Progress management device and progress management method

Abstract

To improve efficiency of work.SOLUTION: A progress management apparatus includes a processor for executing a program, and a storge device storing the program. The progress management apparatus can access work plan data indicating plans for multiple time-series tasks that are performed by a group including a plurality of workers. The processor executes: a specifying process to specify progress of the task on the basis of voice data pertaining to the worker in the group and the time at which the group should perform the task in the work plan data; a correction process to correct the work plan data on the basis of the progress specified by the specifying process; and an output process to output a result of the correction process.SELECTED DRAWING: Figure 1

Claims

1. A progress management device having a processor for executing a program and a storage device for storing the program, It is possible to access a work plan database that stores work plan data showing the time-series plan of multiple tasks performed by a group including multiple workers, a conversation recognition database containing terminology related to the tasks, and a machine learning model that has been trained to output the tasks when two consecutive audio paragraphs are input. The aforementioned processor, A process for acquiring voice data relating to the worker within the group, A speech recognition process that recognizes the audio data acquired by the aforementioned acquisition process and outputs a group of audio paragraphs, A first identification process involves referring to the conversation recognition database to determine whether each audio paragraph in the group of audio paragraphs processed by the speech recognition process contains terms related to the work, and if two consecutive audio paragraphs with different workers in the group both contain terms related to the work, then identifying the two consecutive audio paragraphs with different workers as a pattern in which a specific conversation using terms related to the work is formed. A second identification process identifies the current work being performed by the group based on the output from the machine learning model, which is the result of inputting two different consecutive audio paragraphs into the machine learning model, provided that the worker whose work corresponds to a pattern in which a specific conversation is established by the first identification process is a worker. If the current task identified by the second identification process differs from the task that the group should perform in the work plan data, a correction process is performed to adjust the time for subsequent tasks after the current task in the work plan data. An output process that outputs the correction result obtained by the above correction process, A progress management device characterized by performing the following actions.

2. A progress management device according to Claim 1, The equipment used in the aforementioned work has access to a noise acoustic database that stores acoustic patterns that identify the acoustics of the noise emitted in its operation steps as combinations of the equipment and the operation steps. The aforementioned work plan database stores work plan data that shows the time-series plans for multiple tasks performed by the group using the equipment. The machine learning model is further trained to output the work and the equipment when the acoustic pattern is input. In the acquisition process described above, the processor acquires acoustic data relating to the equipment, In the second identification process, the processor refers to the noise acoustic database to identify acoustic patterns corresponding to the acoustic data acquired by the acquisition process and inputs them into the machine learning model to identify the current work and equipment being used by the group. A progress management device characterized by the following features.

3. A progress management device according to Claim 1, For each of the multiple speech sentences, it is possible to access a prosodic pattern database that stores multiple speech durations of the speech and multiple prosodic patterns that identify the prosodicity when the speech was uttered. The aforementioned processor, Referencing the aforementioned prosodic pattern database, the system identifies the sentence corresponding to the speech paragraph recognized from the speech data, sets a first score based on the speech duration of the identified sentence and the speech duration of the speech data, sets a second score for the prosodic pattern of the identified sentence that corresponds to the prosodic pattern of the speech data, and performs a calculation process to calculate the urgency of the worker who spoke the speech data based on the first and second scores. In the correction process, the processor corrects the time in the work plan data for performing the subsequent work according to the urgency calculated by the calculation process. A progress management device characterized by the following features.

4. A progress management device according to Claim 1, In the first identification process, the processor refers to the conversation recognition database and identifies that if one worker's voice paragraph contains question terms signifying a question to the other worker, and the other worker's voice paragraph contains response terms signifying a response to the question, then two consecutive voice paragraphs from different workers constitute a pattern in which a specific conversation using terms related to the work is formed. A progress management device characterized by the following features.

5. A progress management device according to Claim 1, In the first identification process, the processor refers to the conversation recognition database and identifies that if one worker's voice paragraph contains question terms meaning a report to the other worker, and the other worker's voice paragraph contains response terms meaning a response to the report, then two different consecutive voice paragraphs from the worker constitute a pattern in which a specific conversation using terms related to the work is formed. A progress management device characterized by the following features.

6. A progress management device according to claim 1, In the first identification process, the processor refers to the conversation recognition database and identifies that if one worker's voice paragraph contains question terms meaning a question to the other worker, and the other worker's voice paragraph contains command terms meaning a command in response to the question, then two consecutive voice paragraphs from different workers constitute a pattern in which a specific conversation using terms related to the work is formed. A progress management device characterized by the following features.

7. A progress management device according to Claim 1, In the first identification process, the processor refers to the conversation recognition database and identifies that if one worker's voice paragraph contains command terms meaning instructions to the other worker, and the other worker's voice paragraph contains response terms meaning responses to the instructions, then two consecutive voice paragraphs from the workers that are different constitute a pattern in which a specific conversation using terms related to the work is formed. A progress management device characterized by the following features.

8. A progress management method performed by a progress management device having a processor for executing a program and a storage device for storing the program, The progress management device has access to a work plan database that stores work plan data showing a time-series plan for multiple tasks performed by a group including multiple workers, a conversation recognition database containing terms related to the tasks, and a machine learning model trained to output the tasks when two consecutive audio paragraphs are input. The aforementioned processor, A process for acquiring voice data relating to the worker within the group, A speech recognition process that recognizes the audio data acquired by the aforementioned acquisition process and outputs a group of audio paragraphs, A first identification process involves referring to the conversation recognition database to determine whether each audio paragraph in the group of audio paragraphs processed by the speech recognition process contains terms related to the work, and if two consecutive audio paragraphs with different workers in the group both contain terms related to the work, then identifying the two consecutive audio paragraphs with different workers as a pattern in which a specific conversation using terms related to the work is formed. A second identification process identifies the current work being performed by the group based on the output from the machine learning model, which is the result of inputting two different consecutive audio paragraphs into the machine learning model, provided that the worker whose work corresponds to a pattern in which a specific conversation is established by the first identification process is a worker. If the current task identified by the second identification process differs from the task that the group should perform in the work plan data, a correction process is performed to adjust the time for subsequent tasks after the current task in the work plan data. An output process that outputs the correction result obtained by the above correction process, A progress management method characterized by performing the following actions.

Images