JP6945461B2 - Safety information provision system - Google Patents

Description

The present invention relates to a safety information providing system that provides safety information related to construction safety.

Conventionally, it has been proposed that information on disasters at construction sites be used for the safety of future work. For example, Patent Document 1 describes a system that stores information on disasters that have occurred in the past and displays the information in response to input of data for work to be performed in the future.

Japanese Unexamined Patent Publication No. 6-44211

By storing information indicating past disaster cases in the database, entering keywords related to the work, and searching for similar disaster cases from the database based on the keywords, the past disasters and countermeasures due to the work can be confirmed. Will be possible. However, if the input keyword itself is not included in the information indicating the disaster case, even if the disaster case is related to the keyword, the information is not extracted by the search. That is, there is a risk that appropriate information cannot be output depending on the keyword. It is conceivable to prepare synonyms for keywords in advance, but it is not easy to prepare appropriate synonyms for searching for disaster cases, and a large amount of time and effort is required.

The present invention has been made in view of the above, and an object of the present invention is to provide a safety information providing system capable of providing appropriate safety information for input.

In order to achieve the above object, the safety information providing system according to the present invention is a safety information providing system that provides safety information related to construction safety, and includes a providing database that stores the safety information to be provided. A learning database that is used for learning related words and stores learning information related to construction, and a related word identification means that identifies a set of related words that are related to each other by machine learning from the learning information stored in the learning database. , By the search information input means for inputting the keyword and the related word specifying means, by the related keyword specifying means for specifying the related keyword which is the related word of the keyword input by the search information input means, and the related keyword specifying means. Using the specified related keywords, a search means for searching the safety information stored in the providing database and extracting the provided safety information, and an output means for outputting the safety information extracted by the search means are used. Be prepared.

In the safety information providing system according to the present invention, the related words of the input keyword among the related words specified by machine learning from the learning information are used as related keywords to search for safety information. Therefore, even if the entered keyword itself is not included in the safety information, appropriate safety information is extracted by the search. Thereby, according to the safety information providing system according to the present invention, it is possible to provide appropriate safety information for the input.

The search information input means may input the information related to the weather, and the search means may also perform the search using the information related to the weather input by the search information input means. According to this configuration, it is possible to provide appropriate safety information according to the weather at a construction site or the like.

Search information input means, a plurality of setting keyword corresponding to the type of disaster with stored set for each type of advance time and disaster inputs to obtain the set keywords depending on the time of day, the search means It is set keyword inputted also intends row a search using the search information input unit. According to this configuration, it is possible to prevent the search results from becoming a rut.

According to the present invention, even if the input keyword itself is not included in the safety information, the appropriate safety information is extracted by the search, so that the appropriate safety information can be provided for the input.

It is a figure which shows the structure of the local KY support system which is the safety information provision system which concerns on embodiment of this invention. It is a figure which shows the example of the disaster information stored in the disaster case database. It is a figure which shows the example of the input screen of the information about the work scheduled for the next day. This is a table showing an example of setting keywords stored in advance. It is a figure which shows the example of the sent e-mail. It is a flowchart which shows the process executed in the local KY support system which is the safety information provision system which concerns on embodiment of this invention.

Hereinafter, embodiments of the safety information providing system according to the present invention will be described in detail together with the drawings. In the description of the drawings, the same elements are designated by the same reference numerals, and duplicate description will be omitted.

FIG. 1 shows a local KY support system 10 which is a safety information providing system according to the present embodiment. This is a system that supports KY (danger prediction) at the construction site of the local KY support system 10 such as construction or civil engineering. The local KY support system 10 provides construction site workers or work managers with safety information related to construction safety. The safety information provided is disaster information related to past disaster cases.

At the construction site, for example, the work is performed as follows. Employees of a construction company's partner company perform work at the construction site. Usually, a plurality of subcontractors participate in one construction site, and each subcontractor is assigned to each work at the construction site. An employee of a construction company becomes a manager of work, that is, a manager of a partner company. The work to be managed (in charge) for each employee of the construction company, that is, the cooperating company is assigned.

Under the above circumstances, the local KY support system 10 is used, for example, as follows. The local KY support system 10 inputs information on the work scheduled for the next day at the construction site every day from the input terminal 20 in response to the input operation of the foreman who is the leader of the employee of the partner company. The local KY support system 10 searches for disaster information that contributes to the KY of the work scheduled for the next day, based on the input information. The local KY support system 10 transmits the information extracted by the search to the receiving terminal 30 owned by the employee of the construction company that manages the partner company. Employees of the construction company refer to the information and inform and instruct the employees of the partner company who are the workers about the disaster case at the morning assembly on the day when the work is done and the TBM (toolbox meeting) after the morning assembly. Raise awareness of the crisis.

Specifically, the local KY support system 10 is composed of a server device which is a computer including hardware such as a CPU (Central Processing Unit) and a memory. Each function described later in the local KY support system 10 is exhibited by operating these components by a program or the like. The local KY support system 10 may be realized by one computer, or may be realized by a computer system in which a plurality of computers are connected to each other by a network.

The local KY support system 10 can send and receive information between the input terminal 20 and the receiving terminal 30 via a wireless communication network and a network N such as the Internet. The input terminal 20 is, for example, a terminal used by the foreman of a construction site to input information used for searching disaster information. The receiving terminal 30 is, for example, a terminal used by an employee of a construction company to receive information on a disaster case. Specifically, the input terminal 20 and the receiving terminal 30 are computers such as a PC (personal computer), a tablet terminal, and a smartphone.

Subsequently, the function of the local KY support system 10 according to the present embodiment will be described. As shown in FIG. 1, the local KY support system 10 includes a disaster case database 11, a learning database 12, a related word identification unit 13, a search information input unit 14, a related keyword identification unit 15, and a search unit. 16 and an output unit 17 are provided.

The disaster case database 11 is a providing database that stores disaster information to be provided. FIG. 2 shows an example of disaster information stored in the disaster case database 11. The disaster information relates to, for example, a disaster case that actually occurred at a past construction site undertaken by a construction company. One disaster information indicates one disaster case, and the disaster case database 11 stores a plurality of disaster information. Disaster information should be related to various disasters. For example, the disaster information should include not only those related to fatal accidents and injuries but also those related to serious cases of hiyari hats.

As shown in FIG. 2, the disaster information includes a character string (text) related to the disaster. As shown in Fig. 2, the character string indicates what kind of construction and work caused the disaster (accident), the type of disaster, the date and time of the disaster, the weather, the workers involved in the disaster, and the status of the disaster. It shows any matters related to disasters such as explanations of disasters, explanations of causes of disasters, and explanations of countermeasures against disasters. In addition, the disaster information may include drawings showing the situation of the disaster and the like. The disaster information stored in the disaster case database 11 is generated in advance by, for example, the administrator of the local KY support system 10 and input to the local KY support system 10.

The learning database 12 is a database used for learning related words and storing learning information related to construction. The learning database 12 stores, for example, disaster information stored in the disaster case database 11 as learning information. Further, the learning database 12 may store disaster information (disaster information stored in a public occupational accident database) other than the disaster information to be provided. In addition to the disaster information, the learning database 12 may store information related to construction as learning information. For example, the learning database 12 may store information on minutes, educational materials, and on-site patrols in a construction company as learning information. The learning information stored in the learning database 12 includes a character string so that it can be used for learning related words. The learning information stored in the learning database 12 should be prepared as much as possible so that appropriate learning can be performed.

If only the disaster information stored in the disaster case database 11 is used as learning information for learning related words, it is not necessary to provide the learning database 12 separately from the disaster case database 11 (in this case, the disaster case). The database 11 also serves as the learning database 12). Further, the learning database 12 does not necessarily have to store the disaster information stored in the disaster case database 11. The learning information stored in the learning database 12 is, for example, acquired in advance by the administrator of the local KY support system 10 and input to the local KY support system 10.

The related word identification unit 13 is a related word identification means for identifying (learning) a set of related words related to each other by machine learning from the learning information stored in the learning database 12. The related word identification unit 13 reads out the learning information stored in the learning database 12, performs machine learning using the read learning information, and specifies a set of related words related to each other. The related word identification unit 13 identifies a set of related words by a conventional machine learning method (AI (artificial intelligence) program). A set of related words is a combination of words that are related to each other. A set of related words contains multiple words. Specifically, for example, the related word identification unit 13 calculates the meaning vector of each word by the machine learning program “Word2Vec”, and calculates the similarity (score) between words from the calculated meaning vector. The related word identification unit 13 sets a group of words whose similarity is equal to or higher than a preset threshold value as a set of related words. For example, for the word "H steel", the word groups "SRC" and "further edge" are set as related words. Alternatively, for the word "fall", the word groups "shiri" and "fall" are set as related words.

As described above, by learning related words using construction-related learning information, a set of related words appropriate for searching construction-related information, that is, safety information, rather than general related words. Can be obtained. The related words do not necessarily have to be words (synonyms) having similar meanings, and may be any related words found by machine learning. The set of related words identified by the related word identification unit 13 is stored in the local KY support system 10 and referred to by the related keyword identification unit 15.

The search information input unit 14 is a search information input means for inputting information used for searching disaster information in the disaster case database 11. The search information input unit 14 inputs a keyword. The search information input unit 14 inputs information related to the weather. The search information input unit 14 inputs preset keywords according to the time. Specifically, the search information input unit 14 inputs each information as follows.

The search information input unit 14 inputs a keyword from the input terminal 20 as follows. The input terminal 20 requests the local KY support system 10 to input information regarding the work scheduled for the next day by the operation of the foreman. The local KY support system 10 transmits the input form shown in FIG. 3 to the input terminal 20 in response to the request (in the example of FIG. 3, each information has already been input to the form). The input terminal 20 transmits the information entered in the form by the operation of the foreman to the local KY support system 10.

The information entered in the form is information about the work scheduled for the next day at the construction site. As shown in FIG. 3, the information includes work location (building, floor, area), work content, expected danger, safety instructions, precautions for other jobs, and the like. The information includes information input as a character string. In FIG. 3, the information input as a character string is the information in column F of the work content, expected danger, safety instructions, and precautions for other jobs. The character string is freely input by, for example, the foreman (a history of past input may be used to facilitate input).

The search information input unit 14 receives and inputs the information transmitted from the input terminal 20. The search information input unit 14 extracts a keyword from the character string included in the input information and inputs it. For example, the search information input unit 14 performs morphological analysis on the input character string, divides it into words, excludes words set as words to be excluded in advance from the words, and uses the remaining words as keywords. Extract as. The search information input unit 14 outputs the extracted keyword to the related keyword identification unit 15.

Information on the work scheduled for the next day may be input and managed by a system different from the local KY support system 10 (for example, a system that manages the work at the construction site). In that case, the search information input unit 14 may acquire information on the work scheduled for the next day from the other system.

The search information input unit 14 acquires and inputs information related to the weather from an external weather forecast service using an API (application programming interface). The acquired weather information is for the day and place where the searched disaster information is used. For example, when the search is performed the day before the disaster information is used, the search information input unit 14 acquires the information related to the weather on the next day of the construction site. The acquired weather information is in a format that can be used to search for disaster information, and is, for example, a character string such as "rain", "fine", "typhoon", or "strong wind". In addition, the information related to the weather may include a character string indicating an event or the like caused by the weather as the main cause such as "heat stroke". Information on the weather may be input by a method other than the above (for example, manual input).

The search information input unit 14 acquires and inputs the set keywords from the table of the set keywords stored in advance according to the time at the time of acquisition. As shown in FIG. 4, the set keywords are stored for each classification and day. The classification indicates the type of disaster, and as shown in FIG. 4, there are four types: a fall type, a heavy equipment contact type, a flying fall type, and a fire / fire type. In addition, there may be a classification other than the above four. The setting keywords stored in the table are input to the local KY support system 10 in advance by the administrator of the local KY support system 10, for example. The search information input unit 14 acquires the setting keywords of each classification on a daily basis according to the acquisition date. For example, if the setting keyword of "Day-1" was acquired on the previous day, the setting keyword of "Day-2" is acquired.

The search information input unit 14 outputs the acquired information related to the weather and setting keywords to the search unit 16. It should be noted that the input of each information by the search information input unit 14 may be performed at the timing before the search by the search unit 16 is performed.

The related keyword specifying unit 15 is a related keyword specifying means for specifying a related keyword as a related word of the keyword input by the search information input unit 14 by the related word specifying unit 13. The related keyword identification unit 15 inputs a keyword from the search information input unit 14. The related keyword specifying unit 15 reads out a set of related words specified by the related word specifying unit 13. The related keyword identification unit 15 identifies a word that is a related word with the input keyword as a related keyword in the read related word set. For example, if "H steel" is an input keyword, "SRC" and "further edge" are specified as related keywords. When there are a plurality of keywords input from the search information input unit 14, the related keyword specifying unit 15 specifies the related keyword for each of the input keywords. Also, there may be one related keyword (rather than plural) for one keyword.

Further, the related keyword specifying unit 15 may specify the related keyword from the keyword by using a related word other than the set of the related word specified by the related word specifying unit 13. For example, a synonym dictionary may be stored in advance, and related keywords may be specified from the keywords by the synonym dictionary. The related keyword specifying unit 15 outputs the input keyword and the specified related keyword to the search unit 16.

The search unit 16 is a search means for extracting disaster information provided by searching for disaster information stored in the disaster case database 11. The search unit 16 uses the related keyword specified by the related keyword specifying unit 15 for the search. In addition, the search unit 16 also performs a search using the information related to the weather and the setting keywords input by the search information input unit 14.

The search unit 16 inputs a keyword and a related keyword from the related keyword specifying unit 15. The search unit 16 inputs information related to the weather and setting keywords from the search information input unit 14. The search unit 16 uses all the input information as search keywords to search for disaster information stored in the disaster case database 11 for each disaster classification (search for a character string included in the disaster information). The search unit 16 uses the keywords, related keywords, and information related to the weather to search all categories. The search unit 16 uses the set keywords to search for the corresponding classification.

The search by the search unit 16 can be performed by a conventional search engine that performs a text search (for example, a full-text search engine such as Fess). The search unit 16 calculates, for example, the search score (the degree to which the search keyword is hit) of each disaster information by the conventional search engine. The search unit 16 extracts the disaster information having the highest score for each disaster classification as a search result. The search unit 16 outputs information indicating disaster information extracted as a search result for each disaster classification to the output unit 17.

The above search is performed every time the foreman of a partner company inputs information about the work scheduled for the next day. That is, the disaster information extracted by the search corresponds to the partner company (the work). Further, as the search keyword used for the search by the search unit 16, in addition to the above, a keyword other than the above may be input and used. For example, a keyword indicating the occupation of the worker may be used in the search.

The use of the related keyword in the search by the search unit 16 may be performed by a method other than the above. For example, first, when a search is performed without using related keywords and disaster information cannot be extracted (when disaster information does not hit), or when there is no disaster information that exceeds a certain score. A search may be performed using a related keyword as a search keyword.

The output unit 17 is an output means for outputting the disaster information extracted by the search unit 16. The output unit 17 outputs the disaster information to the output destination according to the input source of the keyword by the search information input unit 14.

The output unit 17 inputs information indicating disaster information from the search unit 16. The output unit 17 transmits information indicating the input disaster information to the receiving terminal 30. For example, the output unit 17 generates an e-mail including information indicating the disaster information and transmits it to the receiving terminal 30. An example of the e-mail is shown in FIG. The e-mail includes a link L (for example, a URL (Uniform Resource Locator)) to the disaster information for each disaster classification as information indicating the disaster information. The text of the e-mail is stored in advance in the output unit 17, and is used for generating the e-mail. When the receiving terminal 30 that has received the e-mail accesses the URL, the output unit 17 transmits the disaster information corresponding to the URL to the receiving terminal 30 in response to the access. As a result, the receiving terminal 30 displays the disaster information as shown in FIG.

Further, as shown in FIG. 5, it may be possible to input feedback information as to whether or not the disaster information has been referred from the e-mail. The feedback information can be used, for example, when determining the display order of disaster information. For example, disaster information with a large number of references is displayed at the top, and disaster information with a large number of non-references is displayed at the bottom.

The output unit 17 transmits the e-mail at a preset time. For example, the output unit 17 sends an e-mail at 7:00 am on the day when the work is performed. The e-mails sent by the output unit 17 are sent all at once by the administrator of the local KY support system 10 after checking the destination and the contents of the e-mails to be sent, triggered by the operation of the administrator of the local KY support system 10. Alternatively, it may be performed individually. In addition, the content of the e-mail may be edited by the administrator of the local KY support system 10.

As described above, the disaster information extracted by the search is for each partner company (for each work). The output unit 17 transmits the above-mentioned e-mail related to the disaster information to the receiving terminal 30 of the employee of the construction company who manages the cooperating company corresponding to the extracted disaster information. The correspondence between the partner company and the employees of the construction company to be managed is stored in advance in the output unit 17. For employees of a construction company who are not assigned a subcontractor, the output unit 17 may send an e-mail of disaster information corresponding to the subcontractor at high risk. For example, since there is statistical data that a subcontractor who has just entered the site is prone to accidents, remember the registration date of the subcontractor at the construction site and respond to the subcontractor with the shortest number of days from the registration date. Send an e-mail of disaster information. The above is the function of the local KY support system 10 according to the present embodiment.

Subsequently, the process executed by the local KY support system 10 according to the present embodiment will be described with reference to the flowchart of FIG. In this process, first, the related word identification unit 13 specifies (learns) a set of related words related to each other by machine learning from the learning information stored in the learning database 12 (S01). The specified set of related words is stored in the local KY support system 10 and referred to by the subsequent processing by the related keyword specifying unit 15. The above process (S01) may be performed only once in advance before the subsequent series of processes (S02 to S05), and does not necessarily have to be performed in each process.

Subsequently, the search information input unit 14 inputs information used for searching disaster information in the disaster case database 11 (S02). Specifically, the information transmitted from the input terminal 20 is received by the operation of the foreman, and the keyword based on the information is extracted and input. In addition, information related to the weather and setting keywords are acquired and input. Subsequently, the related keyword specifying unit 15 identifies the related keyword as the related word of the input keyword (S03).

Subsequently, the search unit 16 uses the keyword, the related keyword, the information related to the weather, and the set keyword as the search keyword, and searches for the disaster information stored in the disaster case database 11 (S04). Subsequently, the disaster information extracted by the search is transmitted by the output unit 17 to the receiving terminal 30 owned by the employee of the construction company (S05). The above series of processes (S02 to S05) are repeated. For example, disaster information (e-mail notifying) is transmitted to the receiving terminal 30 every day. The above is the process executed by the local KY support system 10 according to the present embodiment.

In the above-described embodiment, among the related words specified by machine learning from the learning information stored in the learning database 12, the related words of the input keyword are used as the related keywords in the search for disaster information. Therefore, even if the input keyword itself is not included in the disaster information, appropriate disaster information is extracted by the search. Thereby, according to the embodiment, it is possible to provide appropriate disaster information for the input.

When it comes to a large construction site, the number of employees of the construction company and the foreman of the partner company may be several hundred, and some of them have not actually experienced past disasters, and past disaster cases are efficient. And some people cannot search properly. In addition, as the number of employees of construction companies increases, there is a limit to the transfer of knowledge and know-how from veteran employees. In such a situation, it is difficult to disseminate and guide appropriate disaster cases and perform KY appropriately. On the other hand, in the present embodiment, since appropriate disaster information is distributed to the employees of the construction company, it is appropriate and efficient by informing and instructing the workers of the subcontractor who performs the work at the construction site. Becomes possible.

Further, as in the above-described embodiment, the information related to the weather may be input and used for the search. According to this configuration, it is possible to provide appropriate disaster information according to the weather at a construction site or the like. However, the information related to the weather does not necessarily have to be used, and the information related to the weather may not be used.

Further, as in the present embodiment, a daily setting keyword may be input and used for the search. According to this configuration, even if there is no change in the information regarding the work scheduled for the next day entered by the foreman, the disaster information searched every day can be changed. That is, according to this configuration, it is possible to prevent the search result from becoming a rut. The setting keywords do not necessarily have to change daily, but may change according to the time of day. Further, the setting keyword does not necessarily have to be used, and the setting keyword may not be used.

Further, as in the present embodiment, the disaster information may be transmitted to the input source of the keyword, that is, the destination according to the foreman (cooperating company) who has input the information related to the work scheduled for the next day. According to this configuration, disaster information can be output to an appropriate output destination, that is, an employee of an appropriate construction company. However, it is not always necessary to adopt such a configuration, and the disaster information may be output to a fixed output destination set in advance.

In the present embodiment, the provided safety information is assumed to be disaster information related to past disaster cases, but the safety information does not necessarily have to be disaster cases, and any information related to construction safety should be used. Information may be used. Further, the provided safety information may include, for example, moving image, audio information, and the like.

Further, in the present embodiment, disaster information is transmitted to the employees of the construction company based on the information input by the foreman of the cooperating company, but the present invention is not necessarily limited to the above background and the input information. It can be applied to any background as long as it outputs disaster information using.

10 ... Local KY support system, 11 ... Disaster case database, 12 ... Learning database, 13 ... Related word identification unit, 14 ... Search information input unit, 15 ... Related keyword identification unit, 16 ... Search unit, 17 ... Output unit , 20 ... Input terminal, 30 ... Receiving terminal.

Claims (2)

It is a safety information providing system that provides safety information related to construction safety.
A database for providing safety information to be provided and a database for providing
A learning database that is used for learning related words and stores learning information related to construction,
A related word identification means for identifying a set of related words related to each other by machine learning from the learning information stored in the learning database, and
Search information input means for entering keywords and
A related keyword specifying means for specifying a related keyword as a related word of a keyword input by the search information input means by the related word specifying means, and a related keyword specifying means.
Using the related keyword specified by the related keyword specifying means, the search means for extracting the provided safety information by searching for the safety information stored in the providing database, and the search means.
An output means that outputs the safety information extracted by the search means, and an output means.
Equipped with a,
The search information input means is set and stored in advance for each time and disaster type, and from a plurality of setting keywords corresponding to the disaster type, the setting keywords are acquired and input according to the time.
The search means is a safety information providing system that performs a search using a set keyword input by the search information input means. The search information input means inputs information related to the weather and
The safety information providing system according to claim 1, wherein the search means also performs a search using information related to the weather input by the search information input means.

Images