Methods, systems, kits, and apparatus for monitoring and controlling industrial environments

Abstract

To provide a component, a system, and a method for monitoring various industrial settings.SOLUTION: A sensor kit deployed to an industrial settings includes: a set of an IoT sensor 102 to be deployed to a middle side, an upper side, or around an industrial component, kinds of the industrial component (e.g., a turbine, a power generator, a fan, a pump, a valve, an assembling line, a pipe or a pipe line, a food inspection line, and a server rack, or the like), industrial settings 120 and / or kinds of the industrial settings 120 (e.g., an indoor, an outdoor, a manufacturing, a digging, an excavation, a resource digging, an underground, and an underwater, or the like); and a set of an edge device for processing an input of the sensor and providing a network based communication.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] (Cross - Reference to Related Applications) This application claims priority to each of U.S. Provisional Patent Application No. 62 / 791,878, filed on January 13, 2019; U.S. Provisional Patent Application No. 62 / 827,166, filed on March 31, 2019; U.S. Provisional Patent Application No. 62 / 869,011, filed on June 30, 2019; and U.S. Provisional Patent Application No. 62 / 914,998, filed on October 14, 2019, each of which is entitled "METHODS, SYSTEMS, KITS, AND APPARATUSES FOR MONITORING INDUSTRIAL SETTINGS". Each of the above applications is hereby incorporated by reference in its entirety as if fully set forth herein.

[0002] The present disclosure relates to various configurations of an Internet of Things (IoT) system in a conveniently deployed kit that uses various configurations of sensors, edge computing devices, networking systems, and artificial intelligence to monitor or manage an industrial environment.

Background Art

[0003] The Internet of Things (IoT) is a network of connected devices, systems, components, services, programs, vehicles, home appliances, machinery, and other electronic items that communicate through a set of communication networks, interfaces, and protocols. Much of the development in the field of IoT has focused on consumer products such as wearable devices, home monitoring systems, and smart appliances, but IoT devices and systems have many industrial applications, including the embodiments described throughout this disclosure and the literature incorporated herein. For example, IoT sensors can be used to monitor industrial facilities such as factories, refineries, oil and gas fields, manufacturing lines, energy production facilities, and mining environments, as well as many machines and systems deployed in such environments. While some machines incorporate sensors and measuring instruments such as onboard diagnostic systems, many machines lack such sensors, or only have a limited set of sensors. Therefore, there is a need and opportunity to collect vast amounts of data by installing numerous heterogeneous sensors of various types on, in, and around machines in industrial environments, either temporarily (portable or mobile data acquisition devices as described in the referenced literature, or drones, autonomous vehicles, etc.), semi-permanently (modular interfaces for convenient connection and disconnection, etc.), or permanently.

[0004] However, various problems arise in industrial IoT environments. For example, many industrial IoT devices may be configured to communicate using cellular protocols such as 3G, 4G, LTE, or 5G, but these protocols may not be natively suitable for communication in industrial environments because heavy machinery and dense structures can negatively affect communication between devices. Wi-Fi systems can also provide network connectivity within a facility, but Wi-Fi systems can also face challenges due to the unfavorable physical environment of industrial settings. For example, Wi-Fi systems are not typically designed to communicate through obstacles such as concrete or brick slabs. Also, because many devices are moving around in industrial environments, it can be difficult for Wi-Fi and cellular systems to determine which devices are communicating at any given time.

[0005] Another issue concerns bandwidth. When hundreds or thousands of sensors are installed in a monitored area (such as a factory, assembly line, or oil field), and these sensors acquire multiple measurements per second, the amount of data collected can strain the computing resources of even the most robust computing systems. Therefore, there is a need for methods and systems to address the challenge of efficient and effective bandwidth utilization.

[0006] Another issue is security. IoT devices can be perceived as a security risk when connected to computer networks used to operate mission-critical machinery. Historically, IoT devices have had security vulnerabilities and have frequently been points of attack on networks and devices.

[0007] Concerns about bandwidth, reliability, latency, and / or security can make businesses hesitant to integrate IoT sensor systems into industrial environments and computer networks. Therefore, there is a need for systems that deliver the benefits of IoT while addressing network needs and security risks.

[0008] Another challenge for companies considering IoT implementation is the need for sophisticated integration between IoT devices and network systems, as well as platforms (such as cloud platforms) where the collected IoT data is analyzed, providing both human and automated control in industrial environments. Companies may lack the diverse expertise and available staff necessary for effective IoT integration. Therefore, there is a need for a simple implementation system that delivers the benefits of IoT. [Overview of the project]

[0009] This specification provides methods and systems for monitoring and managing industrial environments through a variety of configurable kits that are ready to use and offer self-configuration and auto-provisioning capabilities, while mitigating issues of complexity, integration, bandwidth, latency, and security. Actual implementations of IoT solutions may include a set of components that constitute a set of edge computing devices and communication functions (including various protocols, ports, gateways, connectors, interfaces, etc.) that collectively provide an appropriate set of sensors, a set of communication devices, an automatically configured and / or pre-configured process, and the transmission of sensor data from the sensor kits to a set of backend systems (e.g., cloud-based or on-premise systems) via appropriate protocols, each configured for various industrial environments, as well as a set of backend systems that are automatically configured and / or pre-configured and provide the owners and operators of the industrial environment with monitoring and / or management information from specific sensor kits registered in the industrial environment. As used herein, a “set” may include a set of single members. References to “monitoring” and / or “managing” should be understood to encompass a variety of actions and activities that may benefit from information shared via the IoT, such as monitoring machine performance, reporting status, condition, or conditions, managing status, conditions, and parameters, undertaking remote control, supporting autonomous functions that depend on status or condition information, supporting analysis, supporting self-configuration, supporting artificial intelligence, and supporting machine learning, unless the context indicates otherwise.

[0010] According to some embodiments of this disclosure, a sensor kit configured for monitoring an industrial environment is disclosed. In the embodiments, the sensor kit includes an edge device and a plurality of sensors, i.e., a set of sensors, that capture sensor data and transmit the sensor data over a self-configured sensor kit network. The plurality of sensors includes one or more sensors of a first sensor type and one or more sensors of a second sensor type. At least one of the plurality of sensors includes a sensing component that acquires sensor measurements and outputs instances of sensor data, a processing unit that generates and outputs report packets based on one or more instances of the sensor data, each report packet including routing data and one or more instances of sensor data, and a communication device configured to receive report packets from the processing unit and transmit report packets to the edge device over a self-configured sensor kit network according to a first communication protocol. The edge device includes a communication system having a first communication device that receives report packets from the plurality of sensors over a self-configured sensor kit network and a second communication device that transmits sensor kit packets to a backend system over a public network. The edge device further includes a processing system having one or more processors, the one or more processors executing computer-executable instructions causing the processing system to perform the following operations: receiving a report packet from a communication system; performing one or more edge operations on instances of sensor data in the report packet; generating a sensor kit packet based on the instances of sensor data, wherein each sensor kit packet contains at least one instance of sensor data; and outputting the sensor kit packets to the communication system, the communication system transmitting the report packet to the backend system via a public network.

[0011] In some embodiments, the sensor kit further includes a gateway device configured to receive sensor kit packets from an edge device via a wired communication link and, on behalf of the edge device, transmit the sensor kit packets to a backend system via a public network. In some of these embodiments, the gateway device includes a satellite terminal device configured to transmit the sensor kit packets to a satellite that routes the sensor kit to a public network. Alternatively, in some embodiments, the gateway device includes a pre-configured cellular chipset to transmit the sensor kit packets to a cell phone base station of a pre-selected cellular provider.

[0012] In some embodiments, the edge device's second communication device is a satellite terminal device configured to transmit sensor kit packets to a satellite that routes the sensor kit to a public network.

[0013] In some embodiments, the edge device further includes one or more storage devices that store a sensor data store, which stores instances of sensor data captured by multiple sensors of the sensor kit.

[0014] In some embodiments, the edge device further includes one or more storage devices that store a model datastore, each storing one or more machine learning models, each trained to predict or classify the state of an industrial environment and / or industrial components of an industrial environment based on a set of features derived from instances of sensor data captured by one or more of the multiple sensors. In some of these embodiments, performing one or more edge operations includes: generating a feature vector based on one or more instances of sensor data received from one or more of the multiple sensors; inputting the feature vector into a machine learning model to obtain a prediction or classification related to the state of an industrial environment or a particular industrial component of an industrial environment, and a confidence level corresponding to the prediction or classification; and selectively encoding one or more instances of sensor data based on the state or prediction before sending it to a backend system. In some of these embodiments, selectively encoding one or more instances of sensor data includes compressing one or more instances of sensor data using a lossy codec in response to obtaining one or more predictions or classifications related to the state of an industrial environment and each industrial component of an industrial environment that collectively indicate that there is a high probability that there are no problems related to the industrial environment and any of the industrial components of the industrial environment. In some of these embodiments, compressing one or more instances of sensor data using a lossy codec includes normalizing one or more instances of sensor data to their respective pixel values; encoding each pixel value into a video frame; and compressing blocks of video frames using a lossy codec, wherein the lossy codec is a video codec, and the blocks of video frames include a video frame and one or more other video frames containing the normalized pixel values ​​of other instances of sensor data.In some embodiments, selectively encoding one or more instances of sensor data includes compressing one or more instances of sensor data using a lossless codec in response to obtaining predictions or classifications related to the state of a particular industrial component or industrial environment that indicate a high probability of a problem related to that particular industrial component or industrial environment. In some embodiments, selectively encoding one or more instances of sensor data includes refraining from compressing one or more instances of sensor data in response to obtaining predictions or classifications related to the state of a particular industrial component or industrial environment that indicate a high probability of a problem related to that particular industrial component or industrial environment. In some embodiments, performing one or more edge operations includes generating feature vectors based on one or more instances of sensor data received from one or more sensors out of a plurality of sensors; inputting the feature vectors into a machine learning model to obtain predictions or classifications related to the state of a particular industrial component or industrial environment in the industrial environment, and confidence levels corresponding to the predictions or classifications; and selectively storing one or more instances of sensor data in a storage device of an edge device based on the predictions or classifications. In some embodiments, selectively storing one or more instances of sensor data includes storing one or more instances of sensor data on a storage device with an expiration date in response to obtaining one or more predictions or classifications relating to the state of the industrial environment and each of the industrial components of the industrial environment, which collectively indicate that there is a high probability that there are no problems relating to the industrial environment and any of the industrial components of the industrial environment, and purging one or more instances of sensor data from the storage device according to the expiration date.In some embodiments, selectively storing one or more instances of sensor data includes storing one or more instances of sensor data indefinitely in a storage device in response to obtaining predictions or classifications related to the state of a particular industrial component or industrial environment that indicate a high probability of a problem related to that particular industrial component or industrial environment.

