IoT-based network architecture for detecting faults using vibration measurement data
In one embodiment, a device in a network receives a machine learning encoder and decoder trained by a supervisory service. The service trains the encoder and decoder using vibration measurement data sent to the service by a plurality of devices. The device trains, based on the received encoder, a classifier to determine whether vibration measurement data is indicative of a behavioral anomaly. The device receives vibration measurement data captured by a particular set of one or more vibration sensors of a monitored system. The device evaluates, using the trained decoder, the received vibration measurement data to determine whether the data is indicative of a structural anomaly in the monitored system. The device evaluates, using the trained classifier, the received vibration measurement data to determine whether the data is indicative of a behavioral anomaly in the monitored system.
1. A method comprising:
receiving, at a device in a network, a machine learning encoder and decoder of a cloud-based supervisory service, wherein the cloud-based supervisory service trains the encoder and decoder using vibration measurement data sent to the cloud-based supervisory service by a plurality of devices in the network that receive the vibration measurement data captured by one or more vibration sensors of a monitored system;
training, by the device and based on the received encoder, a classifier to determine whether the vibration measurement data is indicative of a behavioral anomaly;
receiving, at the device, particular vibration measurement data captured by a particular set of the one or more vibration sensors;
evaluating, by the device and using the trained decoder, the particular vibration measurement data to determine whether the particular vibration measurement data is indicative of a structural anomaly in the monitored system; and
evaluating, by the device and using the trained classifier, the particular vibration measurement data to determine whether the particular vibration measurement data is indicative of a behavioral anomaly in the monitored system.
2. The method as in claim 1 , wherein the device is a router in the network.
3. The method as in claim 1 , wherein the plurality of devices comprises the device in the network.
4. The method as in claim 1 , wherein the devices of the plurality are located in two or more networks.
5. The method as in claim 1 , wherein training the classifier, based on the received encoder, to determine whether the vibration measurement data is indicative of a behavioral anomaly:
receiving, at the device, vibration measurement data captured by one or more vibration sensors outside of the particular set; and
using, by the device, a transfer learning approach to train the classifier based in part on the vibration measurement data captured by the one or more vibration sensors outside of the particular set.
6. The method as in claim 1 , wherein the encoder comprises a long short-term memory (LSTM) layer.
7. The method as in claim 1 , wherein evaluating, using the trained decoder, the particular vibration measurement data to determine whether the particular vibration measurement data is indicative of a structural anomaly in the monitored system comprises:
determining that the particular vibration measurement data is indicative of a structural anomaly in the monitored system; and
sending, via the network, an alert regarding the structural anomaly in the monitored system.
8. The method as in claim 1 , wherein evaluating, using the trained classifier, the particular vibration measurement data to determine whether the particular vibration measurement data is indicative of a behavioral anomaly in the monitored system comprises:
determining that the particular vibration measurement data is indicative of a behavioral anomaly in the monitored system; and
sending, via the network, an alert regarding the behavioral anomaly in the monitored system.
9. An apparatus, comprising:
one or more network interfaces to communicate with a network;
a processor coupled to the network interfaces and configured to execute one or more processes; and
a memory configured to store a process executable by the processor, the process when executed configured to:
receive a machine learning encoder and decoder of a cloud-based supervisory service, wherein the cloud-based supervisory service trains the encoder and decoder using vibration measurement data sent to the cloud-based supervisory service by a plurality of devices in the network that receive vibration measurement data captured by one or more vibration sensors of a monitored system;
train, based on the received encoder, a classifier to determine whether the vibration measurement data is indicative of a behavioral anomaly;
receive particular vibration measurement data captured by a particular set of the one or more vibration sensors;
evaluate, using the trained decoder, the particular vibration measurement data to determine whether the particular vibration measurement data is indicative of a structural anomaly in the monitored system; and
evaluate, using the trained classifier, the particular vibration measurement data to determine whether the particular vibration measurement data is indicative of a behavioral anomaly in the monitored system.
10. The apparatus as in claim 9 , wherein the apparatus is a router in the network.
11. The apparatus as in claim 9 , wherein the plurality of devices comprises the apparatus.
12. The apparatus as in claim 9 , wherein the devices of the plurality are located in two or more networks.
13. The apparatus as in claim 9 , wherein the apparatus trains the classifier, based on the received encoder, to determine whether the vibration measurement data is indicative of a behavioral by:
receiving vibration measurement data captured by one or more vibration sensors outside of the particular set; and
using a transfer learning approach to train the classifier based in part on the vibration measurement data captured by the one or more vibration sensors outside of the particular set.
14. The apparatus as in claim 9 , wherein the encoder comprises a long short-term memory (LSTM) layer.
15. The apparatus as in claim 9 , wherein the apparatus evaluates, using the trained decoder, the particular vibration measurement data to determine whether the particular vibration measurement data is indicative of a structural anomaly in the monitored system by:
determining that the particular vibration measurement data is indicative of a structural anomaly in the monitored system; and
sending, via the network, an alert regarding the structural anomaly in the monitored system.
16. The apparatus as in claim 9 , wherein the apparatus evaluates, using the trained classifier, the particular vibration measurement data to determine whether the particular vibration measurement data is indicative of a behavioral anomaly in the monitored system by:
determining that the particular vibration measurement data is indicative of a behavioral anomaly in the monitored system; and
sending, via the network, an alert regarding the behavioral anomaly in the monitored system.
17. A tangible, non-transitory, computer-readable medium storing program instructions that cause a device in a network to execute a process comprising:
receiving, at the device in the network, a machine learning encoder and decoder of a cloud-based supervisory service, wherein the cloud-based supervisory service trains the encoder and decoder using vibration measurement data sent to the cloud-based supervisory service by a plurality of devices in the network that receive vibration measurement data captured by one or more vibration sensors of a monitored system;
training, by the device and based on the received encoder, a classifier to determine whether the vibration measurement data is indicative of a behavioral anomaly;
receiving, at the device, particular vibration measurement data captured by a particular set of the one or more vibration sensors;
evaluating, by the device and using the trained decoder, the particular vibration measurement data to determine whether the particular vibration measurement data is indicative of a structural anomaly in the monitored system; and
evaluating, by the device and using the trained classifier, the particular vibration measurement data to determine whether the particular vibration measurement data is indicative of a behavioral anomaly in the monitored system.
18. The computer-readable medium as in claim 17 , wherein the device is a router in the network.
19. The computer-readable medium as in claim 17 , wherein training the classifier, based on the received encoder, to determine whether the vibration measurement data is indicative of a behavioral anomaly:
receiving, at the device, vibration measurement data captured by one or more vibration sensors outside of the particular set; and
using, by the device, a transfer learning approach to train the classifier based in part on the vibration measurement data captured by the one or more vibration sensors outside of the particular set.
20. The computer-readable medium as in claim 17 , wherein the encoder comprises a long short-term memory (LSTM) layer.