A method for analyzing actions that are taking place in a plant or a factory, implemented by one or more processors operatively coupled to a non-transitory computer readable storage device, on which are stored modules of instruction code that when executed cause the one or more processors to perform the steps of:
- monitoring a plurality of data streams originating from a plurality of independent data sources within the plant or factory, wherein the monitoring includes tapping onto production machines and machine sub-units;
- correlating between two or more monitored data streams relating to a specific action occurring within the plant but originating from different data sources, or being of different data types;
- and
- analyzing said data stream correlations to obtain specific parameters and characteristics that indicate the occurrence of specific actions and scenarios that have taken place in the plant;
- wherein said analyzing is performed at a production-machine level and configured to deduce information regarding operation of the production machines and machine sub-units, in respect to the specific actions and scenarios that are taking place in the plant by identifying correlations between different sensory data sources which are correlated to the specific actions and scenarios and provide indications for the functionality of the production machines and machine sub-units,
- wherein the production-machine level analysis is an inter-operability analysis comprising monitoring and analyzing operation of the production machines and machine sub-units and extracting the specific parameters and characteristics relating to mutual influence of the production machines and machine sub-units during performance of the specific actions and scenarios within the plant,
- wherein plant-level operative deductions are achieved by integrating information from a lower-level scenario analysis of production machine and machine sub-unit inter-operability to obtain comprehensive and real-time reflection of the specific actions and scenarios that are being performed in the plant,
- wherein said analyzing analyses video streams to identify human actions including at least one of:
- mouse movement, button clicks, selections, graphic user interface (GUI) elements activation, text insertion, element changes, and image changes,
- wherein said identified human actions are correlated with processes and conditions that take place in the plant, including:
- machine status, production flow state, process status, stages within a process, configuration changes, user intervention, user identification, and misuse of equipment;
- wherein said analyzing is further configured to authenticate correctness of the operation of the production machines and machine sub-units during execution of the specific actions and scenarios, based on correlating between different types of data streams of monitored machines and machine sub-units and data streams of independent indicators and sensors,
- wherein said analyzing further includes the followings steps:
- identifying an occurrence of the actions within the plant;
- identifying the scenarios that are comprised of the actions, said scenarios taking place within the plant at any hierarchical level;
- identifying status of the machines and machine sub-units within said scenarios;
- identifying influence of the operation of the production machines and machine sub-units on measurable physical, chemical or electrical conditions within the plant;
- determining correctness of configuration and operation of the production machines and machine sub-units, in respect to the actions and scenarios that are taking place in the plant;
- and
- assessing probability for a system malfunction or cyber security breach event based on the correction determination;
- wherein said analyzing further includes generating a knowledge base which contains patterns of input data streams that reflect correct functionality on a plant level, wherein the patterns of input data streams indicate correct functionality of an entire assembly line and a specific machine sub-unit, wherein attribution of correct functionality to a specific pattern is obtained through algorithms of machine-learning applied at the specific machine sub-unit and process level;
- and
- wherein deviation of the input data pattern from an expected correct functionality pattern produces an alert.