[0015] In some embodiments, the self-configuring sensor kit network is a star network in which each sensor of a plurality of sensors directly transmits each instance of sensor data to an edge device using a short-range communication protocol. In some of these embodiments, a computer-executable instruction further causes one or more processors in the edge device to initiate the configuration of the self-configuring sensor kit network.

[0016] In some embodiments, the self-configuring sensor kit network is a mesh network in which the communication device of each sensor of a plurality of sensors is configured to establish a communication channel with at least one other sensor of the plurality of sensors, and at least one sensor of the plurality of sensors is configured to receive instances of sensor data from one or more other sensors of the plurality of sensors and route the received instances of sensor data toward edge devices. In some of these embodiments, a computer-executable instruction further causes one or more processors of edge devices to initiate the configuration of the self-configuring sensor kit network, and the plurality of sensors form a mesh network in response to the edge devices initiating the configuration of the self-configuring sensor kit network.

[0017] In some embodiments, the self-configured sensor kit network is a hierarchical network. In some of these embodiments, the sensor kit further includes one or more collection devices configured to receive report packets from one or more of a plurality of sensors and route the report packets to edge devices.

[0018] In this embodiment, the self-configured sensor kit network is a ring network that communicates using a serial data protocol.

[0019] In this embodiment, the sensor kit network is a mesh network.

[0020] In this embodiment, at least one of the sensors in the sensor kit network is a multi-axis vibration sensor.

[0021] In one embodiment, the edge device includes a rule-based network protocol adapter for selecting a network protocol for transmitting sensor kit packets over a public network.

[0022] According to some embodiments of the present disclosure, a method for monitoring an industrial environment is disclosed using a sensor kit having a plurality of sensors and an edge device including a processing system. In embodiments, the method includes: the processing system receiving report packets from one or more of the plurality of sensors, each report packet indicating sensor data transmitted from and captured by the respective sensor; the processing system performing one or more edge operations on one or more instances of the sensor data received in the report packets; the processing system generating one or more sensor kit packets based on the instances of sensor data, each sensor kit packet including at least one instance of the sensor data; and the processing system outputting the sensor kit packets to a backend system over a public network. In some embodiments, the report packets received from one or more of the plurality of sensors include the sensor identifier of the respective sensor. In embodiments, receiving report packets from one or more of the plurality of sensors is performed using a first communication device implementing a first communication protocol, and outputting the sensor kit packets to a backend system is performed using a second communication device implementing a second communication protocol. In some embodiments, the second communication device is a satellite terminal device, and outputting the sensor kit packet includes using the satellite terminal device to transmit the sensor kit packet to a satellite, which then routes the sensor kit packet to a public network. In some embodiments, outputting the sensor kit packet to a backend system includes transmitting the sensor kit packet to a sensor kit gateway device. In some embodiments, transmitting the sensor kit packet to a gateway device includes transmitting the sensor kit packet to the gateway via a wired communication link between the edge device and the gateway device.In some embodiments, the gateway device includes a satellite terminal device configured to transmit sensor kit packets to a satellite that routes the sensor kit to a public network. In some embodiments, the gateway device includes a pre-configured cellular chipset to transmit sensor kit packets to a cellular tower of a pre-selected cellular provider. In some embodiments, the method further includes storing a model data store containing one or more machine learning models in one or more storage devices of the edge device. In some embodiments, one or more machine learning models are trained to predict or classify the state of industrial components in an industrial environment and / or the state of the industrial environment based on a set of features derived from instances of sensor data captured by one or more of the sensors.

[0023] In some embodiments, performing one or more edge operations includes: generating vector features based on one or more instances of sensor data received from one or more sensors among a plurality of sensors; inputting the feature vectors into one or more machine learning models to obtain predictions or classifications related to specific industrial components or states of the industrial environment, and confidence levels corresponding to those predictions or classifications; and selectively encoding one or more instances of sensor data before sending them to a backend system based on the state or prediction. In some embodiments, selectively encoding one or more instances of sensor data includes compressing one or more instances of sensor data using a lossy codec in response to obtaining one or more predictions or classifications regarding the state of each industrial component and industrial environment of the industrial environment, which collectively indicate that there is a high probability that there are no problems related to any of the industrial components and industrial environment of the industrial environment. In some embodiments, compressing one or more instances of sensor data using a lossy codec includes normalizing one or more instances of sensor data to their respective pixel values; encoding each pixel value into a video frame; and compressing blocks of video frames using a lossy codec, wherein the lossy codec is a video codec, and the blocks of video frames include a video frame and one or more other video frames containing the normalized pixel values ​​of other instances of sensor data. In some embodiments, selectively encoding one or more instances of sensor data includes compressing one or more instances of sensor data using a lossless codec in response to obtaining predictions or classifications related to the state of a particular industrial component or industrial environment that indicate a likely problem related to that particular industrial component or industrial environment.In embodiments, selective encoding of one or more instances of sensor data includes refraining from compressing one or more instances of sensor data in response to obtaining predictions or classifications related to the state of a particular industrial component or industrial environment that indicate a high probability of a problem related to that particular industrial component or industrial environment.

[0024] In some embodiments, performing one or more edge operations includes: generating a feature vector based on one or more instances of sensor data received from one or more of a plurality of sensors; inputting the feature vector into a machine learning model to obtain predictions or classifications relating to specific industrial components or states of the industrial environment, and confidence levels corresponding to those predictions or classifications; and selectively storing one or more instances of the sensor data in a storage device of the edge device based on the predictions or classifications. In embodiments, selectively storing one or more instances of the sensor data includes storing one or more instances of the sensor data in a storage device with an expiration date such that one or more instances of the sensor data are purged from the storage device according to their expiration date, and storing one or more instances of the sensor data in a storage device with an expiration date is performed in response to obtaining one or more predictions or classifications relating to the states of each industrial component and industrial environment of the industrial environment, which collectively indicate that there is a high probability that there are no problems relating to any of the industrial components and industrial environment of the industrial environment. In some embodiments, selectively storing one or more instances of sensor data includes storing one or more instances of sensor data indefinitely in a storage device in response to obtaining predictions or classifications related to the state of a particular industrial component or industrial environment that indicate a high probability of a problem related to that particular industrial component or industrial environment.

[0025] In some embodiments, the method further includes capturing sensor measurements by the sensing components of a plurality of sensors; generating one or more report packets based on the captured sensor measurements by the sensor processor; and transmitting one or more report packets to an edge device via a self-configured sensor kit network by the sensor's communication unit. In some of these embodiments, the method further includes initiating the configuration of a self-configured sensor kit network by a processing system, wherein the self-configured sensor kit network is a star network. In some embodiments, the report packets are received directly from each sensor using a short-range communication protocol. In embodiments, the method further includes initiating the configuration of a self-configured sensor kit network by a processing system, wherein the self-configured sensor kit network is a mesh network. In some embodiments, the method further includes establishing a communication channel with at least one other sensor of the plurality of sensors by the communication device of each sensor of the plurality of sensors; receiving instances of sensor data from one or more other sensors of the plurality of sensors by at least one sensor of the plurality of sensors; and routing the received instances of sensor data to an edge device via a mesh network by at least one sensor of the plurality of sensors.

[0026] In some embodiments, the self-configuring sensor kit network is a hierarchical network, and the sensor kit includes one or more collection devices participating in the hierarchical network. In some of these embodiments, the method further includes receiving, by a collection device of one or more collection devices, a report packet from a set of sensors that communicate with the collection device using a first short-range communication protocol; and routing, by one or more collection devices, the report packet to an edge device using either a first short-range communication protocol or a second short-range communication protocol different from the first short-range communication protocol.

[0027] In some embodiments, the edge device includes a rule-based network protocol adapter. In some of these embodiments, the method further includes selecting, by the rule-based network protocol adapter, a network protocol; and transmitting, by the edge device, sensor kit packets over a public network according to the network protocol.

[0028] In some embodiments, the plurality of sensors includes a first set of sensors of a first sensor type and a second set of sensors of a second sensor type.

