically for:

If described judging unit 401 determines that the vehicle flowrate data in described all sections are sparse data, be matrix M, and matrix M is as follows by the vehicle flowrate data construct in described all sections:

Or, the 5th determining unit 403 specifically for:

In embodiments of the present invention, the first computing unit 301 comprises:

Limit collection computing unit 404, for calculating the candidate matches limit collection of each node in each measuring route in the trajectory measurement data of the user got according to described topology information, described candidate matches limit integrates the set in the section that may be positioned at as node;

Feasible path computing unit 405, after obtaining described candidate matches limit collection at described limit collection computing unit 404, utilize the candidate matches limit collection of each node in each measuring route described to build the Hidden Markov Model of each measuring route described, obtain the feasible path set of each measuring route described.

And limit collection computing unit 404 specifically for:

Described matching formula is:

P{ (X _i, Y _i) | (X ' _i, Y ' _i) on e > θ

And feasible path computing unit 405 specifically for:

In embodiments of the present invention, the first determining unit 302 comprises:

6th determining unit 406, for obtain each measuring route described at described first computing unit 301 feasible path set after, determine the set of the number of times S that each section in described road network occurs in all feasible path set of described all measuring route and the section probability that S time occurs, described S is positive integer;

Second computing unit 407, after set for the section probability of the occurrence number S and S appearance that determine described each article of section in described 6th determining unit 406, the set of the section probability occurred according to the number of times S of the appearance in each section described and S time calculates the vehicle flowrate data in each section, comprise the set that section occurrence number is the probability of k in described vehicle flowrate data, the value of described k is 1 to S.

In embodiments of the present invention, the 3rd determining unit 304 comprises:

Acquiring unit 408, for after described 4th determining unit 402 or described 5th determining unit 403, obtains the vehicle flowrate data in each section of each path candidate in described alternative path set from described vehicle flowrate data acquisition;

3rd computing unit 409, after obtaining the vehicle flowrate data in each section of each path candidate in described alternative path set at described acquiring unit 408, the family of functions that pre-sets is utilized to calculate flow distribution value corresponding to the vehicle flowrate data in each section of the path candidate in described alternative path set;

6th determining unit 410, for after the flow distribution value that the vehicle flowrate data obtaining described each article of section at described 3rd computing unit 409 are corresponding, the flow distribution value corresponding according to the vehicle flowrate data in each section of each paths described determines the hotspot path based on probability.

In embodiments of the present invention, the 6th determining unit 410 specifically for:

In embodiments of the present invention, limit collection computing unit 404 in first computing unit 301 calculates the candidate matches limit collection of each node in each measuring route in the trajectory measurement data of the user got according to topology information, concrete: limit collection computing unit 404 determines the candidate matches limit collection of each node of measuring route in the following manner:

Described matching formula is:

P{ (X _i, Y _i) | (X ' _i, Y ' _i) on e > θ

Then, the candidate matches limit collection of each node in each measuring route is utilized to build the Hidden Markov Model (HMM) of each measuring route by the feasible path computing unit 405 in the first computing unit 301, obtain the feasible path set of each measuring route, and concrete: feasible path computing unit 405 determines the feasible path set of each measuring route in the following manner: utilize the candidate matches limit collection of each node on described measuring route A to build Hidden Markov Model (HMM), determine all feasible paths of described measuring route A, the probability in the section on described feasible path is the probability of candidate matches limit centralized node on described section, based on the candidate matches limit collection of the topology information of road network and each node of measuring route A, transition probability when adjacent two nodes get different candidate matches limits in computation and measurement path A, according to the probability in the section on each feasible path of described measuring route A and the matching probability of described transition probability calculating each feasible path described, described matching probability equals the product between the transition probability between the probability in all sections of feasible path and the connected section of described feasible path, from described all feasible paths, select matching probability to come the feasible path set of feasible path as described measuring route A of front N, and the matching probability of each feasible path in described feasible path set is normalized the normalization matching probability obtaining each feasible path described.

