[go: up one dir, main page]

CN106157651B - A traffic radio traffic information broadcast system based on phonetic semantic understanding - Google Patents

A traffic radio traffic information broadcast system based on phonetic semantic understanding Download PDF

Info

Publication number
CN106157651B
CN106157651B CN201610596596.5A CN201610596596A CN106157651B CN 106157651 B CN106157651 B CN 106157651B CN 201610596596 A CN201610596596 A CN 201610596596A CN 106157651 B CN106157651 B CN 106157651B
Authority
CN
China
Prior art keywords
information
road
user
control
broadcast
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.)
Active
Application number
CN201610596596.5A
Other languages
Chinese (zh)
Other versions
CN106157651A (en
Inventor
胡文鹏
郑楠
戴汝为
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Institute of Automation of Chinese Academy of Science
Original Assignee
Institute of Automation of Chinese Academy of Science
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Institute of Automation of Chinese Academy of Science filed Critical Institute of Automation of Chinese Academy of Science
Priority to CN201610596596.5A priority Critical patent/CN106157651B/en
Publication of CN106157651A publication Critical patent/CN106157651A/en
Application granted granted Critical
Publication of CN106157651B publication Critical patent/CN106157651B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/09—Arrangements for giving variable traffic instructions
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/30—Semantic analysis

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Traffic Control Systems (AREA)
  • Navigation (AREA)

Abstract

The present invention discloses a kind of traffic radio traffic information broadcasting system based on voice semantic understanding, including based on speech recognition casting information flag and map control subsystem, mode by voice input carry out casting information flag and map control；Intelligent way asking subsystem, user inputs information and carries out the information feedback of inquired road conditions for identification；Visual history data query subsystem, the road conditions of the historical period for inquiry roads are simultaneously visualized；The present invention facilitates traffic radio host and broadcasts information flag and map control, the road condition query and the inquiry of historical data of the person of being convenient to use.