[0029] According to some embodiments of this disclosure, a sensor kit configured to monitor an industrial environment is disclosed. In embodiments, the sensor kit includes an edge device and a plurality of sensors that capture sensor data and transmit the sensor data over a self-configured sensor kit network. The plurality of sensors includes one or more sensors of a first sensor type and one or more sensors of a second sensor type. At least one of the plurality of sensors includes: a sensing component that acquires sensor measurements and outputs instances of sensor data; a processing unit that generates and outputs report packets based on one or more instances of sensor data, each report packet including routing data and one or more instances of sensor data; and a communication device configured to receive report packets from the processing unit and transmit report packets to the edge device over the self-configured sensor kit network according to a first communication protocol. The edge device includes one or more storage devices that store a model data store that stores a plurality of machine learning models, each trained to predict or classify the state of industrial components of an industrial environment or the industrial environment itself, based on a set of features derived from instances of sensor data captured by one or more of the plurality of sensors. Furthermore, the edge device includes a communication system that uses a first communication protocol to receive report packets from multiple sensors via a self-configured sensor kit network, and uses a second communication protocol different from the first communication protocol to transmit sensor kit packets to a backend system via a public network.Furthermore, the edge device includes a processing system having one or more processors that execute computer-executable instructions, and the computer-executable instructions cause the processing system to: receive a reporting packet from a communication system; generate a set of feature vectors based on one or more instances of sensor data received in the reporting packet; for each feature vector, input the respective feature vector into the respective machine learning model corresponding to the feature vector to obtain a respective prediction or classification regarding the state of each industrial component of the industrial environment or of the industrial environment, and a reliability corresponding to each prediction or classification; selectively encode one or more instances of sensor data to obtain one or more sensor kit packets before transmitting to a backend system based on the respective predictions or classifications output by the machine learning model corresponding to each feature vector; and output the sensor kit packets to the communication system, wherein the communication system transmits the reporting packet to the backend system via a public network.

[0030] In some embodiments, the sensor kit further includes a gateway device configured to receive the sensor kit packets from the edge device via a wired communication link and transmit the sensor kit packets to the backend system via a public network on behalf of the edge device. In some of these embodiments, the gateway device includes a satellite terminal device configured to transmit the sensor kit packets to a satellite that routes the sensor kit to the public network. Alternatively, in some embodiments, the gateway device includes a cellular chipset preconfigured to transmit the sensor kit packets to a cellular tower of a preselected cellular provider.

[0031] In some embodiments, the edge device's second communication device is a satellite terminal device configured to transmit sensor kit packets to a satellite that routes the sensor kit to a public network.

[0032] In the embodiment, the sensor kit includes one or more storage devices that store a sensor data store, which stores instances of sensor data captured by multiple sensors of the sensor kit.

[0033] In some embodiments, selectively encoding one or more instances of sensor data includes compressing one or more instances of sensor data using a lossy codec in response to obtaining one or more predictions or classifications about the state of each industrial component of the industrial environment and the industrial environment, which collectively indicate that there are likely to be no problems related to any industrial component of the industrial environment and the industrial environment. In some embodiments, compressing one or more instances of sensor data using a lossy codec includes normalizing one or more instances of sensor data to their respective pixel values; encoding each pixel value into a video frame; and compressing blocks of video frames using a lossy codec, where the lossy codec is a video codec, and the blocks of video frames include a video frame and one or more other video frames containing the normalized pixel values ​​of other instances of sensor data. In some of these embodiments, selective encoding of one or more instances of sensor data involves compressing one or more instances of sensor data using a lossless codec in response to obtaining a state-related prediction or classification of a particular industrial component or industrial environment that indicates a high probability of a problem related to that particular industrial component or industrial environment.

[0034] In some embodiments, selective encoding of one or more instances of sensor data includes refraining from compressing one or more instances of sensor data in response to obtaining a state-related prediction or classification of a particular industrial component or industrial environment that indicates a high probability of a problem related to that particular industrial component or industrial environment.

[0035] In some embodiments, a computer-executable instruction further causes one or more processors in the edge device to selectively store one or more instances of sensor data in one or more storage devices of the edge device based on their respective predictions or classifications. In some of these embodiments, selectively storing one or more instances of sensor data includes storing one or more instances of sensor data in a storage device with an expiration date in response to obtaining one or more state-related predictions or classifications of each industrial component and the industrial environment of the industrial environment, indicating collectively that there are likely no problems related to any industrial component and the industrial environment of the industrial environment, and purging one or more instances of sensor data from the storage device according to the expiration date. In some embodiments, selectively storing one or more instances of sensor data includes storing one or more instances of sensor data indefinitely in a storage device in response to obtaining a state-related prediction or classification of a particular industrial component or industrial environment, indicating that there is likely to be a problem related to that particular industrial component or industrial environment.

[0036] In some embodiments, the self-configuring sensor kit network is a star network in which each sensor of a plurality of sensors directly transmits its respective instance of sensor data to the edge device using a short-range communication protocol. In some of these embodiments, a computer-executable instruction further causes one or more processors in the edge device to initiate the configuration of the self-configuring sensor kit network.

[0037] In some embodiments, the self-configuring sensor kit network is a mesh network in which the communication device of each of a plurality of sensors is configured to establish a communication channel with at least one other sensor among the plurality of sensors, and at least one sensor among the plurality of sensors is configured to receive instances of sensor data from one or more other sensors among the plurality of sensors and route the received instances of sensor data toward edge devices. In some of these embodiments, a computer-executable instruction further causes one or more processors of the edge devices to initiate the configuration of the self-configuring sensor kit network, and the plurality of sensors form a mesh network in response to the edge devices initiating the configuration of the self-configuring sensor kit network.

[0038] In some embodiments, the self-configured sensor kit network is a hierarchical network. In some of these embodiments, the sensor kit includes one or more collection devices configured to receive report packets from one or more sensors among a plurality of sensors and to route the report packets to edge devices.

[0039] According to some embodiments of the present disclosure, a method for monitoring an industrial environment is disclosed using a sensor kit having an edge device including a plurality of sensors and a processing system. The method includes: the processing system receiving report packets from one or more of the plurality of sensors, each report packet including routing data and instances of one or more sensor data; the processing system generating a set of feature vectors based on each of the one or more instances of sensor data received in the report packets; the processing system inputting the respective feature vectors into each of a plurality of machine learning models, each trained to predict or classify the respective states of industrial components of the industrial environment or the industrial environment, based on the set of features derived from the instances of sensor data captured by one or more of the plurality of sensors; the processing system obtaining the respective predictions or classifications and the corresponding confidence levels from each machine learning model based on the respective feature vectors input to each machine learning model; the processing system selectively encoding one or more instances of sensor data based on the respective predictions or classifications to obtain one or more sensor kit packets; and the processing system transmitting the sensor kit packets to a backend system over a public network. In some embodiments, the sensor kit includes a gateway device configured to receive sensor kit packets from an edge device via a wired communication link and, on behalf of the edge device, transmit the sensor kit packets to a backend system via a public network. In some embodiments, the gateway device includes a satellite terminal device that transmits the sensor kit packets to a satellite that routes the sensor kit packets to a public network. In some embodiments, the gateway device includes a cellular chipset that transmits the sensor kit packets to a cellular tower of a pre-selected cellular provider.In some embodiments, receiving report packets from one or more sensors is performed using a first communication device implementing a first communication protocol, and sending sensor kit packets to a backend system is performed using a second communication device implementing a second communication protocol. In some embodiments, the second communication device of the edge device is a satellite terminal device, and sending sensor kit packets to the backend system includes the satellite terminal device sending the sensor kit packets to a satellite that routes the sensor kit packets to a public network.

[0040] In some embodiments, the method further includes compressing one or more instances of sensor data using a lossy codec in response to the processing system obtaining one or more state-related predictions or classifications of each industrial component of an industrial environment and the industrial environment, indicating collectively that there are likely no problems associated with any particular industrial component of the industrial environment and the industrial environment. In some of these embodiments, compressing one or more instances of sensor data using a lossy codec includes normalizing one or more instances of sensor data to their respective pixel values; encoding each pixel value into a video frame; and compressing blocks of video frames using a lossy codec, where the lossy codec is a video codec, and the block of video frames includes a video frame and one or more other video frames containing the normalized pixel values ​​of other instances of sensor data. In some embodiments, the method includes compressing one or more instances of sensor data using a lossless codec in response to the processing system obtaining state-related predictions or classifications of a particular industrial component or industrial environment, indicating that there are likely to be problems associated with that particular industrial component or industrial environment. In embodiments, the method includes refraining from compressing one or more instances of sensor data in response to the processing system obtaining a state-related prediction or classification of a particular industrial component or industrial environment that indicates a high probability of a problem related to that particular industrial component or industrial environment.

[0041] In some embodiments, the edge communication device includes one or more storage devices that store multiple machine learning models. In some of these embodiments, one or more storage devices store instances of sensor data captured by multiple sensors of the sensor kit. In some embodiments, the method further includes the processing system selectively storing one or more instances of sensor data in one or more storage devices based on their respective predictions or classifications. In embodiments, the method further includes the processing system storing one or more instances of sensor data in a storage device with an expiration date so that one or more instances of sensor data are purged from the storage device according to their expiration date, and the processing system stores one or more instances of sensor data in a storage device with an expiration date in response to obtaining one or more predictions or classifications regarding the state of each industrial component of the industrial environment and the industrial environment, which collectively indicate that there are likely to be no problems related to any of the industrial components of the industrial environment and the industrial environment. In some embodiments, the method further includes storing one or more instances of sensor data indefinitely in a storage device in response to the processing system obtaining a state-related prediction or classification of a particular industrial component or industrial environment that indicates a high probability of a problem related to that particular industrial component or industrial environment.

