| CPC G06Q 10/06398 (2013.01) | 20 Claims |
|
1. A computer-implemented method, comprising:
receiving a set of raw worker data records associated with one or more worker activities of a specific worker having an identity, each respective raw worker data record comprising a computer-readable record comprising first sensor information collected from a first sensor and second sensor information collected from a second sensor, wherein:
the first sensor information comprises information collected from a first sensor, the first sensor information comprising information configured to individually identify the specific worker, wherein the first sensor information comprises information from at least one of: a biometric sensor worn by the specific worker and a biometric sensor with which the specific worker interacts; and
the second sensor information comprises information collected from a second sensor while the specific worker is performing the one or more worker activities, the second sensor comprising at least one of a wearable sensor, a sensor coupled to an entity, a sensor embedded in an entity, and an image sensor, wherein the second sensor information comprises one or more of an image, a video, information indicative of a behavior of the specific worker, information indicative of a position of at least one of the specific worker and the entity, information indicative of a movement of at least one of the specific worker and the entity, and information indicative of an interaction between the specific worker and the entity;
performing an identification analysis on the set of raw worker data records to determine, based on the biometric information, the identity of the specific worker;
persisting, based on the identity of the specific worker, the set of raw worker data records in an encrypted, computer-readable format in a respective personal knowledge repository configured to continually accumulate information about the specific worker;
performing a first analysis on the set of raw worker data records, the first analysis configured to analyze the set of raw worker data records, including second sensor information collected in real-time, as well as the information about the specific worker stored in the respective personal knowledge repository and a training data set, to train a convolution neural network (CNN) to learn one or more behaviors of the specific worker, wherein the training of the CNN is configured to build a dynamic CNN model having expertise that is specific to the specific worker;
performing a second analysis on the set of raw worker data records, wherein the second analysis is configured to perform human behavior recognition (HBR) to classify the set of raw worker data records, including the second sensor information as it is collected in real-time, into one or more types of behaviors;
performing, based on the first analysis and the second analysis, a third analysis, the third analysis configured to dynamically analyze the one or more types of behaviors for one or more undesired conditions;
generating in real-time, based on the dynamic CNN model and at least one of the second analysis and the third analysis, a first output comprising a summary of the one or more worker activities of the specific worker, the summary comprising information indicating whether any of the one or more worker activities comprises the one or more undesired conditions;
persisting the first output, along with results from at least one of the first analysis, the second analysis, and the third analysis, in the respective personal knowledge repository;
providing information from the respective personal knowledge repository to the training data set; and
training the CNN to update the dynamic CNN model in accordance with at least one of the first analysis, the second analysis, and the third analysis.
|