System and method for analyzing periodic task

1. A periodic task analysis system, wherein the periodic task analysis system analyzes a task conducted periodically by an actor, comprising:

a data collecting subsystem, wherein the data collecting subsystem comprises at least one of an imaging device configured to image a state where a plurality of fundamental work operations are performed periodically, and at least one sensing means, wherein the at least one sensing means collects data from the state;

a data receiving subsystem, wherein the data receiving subsystem is operatively coupled to the data collecting subsystem, and the data receiving subsystem is configured to receive at least one of a video file and sensor data, wherein the video file is recorded by the imaging device, and the sensor data is recorded by the at least one sensing means;

a task learning subsystem, wherein the task learning subsystem is operatively coupled to the data receiving subsystem, and the task learning subsystem is configured to identify at least one sub-task from the at least one of the video file and the sensor data, wherein the at least one sub- task follows each other periodically in a loop; wherein the task learning subsystem is configured to generate labeled data by using at least one of frames of the video file and the sensor data, wherein the labeled data indicates a completion percentage of a related sub-task,

a task analyzing subsystem, wherein the task analyzing subsystem is operatively coupled to the data collecting subsystem and the task learning subsystem, and the task analyzing subsystem is configured to

receive a predetermined number of data list, wherein the predetermined number of data list is at least one of a predetermined number of frames of a live video streaming obtained by the imaging device and a predetermined number of sensor data of live streaming sensor data obtained by the at least one sensing means, to evaluate real time operations,

and the task analyzing subsystem is configured to generate a progress position, wherein the progress position indicates the related sub-task, wherein the progress position is generated as a regression output and the related sub-task is being performed by the actor in real time, and wherein the progress position indicates, during performance of the related sub-task, further the completion percentage of the related sub-task based on a continuous evaluation of the predetermined number of data list and the labeled data.

2. The periodic task analysis system according to claim 1 , wherein the task learning subsystem comprises an input unit, wherein the input unit is configured to enable an instructor to determine the labeled data, wherein the labeled data corresponds to at least one of the related frames of the video file and related sensor data of the sensor data.

3. The periodic task analysis system according to claim 1 , wherein the task learning subsystem comprises a convolutional neural network (CNN) and a transformer, wherein the transformer is coupled to the CNN.

4. The periodic task analysis system according to claim 1 , wherein the task analyzing subsystem comprises a convolutional neural network (CNN) and a transformer, wherein the transformer is coupled to the CNN.

5. The periodic task analysis system according to claim 1 , wherein the task analyzing subsystem is configured to provide guidance feedback data in real-time to at least one actor by using at least one type of alerts.

6. The periodic task analysis system according to claim 5 , wherein the task analyzing subsystem is configured to provide an alert to the actor when the task analyzing subsystem determines that the at least one sub-task is performed in an incorrect sequence.

7. A method for analyzing a task conducted by an actor periodically comprising the steps of:

imaging a state where a plurality of fundamental work operations are performed periodically by an imaging device or collecting data from the state by at least one sensing means, by a data collecting subsystem;

receiving a video file or sensor data, wherein the video file is recorded by the imaging device, and the sensor data is recorded by the at least one sensing means, by a data receiving subsystem;

identifying at least one sub-task from the video file or the sensor data by a task learning subsystem, wherein the at least one sub-task follows each other periodically in a loop;

generating labeled data by using at least one of frames of the video file or the sensor data, by the task learning subsystem, wherein the labeled data indicates a completion percentage of a related sub-task;

receiving a predetermined number of data set, wherein the predetermined number of data set is a predetermined number of frames of a live video streaming obtained by the imaging device or a predetermined number of sensor data of live streaming sensor data obtained by the at least one sensing means, to evaluate real time operations, by a task analyzing subsystem,

generating a progress position, wherein the progress position indicates the related sub-task, wherein the progress position is generated as a regression output and the related sub-task is being performed by the actor in real time, and wherein the progress position indicates, during performance of the related sub-task, further the completion percentage of the related sub-task based on a continuous evaluation of the predetermined number of data set and the labeled data, by the task analyzing subsystem.

8. The method according to claim 7 , comprising further the step of:

providing guidance feedback data in real-time to at least one actor by using at least one type of alerts, by the task analyzing subsystem.

9. The method according to claim 7 , comprising further the step of:

enabling an instructor to determine the labeled data by using an input unit of the task learning subsystem.