CN101308487A - Space-time fusion method for natural language expressing dynamic traffic information - Google Patents
Space-time fusion method for natural language expressing dynamic traffic information Download PDFInfo
- Publication number
- CN101308487A CN101308487A CNA2008101155771A CN200810115577A CN101308487A CN 101308487 A CN101308487 A CN 101308487A CN A2008101155771 A CNA2008101155771 A CN A2008101155771A CN 200810115577 A CN200810115577 A CN 200810115577A CN 101308487 A CN101308487 A CN 101308487A
- Authority
- CN
- China
- Prior art keywords
- speech
- road
- information
- vocabulary
- address
- 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
Images
Landscapes
- Navigation (AREA)
Abstract
Disclosed is a temporal-spatial fusion method for a natural language to express dynamic traffic information, comprising the following steps: to make data pre-processing to the dynamic traffic information described in the form of Chinese natural language to eliminate errors or invalid information; to take the pre-processed result as an input parameter to improve the maximum matching algorithm and segment the key words in the algorithm; then to make use of the address positioning vocabulary reconstruction method to process the address positioning vocabularies and organize each key word based on the time, location, direction, offset and the structure of the event so as to express the dynamic traffic information in a formalized way; to process the formally expressed traffic key words through a temporal-spatial matching method facing the information fusion, and to ultimately achieve the purpose of temporal-spatial fusion between the dynamic traffic information and the underlying road network information. The method of the invention has the advantages of intelligence, automation, strong fault-tolerant capability, high fusion accuracy and wide adaptability, which solves the common problem of difficult real-time matching and fusion between the dynamic traffic information described in natural language and the underlying road network spatial information.
Description
Technical field
The present invention relates to a kind of space-time fusion method of natural language expressing dynamic traffic information, be applied to the issue of dynamic information, serve position service system, map web site and professional trip information system platform, logistics dispatching system and the bulletins of emergency traffic prediction scheme such as vehicle mounted guidance and individual mobile navigation service.
Background technology
Along with Chinese society economy and science and technology development, the urban transportation trip requirements constantly increases.Meanwhile, be that the service idea at center is risen rapidly with the public's demand.Information asymmetry situation between government bodies, administrative authority and the public is constantly obtaining changing.The cry of carrying out public information service is more and more higher.Under this background, the trip information service is arisen at the historic moment.City public's trip information service is the public information service that provides real-time trip to instruct towards the public, it is an information source with urban traffic network space and semantic information, real-time traffic and control traffic message and other relevant information, utilize data fusion and integrated, Intelligent treatment and analytical technology, under electronic map platform is supported, realize the trip information service of various ways, comprise position service (LBS) application, map web site and professional trip information system platform, logistics dispatching system and emergency traffic prediction schemes etc. such as vehicle mounted guidance and individual mobile navigation service.
Yet, in present various forms of trip information service systems and platform, all there is a common technical bottleneck---a large amount of dynamic informations are difficult to mate in real time with bottom road network spatial information and merge, be difficult to offer system applies such as vehicle mounted guidance, map web site, logistics scheduling, make public's service of going on a journey also rest on the visual and inquiry phase of static monotype electronic chart substantially, can't serve dynamic many standard trips effectively, practical value is had a greatly reduced quality.Cause the primary crux of this problem to be the natural language description form of transport information.Present dynamic information, the a small amount of video information obtained except the traffic camera and the telecommunication flow information that obtains by Floating Car, inductive coil etc., more be the transport information that adopts the natural language formal description, comprise that traffic administration person and participant are transmitted by phone, short message, website etc. and the traffic of announcement and control information etc.Have natural language description qualitative, fuzzy characteristics and meet the custom of people, but be difficult to directly be understood by existing computer system and handle, seriously restricted the level of utilizing of transport information transport information or event description.Natural language understanding achievement in research both domestic and external helps to address this problem.But, dynamic information and road network spatial information fusion application have only be concerned about the traffic key vocabularies understand and ignore irrelevant vocabulary, priority of long word cutting, support fuzzyly to write, characteristics such as situs ambiguus appointment, conventional natural language understanding method can't meet the demands.In addition, even if how the transport information understood of success also has been present insoluble technical bottleneck problem with the real-time coupling of road network spatial information.Most of at present research also rests on static state or half dynamic fusion stage, is mating fusion with the road network spatial information on static coupling or the artificial cognition basis in advance.The problem reason is that there is the isomerism on the system in the two-dimensional space coordinate setting mode of multiple linear orientation mode and road network spatial information in the transport information.These problems make a large amount of at present dynamically issues, transport information that have the different time action effect, the natural language formal description can't be able to effective processing, and with road network spatial information Dynamic matching and fusion, be applied to the service of public's trip information.In the face of the requirement of various forms of trip information service fast developments, press for the space-time fusion method of a kind of natural language expressing dynamic traffic information of development.
Summary of the invention
The technical matters that the present invention solves is: but single at the dynamic information source of present digital applications, dynamic information with Chinese natural language description also rests on the artificial speech broadcasting stage in a large number, be difficult to mate in real time with bottom road network spatial information and merge, and then can't be used for the trip information service platform, dynamic onboard navigation system, the present situation of application systems such as map web site, a kind of space-time fusion method of natural language expressing dynamic traffic information is proposed, this method solved Chinese natural language description dynamic information intelligent handle and dynamic information in the isomerism difficult problem of two-dimensional space coordinate setting mode of multiple linear orientation mode and road network spatial information, can directly apply to the trip information service, application systems such as dynamic vehicle mounted guidance.
Technical solution of the present invention is: a kind of space-time fusion method of natural language expressing dynamic traffic information, and at first the dynamic information to Chinese natural language formal description carries out the data pre-service, removes mistake or invalid information; After with the pre-service result as input parameter, adopt key vocabularies and formalization in the natural language processing method resolving information of information fusion to express transport information; And then utilize the space-time matching technique to handle formal transport information, and realizing that finally dynamic information and bottom road network information Space merge purpose, concrete steps are as follows:
(1) to obtain in real time by modes such as phone, short message, video decipher, literal issues, carry out the data pre-service with the transport information of Chinese natural language formal description, eliminate mistake or invalid information;
(2) with the pre-service result as input parameter, be syncopated as wherein key vocabularies to improve maximum matching method, and then use address location vocabulary reconstruct method to handle address location vocabulary, and according to each key vocabularies of structure organization of time, place, direction, incident, side-play amount, dynamic information is expressed in formalization;
(3) to through formal dynamic information, utilize space-time matching process, realize the temporal-spatial fusion of dynamic information and bottom road network information towards information fusion.
