Vehicle inspection and maintenance record information creation system and vehicle inspection and maintenance information creation program

Abstract

This technology provides an improved efficiency and accuracy in creating vehicle inspection and maintenance records. [Solution] A vehicle inspection and maintenance record information creation system comprising a maintenance worker terminal and a system computer 3, comprising: a maintenance item information presentation unit 6 that presents maintenance item information to the maintenance worker; a voice response system unit 8 that processes the voice of the maintenance worker's response regarding the maintenance items; a maintenance item information storage unit that stores maintenance items related to the vehicle; a maintenance worker information storage unit that stores maintenance worker information related to the maintenance worker; a maintenance term grouping information storage unit that stores maintenance term grouping information which is a grouping of multiple maintenance term pronunciation patterns pronounced for maintenance terms related to the maintenance item information; and a voice recognition unit that performs voice recognition based on the maintenance term grouping information.

Description

Technical Field

[0004] , , , , , , , , ,

[0005] , , , ,

[0001] The present invention relates to a technique for creating vehicle inspection and maintenance record information such as a vehicle inspection and maintenance record book.

Background Art

[0006] The technical terms that appear in vehicle inspection and maintenance records vary depending on the region of the repair shop in Japan, the race of the workers, and the customs of the repair shop itself. Therefore, simply creating a unified vehicle inspection standard does not guarantee that mechanics will adhere to it, and the process is ineffective. In particular, when there are many foreign mechanics, language communication is difficult, and in repair shops where there is a lot of noise and multiple languages ​​are spoken, creating vehicle inspection and maintenance records is currently a time-consuming process.

[0007] From the perspective of traceability of maintenance work, the conventional manual method of creating vehicle inspection and maintenance records, which is prone to insufficient fact-checking and fraudulent entries, was inadequate in terms of ensuring maintenance quality.

[0008] This invention was made to solve the problems of the prior art described above. While this invention aims to solve problems from various perspectives, some representative examples are as follows. (1) To provide technology that can improve the efficiency and accuracy of creating vehicle inspection and maintenance records. (2) When creating vehicle inspection and maintenance records, we will provide practical technology that can utilize speech recognition by resolving the difficulties of speech recognition due to specialized terminology and the languages ​​of foreign workers. (3) To provide technology that can ensure the traceability of maintenance work by improving conventional record-keeping procedures, which are prone to fraudulent entries in fact-checking. [Means for solving the problem]

[0009] The vehicle inspection and maintenance record information creation system according to the first aspect of the present invention is A vehicle inspection and maintenance record information creation system comprising a maintenance worker terminal and a system computer, A maintenance item information display unit that presents maintenance item information to the maintenance worker, A voice response system unit that processes the voice of the maintenance worker's response regarding the maintenance items, A maintenance item information storage unit that stores maintenance items related to the vehicle, A maintenance information storage unit that stores maintenance information relating to the aforementioned maintenance personnel, A maintenance term grouping information storage unit stores maintenance term grouping information, which is a grouping of multiple maintenance term pronunciation patterns pronounced for maintenance terms related to the maintenance item information. A speech recognition unit that performs speech recognition based on the aforementioned maintenance term grouping information, It is characterized by having the following:

[0010] A second aspect of the present invention is a vehicle inspection and maintenance record information creation system as described in the first aspect, The system has a response word grouping information storage unit that stores response word grouping information, which is a grouping of multiple response word pronunciation patterns pronounced in response to the response words of the mechanic related to the maintenance item information. The voice recognition unit is characterized in that it performs voice recognition based on the grouped response word information.

[0011] A third aspect of the present invention is a vehicle inspection and maintenance record information creation system as described in the first aspect, The maintenance item information presentation unit is characterized by having a display screen response system unit that displays the maintenance item information and allows the maintenance worker to input the maintenance inspection results, and a system unit linkage unit that links the response status of the voice response system unit to the display screen response system unit.

[0012] A fourth aspect of the present invention is a vehicle inspection and maintenance record information creation system as described in the first aspect, The maintenance item information display unit is comprised solely of the voice response system unit, and the voice response system unit is configured to play voice guidance regarding the maintenance item information, to which the maintenance worker responds with the voice response.

[0013] A fifth aspect of the present invention is a vehicle inspection and maintenance record information creation system as described in the second aspect, The maintenance worker information is characterized in that it includes language information spoken by the maintenance worker, and the maintenance term grouping information and the response word grouping information are created based on the language information.

[0014] A sixth aspect of the present invention is a vehicle inspection and maintenance record information creation system as described in the second aspect, The maintenance personnel information includes the maintenance personnel's native language information, and the maintenance term grouping information and the response word grouping information are created based on the native language information.

[0015] A seventh aspect of the present invention is a vehicle inspection and maintenance record information creation system as described in the first aspect, The maintenance item information is divided into stages according to the magnitude of the concept, and the maintenance term grouping information is created for each of the lowest-level maintenance item information categories.

[0016] An eighth aspect of the present invention is a vehicle inspection and maintenance record information creation system as described in the first aspect, The maintenance item information is divided into stages according to the magnitude of the concept, and the maintenance term grouping information is created for each of the highest-level maintenance item information categories.

[0017] A ninth aspect of the present invention is a vehicle inspection and maintenance record information creation system as described in the second aspect, The maintenance terminal is characterized in that it has a motion detection unit that detects movements corresponding to the maintenance worker's response words, and a motion recognition unit that recognizes the response words based on the movements of the motion detection unit.

[0018] A tenth aspect of the present invention is the vehicle inspection and maintenance record information creation system according to the first aspect, which has a maintenance tracer unit that traces the maintenance work performed by the maintainer, wherein the maintenance tracer unit includes a maintenance-related information receiving unit that receives the obtained maintenance-related information, a serialization unit that converts the maintenance-related information into serialized information, and a serialized information storage unit that stores the serialized information, and is characterized by this.

[0019] An eleventh aspect of the present invention is the vehicle inspection and maintenance record information creation system according to the tenth aspect, wherein the maintenance-related information includes the maintainer information, the maintenance item information, and time information, and further includes at least one of the on-site location information of the maintainer, the tool and equipment information used, and the speech information, and is characterized by this.

[0020] A twelfth aspect of the present invention is the vehicle inspection and maintenance record information creation system according to the tenth aspect, wherein the maintenance tracer unit has a maintenance consistency determination unit that determines the consistency between each piece of information in the maintenance-related information, and is characterized by this.

[0021] A thirteenth aspect of the present invention is the vehicle inspection and maintenance record information creation system according to the second aspect, which has a sequential learning processing unit that improves the accuracy of speech recognition of the maintenance term grouping information and the response word grouping information based on the teaching information input by the maintainer obtained from the maintainer terminal, and is characterized by this.

[0022] A fourteenth aspect of the present invention is the vehicle inspection and maintenance record information creation system according to the second aspect, which has an automatic translation unit that translates the conversation between the maintainer and the maintainer or the maintainer and another worker into a predetermined language based on the maintenance term grouping information and the response word grouping information, and is characterized by this.

[0023] The invention according to the 15th aspect of the present invention is a program for creating vehicle inspection and maintenance record information that causes the computer of the maintenance terminal and the system computer to execute the vehicle inspection and maintenance record information creation system described in any one of the 1st to 14 aspects. [Effects of the Invention]

[0024] This invention can improve the efficiency and accuracy of creating vehicle inspection and maintenance records. It provides a practical technology that utilizes speech recognition, overcoming the difficulties of recognizing technical terms and the languages ​​of foreign workers when creating vehicle inspection and maintenance records. Furthermore, it provides a technology that improves upon conventional record-keeping processes, which are prone to inaccurate entries, and ensures the traceability of maintenance work. [Brief explanation of the drawing]

[0025] [Figure 1] This is a schematic block diagram showing one configuration of the vehicle inspection and maintenance record information creation system according to this embodiment. [Figure 2] This is a block diagram showing an example of a system computer configuration. [Figure 3] This figure shows an example of a screen for entering registration items for maintenance item information. [Figure 4] This is an explanatory diagram illustrating an example of how to group together the term "axle load," a technical term that is an example of maintenance terminology. [Figure 5] This figure shows an example of a teaching learning window used for initial grouping learning. [Figure 6] Figures (a) and (b) show examples of cases where maintenance results are provided by operating a display screen response terminal. [Figure 7] This is a block diagram showing an example of the configuration of the maintenance tracer unit. [Figure 8] This flowchart shows an example of processing at a maintenance terminal. [Figure 9] This diagram illustrates an example of the vehicle inspection procedure according to this embodiment. [Figure 10]This diagram shows the maintenance item information displayed on a smartphone screen. [Figure 11] This figure shows the display screen of a smartphone after a response has been submitted via voice or manual input. [Figure 12] This flowchart shows an example of the processes performed by the system computer. [Figure 13] This block diagram shows the components involved in the process from group learning processing and sequential learning processing by each maintenance worker's terminal to recording vehicle inspection and maintenance record information using speech recognition. [Figure 14] This diagram shows some of the inspection items for a vehicle inspection. [Modes for carrying out the invention]

[0026] Hereinafter, with reference to the drawings, a vehicle inspection and maintenance record information creation system, which is one embodiment of the present invention, will be described. Figure 1 is a schematic block diagram showing one example configuration of a vehicle inspection and maintenance record information creation system. In Figure 1, the system 1 comprises a mechanic's terminal 2 held by the mechanic performing the vehicle inspection, a system computer 3, and a maintenance tracer unit 4. The mechanic's terminal 2, the system computer 3, and the maintenance tracer unit 4 are each connected to a network 5 that transmits and receives voice, information, etc., within the repair shop. Network 5 can be either an internal network within the repair shop or a wide-area network such as the internet. The maintenance worker terminal 2 has a maintenance item information display unit 6. The maintenance item information display unit 6 has at least a voice response system unit 8. In a typical configuration, the maintenance worker terminal 2 has a voice response system unit 8, a display screen response system unit 9, a system unit linkage unit 10, and an operation detection unit 11.

[0027] The voice response system unit 8 is necessary for the mechanic to respond to maintenance inspection results by voice during maintenance, and is to be worn at all times during maintenance. Typically, the voice response system unit 8 has earphones (headphones) and a microphone. The voice response system unit 8 is also connected to the network 5 and has the function of responding by voice to an administrator terminal (not shown) to which the administrator sends instructions, etc. to the mechanic.

[0028] The display screen response system unit 9 has a display device that plays voice guidance and displays maintenance item information on the display screen, and is configured to allow users to input answers regarding maintenance by touching the display screen or performing other inputs. The display screen response system unit 9 can be configured by installing the necessary programs on a display screen response terminal 12, such as a smartphone, which can be carried by the maintenance worker as they move around. However, this does not preclude the configuration of a new terminal specifically for maintenance work as the display screen response terminal 12.

[0029] Typically, each system unit linkage unit 10 consists of a program that links with maintenance progress information, such as the status of responses from the voice response system unit 8 and the display screen response system unit 9, including the input of maintenance personnel and the input of maintenance results. The results of the maintenance personnel's voice responses are fed back and updated in real time on the display screen information of the display screen response system unit 9.

[0030] The motion detection unit 11 is a functional unit that detects the operation of the voice response system unit 8 or the display screen response terminal 12. It also consists of a sensor that detects the position or movement of the mechanic and a recognition program unit. The motion detection unit 11 provided in the mechanic terminal 2 may work in conjunction with a camera and its recognition system that can detect the position and movement of mechanics in a repair shop or similar facility to detect the mechanic's position and movement in detail.

[0031] Figure 2 is a block diagram showing an example of the configuration of system computer 3. The system computer 3 refers to the computer that executes the system, and various information processing resources such as servers, clients, and storage devices are distributed across the network 5. The computer that controls and manages the maintenance terminal 2 is also included in the system, and the system computer 3 that constitutes this system has a maintenance information storage unit 26, a maintenance item information storage unit 27, a maintenance term grouping information storage unit 28, a response word grouping information storage unit 29, an initial grouping learning unit 30, a sequential learning processing unit 31, and a speech recognition unit 32. The system computer 3 also has a maintenance inspection result information writing unit 33, a vehicle inspection maintenance record information storage unit 34, an operation recognition unit 35, an automatic translation unit 36, and a receiving unit 25.

[0032] The receiving unit 25 transmits and receives data between the service technician terminal 2 and the service tracer unit 4 via the network 5. The maintenance personnel information storage unit 26 stores maintenance personnel information, which includes not only general information such as the employee number and name of the maintenance personnel, but also nationality, native language information of the maintenance personnel, language information of the maintenance personnel, and years of maintenance experience. The maintenance item information stored in the maintenance item information storage unit 27 is information that stores the items to be inspected and maintained during a vehicle inspection. The maintenance item information is a unit of major maintenance work, and it stores the contents of the vehicle inspection items down to the higher-level items related to higher-level concepts, intermediate items related to intermediate concepts, and lower-level items related to lower-level concepts.

[0033] Figure 3 shows an example of a screen for registering and setting maintenance items in the system. On this screen, engine oil, tires, and notes are registered as higher-level maintenance item information. Gauge is registered as an intermediate item for engine oil, and appearance as an intermediate item for tires. In addition, boundary lines and oil quality are registered as lower-level items for engine oil, and slit value is registered as a lower-level item for tires. In this way, maintenance item information is divided into stages according to the size of the concept. In addition to these, there are numerous other maintenance items, and the system stores information on each maintenance item related to the vehicle inspection and maintenance record book as stipulated by law. It is also possible to add unique vehicle inspection and maintenance items specific to each repair shop as needed. Furthermore, it is possible for each repair shop to create and use its own company-specific vehicle inspection and maintenance record book for internal use.

[0034] The maintenance item information includes defined data types, such as integer data, real number data, Yes / No boolean values, decimal data, and string data. Furthermore, a descriptive text (500 characters or less) describing the inspection performed for each maintenance item is stored, allowing the maintenance worker to retrieve and read it as needed, or for the system to automatically read it aloud. The font size setting determines the size of the characters displayed on the display screen of the display screen response system unit 9.

[0035] The maintenance term grouping information stored in the maintenance term grouping information storage unit 28 is a characteristic component of this embodiment. Maintenance term grouping information is information that groups together multiple pronunciation patterns of maintenance terms related to maintenance item information for each maintenance term.

[0036] This information helps us recognize "disk" and "disk" as a single group, even when different mechanics use different terms to express maintenance terminology. For example, some mechanics might say "disk" as "circular disc." It also helps us group specific technical terms in case of variations in expression for the same object or state, or when pronunciation changes depending on the speaker's race, native language, or the language they speak.

[0037] On the other hand, the response word grouping information in the response word grouping information storage unit 29 registers a wide range of utterances such as "okke," "okay," "kay," "maru," "mondainashi," and "kon (Vietnamese)" as groups. In this case as well, multiple response word pronunciation patterns are stored as response word grouping information, grouped together for that response word. The system can register responses for all items in the maintenance terminology information. Furthermore, it can be used not only for voice recognition related to maintenance items, but also for voice recognition of conversations between other mechanics, workers, or system administrators.

[0038] In a repair shop where Japanese is the standard language, if a Chinese person whose native language is Chinese speaks Japanese, their pronunciation will differ from that of a native Japanese speaker, and using a speech recognition system trained on Japanese speakers will reduce the accuracy of speech recognition. Similarly, if a Chinese person whose native language is Chinese speaks English in a repair shop where English is the standard language, their English is expected to differ from that of a native Japanese speaker. In particular, when a foreigner speaks Japanese using specialized terminology related to repairs, such specialized terminology is not registered, and the pronunciation and intonation will differ from that of speech recognition systems trained for general use, making accurate speech recognition even more difficult.

[0039] Figure 4 is a schematic diagram illustrating an example of grouping words with the aforementioned pronunciation changes. While native Japanese speakers can accurately pronounce the technical term "axle load" (jikujū) as "jikujū," a native Korean speaker might pronounce it differently, with some sounds and intonation altered, such as "[shi]kshū" or "shiku[shū]." This alteration in pronunciation and intonation when foreigners speak Japanese significantly reduces the accuracy of speech recognition.

[0040] Furthermore, environmental disturbances can disrupt speech recognition, resulting in incorrect speech recognition such as "Zeekjuu" or "Zeekzayuu," as shown in Figure 4. Examples of environmental disturbances include the speech and noise of other people in the repair shop. In this case, the information obtained by learning the pronunciations of these various foreigners and recording their pronunciation patterns as a group of words corresponding to "axle load" is called "maintenance terminology grouping information."

[0041] The initial grouping learning unit 30 shown in Figure 2 is a functional unit that learns maintenance term grouping information and response word grouping information as initial groupings. Figure 5 shows an example of a display screen 75 (teaching and learning window) used for initial grouping learning in such a learning process.

[0042] The display screen 75 includes a speech recognition display area 67 that displays the speech-recognized text, a settings display area 68 for setting the extraction method for various words and other elements in the speech recognition display area 67, an extracted word display area 69 that displays extracted words automatically extracted from the speech-recognized text of the teaching audio data, a registered word display area 71 that displays registered words, a maintenance item information display area 72 that displays maintenance item information for higher-level items, intermediate items, and lower-level items, and a speech learning start button 70 for performing the operation to start speech learning.

[0043] This display screen 75 is a work display screen that, when a maintenance item is selected in the maintenance item information display field 72, reads aloud a sentence containing technical terms related to the selected maintenance item, automatically performs speech recognition, and registers the recognition result that is deemed correct as the pronunciation for grouping.

[0044] The extracted phrases displayed in the extracted phrase display field 69 are basically extracted by dividing the speech-recognized text into words and performing syntactic analysis, part-of-speech analysis, proper noun selection, etc. The registered word display field 71 is a field that displays the voice-recognized words that are registered as a group. In Figure 5, an example is shown where the registered word 73, "oil," is displayed.

[0045] As shown in Figure 5, when the maintenance item "oil quantity" is selected in the maintenance item information display field 72, the Japanese pronunciation of "oil quantity" is asked to be pronounced by each foreigner, such as a Vietnamese or Korean, and the extracted phrases are displayed in the extracted phrase display field 69. If the words accurately correspond to "oil" or "oil", the word is selected in extracted phrases 1 to 3 and displayed as registered phrase 73 in the registered phrase display field 71. For example, the response sentence "The engine oil quantity is appropriate" is asked to be pronounced by a foreigner, and the extracted phrases "engine", "oil quantity", and "appropriate" are automatically extracted. For the maintenance item "oil quantity", the foreigner's maintenance term utterance pattern (utterance feature data) and speech recognition set are registered as registered phrases along with information in a predetermined language.

[0046] Here, by performing speech recognition on each of the higher-level, intermediate-level, and lower-level items, and then grouping and registering the extracted words, accurate speech recognition becomes possible even for maintenance terms that have various pronunciations. Such grouping can be done in the areas of higher-level, intermediate-level, and lower-level items of the maintenance item information. Maintenance term grouping information can also be created for each of the lowest-level maintenance item information categories. Generally, the accuracy of pronunciation recognition can be improved by registering phrases that group each sub-item of maintenance information into finer categories.

[0047] On the other hand, maintenance terminology grouping information can also be created for each of the highest-level maintenance item information categories. For example, by creating grouping information for terms to be registered for each high-level item such as brakes and engine, accurate recognition can be achieved even for voice responses that cover a wide range. Furthermore, maintenance terms such as wear and remaining disc, which are intermediate items for the high-level item "brakes," can also be grouped together and registered as a maintenance item information group for "brakes."

[0048] In the learning process, it is preferable to register maintenance terminology for each language, such as creating groups for Vietnamese and Korean maintenance terminology. Since the type of language spoken by the maintenance worker at a single maintenance terminal 2 can be identified, the system can be trained to improve speech recognition for that language. In this case, when the corresponding registered words are pronounced, accurate speech recognition can be performed for each country's pronunciation.

[0049] On the other hand, it is also possible to group the data together without separating it into categories for Japanese, Vietnamese, and Koreans. This integrated configuration is simpler from a system perspective, and even with this configuration, an improvement in speech recognition accuracy can be expected.

[0050] The initial grouping learning unit 30 has a display screen, as shown in Figure 5, that allows users to view a list of already registered words and phrases, and has a mode that allows output in the form of a text file or a CSV file. The initial grouping learning unit 30 also has a mode that allows a system administrator or maintenance worker to delete incorrectly registered words and phrases. Therefore, to enable such processing, the initial grouping learning unit 30 is equipped with a display and playback function unit that can display each registered word and play back the pronunciation data of that registered word and phrase.

[0051] It is also possible to adopt a system in which initial grouping learning unit 30 is not given initial instruction, and a speech recognition set (speech recognition model) for general conversation in Japanese or another language is used as the initial AI model, and specialized terminology and other maintenance terms are sequentially and automatically registered using sequential instruction data from the maintenance worker terminal 2, etc., as maintenance is performed. However, the accuracy of the speech recognition system will be improved if the pronunciation of important maintenance terms is learned in advance as initial instruction learning data and stored beforehand.

[0052] With this configuration, since it uses grouped maintenance term information and grouped response word information, even if the creation of vehicle inspection and maintenance records is made more efficient using speech recognition, it has the advantage of preventing the recording of incorrect maintenance inspection results due to a decrease in the accuracy of speech recognition for foreigners.

[0053] The sequential learning processing unit 31 is a functional unit that sequentially learns from the utterance information of the maintenance worker, which is provided by the maintenance worker terminal 2, and improves the accuracy of the maintenance term grouping information and the response word grouping information. This sequential learning processing unit 31 will be described later. The speech recognition unit 32 is a functional unit that performs speech recognition based on a model resulting from initial grouping learning and sequential learning.

[0054] The maintenance inspection result information writing unit 33 shown in Figure 2 consists of a program and the like that writes the maintenance inspection result information answered by the maintenance worker to the vehicle inspection maintenance record information storage unit 34, which corresponds to the vehicle inspection maintenance record book. As described above, the motion recognition unit 35 consists of recognition software that allows the system to recognize the location, movement, and operations of the mechanic in the repair shop from various terminals and cameras carried by the mechanic.

[0055] This type of gesture recognition can be adopted in multi-racial and multilingual environments to pre-determine specific gestures as inspection results in vehicle inspection and maintenance records, thereby enabling more accurate and time-efficient maintenance.

[0056] Figure 6 shows an example of a case where the maintenance results are answered by moving the smartphone, which is the display screen response terminal 12. Figure 6(a) shows that the maintenance result can be communicated as good (OK) by moving the display screen response terminal 12 in a circular motion, and Figure 6(b) shows that the maintenance result can be communicated as poor (NG) by moving the display screen response terminal 12 in a cross shape like an "X".

[0057] The automatic translation unit 36 ​​shown in Figure 2 is a functional unit that recognizes speech spoken by system administrators and maintenance personnel connected via the network 5 and translates those speeches into other languages. Even when one maintenance person speaks a different language, the system translates the spoken content by requesting translation processing from each maintenance person's terminal as needed. In this case as well, the accuracy of speech recognition for technical terms and response words is improved with each use through sequential learning of maintenance term grouping information and response word grouping information, so the translation accuracy can also be improved. With such a configuration, questions and conversations can be conducted smoothly and communication can be improved even in a maintenance factory where many languages ​​are spoken.

[0058] Figure 7 is a block diagram showing an example of the configuration of the maintenance tracer unit 4. The maintenance tracer unit 4 includes a maintenance-related information receiving unit 38 that receives the obtained maintenance-related information, a serialization unit 45 that converts the maintenance-related information into serialized information, a serialized information storage unit 46 that stores the serialized information, a maintenance consistency determination unit 47 that determines the consistency between each piece of maintenance-related information, and a control unit 48 that controls the entire unit.

[0059] The maintenance-related information receiving unit 38 is a functional unit that receives maintenance-related information detected by detectors and other devices placed at various locations. The maintenance-related information includes worker information including maintenance worker information, on-site location information, time information, tool and equipment used information, maintenance item information, and spoken information. The maintenance-related information receiving unit 38 has storage units 39 to 44 for storing each of these pieces of information. Preferably, the maintenance-related information includes at least maintenance worker information, maintenance item information, and time information, and further includes at least one of the maintenance worker's on-site location information, tool and equipment used information, and spoken information.

[0060] Time information includes the date and time on the calendar, the time interval for performing a certain maintenance task, the time interval until a voice response is received, and the time interval extracted by a predetermined program from the operation time of the display screen response system unit 9.

[0061] Speech information includes voice guidance information, instructions and warnings issued by administrators, voice responses from maintenance staff, and answers to questions. The serialization unit 45 is a functional unit that serializes each piece of maintenance-related information. The serialization method is determined by which elements of the time flow are emphasized when this system performs tracing. The serialization information storage unit 46 is a storage unit that stores such serialization information.

[0062] The maintenance consistency determination unit 47 is a functional unit that determines the consistency of each piece of maintenance-related information. For example, if a worker is performing a maintenance task, and the task can only be performed in a designated area within the maintenance factory, but a worker in another location transmits maintenance result information, the unit can determine that there is a lack of consistency. Other examples include cases where a maintenance inspection requires specific tools and equipment, but the location of those tools and equipment has not changed when the worker transmits maintenance result information, or where a specific maintenance inspection task requires a certain amount of work time, but the worker transmits maintenance result information in a very short time.

[0063] The maintenance consistency determination unit 47 consists of a program unit that determines such consistency. Furthermore, by including at least time information, and combining that time information with location and orientation information such as on-site location information and information on tools and equipment used, as well as speech information, precise tracing becomes possible.