Description

A kind of traffic radio traffic information broadcasting system based on voice semantic understanding
Technical field
The present invention relates to information processing and natural language processing field, more particularly to a kind of based on voice semantic understanding Traffic radio traffic information broadcasting system.
Background technique
Automation refers to machinery equipment, system or process (production, management process) in the direct ginseng of nobody or less people With under, according to the requirement of people, by detecting automatically, information processing, analytical judgment, manipulation and control, the mistake of the set goal is realized Journey.It is intelligent: refer to collected by modern communications and information technology, computer networking technology, industry technology, intelligent control technology and At for the application in terms of some.
With the continuous development of network technology and natural language processing technique, people get used to utilizing various instant messagings Software exchanges, and can obtain oneself desired information in time by these media, wherein wechat be exactly one it is all well and good i.e. When means of communication, spatially communication disorder can be overcome by these tools people, real time information can be obtained by staying indoors, such as Message in this way, at the aspect that spreads news, can be communicated to user in real time by news information, traffic information etc., be solved The disadvantages of hysteresis quality of conventional information of having determined transmitting.It brings for people's lives and significantly facilitates.
With the development of society, the quantity of private car is more and more, traffic jam issue has become each large, medium and small city The significant problem that city is faced.So road condition analyzing and real-time road condition information are broadcasted, are predicted the jam situation of future trajectory Through the important basis for selecting route as taxi driver, private car car owner, traveler.
This system promotes the parties concerned such as the coast is clear to play an important role mediating big city congestion in road；To governments at all levels Department reduces the capital investment of human and material resources, alleviates traffic pressure, environment purification, reduce haze, promote social development etc. With positive effect.This intelligent information broadcasting system, fusion is intelligent, automates, human oriented design, and broadcasting station is facilitated to broadcast While, the casting function of multiple channel, such as wechat intelligent answer may be implemented, user is allowed to obtain accurate traffic letter Breath alleviates public travel status effectively to utilize road resources.
Summary of the invention
In order to allow user more easily to obtain the data of traffic, it is semantic based on voice that the invention proposes one kind The traffic radio traffic information broadcasting system of understanding facilitates traffic radio host and broadcasts information flag and map control, side User road condition query and the inquiry of historical data.
A kind of traffic radio traffic information broadcasting system based on voice semantic understanding proposed by the present invention, including it is based on language Sound identification casting information flag and map control subsystem, the subsystem include speech recognition module, message processing module and Information display module, the subsystem carry out casting information flag and map control for user's mode by voice input；
Speech recognition module is for identifying the voice messaging that user inputs and exporting recognition result；
Information display module is the display device with human-computer interaction interface；
The recognition result that message processing module is used to export speech recognition module is handled, and specifically includes following step It is rapid:
Step A1: the recognition result of traversal speech recognition module output is detected whether with map control order, if having A3 is gone to step, enters step A2 if nothing；
Step A2: recognition result and current human-computer interaction interface to speech recognition module output show that content does fuzzy It with comparison, finds out user speech and inputs the traffic information of corresponding road, and do with read flag and casting label processing；
Step A3: carrying out map control order pretreatment, detects the Control Cooling and control parameter of map control order, will Control command is sent to map end and carries out relevant control.
A kind of traffic radio traffic information broadcasting system based on voice semantic understanding of the invention, further includes intelligent way asking Subsystem, user inputs information and carries out the information feedback of inquired road conditions the system for identification；The system includes temporary Database, for recording the temporary information of each User ID；
In the system information processing method specifically includes the following steps:
Step B1: request of asking the way is submitted by wechat platform；
Step B2: record User ID, and to solicited message of asking the way carry out processing and information extraction, extracted information include Road name, place name, direction；Place name is mapped to road name；
Step B3: it if there is no road name and direction in extracted information, and executes step B1 request user and passes through wechat Platform ask the way the information completion of request；If comprising road name in extracted information, but not including direction, then follow the steps B4；If in extracted information including direction, but do not include place name, thens follow the steps B5；If wrapped in extracted information Containing road name, place name, direction, B6 is thened follow the steps；