[0042] In some embodiments, the method further includes the steps of acquiring sensor data by a plurality of sensors, and transmitting the sensor data by the plurality of sensors over a self-configured sensor kit network. In some of these embodiments, transmitting the sensor data over a self-configured sensor kit network includes each of the plurality of sensors directly transmitting an instance of the sensor data to an edge device using a short-range communication protocol, and the self-configured sensor kit network is a star network. In some embodiments, the method further includes a processing system initiating the configuration of the self-configured sensor kit network. In embodiments, the self-configured sensor kit network is a mesh network, and each of the plurality of sensors includes a communication device. In embodiments, the method further includes the steps of establishing a communication channel with at least one other sensor of the plurality of sensors by the communication device of each of the plurality of sensors, receiving an instance of sensor data from one or more other sensors of the plurality of sensors by at least one sensor of the plurality of sensors, and routing the received instance of sensor data to an edge device by at least one sensor of the plurality of sensors.

[0043] In some embodiments, the self-configured sensor kit network is a hierarchical network, and the sensor kit includes one or more collection devices. In some of these embodiments, the method further includes the steps of: receiving report packets from one or more sensors among a plurality of sensors by at least one of the plurality of collection devices; and routing the report packets to an edge device by at least one of the plurality of collection devices.

[0044] In this embodiment, the plurality of sensors includes a first set of sensors of a first sensor type and a second set of sensors of a second sensor type.

[0045] According to some embodiments of this disclosure, a sensor kit configured for monitoring an industrial environment is disclosed. In the embodiments, the sensor kit includes an edge device and a plurality of sensors that acquire sensor data and transmit the sensor data over a self-configured sensor kit network. The plurality of sensors includes one or more sensors of a first sensor type and one or more sensors of a second sensor type. At least one of the plurality of sensors includes a sensing component that acquires sensor measurements and outputs instances of sensor data, and a processing unit that generates and outputs report packets based on one or more instances of the sensor data, each report packet including routing data and one or more instances of sensor data, and at least one of the plurality of sensors also includes a communication device configured to receive report packets from the processing unit and transmit report packets to the edge device over a self-configured sensor kit network according to a first communication protocol. The edge device includes a first communication device that receives report packets from the plurality of sensors over a self-configured sensor kit network, and a second communication device that transmits sensor kit packets to a backend system over a public network. The edge device further includes a processing system having one or more processors that execute computer executable instructions to process: receiving a report packet from a communication system; generating a block of media content frames; compressing the block of media content frames using a media codec; generating one or more server kit packets based on the block of media content frames; and sending one or more server kit packets to a backend system over a public network, wherein each media content frame contains multiple frame values, and each frame value represents each instance of sensor data.

[0046] In some embodiments, the sensor kit further includes a gateway device configured to receive sensor kit packets from an edge device via a wired communication link and, on behalf of the edge device, transmit the sensor kit packets to a backend system via a public network. In some of these embodiments, the gateway device includes a satellite terminal device configured to transmit the sensor kit packets to a satellite that routes the sensor kit to a public network. Alternatively, in some embodiments, the gateway device includes a cellular chipset pre-configured to transmit the sensor kit packets to a cell phone tower of a pre-selected cellular provider.

[0047] In some embodiments, the edge device's second communication device is a satellite terminal device configured to transmit sensor kit packets to a satellite that routes the sensor kit to the public network.

[0048] In the embodiment, the edge device further includes one or more storage devices that store a sensor data store that stores instances of sensor data acquired by multiple sensors of the sensor kit.

[0049] In some embodiments, the edge device further includes one or more storage devices, each containing a model data store, which stores one or more machine learning models, each trained to predict or classify industrial components and / or industrial environment states in an industrial environment based on feature sets derived from instances of sensor data acquired by one or more of a plurality of sensors. In some embodiments, performing one or more edge operations includes the steps of: generating a feature vector based on one or more instances of sensor data received from one or more of a plurality of sensors; inputting the feature vector into a machine learning model to obtain predictions or classifications related to specific industrial components or industrial environment states in an industrial environment, and confidence levels corresponding to the predictions or classifications; and selecting a codec to be used to compress blocks of media frames based on the state or prediction. In some embodiments, selecting a codec includes selecting a lossy codec in response to obtaining one or more predictions or classifications regarding each industrial component and industrial environment state in an industrial environment that collectively indicate that any of the industrial components and industrial environment states in the industrial environment are likely to be problem-free. In some of these embodiments, selective encoding of one or more instances of sensor data involves selecting a lossless codec in response to obtaining predictions or classifications related to the state of a particular industrial component or industrial environment that indicate a high probability of a problem relating to that particular industrial component or industrial environment.

[0050] In some embodiments, performing one or more edge operations includes the steps of: generating a feature vector based on one or more instances of sensor data received from one or more sensors among a plurality of sensors; inputting the feature vector into a machine learning model to obtain predictions or classifications related to specific industrial components and industrial environment conditions in an industrial environment, and confidence levels corresponding to the predictions or classifications; and selectively storing one or more instances of sensor data in a storage device of the edge device based on the predictions or classifications. In some of these embodiments, selectively storing one or more instances of sensor data includes the step of storing one or more instances of sensor data in a storage device with an expiration date in response to obtaining one or more predictions or classifications related to each industrial component and industrial environment condition in an industrial environment that collectively indicate that there is a high probability that there are no problems related to any of the industrial components and industrial environment conditions in the industrial environment, and the one or more instances of sensor data are erased from the storage device according to the expiration date. In some embodiments, selectively storing one or more instances of sensor data includes the step of storing one or more instances of sensor data indefinitely in a storage device in response to obtaining predictions or classifications related to specific industrial components or industrial environment conditions that indicate a high probability that there are problems related to those specific industrial components or industrial environment conditions.

[0051] In some embodiments, the self-configured sensor kit network is a star network in which each sensor of a plurality of sensors directly transmits each instance of sensor data to the edge device using a short-range communication protocol. In some of these embodiments, a computer executable instruction further causes one or more processors in the edge device to initiate the configuration of the self-configured sensor kit network.

[0052] In some embodiments, the self-configuring sensor kit network is a mesh network. The mesh network is configured such that the communication device of each of the multiple sensors establishes a communication channel with at least one other sensor among the multiple sensors, and at least one sensor among the multiple sensors is configured to receive instances of sensor data from one or more other sensors among the multiple sensors and route the received instances of sensor data toward edge devices. In some of these embodiments, a computer executable instruction further causes one or more processors in the edge devices to initiate the configuration of the self-configuring sensor kit network, and in response to the edge devices initiating the configuration of the self-configuring sensor kit network, the multiple sensors form a mesh network.

[0053] In some embodiments, the self-configured sensor kit network is a hierarchical network. In some of these embodiments, the sensor kit includes one or more collection devices configured to receive report packets from one or more sensors among a plurality of sensors and route the report packets to edge devices.

[0054] In some embodiments, generating a block of media frames includes the steps of: normalizing each instance of sensor data to be included in a media frame to a normalized media frame value that is within the range of media frame values ​​permitted by the encoding standard corresponding to the media frame; and embedding each normalized media frame value into the media frame. In some of these embodiments, each media frame is a video frame containing multiple pixels, and each normalized media frame value is a pixel value. In some embodiments, embedding each normalized media frame value into the media frame includes the steps of: determining the pixels of the multiple pixels corresponding to each normalized media frame based on a mapping that maps each sensor of the multiple sensors to each pixel of the multiple pixels; and setting the determined pixel value to be identical to each normalized media frame value. In embodiments, the codec is an H.264 / MPEG-4 codec. In embodiments, the codec is an H.265 / MPEG-H codec. In embodiments, the codec is an H.263 / MPEG-4 codec.

[0055] According to some embodiments of this disclosure, a method for monitoring an industrial environment is disclosed using a sensor kit having a plurality of sensors and an edge device including a processing system. The method includes the steps of: the processing system receiving report packets from each of one or more of the plurality of sensors, each report packet including routing data and one or more instances of sensor data; the method including the processing system generating a block of media content frames, each media content frame including a plurality of frame values, each frame value representing each instance of sensor data; the method including the processing system compressing the block of media content frames using a media codec to obtain a compressed block; the processing system generating one or more server kit packets based on the compressed block; and the processing system transmitting one or more server kit packets to a backend system over a public network. In some embodiments, the sensor kit includes a gateway device configured to receive sensor kit packets from an edge device over a wired communication link and, on behalf of the edge device, transmit the sensor kit packets to a backend system over a public network. In embodiments, the gateway device includes a satellite terminal device configured to transmit the sensor kit packets to a satellite that routes the sensor kit to a public network. In some embodiments, the gateway device includes a cellular chipset pre-configured to transmit sensor kit packets to a cell phone tower of a pre-selected cellular provider.

[0056] In some embodiments, the reception of report packets from one or more sensors is performed using a first communication device that receives report packets from multiple sensors via a self-configured sensor kit network, and the transmission of sensor kit packets to a backend system is performed using a second communication device. In some of these embodiments, the second communication device of the edge device is a satellite terminal device configured to transmit sensor kit packets to a satellite that routes the sensor kit to a public network. In some embodiments, the method further includes the steps of acquiring sensor data by multiple sensors and transmitting the sensor data to an edge device via a self-configured sensor kit network by multiple sensors. In some embodiments, transmitting sensor data via a self-configured sensor kit network includes each sensor of the multiple sensors directly transmitting an instance of the sensor data to the edge device using a short-range communication protocol, and the self-configured sensor kit network is a star network. In some embodiments, the method further includes the step of a processing system initiating the configuration of the self-configured sensor kit network.

[0057] In some embodiments, the self-configured sensor kit network is a mesh network, and each of the multiple sensors includes a communication device. In some of these embodiments, the method further includes the steps of: establishing a communication channel with at least one other sensor of the multiple sensors by the communication device of each sensor of the multiple sensors; receiving instances of sensor data from one or more other sensors of the multiple sensors by at least one sensor of the multiple sensors; and routing the received instances of sensor data toward an edge device by at least one sensor of the multiple sensors.

[0058] In some embodiments, the self-configured sensor kit network is a hierarchical network, and the sensor kit includes one or more collection devices. In some of these embodiments, the method further includes the steps of: receiving report packets from one or more sensors among a plurality of sensors by at least one of the plurality of collection devices; and routing the report packets to an edge device by at least one of the plurality of collection devices.

[0059] In some embodiments, the method further includes the step of storing instances of sensor data acquired by multiple sensors of the sensor kit in one or more storage devices of the edge device.

[0060] In embodiments, the edge device further includes one or more storage devices that store a model data store that stores one or more machine learning models, each trained to predict or classify industrial components and / or industrial environment states in an industrial environment based on feature sets derived from instances of sensor data acquired by one or more of a plurality of sensors. In some of these embodiments, the method further includes the steps of: the processing system generating feature vectors based on one or more instances of sensor data received from one or more of a plurality of sensors; the processing system inputting the feature vectors into a machine learning model to obtain predictions or classifications related to specific industrial components or industrial environment states in an industrial environment, and confidence levels corresponding to the predictions or classifications; and, based on the classifications or predictions, selecting a media codec to be used to compress blocks of media content frames. In some embodiments, the selection of a media codec includes the step of selecting a lossy codec in response to obtaining one or more predictions or classifications related to each industrial component and industrial environment state in an industrial environment that collectively indicate that there are likely to be no problems related to any of the industrial components and industrial environment states in the industrial environment. In embodiments, selecting a media codec includes the step of selecting a lossless codec in response to obtaining a prediction or classification related to the condition of a particular industrial component or industrial environment that indicates a high probability of problems relating to that particular industrial component or industrial environment.

[0061] In some embodiments, the method includes the steps of: a processing system generating a feature vector based on one or more instances of sensor data received from one or more sensors among a plurality of sensors; the processing system inputting the feature vector into a machine learning model to obtain predictions or classifications related to specific industrial components or states of the industrial environment and confidence levels corresponding to those predictions or classifications; and the processing system selectively storing one or more instances of sensor data in a storage device of an edge device based on the predictions or classifications. In embodiments, selectively storing one or more instances of sensor data in a storage device includes the step of storing one or more instances of sensor data in a storage device with an expiration date such that one or more instances of sensor data are deleted from the storage device according to the expiration date, and storing one or more instances of sensor data in a storage device with an expiration date is performed in response to obtaining one or more predictions or classifications regarding each industrial component and state of the industrial environment that collectively indicate that there is a high probability that any of the industrial components and states of the industrial environment are free from problems. In some embodiments, selectively storing one or more instances of sensor data in a storage device includes the step of storing one or more instances of sensor data indefinitely in a storage device in response to obtaining predictions or classifications related to the state of a particular industrial component or industrial environment that indicate a high probability of a problem relating to that particular industrial component or industrial environment.

[0062] In some embodiments, generating a block of media content frames includes the steps of: normalizing each instance of sensor data to be included in a media content frame to a normalized media content frame value that is within the range of media content frame values ​​permitted by the encoding standard corresponding to the media content frame; and embedding each normalized media content frame value into the media content frame. In some of these embodiments, each media content frame is a video frame containing a plurality of pixels, and each normalized media frame value is a pixel value. In embodiments, embedding each normalized media content frame value into the media content frame includes the steps of: determining the pixels of the plurality of pixels corresponding to each normalized media content frame based on a mapping that maps each sensor of the plurality of sensors to each pixel of the plurality of pixels; and setting the value of the determined pixel to be identical to each normalized media content frame value. In some embodiments, the codec is an H.264 / MPEG-4 codec. In some embodiments, the codec is an H.265 / MPEG-H codec. In some embodiments, the codec is the H.263 / MPEG-4 codec.

[0063] In this embodiment, the plurality of sensors includes a first set of sensors of a first sensor type and a second set of sensors of a second sensor type.

[0064] A system is disclosed according to several embodiments of the present disclosure. The system includes a backend system and a sensor kit configured to monitor an industrial environment. The sensor kit includes a plurality of sensors that acquire sensor data and transmit the sensor data over a self-configured sensor kit network, the plurality of sensors including one or more sensors of a first sensor type and one or more sensors of a second sensor type, at least one of the plurality of sensors including a sensing component that acquires sensor measurements and outputs instances of sensor data, and a processing unit that generates and outputs report packets based on one or more instances of the sensor data, each report packet including routing data and one or more instances of sensor data, and at least one of the plurality of sensors includes a communication device configured to receive report packets from the processing unit and transmit report packets to an edge device over the self-configured sensor kit network in accordance with a first communication protocol. The edge device includes a communication system, the communication system having a first communication device that receives report packets from the plurality of sensors over the self-configured sensor kit network and a second communication device that transmits sensor kit packets to the backend system over a public network. The edge device includes a processing system, which has one or more processors that execute computer executable instructions causing the processing system to perform the following: receiving a report packet from a communication system; performing one or more edge operations on instances of sensor data in the report packet; and generating a sensor kit packet based on the instances of sensor data, each sensor kit packet containing at least one instance of sensor data, and outputting the sensor kit packet to the communication system, which then transmits the report packet to the backend system via a public network.The backend system includes a backend storage system that stores a sensor kit data store that stores sensor data received from one or more sensor kits, including a sensor kit, and a backend processing system having one or more processors that execute computer executable instructions causing the backend processing system to perform the following: receiving sensor kit packets from a sensor kit; determining sensor data collected by the sensor kit based on the sensor kit packets; performing one or more backend operations on the sensor data collected by the sensor kit; and storing the sensor data collected by the sensor kit in the sensor kit data store.

[0065] In some embodiments, the sensor kit further includes a gateway device configured to receive sensor kit packets from an edge device via a wired communication link and, on behalf of the edge device, transmit the sensor kit packets to a backend system via a public network. In some of these embodiments, the gateway device includes a satellite terminal device configured to transmit the sensor kit packets to a satellite that routes the sensor kit to a public network. Alternatively, in some embodiments, the gateway device includes a cellular chipset pre-configured to transmit the sensor kit packets to a cell phone tower of a pre-selected cellular provider.

[0066] In some embodiments, the edge device's second communication device is a satellite terminal device configured to transmit sensor kit packets to a satellite that routes the sensor kit to the public network.

[0067] In the embodiment, the edge device further includes one or more storage devices that store a sensor data store that stores instances of sensor data acquired by multiple sensors of the sensor kit.

[0068] In embodiments, the edge device further includes one or more storage devices that store a model data store that stores one or more machine learning models, each trained to predict or classify industrial components and / or industrial environment states in an industrial environment based on feature sets derived from instances of sensor data acquired by one or more of a plurality of sensors. In some of these embodiments, performing one or more edge operations includes the steps of: generating a feature vector based on one or more instances of sensor data received from one or more of a plurality of sensors; inputting the feature vector into a machine learning model to obtain a prediction or classification related to a specific industrial component or industrial environment state in an industrial environment, and a confidence level corresponding to the prediction or classification; and selectively encoding one or more instances of sensor data before sending them to a backend system based on the state or prediction. In some embodiments, selectively encoding one or more instances of sensor data includes compressing one or more instances of sensor data using a lossy codec in response to obtaining one or more predictions or classifications about each industrial component and industrial environment state in an industrial environment that collectively indicate that any of the industrial components and industrial environment states in the industrial environment are likely to be problem-free. In some embodiments, compressing one or more instances of sensor data using a lossy codec includes the steps of: normalizing one or more instances of sensor data to their respective pixel values; encoding each pixel value into a video frame; and compressing the block of video frames using a lossy codec to obtain a compressed block of frames, wherein the lossy codec is a video codec, and the block of video frames includes a video frame and one or more other video frames containing the normalized pixel values ​​of other instances of sensor data.In some embodiments, the backend system receives a compressed block of frames in one or more sensor kit packets and determines the sensor data collected by the sensor kit by decompressing the compressed block of frames using a lossy codec. In some embodiments, selective encoding of one or more instances of sensor data includes the step of compressing one or more instances of sensor data using a lossless codec in response to obtaining a prediction or classification related to the state of a particular industrial component or industrial environment that indicates a high probability of a problem related to that particular industrial component or industrial environment. In some embodiments, selective encoding of one or more instances of sensor data includes the step of stopping the compression of one or more instances of sensor data in response to obtaining a prediction or classification related to the state of a particular industrial component or industrial environment that indicates a high probability of a problem related to that particular industrial component or industrial environment. In some embodiments, selective encoding of one or more instances of sensor data includes the step of selecting a stream of sensor data instances for uncompressed transmission. In the embodiment, performing one or more edge operations includes the steps of: generating a feature vector based on one or more instances of sensor data received from one or more sensors among a plurality of sensors; inputting the feature vector into a machine learning model to obtain a prediction or classification related to a specific industrial component of an industrial environment or the state of the industrial environment, and a confidence level corresponding to the prediction or classification; and selectively storing one or more instances of sensor data in a storage device of the edge device based on the prediction or classification.In some of these embodiments, selectively storing one or more instances of sensor data includes the step of storing one or more instances of sensor data on a storage device with an expiration date in response to obtaining one or more predictions or classifications related to the state of each industrial component and industrial environment in the industrial environment that collectively indicate that there is a high probability that there are no problems with any industrial component and industrial environment in the industrial environment, and the one or more instances of sensor data are erased from the storage device according to the expiration date. In some embodiments, selectively storing one or more instances of sensor data includes the step of storing one or more instances of sensor data on a storage device indefinitely in response to obtaining predictions or classifications related to the state of a particular industrial component or industrial environment that indicate a high probability that there are problems with that particular industrial component or industrial environment.

