Robot system assisting work of worker, control method, machine learning apparatus, and machine learning method
A robot system assisting a worker to allow the worker to perform a work more smoothly. The robot system includes a robot, a detection device configured to detect motion of a worker when the worker is performing a predetermined work, an end determination section configured to determine whether the work is ended or not on the basis of detection data from the detection device, and a robot controller configured to cause the robot to carry out an article-feed operation or an article-fetch operation when the end determination section determines that the work is ended, the article-feed operation transporting an article for the work to a predetermined position to feed the article to the worker, the article-fetch operation fetching the article used for the work and transporting the article to a predetermined storage location.
1. A robot system configured to assist a work by a worker, the robot system comprising:
a robot;
a detection device configured to detect a motion of a portion of the worker's body continuously in time-series by motion capturing, in response to the worker performing a predetermined work onto a plurality of working locations in a predetermined order;
a processor configured to:
calculate a distance between the portion of the worker's body and a last working location of the plurality of working locations, which is to be worked last, based on detection data of the detection device;
determine that the work has ended when the distance is equal to or smaller than a predetermined threshold; and
in response to the processor determining that the work has ended, cause the robot to perform:
an article-feed operation to transport an article for the work to a predetermined position in order to feed the article to the worker; or
an article-fetch operation to fetch an article which has been used for the work and transport the article to a predetermined storage location.
2. The robot system of claim 1 , wherein the processor is configured to:
monitor whether or not the motion detected by the detection device matches a reference motion pattern predetermined as a reference motion of the worker in response to the worker performing the work; and
determine that the work has ended in response to the motion matching the reference motion pattern.
3. The robot system of claim 1 , wherein the processor is configured to determine whether or not the work has ended, further based on a learning model representing a correlation between the motion and a stage of the work or a time needed for the work.
4. The robot system of claim 1 , wherein, after performing the article-feed operation, the processor is configured to cause the robot to perform the article-fetch operation to fetch the article used for the most-recently performed work and transport the article to the predetermined storage location.
5. The robot system of claim 1 , wherein, in response to the processor determining that the work has ended, the processor is configured to cause the robot to perform the article-feed operation.
6. The robot system of claim 1 , wherein, in response to the processor determining that the work has ended, the processor is configured to cause the robot to perform the article-fetch operation.
7. The robot system of claim 1 , wherein the processor is configured to monitor whether or not the motion detected by the detection device matches a reference motion pattern predetermined as a reference motion of the worker when the worker performs the work onto the last working location.
8. The robot system of claim 7 , wherein the processor is configured to determine that the work is ended when the motion matches the reference motion pattern.
9. A method of controlling a robot configured to assist a work by a worker, the method comprising:
detecting, by a detection device, a motion of a portion of the worker's body continuously in time-series by motion capturing, in response to the worker performing a predetermined work onto a plurality of working locations in a predetermined order;
calculating a distance between the portion of the worker's body and a last working location of the plurality of working locations, which is to be worked last, based on detection data of the detection device;
determining that the work has ended when the distance is equal to or smaller than a predetermined threshold; and
in response to the determination that the work has ended, causing the robot to perform:
an article-feed operation to transport an article for the work to a predetermined position in order to feed the article to the worker; or
an article-fetch operation to fetch an article which has been used for the work and transport the article to a predetermined storage location.
10. A machine learning apparatus configured to learn a timing of when a work by a worker has ended, the worker performing the work onto a plurality of working locations in a predetermined order, the machine learning apparatus comprising a processor configured to:
acquire, as a learning data set,
detection data of a detection device detecting a motion of the worker during performing the work onto at least one of the plurality of working locations, and
from a time point when the work onto the at least one of the plurality of working locations is ended to a time point when the work onto a last one of the plurality of working locations, which is to be worked last, is ended; and
generate a learning model representing a correlation between the motion and the time, using the learning data set, wherein
the learning model is configured to receive an input of the detection data, and output the time which corresponds to the motion indicated by the input detection data, in order to determine the timing of when the work has ended, and
in response to the determination that the work has ended, a robot is caused to perform:
an article-feed operation to transport an article for the work to a predetermined position in order to feed the article to the worker; or
an article-fetch operation to fetch an article which has been used for the work and transport the article to a predetermined storage location.
11. A machine learning method of learning a timing of when a work by a worker has ended, the worker performing the work onto a plurality of working locations in a predetermined order, the machine learning method comprising:
acquiring, as a learning data set,
detection data of a detection device detecting a motion of the worker during performing the work onto at least one of the plurality of working locations, and
a time from a time point when the work onto the at least one of the plurality of working locations is ended to a time point when the work onto a last one of the plurality of working locations, which is to be worked last, is ended; and
generating a learning model representing a correlation between the motion and the time, using the learning data set, wherein
the learning model is configured to receive an input of the detection data, and output the time which corresponds to the motion indicated by the input detection data, in order to determine the timing of when the work has ended, and
in response to the determination that the work has ended, a robot is caused to perform:
an article-feed operation to transport an article for the work to a predetermined position in order to feed the article to the worker; or
an article-fetch operation to fetch an article which has been used for the work and transport the article to a predetermined storage location.