Step B4: retrieving the temporary information of corresponding User ID, if thening follow the steps B6 comprising direction in temporary information, if Temporal data library then is added in extracted information not comprising direction, and executes step B1 request user and is carried out by wechat platform It asks the way the information completion of request；
Step B5: retrieving the temporary information of corresponding User ID, if thening follow the steps B6 containing place name in temporary information, if Temporal data library then is added in extracted information not comprising place name, and executes step B1 request user and is carried out by wechat platform It asks the way the information completion of request；
Step B6: according to the corresponding extracted information of User ID and temporary information, the inspection of newest traffic road condition data is carried out Rope, and natural language is generated according to text template and replys user.
A kind of traffic radio traffic information broadcasting system based on voice semantic understanding of the invention further includes that visualization is gone through History data query subsystem, the system are used for the road conditions of the historical period of inquiry roads and are visualized, specific method The following steps are included:
Step C1: input inquiry information, and the query information of input is pre-processed, extract road name, query time section
Extract road it is entitled multiple when, if wherein one with other two for starting point, be judged as the road of non-starting point For road to be checked；Otherwise multiple roads are road to be checked；
When road to be checked is multiple, following steps are executed respectively for each road to be checked；
Step C2: the Lu Mingyu standard road famous-brand clock of the road to be checked obtained in step C1 is subjected to fuzzy matching, and is led to It crosses mapping table and retrieves the inquiry road Lu Mingyu chain number；
Step C3: using the information of the step C2 road chain number obtained, the corresponding storage ID of road chain number is extracted as index, benefit Index of reference goes in database to retrieve the historical data of institute's query time section；
Step C4: the historical data that inquiry is obtained is classified by road chain number, is spliced in chronological order, and passes through line chart Mode carries out the visual presentation of historical data.
Wherein, the casting information flag based on speech recognition and map control subsystem further include user-oriented dictionary module, should For storing everyday words and matching for speech recognition module to identified voice messaging, the everyday words includes module Common control command keyword, Beijing road name.
Wherein, when the road extracted in step B2 is entitled multiple, if wherein one with other two for starting point, be judged as The road of non-starting point is road to be checked；Otherwise multiple roads are road to be checked；
When road to be checked is multiple, step B3 is executed respectively for each road to be checked.
Wherein, decision tree is detected by control command in step A1 to detect whether with map control order, the control The main control node of order detection decision tree is as follows:
String length to be identified is less than n；
Whether containing the keyword in order keywords database, and its matching rate is greater than threshold value beta；
Whether it is greater than threshold θ with the matching rate of the road name in the keyword or road name dictionary in control parameter dictionary.
Wherein, the method for fuzzy matching is to show that content is weighted by history casting sequence to human-computer interaction interface in step A2 Fuzzy matching, formula are as follows:
Wherein Tnow、TbroadcastRespectively the time of the now time and corresponding history casting record broadcast, it is with minute Unit, TpFor time coefficient；Count () is counting function, and VRR is speech recognition result collection.
Wherein, decision tree is detected by Control Cooling and control parameter in step A3 and detects Control Cooling and control parameter,
The main control node of the Control Cooling and control parameter detection decision tree is as follows:
String length to be identified is less than n；
Whether containing the keyword in order keywords database, and its matching rate be greater than threshold value beta '；
With in control parameter dictionary keyword or road name dictionary in road name matching rate whether be greater than threshold θ '.
Wherein, threshold value beta, θ, β ', the determination method of θ ' are as follows: use information ratio of profit increase selects optimal classification threshold value, and formula is such as Shown in lower:
Δ Gini (A)=Gini (D)-GiniA(D)
Wherein piFor assuming that element belongs to order keyword or control parameter in D under conditions of given classification thresholds Probability uses | Ci,D|/| D | estimation；D1、D2The two class sample sets that sample space is divided into respectively under given threshold value.
A kind of traffic radio traffic information broadcasting system based on voice semantic understanding proposed by the present invention, contains and is based on The casting information flag of speech recognition and map control subsystem, intelligent way asking subsystem, visual history data query subsystem System realizes traffic radio host and carries out casting information flag and map control, the person of being convenient to use by way of voice Road condition query and the inquiry of historical data.
Detailed description of the invention
Fig. 1: overall system of the present invention forms block schematic illustration；
Fig. 2: casting information flag and map control subsystem information processing flow schematic diagram based on speech recognition；
Fig. 3: intelligent way asking subsystem information processing flow schematic diagram.
Fig. 4: visual history data query subsystem information processing flow schematic diagram.
Specific embodiment
In order to enable advantages of the present invention, technical solution, goal of the invention be more clearly understood that it is clear, below with reference to example and attached Figure carries out further point to the present invention and elaborates.Wherein, specific implementation example used herein is used only for explaining this hair It is bright, it is not intended to limit the present invention.
In order to realize the method for the present invention, embodiment in view of processing step and the number of users and resource quantity that are related to, If should preferably guarantee that 16 core of processor, memory are not less than 32GB in single machine realization, can be write using any language, this example It is realized using java language.
A kind of traffic radio traffic information broadcasting system based on voice semantic understanding of the present embodiment, using optimal side Case is described in detail, and the system of the present embodiment includes it is characterised in that it includes the casting based on speech recognition is believed as shown in Figure 1 Breath label and map control subsystem, intelligent way asking subsystem, visual history data query subsystem.
1, at the casting information flag based on speech recognition and map control subsystem, including speech recognition module, information Module and information display module are managed, which carries out casting information flag and map for user's mode by voice input Control；
Speech recognition module is for identifying the voice messaging that user inputs and exporting recognition result；
Information display module is the display device with human-computer interaction interface；
The recognition result that message processing module is used to export speech recognition module is handled, as shown in Fig. 2, specific packet Include following steps:
Step A1: the recognition result of traversal speech recognition module output is detected whether with map control order, if having A3 is gone to step, enters step A2 if nothing；
Step A2: recognition result and current human-computer interaction interface to speech recognition module output show that content does fuzzy It with comparison, finds out user speech and inputs the traffic information of corresponding road, and do with read flag and casting label processing；
Step A3: carrying out map control order pretreatment, detects the Control Cooling and control parameter of map control order, will Control command is sent to map end and carries out relevant control.
In the present embodiment, audio recognition method uses Iflytek speech recognition open platform.
In the present embodiment, casting information flag and map control subsystem based on speech recognition further include user-oriented dictionary mould Block, for the module for storing everyday words and matching for speech recognition module to identified voice messaging, described is common Word includes common control command keyword, Beijing road name.
In the present embodiment, decision tree is detected by control command in step A1 and is detected whether with map control order, institute The main control node for stating control command detection decision tree is as follows:
String length to be identified is less than n；
Whether containing the keyword in order keywords database (Command dictionary), and its matching rate is greater than threshold value β；
Whether it is greater than threshold θ with the matching rate of the road name in the keyword or road name dictionary in control parameter dictionary.
In the present embodiment, the method for fuzzy matching is to show that content is suitable by history casting to human-computer interaction interface in step A2 Sequence weighted fuzzy matching, shown in formula such as formula (1):
Wherein Tnow,TbroadcastRespectively the time of the now time and corresponding history casting record broadcast, it is with minute Unit, TpFor time coefficient；Count () is counting function, and VRR is speech recognition result collection.
In the present embodiment, decision tree is detected by Control Cooling and control parameter in step A3 and detects Control Cooling and control Parameter,
The main control node of the Control Cooling and control parameter detection decision tree is as follows:
String length to be identified is less than n；
Whether containing the keyword in order keywords database (Command dictionary), and its matching rate is greater than threshold value β'；
With in control parameter dictionary keyword or road name dictionary in road name matching rate whether be greater than threshold θ '.
In the present embodiment, threshold value beta, θ, β ', the determination method of θ ' are as follows: use information ratio of profit increase selects optimal classification threshold value, Formula such as formula (2), (3), (4) are shown:
Δ Gini (A)=Gini (D)-GiniA(D) (4)
Wherein piFor assuming that element belongs to order keyword or control parameter in D under conditions of given classification thresholds Probability uses | Ci,D|/| D | estimation；D1,D2The two class sample sets that sample space is divided into respectively under given threshold value.
2, intelligent way asking subsystem, user inputs information and carries out the information feedback of inquired road conditions for identification,；This is System includes temporal data library, for recording the temporary information of each User ID；
As shown in figure 3, in the system information processing method specifically includes the following steps:
Step B1: request of asking the way is submitted by wechat platform；
Step B2: record User ID, and to solicited message of asking the way carry out processing and information extraction, extracted information include Road name, place name, direction；And place name is mapped to road name；
Step B3: it if there is no road name and direction in extracted information, and executes step B1 request user and passes through wechat Platform ask the way the information completion of request；If comprising road name in extracted information, but not including direction, then follow the steps B4；If in extracted information including direction, but do not include place name, thens follow the steps B5；If wrapped in extracted information Containing road name, place name, direction, B6 is thened follow the steps；
Request of asking the way in the step is the voice or text request that wechat user issues to intelligent way asking subsystem, system The information such as the road to be inquired of identification user name and direction lack road name and/or place name and/or directional information, are if incomplete System can generate voice or text information according to identification situation and inquire to user, and the same of user uses voice or text in wechat Platform is answered, and system can identify the relevant information of the inside.
Step B4: retrieving the temporary information of corresponding User ID, if thening follow the steps B6 comprising direction in temporary information, if Temporal data library then is added in extracted information not comprising direction, and executes step B1 request user and is carried out by wechat platform It asks the way the information completion of request；
Step B5: retrieving the temporary information of corresponding User ID, if thening follow the steps B6 containing place name in temporary information, if Temporal data library then is added in extracted information not comprising place name, and executes step B1 request user and is carried out by wechat platform It asks the way the information completion of request；
Step B6: according to the corresponding extracted information of User ID and temporary information, the inspection of newest traffic road condition data is carried out Rope, and natural language is generated according to text template and replys user.
In the present embodiment, when the road extracted in step B2 is entitled multiple, if wherein one with other two for starting point, The road for being judged as non-starting point is road to be checked；Otherwise multiple roads are road to be checked；
When road to be checked is multiple, step B3 is executed respectively for each road to be checked.
In the present embodiment, in step B2 to solicited message of asking the way carry out processing and information extraction method are as follows: using participle System carries out participle and part-of-speech tagging to request of asking the way, and extracts road name, place name, direction.It is asked if wechat platform submission is asked the way It asks as voice messaging, then needs to be handled and extracted by Words partition system again after first carrying out speech recognition.
3, visual history data query subsystem, the road conditions of the historical period for inquiry roads simultaneously carry out visualization exhibition Show, as shown in figure 4, specific method the following steps are included:
Step C1: input inquiry information, and the query information of input is pre-processed, extract road name, query time section
Extract road it is entitled multiple when, if wherein one with other two for starting point, be judged as the road of non-starting point For road to be checked；Otherwise multiple roads are road to be checked；
When road to be checked is multiple, following steps are executed respectively for each road to be checked；
Step C2: the Lu Mingyu standard road famous-brand clock of the road to be checked obtained in step C1 is subjected to fuzzy matching, and is led to It crosses mapping table and retrieves the inquiry road Lu Mingyu chain number；
Step C3: using the information of the step C2 road chain number obtained, the corresponding storage ID of road chain number is extracted as index, benefit Index of reference goes in database to retrieve the historical data of institute's query time section；
Step C4: the historical data that inquiry is obtained is classified by road chain number, is spliced in chronological order, and passes through line chart Mode carries out the visual presentation of historical data.
In the present embodiment, since data volume is huge, when store historical data, traffic information is divided into 24 by the hour A tables of data, corresponding storage information database are made of 24 tables, and daily 24 hours corresponding, it is small that each table records corresponding one When in load conditions data.Using the design mainly since data volume is big, database is imitated when excavating to historical data Rate is lower, in order to reduce database length, convenient for holding analysis between data timesharing, and is conducive to parallel processing, therefore according to daily 24 hours, database is designed as 24 tables, each table records the data of this hour daily, and system is only needed using latest data Its corresponding tables of data is found by current time, then reads newest data.Therefore it is adopted in the step C4 of the present embodiment With table querying method is divided, according to query time segment information never with progress multilist retrieval and data fusion in tables of data.
The rule of the casting based on traffic information has also been devised in the present embodiment, according to user's listening situation and forgets situation, letter Breath propagates range, different brackets traffic information is limited different broadcast frequencies, it is characterised in that: primary information is broadcasted in 3 minutes For the first time, second of casting in 10 minutes；It is broadcasted in second-level message 10 minutes primary；It is broadcasted in three-level information 15 minutes primary.
Particular embodiments described above has carried out further in detail the purpose of the present invention, technical scheme and beneficial effects It describes in detail bright, it should be understood that the above is only a specific embodiment of the present invention, is not intended to restrict the invention, it is all Within the spirit and principles in the present invention, any modification, equivalent substitution, improvement and etc. done should be included in guarantor of the invention Within the scope of shield.

Claims (8)

1.一种基于语音语义理解的交通电台路况信息播报系统，其特征在于，包括基于语音识别的播报信息标记与地图控制子系统，该基于语音识别的播报信息标记与地图控制子系统包括语音识别模块、信息处理模块和信息显示模块，该基于语音识别的播报信息标记与地图控制子系统用于用户通过语音输入的方式进行播报信息标记和地图控制；1. a traffic station road condition information broadcast system based on speech semantic understanding, it is characterized in that, comprise the broadcast information mark and map control subsystem based on speech recognition, this broadcast information mark and map control subsystem based on speech recognition comprise speech recognition module, information processing module and information display module, the broadcast information marking and map control subsystem based on voice recognition is used for users to carry out broadcast information marking and map control by means of voice input; 语音识别模块用于对用户输入的语音信息进行识别并输出识别结果；The speech recognition module is used to recognize the speech information input by the user and output the recognition result; 信息显示模块为具有人机交互界面的显示装置；The information display module is a display device with a human-computer interaction interface; 信息处理模块用于对语音识别模块输出的识别结果进行处理，具体包括以下步骤：The information processing module is used to process the recognition result output by the speech recognition module, which specifically includes the following steps: 步骤A1：遍历语音识别模块输出的识别结果，检测是否具有地图控制命令，若有则转步骤A3，若无则进入步骤A2；Step A1: traverse the recognition result output by the speech recognition module, detect whether there is a map control command, if so, go to Step A3, if not, enter Step A2; 步骤A2：对语音识别模块输出的识别结果与当前人机交互界面显示内容做模糊匹配对比，找出用户语音输入对应道路的路况信息，并做跟读标记与播报标记处理；Step A2: compare the recognition result output by the speech recognition module with the current display content of the human-computer interaction interface, find out the road condition information of the road corresponding to the user's voice input, and perform follow-up mark and broadcast mark processing; 步骤A3:进行地图控制命令预处理，检测地图控制命令的控制类型和控制参数，将控制命令发送到地图端进行相关控制；Step A3: carry out map control order preprocessing, detect the control type and the control parameter of map control order, control order is sent to map terminal and carry out relevant control; 所述基于语音语义理解的交通电台路况信息播报系统还包括可视化历史数据查询子系统，该可视化历史数据查询子系统用于查询道路的历史时段的路况并进行可视化展示，具体方法包括以下步骤：The traffic radio road condition information broadcasting system based on speech semantic understanding also includes a visual historical data query subsystem, which is used to query the road conditions of the road in the historical period of time and visualize it, and the specific method includes the following steps: 步骤C1:输入查询信息，并对输入的查询信息进行预处理，提取路名、查询时间段；Step C1: input query information, and preprocess the input query information, extract road name, query time period; 提取路名为多个时，若其中一条以其他两条为起始点，则判断为非起始端的道路为待查询道路；否则多个道路均为待查询道路；When multiple road names are extracted, if one of them takes the other two as the starting point, it is determined that the non-starting road is the road to be queried; otherwise, the multiple roads are all roads to be queried; 当待查询道路为多个时，针对每个待查询道路分别执行以下步骤；When there are multiple roads to be queried, perform the following steps for each road to be queried; 步骤C2:将步骤C1中获得的待查询道路的路名与标准路名表进行模糊匹配，并通过映射表检索出查询路名与路链号；Step C2: carry out fuzzy matching with the road name of the road to be queried obtained in step C1 and the standard road name table, and retrieve the query road name and the road link number by the mapping table; 步骤C3:使用步骤C2获得的路链号的信息，提取路链号对应的存储ID作为索引，利用索引去数据库中检索所查询时间段的历史数据；Step C3: use the information of the road link number obtained in step C2, extract the storage ID corresponding to the road link number as an index, and utilize the index to retrieve the historical data of the queried time period in the database; 步骤C4:将查询得到的历史数据按路链号分类，按时间顺序拼接，并通过折线图的方式进行历史数据的可视化展示。Step C4: Classify the historical data obtained by the query according to the link number, splicing in chronological order, and visualize the historical data by means of a line graph. 2.如权利要求1所述的基于语音语义理解的交通电台路况信息播报系统，其特征在于，还包括智能问路子系统，该智能问路子系统用于识别用户输入信息并进行所查询路况的信息反馈；该智能问路子系统包含有暂存数据库，用于记录各用户ID的暂存信息；2. The traffic radio road condition information broadcasting system based on speech semantic understanding as claimed in claim 1, characterized in that, further comprising an intelligent way-questioning subsystem, which is used to identify user input information and carry out the information of the queried road conditions Feedback; the intelligent route-questioning subsystem includes a temporary storage database for recording the temporary storage information of each user ID; 该智能问路子系统中信息处理的方法具体包括以下步骤：The method for information processing in the intelligent route-questioning subsystem specifically includes the following steps: 步骤B1:通过微信平台提交问路请求；Step B1: submit a request for directions through WeChat platform; 步骤B2:记录用户ID，并对问路请求信息进行处理和信息提取，所提取的信息包括路名、地名、方向；Step B2: record the user ID, and process and extract the information of the request for directions, and the extracted information includes road name, place name, direction; 步骤B3:如果所提取的信息中没有路名和方向，并执行步骤B1请求用户通过微信平台进行问路请求的信息补全；如果所提取的信息中包含路名，但不包含方向，则执行步骤B4；如果所提取的信息中包含方向，但不包含地名，则执行步骤B5；如果所提取的信息中包含路名、地名、方向，则执行步骤B6；Step B3: if there is no road name and direction in the extracted information, and perform step B1 to request the user to complete the information requested by asking for directions through the WeChat platform; if the extracted information contains road names, but does not contain direction, then perform steps B4; If the extracted information contains directions, but does not contain place names, then execute step B5; if the extracted information includes road names, place names, and directions, execute step B6; 步骤B4:检索对应用户ID的暂存信息，如果暂存信息中包含方向则执行步骤B6，若不包含方向则将所提取的信息加入暂存数据库，并执行步骤B1请求用户通过微信平台进行问路请求的信息补全；Step B4: retrieve the temporary storage information of the corresponding user ID, if the temporary storage information contains the direction, then execute step B6, if the direction is not included, the extracted information is added to the temporary storage database, and step B1 is performed to request the user to ask the user through the WeChat platform. Information completion for road requests; 步骤B5:检索对应用户ID的暂存信息，如果暂存信息中含有地名则执行步骤B6，若不包含地名则将所提取的信息加入暂存数据库，并执行步骤B1请求用户通过微信平台进行问路请求的信息补全；Step B5: retrieve the temporary storage information of the corresponding user ID, if the temporary storage information contains a place name, then execute step B6, if the place name is not included, then the extracted information is added to the temporary storage database, and step B1 is performed to request the user to ask questions through the WeChat platform. Information completion for road requests; 步骤B6:依据对应用户ID所提取的信息和暂存信息，进行最新交通路况数据的检索，并根据文字模版生成自然语言回复用户。Step B6: According to the extracted information and the temporarily stored information corresponding to the user ID, retrieve the latest traffic road condition data, and generate a natural language reply to the user according to the text template. 3.如权利要求1～2中任一项所述的基于语音语义理解的交通电台路况信息播报系统，其特征在于，基于语音识别的播报信息标记与地图控制子系统还包括用户词典模块，该模块用于存储常用词和供语音识别模块对所识别出语音信息进行匹配，所述的常用词包括常用的控制命令关键词、北京市路名。3. The traffic radio road condition information broadcast system based on speech semantic understanding according to any one of claims 1 to 2, wherein the speech recognition-based broadcast information marking and map control subsystem further comprises a user dictionary module, the The module is used for storing common words and for the speech recognition module to match the recognized voice information, and the common words include commonly used control command keywords and Beijing road names. 4.如权利要求2所述的基于语音语义理解的交通电台路况信息播报系统，其特征在于，步骤B2中提取的路名为多个时，若其中一条以其他两条为起始点，则判断为非起始端的道路为待查询道路；否则多个道路均为待查询道路；4. the traffic radio road condition information broadcasting system based on speech semantic understanding as claimed in claim 2, is characterized in that, when the road name extracted in step B2 is multiple, if one of them takes other two as starting point, then judge The road that is not the starting end is the road to be queried; otherwise, multiple roads are the road to be queried; 当待查询道路为多个时，针对每个待查询道路分别执行步骤B3。When there are multiple roads to be queried, step B3 is respectively executed for each road to be queried. 5.如权利要求1～2中任一项所述的基于语音语义理解的交通电台路况信息播报系统，其特征在于，步骤A1中通过控制命令检测决策树检测是否具有地图控制命令，所述控制命令检测决策树的主要控制节点如下：5. The traffic station road condition information broadcasting system based on speech semantic understanding according to any one of claims 1 to 2, wherein in step A1, whether there is a map control command is detected by a control command detection decision tree, and the control The main control nodes of the command detection decision tree are as follows: 待识别字符串长度小于n；The length of the string to be recognized is less than n; 是否含有命令关键词库中的关键词，且其匹配率大于阈值β；Whether it contains keywords in the command keyword library, and its matching rate is greater than the threshold β; 与控制参数词典中的关键字或路名词典中的路名的匹配率是否大于阈值θ。Whether the matching rate with the keyword in the control parameter dictionary or the road name in the road name dictionary is greater than the threshold θ. 6.如权利要求5所述的基于语音语义理解的交通电台路况信息播报系统，其特征在于，步骤A2中模糊匹配的方法为对人机交互界面显示内容按历史播报顺序加权模糊匹配，其公式如下所示：6. the traffic station road condition information broadcast system based on speech semantic understanding as claimed in claim 5, is characterized in that, the method for fuzzy matching in step A2 is the weighted fuzzy matching by historical broadcast order to human-computer interaction interface display content, and its formula As follows: 其中Tnow、Tbroadcast分别为现在的时间和对应历史播报记录播出的时间，以分钟为单位，Tp为时间系数；count()为计数函数，VRR为语音识别结果集。Among them, T now and T broadcast are the current time and the broadcast time of the corresponding historical broadcast record, respectively, in minutes, and T p is the time coefficient; count() is the counting function, and VRR is the speech recognition result set. 7.如权利要求6所述的基于语音语义理解的交通电台路况信息播报系统，其特征在于，步骤A3中通过控制类型和控制参数检测决策树检测控制类型和控制参数，7. the traffic radio road condition information broadcasting system based on speech semantic understanding as claimed in claim 6, is characterized in that, in step A3, detects control type and control parameter by control type and control parameter detection decision tree, 所述控制类型和控制参数检测决策树的主要控制节点如下：The main control nodes of the control type and control parameter detection decision tree are as follows: 待识别字符串长度小于n；The length of the string to be recognized is less than n; 是否含有命令关键词库中的关键词，且其匹配率大于阈值β'；Whether it contains keywords in the command keyword library, and its matching rate is greater than the threshold β'; 与控制参数词典中的关键字或路名词典中的路名的匹配率是否大于阈值θ'。Whether the matching rate with the keyword in the control parameter dictionary or the road name in the road name dictionary is greater than the threshold θ'. 8.如权利要求7所述的基于语音语义理解的交通电台路况信息播报系统，其特征在于，阈值β、θ、β'、θ'的确定方法为：使用信息增益率，选择最优分类阈值，公式如下所示：8. The traffic radio road condition information broadcasting system based on speech semantic understanding as claimed in claim 7, wherein the method for determining the thresholds β, θ, β', θ' is: using the information gain rate, selecting the optimal classification threshold , the formula is as follows: ΔGini(A)＝Gini(D)-GiniA(D)ΔGini(A)=Gini(D)-Gini A (D) 其中pi为在假设给定分类阈值的条件下D中元素属于命令关键词或者控制参数的概率，使用|Ci,D|/|D|估计；D1,D2分别为在给定阈值下样本空间分成的两类样本集。where p i is the probability that the elements in D belong to command keywords or control parameters under the assumption of a given classification threshold, estimated using |C i,D |/|D|; D 1 , D 2 are respectively at the given threshold The lower sample space is divided into two types of sample sets.
CN201610596596.5A 2016-07-26 2016-07-26 A traffic radio traffic information broadcast system based on phonetic semantic understanding Active CN106157651B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610596596.5A CN106157651B (en) 2016-07-26 2016-07-26 A traffic radio traffic information broadcast system based on phonetic semantic understanding

