SENSOR KITS FOR GENERATING FEATURE VECTORS FOR MONITORING AND MANAGING INDUSTRIAL SETTINGS

1 . A sensor kit configured for monitoring an industrial setting, the sensor kit comprising:

an edge device; and

a plurality of sensors that capture sensor data and transmit the sensor data via a self- configuring sensor kit network, wherein the plurality of sensors includes one or more sensors of a first sensor type and one or more sensors of a second sensor type, wherein at least one sensor of the plurality of sensors comprises:

a sensing component that captures sensor measurements and outputs instances of sensor data;

a processing unit that generates reporting packets based on one or more instances of sensor data and outputs the reporting packets, wherein each reporting packet includes routing data and one or more instances of sensor data; and

a communication device configured to receive reporting packets from the processing unit and to transmit the reporting packets to the edge device via the self- configuring sensor kit network in accordance with a first communication protocol;

wherein the edge device comprises:

one or more storage devices that store a model data store that stores a plurality of machine-learned models that are each trained to predict or classify a condition of an industrial component of the industrial setting or of the industrial setting based on a set of features that are derived from instances of sensor data captured by one or more of the plurality of sensors;

a communication system that receives reporting packets from the plurality of sensors via the self-configuring sensor kit network using a first communication protocol and that transmits sensor kit packets to a backend system via a public network using a second communication protocol that is different from the first communication protocol; and

a processing system having one or more processors that execute computer- executable instructions that cause the processing system to:

receive the reporting packets from the communication system;

generate a set of feature vectors based on one or more respective instances of sensor data received in the reporting packets;

for each respective feature vector, input the respective feature vector into a respective machine-learned model that corresponds to the feature vector to obtain a respective prediction or classification relating to a condition of a respective industrial component of the industrial setting or the industrial setting and a degree of confidence corresponding to the respective prediction or classification;

selectively encode the one or more instances of sensor data prior to transmission to the backend system based on the respective predictions or classifications outputted by the machine-learned models in response to the respective feature vector to obtain one or more sensor kit packets; and

output the sensor kits packets to the communication system, wherein the communication system transmits the sensor kit packets to the backend system via the public network.

2 . The sensor kit of claim 1 , further comprising a gateway device, wherein the gateway device is configured to receive sensor kit packets from the edge device via a wired communication link and transmit the sensor kit packets to the backend system via the public network on behalf of the edge device.

3 . The sensor kit of claim 2 , wherein the gateway device includes a satellite terminal device that is configured to transmit the sensor kit packets to a satellite that routes the sensor kit packets to the public network.

4 . The sensor kit of claim 2 , wherein the gateway device includes a cellular chipset that is pre- configured to transmit the sensor kit packets to a cellphone tower of a preselected cellular provider.

5 . The sensor kit of claim 1 , wherein the second communication device of the edge device is a satellite terminal device that is configured to transmit the sensor kit packets to a satellite that routes the sensor kits to the public network.

6 . The sensor kit of claim 1 , wherein the one or more storage devices store a sensor data store that stores instances of sensor data captured by the plurality of sensors of the sensor kit.

7 . The sensor kit of claim 1 , wherein selectively encoding the one or more instances of sensor data includes: in response to obtaining one or more predictions or classifications relating to conditions of respective industrial components of the industrial setting and the industrial setting that collectively indicate that there are likely no issues relating to any industrial component of the industrial setting and the industrial setting, compressing the one or more instances of sensor data using a lossy codec.

8 . The sensor kit of claim 7 , wherein compressing the one or more instances of sensor data using the lossy codec includes:

normalizing the one or more instances of sensor data into respective pixel values;

encoding the respective pixel values into a video frame;

compressing a block of video frames using the lossy codec, wherein the lossy codec is a video codec and the block of video frames includes the video frame and one or more other video frames that include normalized pixel values of other instances of sensor data.

9 . The sensor kit of claim 7 , wherein selectively encoding the one or more instances of sensor data includes:

in response to obtaining a prediction or classification relating to a condition of a particular industrial component or the industrial setting that indicates that there is likely an issue relating to the particular industrial component or the industrial setting, compressing the one or more instances of sensor data using a lossless codec.

10 . The sensor kit of claim 7 , wherein selectively encoding the one or more instances of sensor data includes:

in response to obtaining a prediction or classification relating to a condition of a particular industrial component or the industrial setting that indicates that there is likely an issue relating to the particular industrial component or the industrial setting, refraining from compressing the one or more instances of sensor data.

11 . The sensor kit of claim 1 , wherein the computer-executable instructions further cause the one or more processors of the edge device to selectively store the one or more instances of sensor data in the one or more storage devices of the edge device based on the respective predictions or classifications.

12 . The sensor kit of claim 11 , wherein selectively storing the one or more instances of sensor data includes: in response to obtaining one or more predictions or classifications relating to conditions of respective industrial components of the industrial setting and the industrial setting that collectively indicate that there are likely no issues relating to any industrial component of the industrial setting and the industrial setting, storing the one or more instances of sensor data in the storage device with an expiry, such that the one or more instances of sensor data are purged from the storage device in accordance with the expiry.

13 . The sensor kit of claim 11 , wherein selectively storing the one or more instances of sensor data includes:

in response to obtaining a prediction or classification relating to a condition of a particular industrial component or the industrial setting that indicates that there is likely an issue relating to the particular industrial component or the industrial setting, storing the one or more instances of sensor data in the storage device indefinitely.