The present invention's advantage compared with prior art is: the transport information that the present invention overcomes existing natural language description is difficult to the shortcoming of mating in real time and merging with bottom road network spatial information, transform specialized vocabulary library structure and data organizational structure based on the information fusion characteristics, homophonic fuzzy matching technology is embedded in the maximum matching algorithm, utilize address location vocabulary reconstruct method to handle address location vocabulary, and according to the time, the place, direction, side-play amount, each key vocabularies of the structure organization of incident, dynamic information is expressed in formalization, and integrated ageing processing, the road network associating information, the positioning reference inquiry, the width of cloth search of heuristic road, space-time coupling means such as space length conversion, the isomery of the two-dimensional space coordinate setting mode of multiple linear orientation mode and road network spatial information in the breakthrough transport information has successfully made up the dynamic information of natural language description and the integrated space-time of road network spatial information and has mated and integration program.
Description of drawings
Fig. 1 is the process flow diagram of improvement maximum matching algorithm of the present invention;
Fig. 2 is the process flow diagram of the space-time matching process towards information fusion of the present invention.
Embodiment
Concrete implementation step of the present invention is as follows:
1, to obtain in real time by modes such as phone, short message, video decipher, literal issues, the original traffic information of Chinese natural language formal description is carried out simply fast artificial pre-service, promptly delete mess code information and imperfect information, the tangible misspelling information of editor, the redundant information that deletion repeats is to eliminate apparent error or the invalid data in the source-information as far as possible;
2, with the pre-service result as input parameter, employing improvement maximum matching algorithm is syncopated as key vocabularies wherein, and then use address location vocabulary reconstruct method to handle address location vocabulary, and according to each key vocabularies of structure organization of time, place, direction, side-play amount, incident, dynamic information is expressed in formalization.
As shown in Figure 1, the improvement maximum matching algorithm specific implementation method among the present invention is as follows:
(1) characteristics of using according to information fusion are organized specialized dictionary and data structure, and specialized vocabulary is imported in the desired data structure.
Specialized dictionary comprises Chinese library and phonetic storehouse.Chinese vocabulary bank comprises address base, direction storehouse, event base and attached location dictionary again.Address base has comprised the title of all atural objects in some specific regions; The direction library storage the various expression of directional information in the traffic events; Event base has been stored various description of status information in the traffic events; The vocabulary that is used in combination, points to final positioning address with address vocabulary but itself can not location-independent stored in attached location dictionary.The phonetic library storage with Chinese vocabulary bank in the corresponding phonetic of each vocabulary describe, comprise phonetic storehouse, address, direction phonetic storehouse, incident phonetic storehouse and speech phonetic storehouse, attached location.
The store data structure of Chinese vocabulary bank adopts multilayered schema.Ground floor storage individual character, second layer storage is the double word or the two-character word of prefix with the word string of ground floor, having stored with the word string of the second layer for the 3rd layer is three words or three words of prefix, by that analogy.The data structure of pinyin lexicon is consistent with Chinese vocabulary bank, but the minimum memory unit is the phonetic rather than the Chinese individual character of single word.
(2) with the input of transport information as the improvement maximum matching algorithm, be syncopated as the traffic key vocabularies, concrete cutting step is as follows:
1.. choose first word C in the transport information
1As current word string, in Chinese vocabulary bank, search C
1Whether exist;
2.. if do not have C in the Chinese vocabulary bank
1, then with C
1Be converted to phonetic, and in pinyin lexicon, search with C
1For whether the speech of phonetic exists;
3.. if do not have C in the pinyin lexicon
1, cutting C then
1, mark C
1For non-vocabulary, from C
2Begin participle next time;
4.. if having C in the pinyin lexicon
1, get next word C
2, current word string is composed and is C
1C
2
5.. in pinyin lexicon, further judge with current word string C
1C
2For whether the speech of phonetic exists;
6.. do not exist, then the last position of the current word string of cutting word string C
1, record C
1Be subordinate to the sign of dictionary, participle finishes, change over to step (
);
7.. otherwise, according to becoming the current word string C of speech marker for judgment
1C
2Whether become speech.If become speech, preserve C
1C
2Express, change step (8.) over to; Do not become speech, directly change step (8.) over to;
8.. constantly next word C is got in circulation
i, judge in the pinyin lexicon with current word string C
1C
2C
iFor whether the pinyin word of prefix exists;
9.. if do not exist, obtain the expression C that the last time can become speech
1C
2C
I-kPhonetic, return C
1C
2C
I-kThe pairing standard Chinese vocabulary of phonetic, and record C
1C
2C
I-kBe subordinate to the mark of dictionary;
10.. otherwise, current word string is made as C
1C
2C
i, change step (7.) over to;
Otherwise, according to becoming the current word string C of speech marker for judgment
1C
2Whether become speech.If become speech, preserve C
1C
2Express, change over to step (
); Do not become speech, directly change over to step (
);
Otherwise, judge current word string C
1C
2C
iWhether become speech.If become speech, preserve C
1C
2C
iExpress, change over to step (
); Do not become speech, directly change over to step (
);
(3) character string of the cutting that fails in the source-information is carried out digital judgment processing, the disposable numeric type information that extracts wherein, with this as side-play amount.
Address location vocabulary reconstruct method specific implementation method is as follows among the present invention:
A. determine to judge the method for the reference element type that transport information was subordinate to, ending character with positioning address name in the transport information is characterized as the basis, by enumerating the reference element type that the mode inferential information is subordinate to, the reference element type is road, bridge, crossing and POI point four big classes.If positioning address positioning address word combination by name is then got crucial ending character of locating speech in the combination.Described road shows as single positioning address word form, Yi Lu, lining, bar, street, alleyway, road, section, lane, ring, ring East Road, ring West Road, ring South Road, ring North Road or east side road, west side road, southern side road, the ending of north side road.Described bridge shows as single positioning address speech or positioning address word combination form, ends up on Yi Qiao, the bridge or under the bridge.There are the two kinds of forms in road terminus crossing and middle crossing in described crossing.Road terminus crossing shows as single positioning address speech or positioning address word combination form, to the east of mouth, western entrance, Nan Kou, Bei Kou ending.The crossing shows as positioning address word combination form in the middle of the road, is subdivided into common crossing, loop crossing and special crossing again.Common crossing Yi Lu, lining, bar, street, alleyway, road, section, lane ending; The loop crossing is with ring, ring East Road, ring West Road, ring South Road, the ending of ring North Road; Dypass, west side road, southern side road, the ending of north side road to the east of the special crossing.Middle crossing also exists with " crossing " ending form, and need obtain and be positioned at " crossing " information representation before this moment, further judges the crossing type according to the ending at three kinds of middle crossings.Described POI point shows as single positioning address speech or positioning address word combination form, and except that above-mentioned several types ending, other all is defaulted as the POI point.
B. according to the order that is syncopated as vocabulary in the address base, fetch bit all address vocabulary before direction vocabulary are if all preceding address vocabulary of first incident vocabulary then got in directionless vocabulary.
C. handle respectively according to the number of address vocabulary, thereby reorganize address vocabulary, form the positioning address structure that meets the information fusion requirement.Method for organizing is:
If only there is an address vocabulary A
1A
2A
n, then get A
1A
2A
nPositioning address name for the carrying traffic events;
If there is A
1A
2A
nAnd B
1B
2B
mTwo address speech need further to judge.If exist between two addresses, toward, to these three of pointing in the verbs, then previous vocabulary A
1A
2A
nBe the positioning address name of carrying traffic events starting point, a back vocabulary B
1B
2B
mPositioning address name for carrying traffic events terminal point; Otherwise two address names only are presented as an independent positioning address in essence.At this moment, adopt a back vocabulary B
1B
2B
mAs key location speech, A
1A
2A
nAs the auxiliary positioning speech, constitute crucial location speech, the positioning address word combination of auxiliary positioning word form, i.e. B
1B
2B
m, A
1A
2A
n, and with this as the positioning address name;
If there is A
1A
2A
n, B
1B
2B
mAnd C
1C
2C
kThree address speech, then first vocabulary A
1A
2A
nBe the auxiliary positioning speech, second vocabulary B
1B
2B
mBe the key location speech of bearing event starting point, the 3rd vocabulary C
1C
2C
kKey location speech for the bearing event terminal point.Further constitute B this moment respectively
1B
2B
m, A
1A
2A
nAnd C
1C
2C
k, A
1A
2A
nTwo groups address location word combination, and with this positioning address name as traffic events starting point and terminal point;
If more than three address speech, then get first three vocabulary, handle according to the disposal route of three address speech.
D. according to the method for the reference element type that transport information was subordinate to handling with the positioning address name of obtaining.If single address lexical form is then directly judged and is write down pairing key element type.If address location word combination form is then got the ending character of crucial location speech in the combination and is judged, and writes down pairing key element type.If road terminus crossing, POI vertex type, then address location word combination locating effect can be reached by crucial location vocabulary in the combination, the auxiliary positioning speech becomes redundant vocabulary, and the crucial location of address location word combination this moment speech, auxiliary positioning speech change independent crucial location vocabulary into and reach.
E. further according to the cutting order of attached location speech, obtain the positioning address name that is positioned at the last position of attached location speech, again the two is constituted same group of lexical representation, promptly crucial location speech, attached location speech or crucial location speech, auxiliary positioning speech, attached location speech forms final positioning address vocabulary.
3, to through formal dynamic information, utilize space-time matching process, realize the temporal-spatial fusion of dynamic information and bottom road network information towards information fusion.
As shown in Figure 2, the space-time matching process towards information fusion of the present invention is specific as follows:
(1) transport information of formalization expression is imported ageing processing unit and is carried out Ageing Treatment as parameter.With the current time as the time reference benchmark, a time threshold at the Real-time Traffic Information effective range is set, thus Time Created the action effect constraint set, be positioned at the transport information outside the constraint set on the rejecting time, guarantee the ageing of source-information.
(2) the time efficient transport information is sent to road network associating information cell processing and carries out association process.Road network associating information unit by using character string matching method mates positioning address vocabulary and road attribute information in the transport information, sets up the incidence relation between the two.String matching is divided into accurate coupling and fuzzy matching, and it is in full accord that accurately coupling requires two contrast word string information representations.Fuzzy matching can be handled and lack the word phenomenon between two word strings, adopts the database Fuzzy Query Technology to be achieved.In the string matching, the selected match objects of multi-form positioning address vocabulary is different with matching process.
If transport information is described with single key location word form, then directly will locate speech and road attribute information coupling;
If transport information is located speech with key, the auxiliary positioning word form is described, and then gets location speech and road attribute information coupling one by one according to the sequencing of combination.If crucial location speech successfully mates, then continue coupling auxiliary positioning speech; Otherwise directly withdraw from coupling, the failure of road network associating information cell processing;
If transport information is located speech with key, the word form of attached location is described, and then key is located the direct and road attribute information coupling of speech.If crucial location speech successfully mates, then write down attached location speech; Otherwise directly withdraw from coupling, the failure of road network associating information cell processing;
If transport information is located speech with key, the auxiliary positioning speech, the word form of attached location is described, and then gets speech and road attribute information coupling one by one according to the order of auxiliary positioning speech behind elder generation's crucial location speech.If crucial location speech successfully mates, then continue coupling auxiliary positioning speech, and write down attached location speech; Otherwise directly withdraw from coupling, the failure of road network associating information cell processing.
(3) transport information that is successfully associated is further imported the positioning reference query unit and is inquired about.According to different reference element types, obtain the pairing road width of cloth or initial sum termination topological node in the road network.If road is then directly searched the road width of cloth corresponding in the road network according to link name, then directly enter space length scaling unit and handle if there is side-play amount this moment; Otherwise, leave it at that towards the space-time matching process of information fusion.If crossing in the middle of the road should travel through all nodes of two road intersections in the road net data storehouse, with their common node as the location node collection.If the road terminus should be according to the volume coordinate coordinate extreme value node of search on the respective direction that put in order, with this as the location node collection.Meet all nodes that the bridge name is described if should in the road net data storehouse, travel through when bridge, bridge are located separately, with this as the location node collection; When bridge and road information co-located, should in the road net data storehouse, travel through all nodes of bridge and co-located road, with their common node as the location node collection.If the POI point, get with the nearest road width of cloth of POI point and oppositely the starting point of the road width of cloth and terminal point as the location node collection.At this moment, if there is attached location speech record, then should determine its key that depends on location speech, and based on key location speech the volume coordinate of corresponding location node, according to the nearest road width of cloth on the bearing sense speech detection range location node corresponding orientation in the speech of attached location, get the shortcut width of cloth and oppositely the starting point of the road width of cloth and terminal point as the location node collection of key location speech and attached location speech co-located.Further the travel direction constraint is judged on this basis, promptly with directional information as constraint condition, remove location node and concentrate and do not meet the node that direction requires, to obtain the true initial sum terminal node of traffic events.If there is not attached location speech, then directly the travel direction constraint is judged.
(4) further location node information is imported heuristic road width of cloth search unit and carried out road width of cloth search.According to the start node and the terminal node of traffic events, the mode of searching for the road width of cloth in road network connected relation tables of data is extracted the road width of cloth between origin-to-destination.Algorithm with link name and direction as heuristic information, according to first link name rear to the priority principle search, to guarantee the correctness and the validity of road width of cloth search.If there is side-play amount, then enters the space length scaling unit and handle; Otherwise, leave it at that towards the space-time matching process of information fusion.
(5) on the basis of road width of cloth search, enter the space length scaling unit and handle.Retrain road width of cloth search with side-play amount as restrictive condition, relativity shift distance under road width of cloth absolute growth and the transport information neutral line reference hierarchy under the continuous conversion road network volume coordinate system, finish search during less than current road width of cloth length up to relative displacement, and calculate relative displacement current relative position on the width of cloth of road, thereby transport information is navigated on certain point on the width of cloth of road or a certain section more exactly.Finally, the space-time that completes successfully between dynamic information and the road network spatial information mates and fusion.
For example, " garage is slow from west to east to North Star West Road mouth for sandy beach, Da Tun north of a road bridge " this transport information, issuing time is " 2008-05-01 14:30 ".By improving the processing of maximum matching algorithm, address location vocabulary reconstruct method, can derive the transport information formalization and express.Promptly, the traffic events time is " 2008-05-01 14:30 ", the traffic events starting point is " Bei Shatanqiao; Da Tun road " and belong to the bridge type, the traffic events terminal point is " North Star West Road; Da Tun road " and belong to crossing type in the middle of the ordinary road, and the traffic events direction is " from west to east ", and the traffic events type is " garage is slow ".Be complementary and merge by the transport information that formalization can be expressed towards the space-time matching process of information fusion and bottom road network spatial information.Concrete steps are as follows: the first, and the transport information that formalization is expressed is imported ageing processing unit and is carried out Ageing Treatment as parameter.According to current system time, 2 minutes time thresholds as the transport information effective range are set, structure and current time differ 2 minutes time effect effect constraint set, and then compare with issuing time " 2008-05-01 14:30 ", remove the outer time invalid information of constraint set.The second, the time efficient transport information is sent to road network associating information cell processing carries out association process.According to different location lexical forms, handle positioning address vocabulary in link name information and the transport information in the string matching mode.In the example, string matching shows as accurate coupling, and positioning address vocabulary shows as crucial location speech in traffic events starting point and the terminal point, and the auxiliary positioning word form is then got speech and link name information matches one by one according to the sequencing of address combination.Be that the traffic events starting point is successively used " Bei Shatanqiao " and " Da Tun road " and road network link name coupling, the traffic events terminal point is successively used " North Star West Road " and " Da Tun road " and road network link name coupling.The coupling of character string success has promptly been set up the incidence relation of the transport information and the road network road width of cloth.The 3rd, the transport information in the association is further imported the positioning reference query unit and is inquired about.The traffic events starting point belongs to the bridge type, and with " Da Tun road " co-located, then should in the road net data storehouse, travel through all nodes of " Bei Shatanqiao " and " Da Tun road ", with their common node as the location node collection.The traffic events terminal point belongs to crossing in the middle of the ordinary road, all nodes of traversal " North Star West Road " and " Da Tun road " in the road net data storehouse then, with their common node as the location node collection.At this moment, further removing location node according to direction " from west to east " concentrates and the incompatible topological node of directional information.The 4th, further location node information is imported heuristic road width of cloth search unit and carried out road width of cloth search.In road network connected relation table, extract start node to the road width of cloth between terminal node with road width of cloth way of search.Algorithm, is searched for according to the priority principle of elder generation " Da Tun road " back " from west to east " as heuristic information with " Da Tun road " and " from west to east ".Do not have side-play amount in the example in the transport information, the road width of cloth that this time channel width of cloth searches be successfully mate and merge on the road width of cloth.If come to exist in the source traffic information side-play amount, then needing to enter the space length scaling unit handles, retrain road width of cloth search with side-play amount as restrictive condition, constantly calculate the relative geometric displacement relation between current bias and the current road width of cloth length, finish search during less than current road width of cloth length up to current side-play amount, and then the relative position of current side-play amount on the width of cloth of current road that further convert, traffic events accurately mated and be fused on certain point on the width of cloth of current road or a certain section.
In a word, it is single that but the present invention has overcome the dynamic information source of present digital applications, can't satisfy the trip information service and the dynamic difficulty of vehicle mounted guidance application demand, advantage such as have intelligent and robotization, fault-tolerant ability is strong, fusion accuracy is high, method applicability is wide; Can be used for solving a large amount of at present natural language description dynamic informations that exist and be difficult to the problem of mating in real time and merging with bottom road network spatial information, the issue of dynamic information be can directly apply to, position service system, map web site and professional trip information system platform, logistics dispatching system and emergency traffic prediction schemes such as vehicle mounted guidance and individual mobile navigation service served.
The content that is not described in detail in the instructions of the present invention belongs to this area professional and technical personnel's known prior art.
The above only is a preferred implementation of the present invention; should be pointed out that for those skilled in the art, under the prerequisite that does not break away from the principle of the invention; can also make some improvements and modifications, these improvements and modifications also should be considered as protection scope of the present invention.
Claims (5)
1, a kind of space-time fusion method of natural language expressing dynamic traffic information is characterized in that step is as follows:
(1) to by obtain in real time, carry out the data pre-service with the transport information of Chinese natural language formal description, eliminate mistake or invalid information;
(2) with the pre-service result as input parameter, be syncopated as wherein key vocabularies to improve maximum matching algorithm, re-use address location vocabulary reconstruct method and handle address location vocabulary, and according to each key vocabularies of structure organization of time, place, direction, side-play amount, incident, dynamic information is expressed in formalization;
(3) to through formal dynamic information, utilize space-time matching process, realize the temporal-spatial fusion of dynamic information and bottom road network information towards information fusion.
2, the space-time fusion method of a kind of natural language expressing dynamic traffic information according to claim 1 is characterized in that: it is as follows to improve maximum matching algorithm in the described step (2):
A. the characteristics of at first using according to information fusion are organized specialized dictionary and data structure, and specialized vocabulary is imported in needed data structure, and specialized dictionary comprises Chinese library and phonetic storehouse; Described Chinese vocabulary bank comprises address base, direction storehouse, event base and attached location dictionary again; Address base has comprised the title of all atural objects in some specific regions; The direction library storage the various expression of directional information in the traffic events; Event base has been stored various description of status information in the traffic events; The vocabulary that is used in combination, points to final positioning address with address vocabulary but itself can not location-independent stored in attached location dictionary; The phonetic library storage with Chinese vocabulary bank in the corresponding phonetic of each vocabulary describe, comprise phonetic storehouse, address, direction phonetic storehouse, incident phonetic storehouse and speech phonetic storehouse, attached location;
B. then with the input of transport information as the improvement maximum matching algorithm, be syncopated as the traffic key vocabularies, concrete cutting step is as follows:
A. choose first word C in the transport information
1As current word string, in described Chinese vocabulary bank, search C
1Whether exist;
B. if do not have C in the Chinese vocabulary bank
1, then with C
1Be converted to phonetic, and in pinyin lexicon, search with C
1For whether the speech of phonetic exists;
C. if do not have C in the pinyin lexicon
1, cutting C then
1, mark C
1For non-vocabulary, from C
2Begin participle next time;
D. if having C in the pinyin lexicon
1, get next word C
2, current word string is composed and is C
1C
2
E. in pinyin lexicon, further judge with current word string C
1C
2For whether the speech of phonetic exists;
F. do not exist, then the last position of the current word string of cutting word string C
1, record C
1Be subordinate to the sign of dictionary, participle finishes, and changes step (q) over to;
G. otherwise, according to becoming the current word string C of speech marker for judgment
1C
2Whether become speech; If become speech, preserve C
1C
2Express, change step (h) over to; Do not become speech, directly change step (h) over to;
H. constantly next word C is got in circulation
i, judge in the pinyin lexicon with current word string C
1C
2C
iFor whether the pinyin word of prefix exists;
I. if do not exist, obtain the expression C that the last time can become speech
1C
2C
I-kPhonetic, return C
1C
2C
I-kThe pairing standard Chinese vocabulary of phonetic, and record C
1C
2C
I-kBe subordinate to the mark of dictionary;
J. otherwise, current word string is made as C
1C
2C
i, change step (g) over to;
K. if having C in the Chinese vocabulary bank
1, judge C
1Whether become speech; If become speech, preserve C
1Get C after the expression
2, do not become speech, directly get C
2, in Chinese vocabulary bank, further judge in the dictionary with C
1C
2For whether the speech of prefix exists;
L. do not exist, then change step (e) over to;
M. otherwise, according to becoming the current word string C of speech marker for judgment
1C
2Whether become speech; If become speech, preserve C
1C
2Express, change step (n) over to; Do not become speech, directly change step (n) over to;
N. constantly next word C is got in circulation
i, judge in the dictionary with C
1C
2C
iFor whether the speech of prefix exists;
O. if do not exist, current word string is made as C
1C
2C
i, change (h) over to;
P. otherwise, judge current word string C
1C
2C
iWhether become speech; If become speech, preserve C
1C
2C
iExpress, change (n) over to; Do not become speech, directly change (n) over to;
Q. from C
I-k+1Begin participle next time, the last character C in determining sentence
n
C. the character string of the cutting that fails in the source-information is carried out digital judgment processing, the disposable numeric type information that extracts wherein, with this as side-play amount.
3, the space-time fusion method of a kind of natural language expressing dynamic traffic information according to claim 2, it is characterized in that: the store data structure of described Chinese vocabulary bank adopts multilayered schema, ground floor storage individual character, second layer storage is the double word or the two-character word of prefix with the word string of ground floor, having stored with the word string of the second layer for the 3rd layer is three words or three words of prefix, by that analogy; The data structure of described pinyin lexicon is consistent with Chinese vocabulary bank, but the minimum memory unit is the phonetic rather than the Chinese individual character of single word.
4, the space-time fusion method of a kind of natural language expressing dynamic traffic information according to claim 1 is characterized in that: address location vocabulary reconstruct method is as follows in the described step (2):
A. determine to judge the method for the reference element type that transport information was subordinate to, ending character with positioning address name in the transport information is characterized as the basis, by enumerating the reference element type that the mode inferential information is subordinate to, the reference element type is road, bridge, crossing and POI point four big classes, if positioning address positioning address word combination by name is then got crucial ending character of locating speech in the combination; Described road shows as single positioning address word form, Yi Lu, lining, bar, street, alleyway, road, section, lane, ring, ring East Road, ring West Road, ring South Road, ring North Road or east side road, west side road, southern side road, the ending of north side road; Described bridge shows as single positioning address speech or positioning address word combination form, ends up on Yi Qiao, the bridge or under the bridge; There are the two kinds of forms in road terminus crossing and middle crossing in described crossing, and road terminus crossing shows as single positioning address speech or positioning address word combination form, to the east of mouth, western entrance, Nan Kou, Bei Kou ending; The crossing shows as positioning address word combination form in the middle of the road, is subdivided into common crossing, loop crossing and special crossing again, common crossing Yi Lu, lining, bar, street, alleyway, road, section, lane ending; The loop crossing is with ring, ring East Road, ring West Road, ring South Road, the ending of ring North Road; Dypass, west side road, southern side road, the ending of north side road to the east of the special crossing; Middle crossing also exists with " crossing " ending form, and need obtain and be positioned at " crossing " information representation before this moment, further judges the crossing type according to the ending at three kinds of middle crossings; The POI point shows as single positioning address speech or positioning address word combination form, and except that above-mentioned several types ending, other all is defaulted as the POI point;
B. according to the order that is syncopated as vocabulary in the address base, fetch bit all address vocabulary before direction vocabulary are if all preceding address vocabulary of first incident vocabulary then got in directionless vocabulary;
C. handle respectively according to the number of address vocabulary, thereby reorganize address vocabulary, form the positioning address structure that meets the information fusion requirement, method for organizing is:
If only there is an address vocabulary A
1A
2A
n, then get A
1A
2A
nPositioning address name for the carrying traffic events;
If there is A
1A
2A
nAnd B
1B
2B
mTwo address speech need further to judge, if exist between two addresses, toward, to these three of pointing in the verbs, then previous vocabulary A
1A
2A
nBe the positioning address name of carrying traffic events starting point, a back vocabulary B
1B
2B
mPositioning address name for carrying traffic events terminal point; Otherwise two address names only are presented as an independent positioning address in essence, at this moment, adopt a back vocabulary B
1B
2B
mAs key location speech, A
1A
2A
nAs the auxiliary positioning speech, constitute crucial location speech, the positioning address word combination of auxiliary positioning word form, i.e. B
1B
2B
m, A
1A
2A
n, and with this as the positioning address name;
If there is A
1A
2A
n, B
1B
2B
mAnd C
1C
2C
kThree address speech, then first vocabulary A
1A
2A
nBe the auxiliary positioning speech, second vocabulary B
1B
2B
mBe the key location speech of bearing event starting point, the 3rd vocabulary C
1C
2C
kKey location speech for the bearing event terminal point further constitutes B this moment respectively
1B
2B
m, A
1A
2A
nAnd C
1C
2C
k, A
1A
2A
nTwo groups address location word combination, and with this positioning address name as traffic events starting point and terminal point;
If more than three address speech, then get first three vocabulary, handle according to the disposal route of three address speech;
D. according to the method for the reference element type that transport information was subordinate to handling, if single address lexical form is then directly judged and write down pairing key element type with the positioning address name of obtaining; If address location word combination form is then got the ending character of crucial location speech in the combination and is judged, and writes down pairing key element type; If road terminus crossing, POI vertex type, then address location word combination locating effect can be reached by crucial location vocabulary in the combination, and the auxiliary positioning speech becomes redundant vocabulary; The crucial location of address location word combination this moment speech, auxiliary positioning speech change independent crucial location vocabulary into and reach;
E. further according to the cutting order of attached location speech, obtain the positioning address name that is positioned at the last position of attached location speech, again the two is constituted same group of lexical representation, promptly crucial location speech, attached location speech or crucial location speech, auxiliary positioning speech, attached location speech forms final positioning address vocabulary.
5, the space-time fusion method of a kind of natural language expressing dynamic traffic information according to claim 1 is characterized in that: the space-time matching process towards information fusion in the described step (3) is as follows:
A. the transport information of formalization expression is imported ageing processing unit and is carried out Ageing Treatment as parameter
With the current time as the time reference benchmark, a time threshold at the Real-time Traffic Information effective range is set, Time Created, the action effect constraint set was positioned at the transport information outside the constraint set on the rejecting time, guaranteed the ageing of source-information;
B. the time efficient transport information is sent to road network associating information cell processing and carries out association process
Through the transport information of ageing check, need mate with road network information, the transport information influence is reflected on the road network, this process is finished by road network associating information unit; It utilizes character string matching method that positioning address vocabulary and road attribute information in the transport information are mated, and sets up the incidence relation between the two; String matching is divided into accurate coupling and fuzzy matching, and it is in full accord that accurately coupling requires two contrast word string information representations, and fuzzy matching can be handled and lack the word phenomenon between two word strings, adopts the database Fuzzy Query Technology to be achieved; In the string matching, the selected match objects of multi-form transport information positioning address vocabulary is different with matching process:
If transport information is described with single key location word form, then directly key is located speech and road attribute information coupling;
If transport information is located speech with key, the auxiliary positioning word form is described, and then gets location speech and road attribute information coupling one by one according to the sequencing of combination;
If transport information is located speech with key, the word form of attached location is described, and then key is located the direct and road attribute information coupling of speech, and writes down attached location speech;
If transport information is located speech with key, the auxiliary positioning speech, the word form of attached location is described, and then gets speech and road attribute information coupling one by one according to the order of auxiliary positioning speech behind elder generation's crucial location speech, and writes down attached location speech;
C. the transport information in the association is further imported the positioning reference query unit and is inquired about
According to different reference element types, obtain the pairing road width of cloth or initial sum termination topological node in the road network; If road is directly searched the road width of cloth corresponding in the road network according to link name, and then directly enter the processing of space length scaling unit; If crossing in the middle of the road should travel through all nodes of two road intersections in the road net data storehouse, with their common node as the location node collection; If the road terminus should be according to the volume coordinate coordinate extreme value node of search on the respective direction that put in order, with this as the location node collection; If bridge, separately should in the road net data storehouse, travel through all nodes that meet the description of bridge name during the location, with this as the location node collection, during with the road co-located, should in the road net data storehouse, travel through all nodes of bridge and corresponding road, with their common node as the location node collection; If the POI point, get with the nearest road width of cloth of POI point and oppositely the starting point of the road width of cloth and terminal point as the location node collection; At this moment, if there is attached location speech, should be based on pairing volume coordinate, the road width of cloth nearest according to the bearing search of attached location speech got the starting point of the road width of cloth and the reverse road width of cloth and terminal point as the location node collection; Further the travel direction constraint is judged on this basis, promptly with directional information as constraint condition, remove location node and concentrate and do not meet the node that direction requires, to obtain the true initial sum terminal node of traffic events; If there is not attached location speech, then directly the travel direction constraint is judged;
D. further location node information is imported heuristic road width of cloth search unit and carried out road width of cloth search
Under the road network connected relation is supported, extract the road width of cloth between origin-to-destination with road width of cloth way of search, with link name and direction as heuristic information, according to first link name rear to the principle search, to guarantee the correctness and the validity of road width of cloth search;
E. on the basis of road width of cloth search, enter the space length scaling unit and handle and carry out space length and convert and finish space-time coupling and fusion between dynamic information and the road network spatial information
Based on side-play amount and road width of cloth length, relativity shift distance under road width of cloth absolute distance and the transport information neutral line reference hierarchy under the conversion road network volume coordinate system, thereby transport information is navigated on certain point on the width of cloth of road or a certain section more exactly, finally complete successfully the space-time coupling between dynamic information and the road network spatial information and merge.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2008101155771A CN100573506C (en) | 2008-06-25 | 2008-06-25 | A kind of space-time fusion method of natural language expressing dynamic traffic information |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2008101155771A CN100573506C (en) | 2008-06-25 | 2008-06-25 | A kind of space-time fusion method of natural language expressing dynamic traffic information |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101308487A true CN101308487A (en) | 2008-11-19 |
| CN100573506C CN100573506C (en) | 2009-12-23 |
Family
ID=40124947
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2008101155771A Expired - Fee Related CN100573506C (en) | 2008-06-25 | 2008-06-25 | A kind of space-time fusion method of natural language expressing dynamic traffic information |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100573506C (en) |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101894151A (en) * | 2010-06-24 | 2010-11-24 | 北京世纪高通科技有限公司 | Method and device for acquiring event information |
| CN102163225A (en) * | 2011-04-11 | 2011-08-24 | 中国科学院地理科学与资源研究所 | A fusion evaluation method of traffic information collected based on micro blogs |
| CN102411597A (en) * | 2011-07-28 | 2012-04-11 | 北京百度网讯科技有限公司 | Equipment and method for providing traffic information results corresponding to inquiry sequence |
| CN102831766A (en) * | 2012-07-04 | 2012-12-19 | 武汉大学 | Multi-source traffic data fusion method based on multiple sensors |
| CN102855272A (en) * | 2012-07-16 | 2013-01-02 | 中国科学院地理科学与资源研究所 | Dempster-shafer (D-S) evidence theory method capable of achieving micro-blog inclusion traffic information fusion |
| CN103049485A (en) * | 2012-11-30 | 2013-04-17 | 浙江工商大学 | City public data flow fusion method based on wavelet timing sequence arborescence outline |
| CN103118427A (en) * | 2011-11-16 | 2013-05-22 | 北京百度网讯科技有限公司 | Location method and system based on location prefetching |
| WO2013167014A1 (en) * | 2012-12-18 | 2013-11-14 | 中兴通讯股份有限公司 | Area positioning system and method |
| US8700295B2 (en) | 2009-12-28 | 2014-04-15 | Nec (China) Co., Ltd. | Method and apparatus for traffic information conversion using traffic information element knowledge base |
| WO2016107611A1 (en) * | 2015-01-04 | 2016-07-07 | 华为技术有限公司 | Method and device for text search for mobile terminal |
| CN103886080B (en) * | 2014-03-25 | 2017-01-25 | 中国科学院地理科学与资源研究所 | Method for extracting road traffic information from Internet unstructured text |
| CN106407469A (en) * | 2016-10-24 | 2017-02-15 | 北京亚控科技发展有限公司 | Method for describing time attributes of things and carrying out search based on description |
| CN106446278A (en) * | 2016-10-24 | 2017-02-22 | 北京亚控科技发展有限公司 | Method for searching data target on basis of spatial-temporal database |
| CN107818688A (en) * | 2017-11-20 | 2018-03-20 | 安徽帆扬通信电子技术有限公司 | A kind of intelligent transportation addressing navigation system based on RDS |
| CN108597223A (en) * | 2018-04-28 | 2018-09-28 | 北京智行者科技有限公司 | A kind of data processing method and system for intelligent vehicle behavior description |
| CN111731207A (en) * | 2020-06-18 | 2020-10-02 | 北京百度网讯科技有限公司 | Method, device and equipment for acquiring vehicle configuration information and storage medium |
| CN112989228A (en) * | 2021-04-25 | 2021-06-18 | 湖南视觉伟业智能科技有限公司 | Distributed space-time query method and system |
| CN113570107A (en) * | 2021-06-08 | 2021-10-29 | 众能联合数字技术有限公司 | Project address positioning method for project rental scene |
| CN115345152A (en) * | 2022-10-19 | 2022-11-15 | 北方健康医疗大数据科技有限公司 | Template library updating method, report analyzing method, device, equipment and medium |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7644054B2 (en) * | 2005-11-23 | 2010-01-05 | Veveo, Inc. | System and method for finding desired results by incremental search using an ambiguous keypad with the input containing orthographic and typographic errors |
| CN101192215B (en) * | 2006-11-24 | 2010-08-11 | 中国科学院声学研究所 | Information aggregation and enquiry method based on geographic coordinates |
| CN100478960C (en) * | 2007-08-21 | 2009-04-15 | 北京大学 | Method for locating unknown place name in network map service |
-
2008
- 2008-06-25 CN CNB2008101155771A patent/CN100573506C/en not_active Expired - Fee Related
Cited By (28)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8700295B2 (en) | 2009-12-28 | 2014-04-15 | Nec (China) Co., Ltd. | Method and apparatus for traffic information conversion using traffic information element knowledge base |
| WO2011160496A1 (en) * | 2010-06-24 | 2011-12-29 | 北京世纪高通科技有限公司 | Method and device for acquiring event information |
| CN101894151A (en) * | 2010-06-24 | 2010-11-24 | 北京世纪高通科技有限公司 | Method and device for acquiring event information |
| CN102163225A (en) * | 2011-04-11 | 2011-08-24 | 中国科学院地理科学与资源研究所 | A fusion evaluation method of traffic information collected based on micro blogs |
| CN102411597A (en) * | 2011-07-28 | 2012-04-11 | 北京百度网讯科技有限公司 | Equipment and method for providing traffic information results corresponding to inquiry sequence |
| CN103118427B (en) * | 2011-11-16 | 2015-01-07 | 北京百度网讯科技有限公司 | Location method and system based on location prefetching |
| CN103118427A (en) * | 2011-11-16 | 2013-05-22 | 北京百度网讯科技有限公司 | Location method and system based on location prefetching |
| WO2013071885A1 (en) * | 2011-11-16 | 2013-05-23 | 北京百度网讯科技有限公司 | Positioning method and system based on position prefetching |
| US9107181B2 (en) | 2011-11-16 | 2015-08-11 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Location-based positioning method and system |
| CN102831766A (en) * | 2012-07-04 | 2012-12-19 | 武汉大学 | Multi-source traffic data fusion method based on multiple sensors |
| CN102855272A (en) * | 2012-07-16 | 2013-01-02 | 中国科学院地理科学与资源研究所 | Dempster-shafer (D-S) evidence theory method capable of achieving micro-blog inclusion traffic information fusion |
| CN102855272B (en) * | 2012-07-16 | 2016-08-17 | 中国科学院地理科学与资源研究所 | A kind of micro-blog contains the D-S evidence theory method of traffic information fusion |
| CN103049485A (en) * | 2012-11-30 | 2013-04-17 | 浙江工商大学 | City public data flow fusion method based on wavelet timing sequence arborescence outline |
| WO2013167014A1 (en) * | 2012-12-18 | 2013-11-14 | 中兴通讯股份有限公司 | Area positioning system and method |
| CN103874016A (en) * | 2012-12-18 | 2014-06-18 | 中兴通讯股份有限公司 | Area positioning system and method |
| CN103886080B (en) * | 2014-03-25 | 2017-01-25 | 中国科学院地理科学与资源研究所 | Method for extracting road traffic information from Internet unstructured text |
| WO2016107611A1 (en) * | 2015-01-04 | 2016-07-07 | 华为技术有限公司 | Method and device for text search for mobile terminal |
| CN106407469A (en) * | 2016-10-24 | 2017-02-15 | 北京亚控科技发展有限公司 | Method for describing time attributes of things and carrying out search based on description |
| CN106446278A (en) * | 2016-10-24 | 2017-02-22 | 北京亚控科技发展有限公司 | Method for searching data target on basis of spatial-temporal database |
| CN106446278B (en) * | 2016-10-24 | 2018-07-20 | 北京亚控科技发展有限公司 | A kind of search method to data object based on space-time database |
| CN106407469B (en) * | 2016-10-24 | 2018-09-14 | 北京亚控科技发展有限公司 | A method of description things time attribute is simultaneously searched based on the description |
| CN107818688A (en) * | 2017-11-20 | 2018-03-20 | 安徽帆扬通信电子技术有限公司 | A kind of intelligent transportation addressing navigation system based on RDS |
| CN108597223A (en) * | 2018-04-28 | 2018-09-28 | 北京智行者科技有限公司 | A kind of data processing method and system for intelligent vehicle behavior description |
| CN111731207A (en) * | 2020-06-18 | 2020-10-02 | 北京百度网讯科技有限公司 | Method, device and equipment for acquiring vehicle configuration information and storage medium |
| CN112989228A (en) * | 2021-04-25 | 2021-06-18 | 湖南视觉伟业智能科技有限公司 | Distributed space-time query method and system |
| CN113570107A (en) * | 2021-06-08 | 2021-10-29 | 众能联合数字技术有限公司 | Project address positioning method for project rental scene |
| CN115345152A (en) * | 2022-10-19 | 2022-11-15 | 北方健康医疗大数据科技有限公司 | Template library updating method, report analyzing method, device, equipment and medium |
| CN115345152B (en) * | 2022-10-19 | 2023-03-14 | 北方健康医疗大数据科技有限公司 | Template library updating method, report analyzing method, device, equipment and medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN100573506C (en) | 2009-12-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN100573506C (en) | A kind of space-time fusion method of natural language expressing dynamic traffic information | |
| US10360228B1 (en) | Ranking of geographic information | |
| CN107145577A (en) | Address standardization method, device, storage medium and computer | |
| Richter et al. | A model for context-specific route directions | |
| CN102395965B (en) | Method for searching objects in a database | |
| CN112347222B (en) | Method and system for converting non-standard address into standard address based on knowledge base reasoning | |
| CN112329467B (en) | Address recognition method and device, electronic equipment and storage medium | |
| US8374791B2 (en) | Method of operating a navigation system to provide route guidance | |
| JP5390840B2 (en) | Information analyzer | |
| CN103471610B (en) | A kind ofly support online, the double mode air navigation aid of off-line | |
| CN104679867B (en) | Address method of knowledge processing and device based on figure | |
| CN102169591B (en) | Line selecting method and drawing method of text note in drawing | |
| CN103870608A (en) | Method for mapping knowledge construction and formalized expression based on map symbols | |
| CN111625732A (en) | Address matching method and device | |
| CN108629443A (en) | Path description is converted to the paths planning method and navigation system of machine readable format | |
| CN105117494A (en) | Spatial entity mapping method in fuzzy linguistic environment | |
| CN109271625B (en) | Pinyin spelling standardization method for Chinese place names | |
| CN115630648A (en) | Address element analysis method and system for man-machine conversation and computer readable medium | |
| CN115878737B (en) | Intersection extraction and topology structure description method based on road network data | |
| CN102955783A (en) | Device and method for map data road search and vehicular system | |
| CN111191084B (en) | Map structure-based place name address resolution method | |
| CN113626536A (en) | News geocoding method based on deep learning | |
| Jain et al. | Public transport ontology for passenger information retrieval | |
| CN104317830A (en) | Method and system for intelligently generating position description based on longitude and latitude | |
| CN116484859A (en) | Police condition space position positioning method and related products |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20091223 Termination date: 20150625 |
|
| EXPY | Termination of patent right or utility model |