Then, the 6th determining unit 406 in first determining unit 302 determines the set of the number of times S that each article of section in described road network occurs in all feasible path set of described all measuring route and the section probability that S time occurs, described S is positive integer; And the set of the section probability occurred according to the number of times S of the appearance in each section described and S time by the second computing unit 407 in the first determining unit 302 calculates the vehicle flowrate data in each section, comprise the set that section occurrence number is the probability of k in described vehicle flowrate data, the value of described k is 1 to S.

Then the reference position that the second determining unit 303 utilizes described topology information, user inputs and destination locations determination alternative path set, and judging unit 401 judges whether the vehicle flowrate data in described all sections are non-sparse data, if described judging unit 401 determines that the vehicle flowrate data in described all sections are non-sparse data, then the 4th determining unit 402 determines that the vehicle flowrate data in described all sections are the vehicle flowrate data acquisition in the section of described road network, if described judging unit 401 determines that the vehicle flowrate data in described all sections are sparse data, then the 5th determining unit 403 is based on the common triage techniques of multiple linear regression, or fill up based on the common triage techniques of described multiple linear regression and the training mechanism of semi-supervised learning the vehicle flowrate data belonging to unavailable section in described all sections, the vehicle flowrate data in all sections after being optimized, using the vehicle flowrate data acquisition of the vehicle flowrate data in all sections after described optimization as the section of described road network, described unavailable section refers to that vehicle flowrate data are less than the section of the threshold value pre-set.

Wherein, concrete, if described judging unit 401 determines that the vehicle flowrate data in described all sections are sparse data,

Or if described judging unit 401 determines that the vehicle flowrate data in described all sections are sparse data, the vehicle flowrate data construct in described all sections is matrix M by the 5th determining unit 403 ₀, and matrix M ₀as follows:

After the 4th determining unit 402 or the 5th determining unit 403 obtain the vehicle flowrate data in each article of section of each article of path candidate in alternative path set, acquiring unit 408 in 3rd determining unit 304 obtains the vehicle flowrate data in each article of section of each article of path candidate in described alternative path set from described vehicle flowrate data acquisition, and the 3rd computing unit 409, utilize the family of functions pre-set to calculate flow distribution value corresponding to the vehicle flowrate data in each section of the path candidate in described alternative path set; And determine the hotspot path based on probability by the flow distribution value that the 6th determining unit 410 is corresponding according to the vehicle flowrate data in each article of section of each paths described.

Wherein, concrete, the 6th determining unit 410 processes in the following manner:

Refer to Fig. 5, be the constructive embodiment of the locking equipment really of hotspot path in the embodiment of the present invention, comprise:

Processor 501, receiving trap 502, dispensing device 503 and storer 504;

Wherein, processor 501 is for carrying out the fuzzy map matching primitives based on Hidden Markov Model (HMM) to each measuring route in the user trajectory measurement data got, obtain the feasible path set of each measuring route described, wherein, comprise normalization matching probability in all feasible paths of a measuring route in feasible path set described in and come the feasible path of front N, described user trajectory measurement data is the trajectory measurement data of the user of the road network travelling on preset regions in preset time period, and described N is positive integer; The described feasible path set of each measuring route and the topology information of described road network is utilized to determine the vehicle flowrate data in all sections of described road network; Utilize described topology information, user input reference position and destination locations determination alternative path set, the path in described alternative path set is all with described reference position for starting point, and described destination locations is the set in the path of terminal; The hotspot path based on probability is determined according to the vehicle flowrate data in described all sections and described alternative path set.

One of ordinary skill in the art will appreciate that all or part of step realized in above-described embodiment method is that the hardware that can carry out instruction relevant by program completes, described program can be stored in a kind of computer-readable recording medium, the above-mentioned storage medium mentioned can be ROM (read-only memory), disk or CD etc.

Above the determination method and apparatus of a kind of hotspot path provided by the present invention is described in detail, for one of ordinary skill in the art, according to the thought of the embodiment of the present invention, all will change in specific embodiments and applications, in sum, this description should not be construed as limitation of the present invention.

