US-12626705-B2 - Apparatus and method for mapping emergency call data manual

US12626705B2US 12626705 B2US12626705 B2US 12626705B2US-12626705-B2

Abstract

According to an embodiment of the present disclosure, an apparatus for An apparatus for mapping an emergency call data manual, the apparatus comprising: an input unit that receives an emergency call in voice form; a preprocessing unit that converts the voice form into text data and preprocesses the text data to generate an emergency call data token; a keyword detection unit that detects main keywords using a manual and generates a mapping rule using the main keywords; a model generation unit that generates a model by training an artificial intelligence model using manual-mapped data; a manual mapping unit that generates a response manual for the text data by performing mapping based on the mapping rule and emergency call data token and mapping based on the model; and a display unit that displays the response manual.

Inventors

Min Jung Lee
Eun Jung KWON
Hyun Ho Park
Sung Won BYON

Assignees

ELECTRONICS AND TELECOMMUNICATIONS RESEARCH INSTITUTE

Dates

Publication Date: 20260512
Application Date: 20240517
Priority Date: 20230522

Claims (10)

1 . An apparatus for mapping an emergency call data manual, the apparatus comprising: an input unit that receives an emergency call in voice form; a preprocessing unit that converts the voice form into text data and preprocesses the text data to generate an emergency call data token; a keyword detection unit that detects main keywords using a stored manual and generates a mapping rule using the main keywords; a model generation unit that generates a model by training an artificial intelligence model using manual-mapped data; a manual mapping unit that generates a response manual for the text data by performing mapping based on the mapping rule and mapping based on the model; and a display unit that displays the response manual, wherein the manual of the keyword detection unit is extracted from a manual DB, and the manual-mapped data of the model generation unit is extracted from a manual-mapped data DB, wherein the keyword detection unit is configured to: refine and tokenize manual text data existing in the manual DB; calculate a frequency of a token appearing in an entire set of manuals for each token; calculate an importance of each token appearing in each manual; and detect a set of main keywords for each manual using the importance, and generate the mapping rule using the set of main keywords, and wherein the importance is calculated by dividing the frequency of each token appearing in each manual by the frequency of the corresponding token appearing in the entire set of manuals stored in the manual DB.
2 . The apparatus of claim 1 , wherein the manual mapping unit includes a rule-based mapping unit that when the rule generated by the keyword detection unit is applied to the emergency call data token, maps the corresponding manual.
3 . The apparatus of claim 2 , wherein the wherein the importance is calculated according to the following equation, in which S is an importance of a token i: S ⁡ ( token i ) = frequency ⁢ of ⁢ token i ⁢ appearing ⁢ in ⁢ manual frequency ⁢ of ⁢ token i ⁢ appearing ⁢ in ⁢ entire ⁢ set ⁢ of ⁢ manuals .
4 . The apparatus of claim 2 , wherein the manual mapping unit includes a model-based mapping unit that performs manual mapping based on the model generated by the model generation unit when the number of the manual-mapped data is greater than or equal to a predetermined value.
5 . The apparatus of claim 1 , wherein the manual mapping unit is further configured to pseudo-map unlabeled emergency call data to which no manual is mapped with a predicted value output from the model.
6 . A method for mapping an emergency call data manual, the method comprising: receiving an emergency call in voice form in an input unit; converting the voice form into text data and preprocessing the text data to generate an emergency call data token in a preprocessing unit; detecting main keywords using a stored manual and generating a mapping rule using the main keywords in a keyword detection unit; generating a model by training an artificial intelligence model using manual-mapped data in a model generation unit; generating a response manual for the text data by performing mapping based on the mapping rule and mapping based on the model in a manual mapping unit; and displaying the response manual in a display unit, wherein the manual of the keyword detection unit is extracted from a manual DB, and the manual-mapped data of the model generation unit is extracted from a manual-mapped data DB, wherein generating the mapping rule includes: refining and tokenizing manual text data existing in the manual DB; calculating a frequency of a token appearing in an entire set of manuals for each token; calculating an importance of each token appearing in each manual; and detecting a set of main keywords for each manual using the importance, and generating the mapping rule using the set of main keywords, and wherein the importance is calculated by dividing the frequency of each token appearing in each manual by the frequency of the token appearing in the entire set of manuals.
7 . The method of claim 6 , wherein the manual mapping unit is further configured to, when the rule generated by the keyword detection unit is applied to the emergency call data token, map the corresponding manual.
8 . The method of claim 7 , wherein the manual mapping unit is further configured to perform manual mapping based on the model generated by the model generation unit when the number of the manual-mapped data is greater than or equal to a predetermined value.
9 . The method of claim 6 , wherein the manual mapping unit is further configured to pseudo-map unlabeled emergency call data to which no manual is mapped with a predicted value output from the model.
10 . The method of claim 6 , wherein the importance is calculated according to the following equation, in which S is an importance of a token i: S ⁡ ( token i ) = frequency ⁢ of ⁢ token i ⁢ appearing ⁢ in ⁢ manual frequency ⁢ of ⁢ token i ⁢ appearing ⁢ in ⁢ entire ⁢ set ⁢ of ⁢ manuals .

Description

CROSS-REFERENCE TO RELATED APPLICATION The present application claims priority to Patent Application No. 10-2023-0065737, filed on in Korea Intellectual Property Office on May 22, 2023, the entire contents of which are incorporated herein by reference. TECHNICAL FIELD The present disclosure relates to an apparatus and method for mapping emergency call data manuals. BACKGROUND The contents described below merely provide background information related to the present disclosure and do not constitute prior art. An emergency call refers to a call for help in the event of an unexpected emergency, such as a call to 119, 112, etc. When an operator receives an emergency call, the ability to understand the situation and respond to the call varies depending on the operator's capability, experience, and real-time conditions. However, in the event of a disaster or incident, the scale of damage may vary depending on the operator's ability to understand the situation and initially respond it. Therefore, there is a need for an apparatus and method that supports rapid and accurate response to emergency calls, regardless of the operator's capabilities and conditions. SUMMARY In view of the above, the present disclosure provides an apparatus and method for supporting rapid and accurate response to emergency calls. The present disclosure provides an apparatus and method for providing a person receiving an emergency call with a response manual appropriate to the content of the call in real time. The objects to be achieved by the present disclosure are not limited to the objects mentioned above, and other objects not mentioned will be clearly understood by one of ordinary skill in the art from the description below. According to an embodiment of the present disclosure, an apparatus for An apparatus for mapping an emergency call data manual, the apparatus comprising: an input unit that receives an emergency call in voice form; a preprocessing unit that converts the voice form into text data and preprocesses the text data to generate an emergency call data token; a keyword detection unit that detects main keywords using a manual and generates a mapping rule using the main keywords; a model generation unit that generates a model by training an artificial intelligence model using manual-mapped data; a manual mapping unit that generates a response manual for the text data by performing mapping based on the mapping rule and emergency call data token and mapping based on the model; and a display unit that displays the response manual. According to an embodiment of the present disclosure, a method for A method for mapping an emergency call data manual, the method comprising: receiving an emergency call in voice form in an input unit; converting the voice form into text data and preprocessing the text data to generate an emergency call data token in a preprocessing unit; detecting main keywords using a manual and generating a mapping rule using the main keywords in a keyword detection unit; generating a model by training an artificial intelligence model using manual-mapped data in a model generation unit; generating a response manual for the text data by performing mapping based on the mapping rule and the emergency call data token and mapping based on the model in a manual mapping unit; and displaying the response manual in a display unit. According to the present disclosure, by providing to a receiver who receives an emergency call with a response manual appropriate to the content of the call in real time, the receiver can perform efficient reception and response. The present disclosure enables effective response and initial response to an incident regardless of the call taker's capabilities and conditions. The present disclosure enables manual mapping and display with rule-based mapping even in the early stages when model-based mapping cannot be accurately performed. According to the present disclosure, manual mapping accuracy can be increased by performing pseudo mapping using model-based mapping and using the data for model learning even on emergency call data that is not mapped to a manual. The effects of the present disclosure are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by one of ordinary skill in the art from the following description. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram of an apparatus for mapping emergency call data manuals according to one embodiment of the present disclosure. FIG. 2 is a flowchart of the operation of a keyword detection unit of the apparatus for mapping emergency call data manuals according to one embodiment of the present disclosure. FIG. 3 is a flowchart of the operation of a manual mapping unit of the apparatus for mapping emergency call data manuals according to one embodiment of the present disclosure. DETAILED DESCRIPTION Hereinafter, some embodiments of the present disclosure will be described in detail with referenc