[0064] The control unit 48 receives, takes in, and stores maintenance-related information, and may also be configured to work in conjunction with the maintenance consistency determination unit 47 and the serialization information storage unit 46 to identify inconsistent work after the recording of one vehicle inspection maintenance record book has been completed over time, and to send warning information to the administrator or maintenance worker. It is also possible to store and display maintenance workers with inconsistent records over a certain period of time. The system is configured to identify inconsistent work or poorly consistent vehicle inspection and maintenance records via the control unit 48 by inputting query information, which is predetermined search information (such as a search expression), and obtaining it as query output information.

[0065] Regarding worker location and time information, general-purpose technologies are often used to identify worker locations from smartphone location detection functions or images captured by cameras. In maintenance workshops, certain inspections can only be performed in specific locations due to equipment and machinery constraints, and such arrangements improve the accuracy of consistency determination through location detection.

[0066] Regarding the detection of the position and orientation of the tools and equipment used as described above, this information can also be obtained by attaching various tags or radio transmitters to the tools used for maintenance, thereby providing means to detect changes in tool position and the type of movement they make. Alternatively, the tool position and orientation can be recognized using multiple cameras installed in the maintenance workshop.

[0067] Furthermore, if we explain the trace function according to this embodiment in the context of a maintenance operation that involves checking the disk as described in Figure 14, the algorithm uses inferences from the fact that a maintenance worker or tool is present at the location where the maintenance result corresponding to the check mark is transmitted, and from the location where the status of the disk is recorded, thereby identifying the location not only proactively but also backward. The aforementioned location information is also used to identify the location where the information is recorded. Subsequently, the computer outputs the results for the identified location.

[0068] In this way, inspection results are linked to worker location information and tool location information, and are used as information to prevent incorrect entries when filling out the vehicle inspection and maintenance record book. Furthermore, inspection results, location information, and time information are recorded in the computer as a form of traceability. In conventional vehicle inspection and maintenance record book creation, inspection content and worker location information and time information were not linked, but the system method of this embodiment is useful in that it links information to work results, and is an important technical approach in improving the quality of maintenance work. In addition, by serializing the maintenance-related information and using it as a trace number to uniquely identify each vehicle inspection operation, tampering is made impossible.

[0069] Figure 8 is a flowchart showing an example of processing at the maintenance terminal 2. Figure 9 illustrates an example of the vehicle inspection work according to this embodiment. Referring to Figure 8, the processes performed by the maintenance terminal 2 will be explained according to the flowchart. In step S1 of Figure 8, the system authenticates the identity of the service provider using a service provider code or similar method, and authenticates the service provider who possesses a service provider terminal. Alternatively, authentication may be performed using individual terminals owned by the service provider, or a dedicated terminal may be set up for each service, and the service provider may authenticate using the dedicated terminal for each service. In step S2, the maintenance work that the mechanic intends to perform is identified.

[0070] In step S3, the service terminal 2 or the system presents the service provider with service item information. This presentation may be done solely by voice, or by displaying the service item information using the smartphone, which is the display screen response terminal 12.

[0071] If the system consists only of the voice response system unit, the configuration is such that voice guidance regarding maintenance item information is played from the earphones or headphones of the voice response system unit 8, and the maintenance worker responds with a voice response. For example, a voice guidance such as "Please inspect the degree of tire wear" is played from the earphones, the maintenance worker performs the necessary work for the inspection, and the maintenance worker responds through the microphone, for example, "The tire wear is normal (○)." Next, in step S4, the system memorizes the voice response.

[0072] For example, if the answer is "Tire wear is ○," the voice recognition unit 32 of the system computer 2 recognizes the voice. Then, each system unit coordination unit 10 coordinates to display the answer field 17 for tire wear (see Figures 10 and 11) on the display screen response terminal 12 so that it shows ○.

[0073] Next, in step S5, the system presents the recognized answer via voice or screen display obtained through speech recognition. In this case, for example, the system might play a voice message saying, "Is the tire wear correct? If so, please answer yes. Otherwise, please answer no." The mechanic would then respond.

[0074] Next, in step S6, the system determines whether there was an audio response of "yes" or whether an approval input was made from the smartphone's display screen. If the input is incorrect and the mechanic responds "no," the system returns to step S3 and again provides audio or screen guidance. On the other hand, if the mechanic responds that the system's recognition result in step S6 is correct, the system instructs the mechanic's terminal 2 to proceed to the next maintenance item in step S7.

[0075] In this series of processes, it is recommended that maintenance personnel use voice-based responses as much as possible, as this allows them to work more efficiently and free up both hands compared to constantly checking their smartphones.

[0076] However, if only voice responses are used, it can be difficult to know how far the maintenance inspection has progressed. Also, when verifying whether the answers were correct or not, it is reassuring to be able to view the maintenance results report in a list format on a smartphone screen, which improves the accuracy of maintenance verification. Therefore, it is preferable for each mechanic to carry a display device such as a smartphone that they always have with them.

[0077] Figure 10 shows maintenance item information displayed on a smartphone screen. Top-level items 13, intermediate items 14, and lower-level items 15 are displayed in columns corresponding to the item names, and various information such as importance related to those items is displayed in the points column 16. However, the various information displayed in the points column 16 can be set so that the maintenance worker cannot see it. The answer column 17 is where the answer results are displayed. The answer column 17 displays ○, ×, a blank space indicating no answer, and numerical values ​​for the maintenance results. The answer column 17 may have multiple columns that can display multiple items such as ○, × results and their numerical values.

[0078] Figure 11 shows the state after the system has recognized the response via voice or manual input from a smartphone. It shows the state where the maintenance result in the engine oil maintenance items (gauge, boundary line, oil quality) is marked with a circle (○) indicating good performance. If the system's recognition of the user's voice response is incorrect, the system is configured to allow the user to correct the result manually by touching the display screen without having to provide another voice response.

[0079] Figure 12 is a flowchart showing an example of the processing performed by system computer 3. As shown in Figure 12, in step S21, the system verifies authentication on the servicer terminal 2. In step S22, the service item information is sent to the servicer terminal 2. This transmission process does not need to be repeated if the service item information has already been downloaded to the servicer terminal 2, but if it is a new service, the service item information is sent to the servicer terminal 2. In step S23, the service tracer unit 4 starts tracing on the servicer terminal 2. In step S24, the service item is presented via voice or screen.

[0080] The mechanic then performs the maintenance work, the audio response from the mechanic is captured in step S25, the display screen response is recognized in step S26, and the actions related to the mechanic are confirmed in step S27 by the operation detection unit 11 of the mechanic terminal 2 and the operation recognition unit 35 of the system computer 2.

[0081] As mentioned above, the processes from step S25 to step S27 are performed either by voice alone or with a display screen response. Furthermore, as mentioned above, the actions of the mechanic in step S27 are recognized by the movement of the smartphone, as well as the movement of cameras and tools in the repair shop. Similarly, the system confirms the position and movement of tools from the tool and equipment information being used.

[0082] Then, in step S28, the system determines whether the mechanic has given OK to the input of the inspection result for the maintenance item (whether or not information indicating OK has reached the system). If it is determined that the mechanic has not given OK, the system prompts the response processing again from step S25 to step S27. On the other hand, if it is determined that the mechanic has given OK, the inspection result is written to the vehicle inspection maintenance record book in step S29. Then, in step S30, this determination process is used as teaching information for sequential learning.

[0083] Figure 13 is a block diagram illustrating an example of the components involved in the process from group learning processing and sequential learning processing by each maintenance terminal 2 to recording vehicle inspection and maintenance record information using speech recognition. In Figure 13, the microphone 54 of each maintenance worker terminal 2 receives the voice response for each maintenance item, and the data collection unit 55 collects the data. The data collected by the data collection unit 55 includes information such as other people's conversations and background noise, so the data collection unit 55 uses various filters to remove noise and other unwanted sounds. The data labeling unit 56 labels the voice response with maintenance worker information related to speech recognition, such as native language information and the language the maintenance worker speaks, and stores this voice response data in the voice response data storage unit 57.

[0084] Then, using the initial grouping learning unit 30, instructional learning is performed as described in Figure 5, and the initial AI model is stored in the initial AI model unit 58. Next, the sequential learning processing unit 31 performs sequential learning as shown in step S30 of Figure 12. Sequential learning is performed based on correct / incorrect judgment inputs, where the maintenance worker inputs whether the maintenance results presented by the system, obtained through response words for inspection results such as ○, ×, OK, good, intermediate, and bad, are correct or incorrect, as well as conversion processing information (which can also be called response process information) related to the conversion process.

[0085] For example, if a mechanic repeatedly pronounces something and it is ultimately not accurately recognized, and the mechanic then enters the inspection results from the smartphone screen, the system will provide instructional data indicating that the speech recognition was incorrect. Conversely, if the system determines that the mechanic correctly recognized the inspection results on the first attempt, the system will provide instructional data indicating that the pronunciation and recognition of the technical term at the corresponding position were accurate, based on syntactic analysis.

[0086] This sequential learning process improves the recognition of grouped pronunciations. This improvement process includes the registration of new pronunciations to grouped information for the Japanese word "jikujuu" (axial weight), as explained in Figure 4. In this sequential learning, for maintenance items in which the word "jikujuu" is used, the artificial intelligence automatically identifies that the pronunciation in question is a new pronunciation related to "jikujuu" based on the order and meaning of syntactic analysis in the maintenanceer's response voice and registers it sequentially.

[0087] The models sequentially learned by the sequential learning processing unit 31 are evaluated by the model evaluation unit 62, and models that meet the evaluation criteria are sequentially updated and stored as sequential learning models in the sequential learning model storage unit 63. Based on such a sequential learning model, the speech recognition unit 32 recognizes speech, and the maintenance inspection result information writing unit 33 stores the maintenance inspection result in the vehicle inspection maintenance record information stored in the vehicle inspection maintenance record information storage unit 34.

[0088] In this embodiment, the focus has been on the vehicle inspection and maintenance record information creation system. However, this technology can be installed on the computer of the maintenance worker's terminal and the system computer, and can be configured as a program for creating vehicle inspection and maintenance record information to cause the computer to execute the vehicle inspection and maintenance record information creation system described above. It can also be recognized as a method for creating vehicle inspection and maintenance record information that performs the predetermined processing described above based on the response of a maintenance worker, worker, or system administrator using the vehicle inspection and maintenance record information creation system.

[0089] Although the present invention has been described above with examples of embodiments, the technical scope of the present invention is not limited to the configurations described in the above embodiments. The technical scope of the present invention should be determined based on the description in the claims, and it goes without saying that various modifications, additions of configurations, or improvements can be made within that scope. [Explanation of symbols]

[0090] 1: Vehicle Inspection and Maintenance Record Information Creation System 2: Maintenance terminal 3: System Computer 4: Maintenance Tracer Department 5: Network 6: Maintenance Item Information Display Section 8: Voice Response System Unit 9: Display screen response system unit 10: Systems Department Liaison Department 11: Motion detection unit 12: Display screen response terminal 26: Maintenance personnel information storage unit 27: Maintenance Item Information Storage Unit 28: Maintenance Terminology Grouping Information Storage Unit 29: Answer word grouping information storage unit 30: Initial Grouping Learning Unit 31: Sequential Learning Processing Unit 32: Speech Recognition Unit 33: Maintenance inspection result information writing section 34: Vehicle inspection and maintenance record information storage unit 35: Motion recognition section 45: Serialization section 46: Serialized Information Storage Unit 47: Maintenance consistency determination unit

Claims

1. A vehicle inspection and maintenance record information creation system comprising a maintenance worker terminal and a system computer, A maintenance item information display unit that presents maintenance item information to the maintenance worker, A voice response system unit that processes the voice of the maintenance worker's response regarding the maintenance items, A maintenance item information storage unit that stores maintenance items related to the vehicle, A maintenance information storage unit that stores maintenance information relating to the aforementioned maintenance personnel, A maintenance term grouping information storage unit stores maintenance term grouping information, which is a grouping of multiple maintenance term pronunciation patterns pronounced for maintenance terms related to the maintenance item information, A speech recognition unit that performs speech recognition based on the aforementioned maintenance term grouping information, A vehicle inspection and maintenance record information creation system characterized by having the following features.

2. The system has a response word grouping information storage unit that stores response word grouping information, which is a grouping of multiple response word pronunciation patterns pronounced in response to the response words of the mechanic related to the maintenance item information. The vehicle inspection and maintenance record information creation system according to claim 1, wherein the voice recognition unit performs voice recognition based on the response word grouping information.

3. The vehicle inspection and maintenance record information creation system according to claim 1, wherein the maintenance item information presentation unit includes a display screen response system unit that displays the maintenance item information and allows the maintenance worker to input the maintenance inspection results, and a system unit linkage unit that links the response status of the voice response system unit to the display screen response system unit.

4. The vehicle inspection and maintenance record information creation system according to claim 1, wherein the maintenance item information presentation unit consists only of the voice response system unit, and the voice response system unit is configured to play voice guidance regarding the maintenance item information, and the maintenance worker responds with the voice response.

5. The vehicle inspection and maintenance record information creation system according to claim 2, wherein the maintenance worker information includes language information spoken by the maintenance worker, and the maintenance term grouping information and the response word grouping information are created based on the language information.

6. The vehicle inspection and maintenance record information creation system according to claim 2, wherein the maintenance personnel information includes the maintenance personnel's native language information, and the maintenance term grouping information and the response word grouping information are created based on the native language information.

7. The vehicle inspection and maintenance record information creation system according to claim 1, wherein the maintenance item information is divided into stages according to the size of the concept, and the maintenance term grouping information is created for each of the lowest-level maintenance item information divisions.

8. The vehicle inspection and maintenance record information creation system according to claim 1, wherein the maintenance item information is divided into stages according to the size of the concept, and the maintenance term grouping information is created for each of the highest level maintenance item information to which it is divided.

9. The vehicle inspection and maintenance record information creation system according to claim 2, wherein the maintenance terminal has a motion detection unit that detects movements corresponding to the maintenance worker's response words, and a motion recognition unit that recognizes the response words based on the movements of the motion detection unit.

10. It has a maintenance tracer unit that traces the maintenance work performed by the aforementioned maintenance worker, The vehicle inspection and maintenance record information creation system according to claim 1, wherein the maintenance tracer unit comprises a maintenance-related information receiving unit that receives obtained maintenance-related information, a serialization unit that converts the maintenance-related information into serialization information, and a serialization information storage unit that stores the serialization information.

11. The vehicle inspection and maintenance record information creation system according to claim 10, wherein the maintenance-related information includes the maintenance worker information, the maintenance item information, and time information, and further includes at least one of the maintenance worker's on-site location information, tool and equipment information used, and spoken information.

12. The vehicle inspection and maintenance record information creation system according to claim 10, wherein the maintenance tracer unit has a maintenance consistency determination unit that determines the consistency between each piece of maintenance-related information.

13. The vehicle inspection and maintenance record information creation system according to claim 2, further comprising a sequential learning processing unit that improves the accuracy of speech recognition of the maintenance term grouping information and the response word grouping information based on the teaching information input by the maintenance worker obtained from the maintenance worker terminal.

14. The vehicle inspection and maintenance record information creation system according to claim 2, further comprising an automatic translation unit that translates a conversation between a maintenance worker and another maintenance worker, or between a maintenance worker and another worker, into a predetermined language based on the maintenance term grouping information and the response word grouping information.

15. A program for creating vehicle inspection and maintenance record information to cause the computer of the maintenance worker terminal and the system computer to execute the vehicle inspection and maintenance record information creation system described in any one of claims 1 to 14.

Images