Accompanying drawing explanation

Fig. 1 is a schematic diagram of the defining method of hotspot path in the embodiment of the present invention;

Fig. 2 is another schematic diagram of the defining method of hotspot path in the embodiment of the present invention;

Fig. 3 is a schematic diagram of the determining device of hotspot path in the embodiment of the present invention;

Fig. 4 is another schematic diagram of the determining device of hotspot path in the embodiment of the present invention;

Fig. 5 is another schematic diagram of the determining device of hotspot path in the embodiment of the present invention.

Embodiment

Below by specific embodiment, be described in detail respectively.

For making goal of the invention of the present invention, feature, advantage can be more obvious and understandable, below in conjunction with the accompanying drawing in the embodiment of the present invention, technical scheme in the embodiment of the present invention is clearly and completely described, obviously, the embodiments described below are only the present invention's part embodiments, and the embodiment of not all.Based on the embodiment in the present invention, those of ordinary skill in the art, not making other embodiments all obtained under creative work prerequisite, belong to the scope of protection of the invention.

Claims

1. a defining method for hotspot path, is characterized in that, comprising:

2. method according to claim 1, is characterized in that, the described vehicle flowrate data according to described all sections and described alternative path set also comprise before determining the hotspot path based on probability:

3. method according to claim 2, is characterized in that, the described common triage techniques based on multiple linear regression fills up the vehicle flowrate data belonging to unavailable section in described all sections, and the vehicle flowrate data in all sections after being optimized, comprising:

4. method according to claim 2, it is characterized in that, the training mechanism of the described common triage techniques based on described multiple linear regression and semi-supervised learning fills up the vehicle flowrate data belonging to unavailable section in described all sections, the vehicle flowrate data in all sections after being optimized, comprising:

5. method according to claim 1 and 2, it is characterized in that, each measuring route in the described user trajectory measurement data to getting carries out the fuzzy map matching primitives based on Hidden Markov Model (HMM), obtains the feasible path set of each measuring route described, comprising:

6. method according to claim 5, is characterized in that, the described topology information according to described road network calculates the candidate matches limit collection of each node in each measuring route described, comprising:

Described matching formula is:

P{ (X _i, Y _i) | (X ' _i, Y ' _i) on e > θ

7. method according to claim 6, it is characterized in that, the candidate matches limit collection of each node in each measuring route described in described utilization builds the Hidden Markov Model of each measuring route described, and the feasible path set obtaining each measuring route described comprises:

8. method according to claim 7, is characterized in that, the vehicle flowrate data in all sections of described road network are determined in the feasible path set of each measuring route described in described utilization, comprising:

9. method according to claim 7, is characterized in that, describedly determines to comprise based on the hotspot path of probability according to described vehicle flowrate data and described alternative path set:

10. method according to claim 9, is characterized in that, the flow distribution value that the vehicle flowrate data in each section of each paths described in described basis are corresponding is determined to comprise based on the hotspot path of probability:

11. 1 kinds of hotspot path locking equipment really, is characterized in that, comprising:

12. equipment according to claim 11, is characterized in that, described equipment also comprises:

13. equipment according to claim 12, is characterized in that, described 5th determining unit specifically for:

14. equipment according to claim 12, is characterized in that, described 5th determining unit specifically for:

15. equipment according to claim 11 or 12, it is characterized in that, described first computing unit comprises:

16. equipment according to claim 15, is characterized in that, described limit collection computing unit specifically for:

Described matching formula is:

P{ (X _i, Y _i) | (X ' _i, Y ' _i) on e > θ

17. equipment according to claim 16, is characterized in that, described feasible path computing unit specifically for:

18. equipment according to claim 17, is characterized in that, described first determining unit comprises:

19. equipment according to claim 17, is characterized in that, described 3rd determining unit comprises:

20. equipment according to claim 19, is characterized in that, described 6th determining unit specifically for: