KR20220077258A - Big data based voice recognition poi system, method for the system processing, computer program for the system, and medium with the computer program - Google Patents

Abstract

The present invention relates to a big data-based voice recognition POI system and the like, and more particularly, to a big data-based system for collecting and utilizing personalized POI (Point of Interest) data through voice recognition, and efficiently managing it. will be.
The big data-based voice recognition POI system of the present invention includes a user request processing unit, a voice recognition processing unit, and a big data processing unit interconnected by a network, wherein the user request processing unit is installed in a plurality of user terminals or downloaded from a server and driven as an agent. , process user authentication, collect voice or text data in conjunction with an application program, or transmit requested user data to a voice recognition processing unit and a big data processing unit, and transmit a response to the user, wherein the voice recognition processing unit serves as the user It extracts and recognizes voice data from the data received from the request processing unit, converts it into text data, and transmits it to the user request processing unit or big data processing unit, wherein the big data processing unit receives the received data from the user request processing unit or the voice recognition processing unit POI data is collected and processed in real time and stored in the data storage unit, or data is transmitted to the user request processing unit in response to a POI data search request.

Description

Big data-based voice recognition POI system, a processing method for the system, a computer program for the system processing, and a medium storing the computer program AND MEDIUM WITH THE COMPUTER PROGRAM}

The present invention relates to a big data-based voice recognition POI system, a processing method for the system, a computer program for processing the system, and a medium storing the computer program, and more particularly, to a personalized point of POI (POI) through voice recognition. Interest) It relates to a big data-based system and processing method for collecting and utilizing data, and to efficiently manage it, and a computer program for processing.

Recently, various location-based information services are being provided. In particular, interest in services that collect and manage data on a personalized point of interest (POI) preferred by each user is very high.

POI (Point of Interest) means a specific location in the real world or on a map or drawing that a specific person is interested in. For example, buildings, major facilities, stations, etc. It is location information about restaurants, tourist attractions, etc.

Conventional systems for providing POI information mainly provide a text-based or icon-click-based user interface, so it is necessary to accompany a rather complicated terminal operation, especially when operating a terminal for registering and calling a point of interest while driving. It causes many inconveniences and dangerous situations.

In addition, since personalized POI data accompanies a lot of data for each user, there is a considerable burden in processing a large amount of big data collected from a number of terminals and servers on the network. have.

As related prior art, Registered Patent No. 10-1544462 (Title of the Invention: Personalized POI information providing method using client terminal information in a location-based information providing system and its client terminal), Registered Patent No. 10-1134883 (invention of Title: System and method for registering open POI), Registered Patent No. 10-2096419 (Title of the invention: travel companion and product providing system and method based on location information and big data analysis), etc.

KR 10-1544462 B1 KR 10-1134883 B1 KR 10-2096419 B1

In order to solve the conventional problems, the present invention provides an invention related to a system, processing method, computer program, etc. capable of safely and simply collecting personalized POI data through voice recognition and effectively managing and utilizing it.

The big data-based voice recognition POI system of the present invention includes a user request processing unit 1, a voice recognition processing unit 2, and a big data processing unit 3 interconnected by a network, wherein the user request processing unit 1 includes a plurality of users As an agent installed in a terminal or downloaded from a server, it processes user authentication, collects voice or text data in conjunction with an application program, or transmits requested user data to the voice recognition processing unit 2 and the big data processing unit 3 and functions to deliver a response to the user, the voice recognition processing unit 2 extracts and recognizes voice data from the data received from the user request processing unit 1, converts it into text data, and the user request processing unit 1 Alternatively, it performs a function of transmitting it to the big data processing unit 3, and the big data processing unit 3 collects the POI data received from the user request processing unit 1 or the voice recognition processing unit 2 in real time and processes the data storage unit It is characterized in that it performs a function of storing data in (34) or transmitting data to the user request processing unit (1) in response to a POI data search request.

In addition, the big data-based voice recognition POI system of the present invention further includes an interactive voice processing unit 4 interlocked with the voice recognition processing unit 2, and the interactive voice processing unit 4 includes the voice recognition processing unit 2 It is characterized in that it interprets and recognizes similar sentences or interactive sentences that are not immediately recognized in the , and functions to respond to the voice recognition processing unit (2).

In addition, the big data processing unit 3 includes a data collection unit 31 , a buffering processing unit 32 , and a data processing unit 33 to store data after distributed processing, and the data storage unit 34 is distributed It includes a database 341 and a relational database 342, wherein the data collection unit 31 collects a large amount of data produced by various agents and transmits it to the buffering processing unit 32, and the buffering processing unit ( 32) buffers the data transmitted from the data collection unit 31 and transmits it to the data processing unit 33, and the data processing unit 33 processes the data transmitted from the buffering processing unit 32 into an appropriate format. Then, it is characterized in that it performs a function of storing the data in the distributed database 341 and the relational database 342 of the data storage unit 34 according to the data type, respectively.

In addition, in the user request processing unit 1, (S1101) waiting to detect a user utterance, (S1102) detecting a user utterance, and (S1103) processing the detected user utterance in the voice recognition processing unit Performing a request and receiving a response, (S1104) generating GPS address conversion and POI data according to the received response, and (S1105) requesting data collection of the generated POI data to a big data processing unit characterized in that

In addition, in the user request processing unit (1), (S1201) waiting for service processing, (S1202) performing GPS-based automatic radius search, (S1203) registration information search request and response to the big data processing unit Receiving, (S1204) providing a response list to the user, (S1205) receiving a user utterance, (S1206) receiving a processing request from a voice recognition processing unit and a response from a big data processing unit, and ( S1207) and linking the application program and POI data.

In addition, in the user request processing unit 1, (S1301) waiting to detect the user's utterance, (S1302) detecting the user's utterance, (S1303) requesting the voice recognition processing unit to process the detected user's utterance and receiving a response, (S1304) receiving a registration information search request and response to the big data processing unit, (S1305) providing a response list to the user, (S1306) receiving a user utterance; , (S1307) receiving a processing request and response from the voice recognition processing unit, and (S1308) linking the POI data with the application program.

In addition, in the big data processing unit 3, (S301) storing the collected data in the form of a file in a designated folder, (S302) detecting a new file in the designated folder, (S303) information from the new file The steps of reading and delivering to the buffering processing unit, (S304) buffering the collected data, (S305) processing the collected data, and (S306) separately storing the data in a distributed database and a relational database characterized by including.

The computer program for processing the big data-based voice recognition POI system stored in the medium of the present invention is characterized in that it includes the processing method.

The medium storing the computer program for processing the big data-based voice recognition POI system of the present invention stores the computer program including the processing method.

The big data-based voice recognition POI system and processing method of the present invention, and a computer program and medium for processing POI data through voice recognition and interactive voice recognition so that users can conveniently use it in registering and utilizing personalized POIs POI data can be collected and registered, and location-based or search-based POI data can be requested and utilized through voice.

In addition, by providing a big data processing system that can effectively process massively collected POI data, each individual participating in the network can efficiently register, manage, and utilize POI data.

1 is a block diagram showing the main components of the big data-based voice recognition POI system of the present invention.
2 is a flowchart showing the processing procedure of the big data processing unit of the big data-based voice recognition POI system of the present invention.
3 is a flowchart illustrating a POI data collection processing procedure of a user request processing unit of the big data-based voice recognition POI system of the present invention.
4 is a sequence diagram illustrating a data flow of a location-based POI registration procedure when collecting POI data of the big data-based voice recognition POI system of the present invention.
5 is a sequence diagram illustrating a data flow of a search-based POI registration procedure when POI data is collected in the big data-based voice recognition POI system of the present invention.
6 is a flowchart illustrating a processing procedure when a location-based POI is utilized by a user request processing unit of the big data-based voice recognition POI system of the present invention.
7 is a sequence diagram illustrating a data flow of a location-based POI utilization procedure when POI is used in the big data-based voice recognition POI system of the present invention.
8 is a flowchart illustrating a processing procedure when a search-based POI is utilized by the user request processing unit of the big data-based voice recognition POI system of the present invention.
9 is a sequence diagram illustrating a data flow of a POI utilization procedure based on a memo search when POI is used in the big data-based voice recognition POI system of the present invention.
10 is a sequence diagram illustrating a data flow of an address search-based POI utilization procedure when POI is used in the big data-based voice recognition POI system of the present invention.

Specific details for carrying out the present invention will be described with reference to the drawings as follows.

1 is a diagram showing the main components of the big data-based voice recognition POI system of the present invention. The big data-based voice recognition POI system of the present invention includes a user request processing unit (1) and a voice recognition processing unit (2) interconnected by a network. and a big data processing unit 3 .

The user request processing unit 1 is an agent that is automatically driven, and functions to process a user's request and a server's response as a part of the system by interworking with the network. A web service that can be installed on terminals such as smart devices and vehicle AVNs used by multiple users, can be downloaded from a server and linked with an application program in a service app or web browser, or implemented as a part of a program, and is provided in the form of an Open API may be implemented as

The user request processing unit 1 transmits the voice or text data collected or requested by the user to the voice recognition processing unit 2 and the big data processing unit 3 by interworking with a user authentication function such as automatic login and an application program, respectively, It performs the function of delivering the data received from the server of the user in conjunction with the application program of the user terminal to the user. In addition, it provides a user interface for performing detailed user requests, such as a GPS-based automatic radius search function, a memo-based search function, and an address-based search function.

The voice recognition processing unit 2 extracts and recognizes voice data from the data received from the user request processing unit 1, converts the recognized voice data into text data, and the user request processing unit 1 or the big data processing unit 3 ) to transmit the function. The main function of the speech recognition processing unit is STT (Speech to Text) or TTS (Text to Speech) function, and for this, a program for speech recognition processing is included. As a voice recognition program, programs such as KALDI, which are provided as open sources, may be used, and various other open programs may be used.

The voice recognition POI system of the present invention may further include an interactive voice processing unit 4 interlocked with the voice recognition processing unit 2 in order to process more complex interactions with the user depending on the situation. The interactive voice processing unit 4 processes such that the user can input various sentences such as “registration of a branch” and “registration of a place” for sentences similar to user request data, for example, “registration of address”, or makes an interactive request. processed to be recognizable. As a program for processing this, the open artificial intelligence library for natural language processing can be used, and the DeepPavlov program, which is open as an open source, can be used.

The big data processing unit 3 collects and processes the POI data received from the user request processing unit 1 or the voice recognition processing unit 2 in real time and then stores it in the data storage unit 34 or responds to a POI data search request It performs a function of transmitting data to the user request processing unit (1).

The big data processing unit 3 may include a data collection unit 31 , a buffering processing unit 32 , and a data processing unit 33 to store data after distributed processing in order to effectively process massive data collected in real time. . At this time, the data storage unit 34 is divided into a distributed database 341 that stores real-time collected data for a point of interest and a relational database 342 that stores basic information such as user information, vehicle additional information, and Open API information. to manage it effectively. The distributed database 341 may utilize HBASE based on Hadoop Distributed File System (HDFS), and the relational database may utilize various tools such as MSSQL and MySQL.

The data collection unit 31 effectively collects a large amount of data produced by various agents and transmits it to the buffering processing unit 32 for management in a distributed file system such as HDFS. When the collected data is saved in the form of a file, it includes a function to detect a new file in a specified folder, and a function to read information from the new file and deliver it to the collected data buffering processing unit. An example is the open source Flume provided by the (Apache) foundation as a log data collection technology. The advantage of using Flume for mass data collection is that it can secure reliability, scale scalability, and function scalability. In addition, it can prevent the loss of collected data even in the event of a failure, and it is very useful because it supports all extension methods of scale-up/scale-out.

The buffering processing unit 32 buffers the data transmitted from the data collection unit 31 and transmits it to the data processing unit 33 . Kafka and zookeeper provided by the Apache Foundation can be used for buffering of the collected data.

The data processing unit 33 processes the data transmitted from the buffering processing unit 32 into an appropriate format and then stores the data in the distributed database 341 and the relational database 342 according to the data type, respectively. You will be able to utilize the Apache Foundation's Storm for this. Apache Storm is a real-time distributed processor that distributes data received from Kafka, etc. and loads it into the final destination storage.

2 is a view for explaining a processing procedure for the big data processing unit 3 to effectively process a large amount of POI data collected in real time. As shown, the method of processing big data in the big data processing unit 3 of the big data-based voice recognition POI system of the present invention includes (S301) storing the collected data in a file form in a designated folder, (S302) Detecting a new file in a designated folder, (S303) reading information from the new file and transmitting it to the buffering processing unit, (S304) buffering the collected data, (S305) processing the collected data and, (S306) separately storing the distributed database and relational data. The steps (S301) to (S303) are processed by the data collection unit 31, the step (S304) is processed by the buffering processing unit 32, and the steps (S304) to (S306) are processed by the data processing unit (33) and the data storage unit (34).

3 is a diagram illustrating a POI data collection processing procedure of the user request processing unit 1 of the big data-based voice recognition POI system of the present invention, and FIG. 4 is a data flow diagram illustrating a location-based POI registration procedure when collecting POI data. It is a sequence diagram, and FIG. 5 is a sequence diagram illustrating a data flow of a search-based POI registration procedure when collecting POI data.

In order to collect personalized POI data, a procedure for processing user data in the user request processing unit 1 is important. To this end, as shown in FIG. 3, in the user request processing unit 1 of the big data-based voice recognition POI system of the present invention, (S1101) waiting to detect a user's speech, (S1102) detecting the user's speech Steps, (S1103) requesting processing of the detected user utterance to the voice recognition processing unit and receiving a response, (S1104) converting a GPS address and generating POI data according to the received response, and (S1105) generated It includes the procedure of performing the step of requesting data collection to the big data processing unit for POI data.

The POI collection method of the big data-based voice recognition POI system of the present invention is a location-based POI registration method (FIG. 4) that directly visits a place to register (restaurant, tourist destination, etc.) There is a search-based POI registration method (FIG. 5) that registers it after doing so.

An example of location-based POI registration is shown in FIG. 4 . When the user makes a voice call to the waiting user request processing unit (agent), the user request processing unit detects the user's utterance. At this time, the voice recognition processing unit 2 may be switched to a Wakeup/Ready state. When the user utters "Register here" to register the current location, the voice recognition processing unit 2 completes the recognition, and responds to the user request processing unit 1 whether or not the recognition has been completed. When the user request processing unit 1 requests a memo to the user, "Please tell me the registration memo," the user ignites a custom memo such as "Yonggye-dong Heukdon", and the user request processing unit 1 converts the voice data into the voice recognition processing unit 2 ) to convert it to text. The user request processing unit 1 converts the data converted into text into a GPS address and then creates a personalized POI along with a memo. Personalized POI data such as GPS, addresses, memos, and utterances are transmitted to the big data processing system and registered.

An example of the search-based POI registration is as shown in FIG. 5, and the same procedure as above is performed except that the address to be registered after the user searches is registered along with a memo.

In order to process complex interactions according to situations such as similar sentences or conversational sentences in the voice processing process, the conversational voice processing unit 4 interlocked with the voice recognition processing unit 1 helps the processing.

6 is a flowchart illustrating a processing procedure when a location-based POI is utilized by the user request processing unit of the big data-based voice recognition POI system of the present invention, and FIG. 7 is a sequence diagram illustrating a data flow of the location-based POI usage procedure.

In order to utilize the location-based POI of the big data-based voice recognition POI system of the present invention, in the user request processing unit 1 (S1201) waiting for service processing, (S1202) performing a GPS-based automatic radius search and (S1203) receiving a registration information search request and a response from the big data processing unit to the big data processing unit, (S1204) providing a response list to the user, (S1205) receiving a user utterance, ( S1206) receiving a processing request and a response from the voice recognition processing unit, and (S1207) linking the application program and POI data (POI-GPS data).

Looking at the location-based POI application example as shown in FIG. 7 , when a user visits a place already registered as POI data, the user request processing unit 1 continuously performs a GPS-based automatic radius search while waiting for the user's request while requesting the big data processing unit 3 to search the registration information of the POI collected along with the memo. The user request processing unit 1 provides the user with the data received from the big data processing unit saying, “There are OO registered P.O.I.” When the user utters “Tell me the list”, the user request processing unit 1 detects the utterance and requests the voice recognition processing unit 2 for voice recognition. When recognition is completed by the voice recognition processing unit, the recognized data is converted into text and transmitted to the big data processing unit 2 , and the big data processing unit 2 transmits a response to the request to the user request processing unit 1 . After receiving the responses of the voice recognition processing unit 2 processing and the big data processing unit 3 for additional requests such as "Let's go to the first memo", the user request processing unit 1 performs the navigation program (map, etc.) and POI - Link GPS data to provide a service to users.

8 is a flowchart illustrating a processing procedure when a search-based POI is utilized by the user request processing unit of the big data-based voice recognition POI system of the present invention. As shown, in the big data-based voice recognition POI system of the present invention, in the user request processing unit 1 (S1301) waiting to detect a user utterance, (S1302) detecting a user utterance, (S1303) ) requesting processing of the sensed user's utterance from the voice recognition processing unit and receiving a response, (S1304) receiving a registration information search request and response from the big data processing unit, (S1305) providing a response list to the user Steps, (S1306) receiving a user's utterance, (S1307) receiving a processing request and response to a voice recognition processing unit, and (S1308) linking the application program and POI data (POI-GPS data) do.

9 and 10 are sequence diagrams illustrating a data flow of a search-based POI utilization procedure when POI is utilized in the big data-based voice recognition POI system of the present invention. is an example of POI utilization procedure through address-based search.

As shown, for the wakeup/ready state of the user request processing unit 1 and the voice recognition processing unit 2 (and the interactive voice processing unit 3), the user utters "Doll P.O.I!". When the user utters "let's find" to the user request processing unit 1 , the voice recognition processing unit 2 directly completes recognition, or completes recognition and responds through the interactive voice processing unit 4 . When the user request processing unit 1 utters "Please tell me a note or a nearby address you want to find", the user says "Memo! Pork" or "Address! Daegu Yonggye-dong" transmits it to the voice recognition processing unit 2, and the voice recognition processing unit 2 converts the received voice data into text and responds. The user request processing unit (1) requests a search from the big data processing unit (3) based on the memo data or address data converted into text, and after receiving a response, “There are OO P.O.Is registered with pork” or “Daegu Yonggye-dong” There are OO registered P.O.I." to the user. Afterwards, when the user utters "Tell me the list~", the following procedure is performed as in the location-based POI utilization method procedure.

The above contents relate to preferred embodiments of the present invention, and based on the above contents, various forms can be applied, modified, and implemented within the scope of the technical idea of the present invention. Accordingly, the scope of the present invention should not be construed as being limited by the above content.

1: User request processing unit
2: Voice recognition processing unit
3: Big data processing unit
31: data collection unit
32: buffering processing unit
33: data processing unit
34: data storage unit
341: distributed database
342: relational database
4: Interactive speech processing unit

Claims (9)

Including a user request processing unit, a voice recognition processing unit, and a big data processing unit interconnected by a network,
The user request processing unit is an agent installed in a plurality of user terminals or downloaded from a server and driven, and processes user authentication, collects voice or text data by interworking with an application program, or collects user data requested by a voice recognition processing unit and a big data processing unit transmits the function to and forwards the response to the user,
The voice recognition processing unit extracts and recognizes voice data from the data received from the user request processing unit, converts it into text data, and transmits it to the user request processing unit or big data processing unit,
The big data processing unit collects and processes the POI data received from the user request processing unit or the voice recognition processing unit in real time and stores it in the data storage unit, or transmits the data to the user request processing unit in response to a POI data search request. Big data-based voice recognition POI system characterized by.
According to claim 1, wherein the big data-based voice recognition POI system further comprises an interactive voice processing unit interlocked with a voice recognition processing unit, wherein the interactive voice processing unit is a similar sentence or conversational sentence that is not immediately recognized by the voice recognition processing unit Big data-based voice recognition POI system, characterized in that it interprets and recognizes them and responds to the voice recognition processing unit.
The method of claim 1, wherein the big data processing unit includes a data collection unit, a buffering processing unit, and a data processing unit to store data after distributed processing, and the data storage unit includes a distributed database and a relational database,
The data collection unit collects a large amount of data produced by various agents and transmits it to the buffering processing unit,
The buffering processing unit buffers the data transmitted from the data collection unit and transmits it to the data processing unit,
Big data-based voice recognition POI, characterized in that the data processing unit processes the data transmitted from the buffering processing unit into an appropriate format and then stores the data in the distributed database and the relational database of the data storage unit according to the data type, respectively system.
In the user request processing unit (1) of the big data-based voice recognition POI system of claim 1,
(S1101) waiting to detect a user's utterance;
(S1102) detecting a user's utterance;
(S1103) requesting the speech recognition processing unit to process the sensed user's utterance and receiving a response;
(S1104) converting a GPS address and generating POI data according to the received response; and
(S1105) Big data-based voice recognition POI system processing method, characterized in that performing the step of requesting the data collection of the generated POI data to the big data processing unit.
In the user request processing unit (1) of the big data-based voice recognition POI system of claim 1,
(S1201) waiting for service processing;
(S1202) performing a GPS-based automatic radius search;
(S1203) receiving a registration information search request and response to the big data processing unit;
(S1204) providing a response list to the user;
(S1205) receiving a user utterance;
(S1206) receiving a processing request and a response from the big data processing unit to the voice recognition processing unit; and
(S1207) Big data-based voice recognition POI system processing method comprising the step of linking an application program and POI data.
In the user request processing unit (1) of the big data-based voice recognition POI system of claim 1,
(S1301) waiting to detect a user's utterance;
(S1302) detecting a user's utterance;
(S1303) requesting the speech recognition processing unit to process the detected user's utterance and receiving a response;
(S1304) receiving the registration information search request and response to the big data processing unit;
(S1305) providing a response list to the user;
(S1306) receiving a user utterance;
(S1307) receiving a processing request and a response to the voice recognition processing unit; and
(S1308) Big data-based voice recognition POI system processing method comprising the step of linking an application program and POI data.
In the big data processing unit (3) of the big data-based voice recognition POI system of claim 1,
(S301) storing the collected data in the form of a file in a designated folder;
(S302) detecting a new file in a designated folder;
(S303) reading information from the new file and transmitting the information to the buffering processing unit;
(S304) buffering the collected data;
(S305) processing the collected data; and
(S306) Big data-based voice recognition POI system processing method comprising the step of separately storing each in a distributed database and a relational database.
A computer program for processing a big data-based voice recognition POI system stored in a medium comprising the processing method of any one of claims 4 to 7.
A medium storing a computer program for processing a big data-based voice recognition POI system comprising the processing method of any one of claims 4 to 7.

Images