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610596596.5A CN106157651B (en) 2016-07-26 2016-07-26 A traffic radio traffic information broadcast system based on phonetic semantic understanding

Publications (2)

Publication Number Publication Date
CN106157651A CN106157651A (en) 2016-11-23
CN106157651B true CN106157651B (en) 2019-06-14

Family

ID=58060719

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610596596.5A Active CN106157651B (en) 2016-07-26 2016-07-26 A traffic radio traffic information broadcast system based on phonetic semantic understanding

Country Status (1)

Country Link
CN (1) CN106157651B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109217885B (en) * 2018-09-21 2022-07-12 法法汽车(中国)有限公司 Vehicle-mounted radio channel searching device and method
CN113609258B (en) * 2021-08-14 2024-12-24 北京庚图科技有限公司 Interactive display method, electronic device and medium based on geographic information platform

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1779673A (en) * 2004-11-25 2006-05-31 株式会社电装 Information service system
CN101109641A (en) * 2006-07-21 2008-01-23 英华达(上海)电子有限公司 Voice controlling and interacting device and method used for GPS device
CN101939740A (en) * 2007-12-11 2011-01-05 声钰科技 Providing a natural language voice user interface in an integrated language navigation service environment
CN102184641A (en) * 2011-05-09 2011-09-14 浙江大学 Running information based road condition management method and system
CN103778797A (en) * 2012-10-25 2014-05-07 北京掌城科技有限公司 Multimode response method of voice transportation information
CN104183144A (en) * 2013-05-20 2014-12-03 北京百度网讯科技有限公司 Real-time traffic condition information generating method and system thereof

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH0321817A (en) * 1989-06-20 1991-01-30 Tokico Ltd On-vehicle travel path guide device

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1779673A (en) * 2004-11-25 2006-05-31 株式会社电装 Information service system
CN101109641A (en) * 2006-07-21 2008-01-23 英华达(上海)电子有限公司 Voice controlling and interacting device and method used for GPS device
CN101939740A (en) * 2007-12-11 2011-01-05 声钰科技 Providing a natural language voice user interface in an integrated language navigation service environment
CN102184641A (en) * 2011-05-09 2011-09-14 浙江大学 Running information based road condition management method and system
CN103778797A (en) * 2012-10-25 2014-05-07 北京掌城科技有限公司 Multimode response method of voice transportation information
CN104183144A (en) * 2013-05-20 2014-12-03 北京百度网讯科技有限公司 Real-time traffic condition information generating method and system thereof

Also Published As

Publication number Publication date
CN106157651A (en) 2016-11-23

Similar Documents

Publication Publication Date Title
CN117952209A (en) Knowledge graph construction method and system
CN112650848A (en) Urban railway public opinion information analysis method based on text semantic related passenger evaluation
CN108460136A (en) Electric power O＆M information knowledge map construction method
CN111813956A (en) Knowledge graph construction method and device, and information penetration method and system
WO2021159834A1 (en) Abnormal information processing node analysis method and apparatus, medium and electronic device
CN109670843A (en) Data processing method, device, computer equipment and the storage medium of complaint business
CN114077705A (en) A method and system for profiling media accounts on social platforms
CN104252507B (en) A kind of business data matching process and device
CN111708774A (en) An industrial analysis system based on big data
CN120216612B (en) Knowledge-driven underground space information retrieval method, system and equipment
CN114265957A (en) Method and system for joint query of multiple data sources based on graph database
WO2021036439A1 (en) Method for responding to complaint, and device
CN104615734B (en) A kind of community management service big data processing system and its processing method
CN105808722A (en) Information discrimination method and system
CN114417974A (en) Model training method, information processing method, device, electronic device and medium
CN112395513A (en) Public opinion transmission power analysis method
CN111538753A (en) Research and judgment platform based on big data analysis
CN118445406A (en) Integration system based on massive polymorphic circuit heritage information
CN106157651B (en) A traffic radio traffic information broadcast system based on phonetic semantic understanding
CN107480270A (en) A kind of real time individual based on user feedback data stream recommends method and system
CN109543712B (en) Method for identifying entities on temporal data set
CN105786929A (en) Information monitoring method and device
CN120124628A (en) A quantitative analysis method and system for modern service industry policies based on knowledge graph
CN112668836A (en) Risk graph-oriented associated risk evidence efficient mining and monitoring method and device
CN116701760A (en) Non-private data acquisition and message push method and system

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant