Dynamic monitoring and securing of factory processes, equipment and automated systems
A system including a deep learning processor obtains response data of at least two data types from a set of process stations performing operations as part of a manufacturing process. The system analyzes factory operation and control data to generate expected behavioral pattern data. Further, the system uses the response data to generate actual behavior pattern data for the process stations. Based on an analysis of the actual behavior pattern data in relation to the expected behavioral pattern data, the system determines whether anomalous activity has occurred as a result of the manufacturing process. If it is determined that anomalous activity has occurred, the system provides an indication of this anomalous activity.
1. A manufacturing system, comprising:
two or more process stations, wherein a first process station is logically positioned upstream of a second process station, wherein each of the first process station and the second process station is configured to perform a step of a multi-step manufacturing process;
a first station controller programmed to control a first operation of the first process station;
a second station controller programmed to control a second operation of the second process station;
a deep learning controller in communication with the two or more process stations, the first station controller, and the second station controller, wherein the deep learning controller is trained to identify anomalous activity in the multi-step manufacturing process based on response data received from the first station controller or the second station controller; and
a signal splitter positioned between the first station controller and the deep learning controller, the signal splitter having a single input, a first output, and a second output, wherein the signal splitter is configured to receive a control value signal output from the first process station via the single input, divide the control value signal output, provide a first instance of the divided control value signal output to the first station controller via the first output and, provide a second instance of the divided control value signal output to the deep learning controller via the second output.
2. The manufacturing system of claim 1 , further comprising:
a second signal splitter positioned between the second station controller and the deep learning controller, the second signal splitter having a further single input, a further first output, and a further second output, wherein the second signal splitter is configured to receive a second control value signal output from the second process station via the further single input, divide the second control value signal output, provide a first instance of the second divided control value signal output to the second station controller via the further first output, and provide a second instance of the second divided control value signal output to the deep learning controller via the further second output.
3. The manufacturing system of claim 1 , further comprising:
training the deep learning controller to identify anomalous activity in the multi-step manufacturing process based on expected behavioral patterns and actual behavior patterns.
4. The manufacturing system of claim 3 , further comprising:
generating a training data set comprising at least one set of actual behavioral patterns comprising unusual frequency patterns.
5. The manufacturing system of claim 2 , wherein the signal splitter is configured to send a first control value signal to the first station controller, wherein the first control value signal is an analog or digital signal, and wherein the second signal splitter is configured to send a second control value signal to the second station controller, wherein the second control value signal is an analog or digital signal.
6. The manufacturing system of claim 1 , wherein the deep learning controller is configured to receive the response data from the first process station and determine whether the response data comprises anomalous activity.
7. The manufacturing system of claim 6 , wherein the deep learning controller is configured to generate a signal indicating that the anomalous activity is identified.
8. A manufacturing system, comprising:
a process station configured to perform a step of a multi-step manufacturing process;
a station controller programmed to control an operation of the process station;
a deep learning controller in communication with the process station and the station controller, wherein the deep learning controller is trained to identify anomalous activity in the multi-step manufacturing process based on response data received from the station controller; and
a signal splitter positioned between the station controller and the deep learning controller, the signal splitter having a single input, a first output, and a second output, wherein the signal splitter is configured to receive a control value signal output from the process station via the single input, divide the control value signal output, provide a first instance of the divided control value signal output to the station controller via the first output, and provide a second instance of the divided control value signal output to the deep learning controller via the second output.
9. The manufacturing system of claim 8 , further comprising:
a second process station positioned downstream of the process station;
a second station controller programmed to control a second operation of the second process station; and
a second signal splitter positioned between the second station controller and the deep learning controller, the second signal splitter having a further single input, a further first output, and a further second output, wherein the second signal splitter is configured to receive a second control value signal output from the second process station via the single input, divide the second control value signal output, provide a first instance of the divided second control value signal output to the second station controller via the further first output, and provide a second instance of the divided second control value signal output to the deep learning controller via the further second output.
10. The manufacturing system of claim 8 , further comprising:
training the deep learning controller to identify anomalous activity in the multi-step manufacturing process based on expected behavioral patterns and actual behavior patterns.
11. The manufacturing system of claim 10 , further comprising:
generating a training data set comprising at least one set of actual behavioral patterns comprising unusual frequency patterns.
12. The manufacturing system of claim 9 , wherein the signal splitter is configured to send a first control value signal to the station controller, wherein the first control value signal is an analog or digital signal, and wherein the second signal splitter is configured to send a second control value signal to the second station controller, wherein the second control value signal is an analog or digital signal.
13. The manufacturing system of claim 8 , wherein the deep learning controller is configured to receive response data from the process station and determine whether the response data comprises anomalous activity.
14. The manufacturing system of claim 13 , wherein the deep learning controller is configured to generate a signal indicating that the anomalous activity is identified.
15. A computer-implemented method, comprising:
receiving, at a single input of a signal splitter, a control value signal output from a process station involved in a multi-step manufacturing process;
splitting, by the signal splitter, the control value signal output from the process station;
providing, by the signal splitter via a first output of the signal splitter, a first instance of the split control value signal output to a deep learning controller trained to identify anomalous activity in the multi-step manufacturing process based on response data;
providing, by the signal splitter via a second output of the signal splitter, a second divided control value signal output to a station controller associated with the process station;
generating, by the deep learning controller, an expected behavioral pattern data and an actual behavioral pattern based at least in part on the first instance of the split control value signal output;
detecting, by the deep learning controller, based on an evaluation of the expected behavioral pattern data and the actual behavioral pattern data, anomalous activity in the manufacturing process, by identifying an unusual frequency pattern in the actual behavioral pattern data in relation to the expected behavioral pattern data; and
providing an indication of the anomalous activity as a result of detection of the anomalous activity in the multi-step manufacturing process.
16. The computer-implemented method of claim 15 , further comprising:
receiving, by a second signal splitter via a further single input, a second control value signal output from a second process station involved in the multi-step manufacturing process;
splitting, by a second signal splitter, the second control value signal output from the second process station;
providing, by the second signal splitter via a further first output of the second signal splitter, a first instance of the second split control value signal output to the deep learning controller trained to identify anomalous activity in the multi-step manufacturing process based on the second control value signal output; and
providing, by the second signal splitter via a further second output of the second signal splitter, a second instance of the second split control value signal output to the second station controller.
17. The computer-implemented method of claim 15 , further comprising:
training the deep learning controller to predict expected behavioral patterns; and
training the deep learning controller to generate actual behavioral patterns.
18. The computer-implemented method of claim 17 , further comprising:
training the deep learning controller to identify anomalous activity in the manufacturing process based on the expected behavioral patterns and the actual behavioral patterns.
19. The computer-implemented method of claim 15 , wherein first divided control value signal is an analog or digital signal and the second control value signal is an analog or digital signal.
20. The computer-implemented method of claim 15 , wherein the deep learning controller is trained based on a training data set comprising at least one set of actual behavioral patterns comprising unusual frequency patterns.