[0069] In some embodiments, the self-configuring sensor kit network is a star network in which each sensor of a plurality of sensors directly transmits each instance of sensor data to an edge device using a short-range communication protocol. In some of these embodiments, a computer executable instruction further causes one or more processors in the edge device to initiate the configuration of the self-configuring sensor kit network.

[0070] In some embodiments, the self-configuring sensor kit network is a mesh network, wherein the communication device of each of the multiple sensors is configured to establish a communication channel with at least one other sensor among the multiple sensors, and at least one sensor among the multiple sensors is configured to receive instances of sensor data from one or more other sensors among the multiple sensors and route the received instances of sensor data toward edge devices. In some of these embodiments, a computer executable instruction further causes one or more processors in the edge devices to initiate the configuration of the self-configuring sensor kit network, and the multiple sensors form a mesh network in response to the edge devices initiating the configuration of the self-configuring sensor kit network.

[0071] In some embodiments, the self-configured sensor kit network is a hierarchical network. In some of these embodiments, the sensor kit includes one or more collection devices configured to receive report packets from one or more sensors among a plurality of sensors and route the report packets to edge devices.

[0072] In the embodiment, the backend operation includes the steps of performing one or more analytical tasks using sensor data, performing one or more artificial intelligence tasks using sensor data, issuing notifications to human users related to the industrial environment based on sensor data, and / or controlling at least one component of the industrial environment based on sensor data.

[0073] According to some embodiments of the present disclosure, a method for monitoring an industrial environment using a sensor kit that communicates with a backend system is disclosed, the sensor kit comprising a plurality of sensors and an edge device. The method comprises the steps of: an edge processing system of the edge device receiving a report packet from each of one or more of the plurality of sensors, each report packet comprising routing data and one or more instances of sensor data; the method comprising the edge processing system performing one or more edge operations on the instances of sensor data in the report packet; the edge processing system generating a plurality of sensor kit packets based on the instances of sensor data, each sensor kit packet comprising at least one instance of sensor data; the method comprising the edge processing system transmitting the sensor kit packets to the backend system over a public network; the backend processing system of the backend system receiving the sensor kit packets from the sensor kit over a public network; the backend processing system determining the sensor data collected by the sensor kit based on the sensor kit packets; the backend processing system performing one or more backend operations on the sensor data collected by the sensor kit; and the backend processing system storing the sensor data collected by the sensor kit in a sensor kit data store located in the backend storage system of the backend system. In some embodiments, the sensor kit further includes a gateway device configured to receive sensor kit packets from edge devices via a wired communication link and, on behalf of the edge devices, transmit the sensor kit packets to a backend system over a public network. In some embodiments, the gateway device includes a satellite terminal device configured to transmit the sensor kit packets to a satellite that routes the sensor kit to a public network.In one embodiment, the gateway device includes a cellular chipset pre-configured to transmit sensor kit packets to a cell phone tower of a pre-selected cellular provider.

[0074] In some embodiments, the reception of report packets from one or more sensors is performed using a first communication device on an edge device that receives report packets from multiple sensors via a self-configured sensor kit network, and the transmission of sensor kit packets to a backend system is performed using a second communication device on the edge device. In some of these embodiments, the second communication device on the edge device is a satellite terminal device configured to transmit sensor kit packets to a satellite that routes the sensor kit to a public network. In some embodiments, the method further includes the steps of acquiring sensor data by multiple sensors and transmitting the sensor data to an edge device via a self-configured sensor kit network by multiple sensors. In some embodiments, transmitting sensor data via a self-configured sensor kit network includes the step of each sensor of the multiple sensors directly transmitting an instance of the sensor data to the edge device using a short-range communication protocol, and the self-configured sensor kit network is a star network. In some embodiments, the method further includes the step of initiating the configuration of the self-configured sensor kit network by an edge processing system. In some embodiments, the self-configured sensor kit network is a mesh network, and each sensor of the multiple sensors includes a communication device. In some embodiments, the method further includes the steps of: establishing a communication channel with at least one other sensor of the plurality of sensors by a communication device of each sensor of the plurality of sensors; receiving an instance of sensor data from one or more other sensors of the plurality of sensors by at least one sensor of the plurality of sensors; and routing the received instance of sensor data to an edge device by at least one sensor of the plurality of sensors.

[0075] In some embodiments, the self-configured sensor kit network is a hierarchical network, and the sensor kit includes one or more collection devices. In some of these embodiments, the method further includes the steps of: receiving report packets from one or more sensors among a plurality of sensors by at least one of the plurality of collection devices; and routing the report packets to an edge device by at least one of the plurality of collection devices.

[0076] In the embodiment, the method further includes the step of storing instances of sensor data acquired by multiple sensors of the sensor kit in one or more storage devices of the edge device.

[0077] In some embodiments, the edge device further includes one or more storage devices that store a model data store that stores one or more machine learning models, each trained to predict or classify industrial components and / or industrial environment conditions in an industrial environment based on feature sets derived from instances of sensor data acquired by one or more of a plurality of sensors. In some of these embodiments, performing one or more edge operations includes the steps of: generating a feature vector by an edge processing system based on one or more instances of sensor data received from one or more of a plurality of sensors; inputting the feature vector into a machine learning model by the edge processing system to obtain predictions or classifications related to specific industrial components or industrial environment conditions in an industrial environment, and confidence levels corresponding to the predictions or classifications; and selectively encoding one or more instances of sensor data by the edge processing system before sending them to a backend system based on the predictions or classifications. In some embodiments, selectively encoding one or more instances of sensor data includes the step of compressing one or more instances of sensor data using a lossy codec in response to obtaining one or more predictions or classifications about each industrial component and industrial environment condition in an industrial environment that collectively indicate that there is a high probability that there are no industrial components and industrial environment-related problems in any of the industrial environments.In some embodiments, compressing one or more instances of sensor data using a lossy codec includes the steps of: normalizing one or more instances of sensor data to individual pixel values ​​by an edge processing system; encoding each pixel value into a media content frame by the edge processing system; and compressing the block of media content frames using a lossy codec to obtain a compressed block, wherein the lossy codec is a video codec, and the compressed block includes a media content frame and one or more other media content frames containing the normalized pixel values ​​of other instances of sensor data. In embodiments, a backend system receives a compressed block of one or more sensor kit packets and determines the sensor data collected by the sensor kit by decompressing the compressed block using a lossy codec.

[0078] In some embodiments, selective encoding of one or more instances of sensor data includes the step of compressing one or more instances of sensor data using a lossless codec in response to obtaining a prediction or classification related to the state of a particular industrial component or industrial environment that indicates a high probability of a problem related to that particular industrial component or industrial environment. In some embodiments, selective encoding of one or more instances of sensor data includes the step of stopping the compression of one or more instances of sensor data in response to obtaining a prediction or classification related to the state of a particular industrial component or industrial environment that indicates a high probability of a problem related to that particular industrial component or industrial environment. In some embodiments, selective encoding of one or more instances of sensor data includes the edge processing system selecting a stream of sensor data instances for uncompressed transmission.

[0079] In some embodiments, performing one or more edge operations includes the steps of: generating a feature vector by an edge processing system based on one or more instances of sensor data received from one or more sensors among a plurality of sensors; inputting the feature vector into a machine learning model by the edge processing system to obtain predictions or classifications related to specific industrial components or states of the industrial environment, and confidence levels corresponding to the predictions or classifications; and selectively storing one or more instances of the sensor data in one or more storage devices by the edge processing system based on the predictions or classifications. In some embodiments, selectively storing one or more instances of the sensor data includes the step of storing one or more instances of the sensor data in a storage device with an expiration date in response to the edge processing system obtaining one or more predictions or classifications related to each industrial component and state of the industrial environment that collectively indicate that there is a high probability that there are no problems with any of the industrial components and states of the industrial environment, and storing one or more instances of the sensor data in a storage device with an expiration date is performed such that one or more instances of the sensor data are deleted from the storage device according to the expiration date. In some embodiments, selectively storing one or more instances of sensor data includes the step of storing one or more instances of sensor data indefinitely in a storage device in response to an edge processing system obtaining a prediction or classification related to the state of a particular industrial component or industrial environment that indicates a high probability of a problem relating to that particular industrial component or industrial environment.

[0080] In some embodiments, the plurality of sensors includes a first set of sensors of a first sensor type and a second set of sensors of a second sensor type.

[0081] According to some embodiments of this disclosure, a sensor kit configured to monitor an indoor agricultural facility is disclosed. The sensor kit includes an edge device and a plurality of sensors that acquire sensor data and transmit the sensor data over a self-configured 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. At least one of the plurality of sensors includes a sensing component that acquires sensor measurements and outputs instances of sensor data, and a processing unit that generates a report packet based on one or more instances of the sensor data and outputs the report packet, each report packet including routing data and one or more instances of sensor data, and at least one of the plurality of sensors includes a communication device configured to receive the report packet from the processing unit and transmit the report packet to the edge device over the self-configured sensor kit network according to a first communication protocol. The plurality of sensors includes two or more sensor types selected from a group including light sensors, humidity sensors, temperature sensors, carbon dioxide sensors, fan speed sensors, weight sensors, and camera sensors. The edge device includes a communication system comprising a first communication device that receives report packets from multiple sensors via a self-configured sensor kit network, and a second communication device that transmits sensor kit packets to a backend system via a public network. The edge device also includes a processing system having one or more processors that execute computer executable instructions causing the processing system to perform the following: receiving report packets from the communication system, performing one or more edge operations on instances of sensor data in the report packets, and generating sensor kit packets based on the instances of sensor data. Each sensor kit packet contains at least one instance of sensor data, and the sensor kit packet is output to the communication system, which transmits the report packets to the backend system via a public network.

[0082] In this embodiment, the sensor kit includes an edge device and a plurality of sensors that capture sensor data and transmit the sensor data via a self-configured sensor kit network. The plurality of sensors includes one or more sensors of a first sensor type and one or more sensors of a second sensor type. At least one of the plurality of sensors includes a sensing component that captures sensor measurements and outputs instances of sensor data. A processing unit generates and outputs report packets based on one or more instances of the sensor data. Each report packet includes routing data and one or more instances of sensor data. A communication device is configured to receive report packets from the processing unit and transmit report packets to the edge device via the self-configured sensor kit network based on a first communication protocol.

[0083] In this embodiment, the edge device further includes one or more storage devices that store a sensor data store that stores instances of sensor data captured by multiple sensors of the sensor kit.

[0084] In this embodiment, the edge device further includes one or more storage devices that store a model data store that stores one or more machine learning models configured to predict or classify the state of components of an indoor farm environment. The indoor farm environment is based on a set of functions derived from instances of sensor data captured by one or more of a plurality of sensors. In some of these embodiments, the performance of one or more edge operations includes: generating a feature vector based on one or more instances of sensor data received from one or more of the plurality of sensors; inputting the feature vector into a machine learning model to obtain predictions or classifications related to the state of the indoor farm environment or a particular component of the indoor farm environment, and confidence levels corresponding to the predictions or classifications; selectively encoding one or more instances of sensor data before sending them to a backend system based on the state or prediction. In some embodiments, selectively encoding one or more instances of sensor data includes compressing one or more instances of sensor data using a lossy codec in response to obtaining one or more predictions or classifications related to the state of each industrial component of the indoor farm environment. The indoor farm environment collectively indicates that there are likely no problems related to the indoor farm environment and any components of the indoor farm environment. In some embodiments, compressing one or more instances of sensor data using a lossy codec includes: Normalizing one or more instances of sensor data to their respective pixel values; encoding each pixel value into a video frame; and compressing blocks of video frames using a lossy codec, where the lossy codec is a video codec, and a block of video frames comprises a video frame and one or more other video frames containing the normalized pixel values ​​of other instances of sensor data.In some embodiments, selectively encoding one or more instances of sensor data includes compressing one or more instances of sensor data using a lossless codec in response to obtaining predictions or classifications related to the state of a particular industrial component or industrial setting. This indicates a likely problem related to a particular industrial component or industrial environment. In this embodiment, selectively encoding one or more instances of sensor data includes refraining from compressing one or more instances of sensor data in response to obtaining predictions or classifications related to the state of a particular component or indoor agricultural environment. This may be a problem related to a particular component or indoor agricultural environment. In this embodiment, performing one or more edge operations includes: generating a feature vector based on one or more instances of sensor data received from one or more sensors of a plurality of sensors; inputting the feature vector into a machine learning model to obtain predictions or classifications related to the state of an indoor agricultural environment or a particular component of an indoor agricultural environment, and confidence levels corresponding to the predictions or classifications; and selectively storing one or more instances of sensor data in a storage device of an edge device based on the predictions or classifications. In some of these embodiments, selectively storing one or more instances of sensor data includes storing one or more instances of sensor data in a storage device with an expiration date in response to obtaining one or more predictions or classifications related to the state of each industrial component. Selectively storing one or more instances of sensor data suggests that, depending on the expiration date of the indoor farming configuration, there are likely no issues related to the indoor farming configuration and its components, so that one or more instances of sensor data are purged from the storage device.In some embodiments, selectively storing one or more instances of sensor data includes storing one or more instances of sensor data indefinitely in a storage device in response to obtaining predictions or classifications related to the state of a particular industrial component or industrial setting. This indicates that there is likely to be a problem related to a particular component or indoor agricultural environment.

[0085] In this embodiment, the self-configured sensor kit network is a star network in which each sensor of a plurality of sensors transmits its respective instance of sensor data directly to the edge device using a short-range communication protocol. In some of these embodiments, a computer executable instruction causes one or more processors in the edge device to initiate the configuration of the self-configured sensor kit network.

[0086] In this embodiment, the self-configured sensor kit network is a mesh network as follows: The communication device of each sensor in the plurality of sensors is configured to establish a communication channel with at least one other sensor in the plurality of sensors. At least one sensor in the plurality of sensors is configured to receive instances of sensor data from one or more other sensors in the plurality of sensors and to route the received instances of sensor data toward edge devices. In some of these embodiments, a computer executable instruction causes one or more processors in the edge devices to initiate the configuration of the self-configured sensor kit network, and in response to the initiation of the edge devices, the plurality of sensors configure the self-configured sensor kit network to form a mesh network.

[0087] In this embodiment, the self-configured sensor kit network is a hierarchical network. In some of these embodiments, the sensor kit further includes one or more collection devices configured to receive report packets from one or more sensors of a plurality of sensors and route the report packets to edge devices. In this embodiment, each collection device is installed in each of the different rooms of the indoor agricultural environment and collects sensor data from the sensors of a plurality of sensors located in each room.

[0088] According to some embodiments of this disclosure, a sensor kit configured to monitor an indoor agricultural environment is disclosed. The sensor kit includes an edge device and a plurality of sensors that capture sensor data and transmit the sensor data over a self-configured sensor kit network, wherein the plurality of sensors includes one or more sensors of a first sensor type and a plurality of sensors of one or a second sensor type. At least one of the plurality of sensors includes: a sensing component that captures sensor measurements and outputs instances of sensor data; a processing unit that generates and outputs report packets based on one or more instances of sensor data, each report packet including routing data and one or more instances of sensor data; and a communication device configured to receive report packets from the processing unit and transmit report packets to the edge device over the self-configured sensor kit network, based on a first communication protocol. The plurality of sensors includes two or more sensor types selected from a group, including infrared sensors, ground penetration sensors, light sensors, humidity sensors, temperature sensors, chemical sensors, fan speed sensors, rotational speed sensors, weight sensors, and camera sensors. The edge device includes a communication system having a first communication device that receives report packets from multiple sensors via a self-configured sensor kit network, and a second communication device that transmits sensor kit packets to a backend system via a public network. The edge device further includes a processing system having one or more processors that execute computer executable instructions causing the processing system to: receive report packets from the communication system; perform one or more edge operations on instances of sensor data in the report packets; and generate sensor kit packets based on instances of sensor data, where each sensor kit packet contains at least one instance of sensor data.The sensor kit packet is then output to the communication system, which in turn sends the report packet to the backend system via the public network.

[0089] In some embodiments, the sensor kit further includes a gateway device configured to receive sensor kit packets from an edge device via a wired communication link and transmit the sensor kit packets to a backend system via a public network on behalf of the edge device. In some of these embodiments, the gateway device includes a satellite terminal device configured to transmit the sensor kit packets to a satellite that routes the sensor kit to a public network. Alternatively, in some embodiments, the gateway device includes a cellular chipset configured to transmit the sensor kit packets to a cell phone tower of a pre-selected cellular provider.

[0090] In some embodiments, the edge device's second communication device is a satellite terminal device configured to send sensor kit packets to a satellite that routes the sensor kit to a public network.

[0091] In this embodiment, the edge device further includes one or more storage devices that store a sensor data store that stores instances of sensor data captured by multiple sensors of the sensor kit.

[0092] In this embodiment, the edge device further includes one or more storage devices that store a model data store that stores one or more machine learning models, each configured to predict or classify the state of components of an indoor farming environment. The indoor farming environment is based on a set of functions derived from instances of sensor data captured by one or more of a plurality of sensors. In some embodiments, the performance of one or more edge operations includes: generating a feature vector based on one or more instances of sensor data received from one or more of the plurality of sensors; inputting the feature vector into a machine learning model to make predictions or classifications related to the state of the indoor farming environment or a particular component of indoor farming, and obtaining confidence levels corresponding to the predictions or classifications; and selectively encoding one or more instances of sensor data before sending them to a backend system based on the state or prediction.

[0093] In this embodiment, selectively encoding one or more instances of sensor data includes compressing one or more instances of sensor data using a lossy codec in response to obtaining one or more predictions or classifications related to the state of each component of the indoor farm environment, which collectively indicates that there are likely no problems related to the indoor farm environment and any component of the indoor farm environment. In this embodiment, compressing one or more instances of sensor data using a lossy codec includes: normalizing one or more instances of sensor data to their respective pixel values; encoding each pixel value into a video frame; and compressing blocks of video frames using a lossy codec, where the lossy codec is a video codec, and a block of video frames includes a video frame and one or more other video frames containing the normalized pixel values ​​of other instances of sensor data. In this embodiment, selectively encoding one or more instances of sensor data includes compressing one or more instances of sensor data using a lossless codec in response to obtaining predictions or classifications related to the state of a particular component or indoor farm environment, which collectively indicates that there are likely to be problems related to a particular component or indoor farm environment. In this embodiment, selectively encoding one or more instances of sensor data includes refraining from compressing one or more instances of sensor data in response to obtaining predictions or classifications related to the state of a particular component or the indoor agricultural environment, which may be related to a particular component or the indoor agricultural environment.

[0094] In some embodiments, performing one or more edge operations includes: generating a feature vector based on one or more instances of sensor data received from one or more sensors among a plurality of sensors; inputting the feature vector into a machine learning model to obtain predictions or classifications related to the state of an indoor farm environment or a particular component of an indoor farm environment, and confidence levels corresponding to the predictions or classifications; and selectively storing one or more instances of sensor data in a storage device of the edge device based on the predictions or classifications. In this embodiment, selectively storing one or more instances of sensor data includes storing one or more instances of sensor data in a storage device with an expiration date in response to obtaining one or more predictions or classifications related to the state of each component of the indoor farm. The setting and indoor farm setting expiration dates collectively indicate that there are likely no problems related to the indoor farm setting and components of the indoor farm setting so that one or more instances of sensor data are purged from the storage device. In this embodiment, selectively storing one or more instances of sensor data includes storing one or more instances of sensor data indefinitely in a storage device in response to obtaining predictions or classifications related to the state of a particular component or indoor farm environment. This indicates that there is likely to be a problem related to a particular component or indoor farm environment.

[0095] In some embodiments, the plurality of sensors include a first set of sensors of a first sensor type selected from the group including speed sensors, weight sensors, and camera sensors, and a second set of sensors of a second sensor type selected from the group including light sensors, humidity sensors, temperature sensors, carbon dioxide sensors, and fans.

[0096] According to some embodiments of this disclosure, a sensor kit configured to monitor pipeline configuration is disclosed. The sensor kit includes an edge device and a plurality of sensors that capture sensor data and transmit the sensor data over a self-configured sensor kit network. The plurality of sensors includes one or more sensors of a first sensor type and one or more sensors of a second sensor type. At least one of the plurality of sensors includes: a sensing component that captures sensor measurements and outputs instances of sensor data; a processing unit that generates and outputs report packets based on one or more instances of sensor data; each report packet containing routing data and one or more instances of sensor data; and a communication device configured to receive report packets from the processing unit and transmit report packets to the edge device over the self-configured sensor kit network based on a first communication protocol. The plurality of sensors include sensors with two or more sensor types selected from a group, such as infrared sensors, metal penetration sensors, concrete penetration sensors, light sensors, strain sensors, rust sensors, biological sensors, humidity sensors, temperature sensors, chemical sensors, valve integrity sensors, vibration sensors, flow sensors, cavitation sensors, pressure sensors, weight sensors, and camera sensors. The edge device includes a communication system having a first communication device that receives report packets from multiple sensors via a self-configured sensor kit network, and a second communication device that transmits sensor kit packets to a backend system via a public network. The edge device further includes a processing system having one or more processors that execute computer executable instructions causing the processing system to: receive report packets from the communication system; perform one or more edge operations on instances of sensor data in the report packets; generate sensor kit packets based on instances of sensor data, where each sensor kit packet contains at least one instance of sensor data.The sensor kit packet is then output to the communication system, which in turn sends the report packet to the backend system via the public network.

[0097] In some embodiments, the sensor kit further includes a gateway device configured to receive sensor kit packets from an edge device via a wired communication link and transmit the sensor kit packets to a backend system via a public network on behalf of the edge device. In some of these embodiments, the gateway device includes a satellite terminal device configured to transmit the sensor kit packets to a satellite that routes the sensor kit to a public network. Alternatively, in some embodiments, the gateway device includes a cellular chipset configured to transmit the sensor kit packets to a cell phone tower of a pre-selected cellular provider.

[0098] In some embodiments, the edge device's second communication device is a satellite terminal device configured to send sensor kit packets to a satellite that routes the sensor kit to a public network.

[0099] In this embodiment, the edge device further includes one or more storage devices that store a sensor data store that stores instances of sensor data captured by multiple sensors of the sensor kit.

[0100] In this embodiment, the edge device further includes one or more storage devices that store a model datastore that stores one or more machine learning models, each configured to predict or classify the state of the pipeline components of the pipeline configuration. The pipeline configuration is based on a set of functions derived from instances of sensor data captured by one or more of a plurality of sensors. In some of these embodiments, performing one or more edge operations includes: generating a feature vector based on one or more instances of sensor data received from one or more of the plurality of sensors; inputting the feature vector into a machine learning model to obtain predictions or classifications related to the state of the pipeline configuration or a particular pipeline component of the pipeline configuration, and confidence levels corresponding to the predictions or classifications; and selectively encoding one or more instances of sensor data before sending them to a backend system based on the state or prediction. In this embodiment, selectively encoding one or more instances of sensor data includes compressing one or more instances of sensor data using a lossy codec in response to obtaining one or more predictions or classifications related to the state of each pipeline component of the pipeline configuration. The pipeline configuration collectively indicates that there are likely no issues related to the pipeline configuration and the pipeline components of the pipeline configuration. In this embodiment, compressing one or more instances of sensor data using a lossy codec includes: Normalizing one or more instances of sensor data to their respective pixel values; encoding each pixel value into a video frame; and compressing blocks of video frames using a lossy codec, where the lossy codec is a video codec, and a block of video frames comprises a video frame and one or more other video frames containing the normalized pixel values ​​of other instances of sensor data.In this embodiment, selectively encoding one or more instances of sensor data includes compressing one or more instances of sensor data using a lossless codec in response to obtaining predictions or classifications related to the state of a particular pipeline component or pipeline configuration, which may be related to a particular pipeline component or pipeline configuration. In this embodiment, selectively encoding one or more instances of sensor data includes refraining from compressing one or more instances of sensor data in response to obtaining predictions or classifications related to the state of a particular pipeline component or pipeline configuration, which may be related to a particular pipeline component or pipeline configuration. In this embodiment, performing one or more edge operations includes generating a feature vector based on one or more instances of sensor data received from one or more sensors among a plurality of sensors, inputting the feature vector into a machine learning model to obtain predictions or classifications related to the state of a particular pipeline configuration or pipeline component of the pipeline configuration, and confidence levels corresponding to the predictions or classifications. Based on the predictions or classifications, this includes selectively storing one or more instances of sensor data in a storage device of the edge device. In this embodiment, selectively storing one or more instances of sensor data includes storing one or more instances of sensor data in a storage device with an expiration date in response to obtaining one or more predictions or classifications related to the state of each pipeline component of the pipeline. The configuration and pipeline configuration collectively indicate that there are likely no issues related to the pipeline configuration and the pipeline components of the pipeline configuration, and that one or more instances of sensor data will be purged from the storage device according to their expiration date.In this embodiment, selectively storing one or more instances of sensor data includes storing one or more instances of sensor data indefinitely in a storage device in response to obtaining predictions or classifications related to the state or pipeline configuration of a particular pipeline component, and indicating that there is a high probability of a problem related to a particular pipeline component or pipeline configuration.

[0101] In this embodiment, the self-configured sensor kit network is a star network in which each sensor of a plurality of sensors transmits its respective instance of sensor data directly to the edge device using a short-range communication protocol. In some of these embodiments, a computer executable instruction further causes one or more processors in the edge device to initiate the configuration of the self-configured sensor kit network.

[0102] In this embodiment, the self-configured sensor kit network is a mesh network as follows: The communication device of each sensor in the plurality of sensors is configured to establish a communication channel with at least one other sensor in the plurality of sensors. At least one sensor in the plurality of sensors is configured to receive instances of sensor data from one or more other sensors in the plurality of sensors and to route the received instances of sensor data toward an edge device. In some of these embodiments, a computer executable instruction further causes one or more processors in the edge device to start configuring the self-configured sensor kit network, and in response to the start of the edge device, causes the plurality of sensors to start configuring the self-configured sensor kit network which forms a mesh network.

[0103] In this embodiment, the self-configured sensor kit network is a hierarchical network. In some of these embodiments, the sensor kit further includes one or more collection devices configured to receive report packets from one or more sensors of a plurality of sensors and route the report packets to edge devices. In this embodiment, each collection device is installed in a different section of the pipeline configuration and collects sensor data from sensors of a plurality of sensors located in their respective rooms.

[0104] According to some embodiments of this disclosure, a method for monitoring pipeline configuration using a sensor kit comprising an edge device and multiple sensors is disclosed. This method involves an edge processing system on the edge device receiving report packets from the multiple sensors via a self-configured sensor kit network, each report packet comprising routing data and one or more instances of sensor data captured by each. The multiple sensors comprise two or more sensor types selected from a group including light sensors, humidity sensors, temperature sensors, carbon dioxide sensors, fan speed sensors, weight sensors, and camera sensors. The edge processing system performs one or more edge operations on the instances of sensor data within the report packets. The edge processing system generates one or more edge operations on the instances of sensor data within the report packets. The edge processing system then transmits the sensor kit packets to the edge communication system on the edge device, which transmits the report packets to a backend system via a public network. In some embodiments, the sensor kit further comprises a gateway device configured to receive sensor kit packets from the edge device via a wired communication link and, in turn, transmit the sensor kit packets to a backend system via a public network. In this embodiment, the gateway device comprises a satellite terminal device configured to transmit the sensor kit packets to a satellite that routes the sensor kit to a public network. In some embodiments, the gateway device includes a cellular chipset configured to transmit sensor kit packets to a cell phone tower of a pre-selected cellular provider. In this embodiment, the reception of reporting packets from one or more respective sensors is performed using a first communication device of the edge device that receives reporting packets from multiple sensors via a self-configured sensor kit network and transmits sensor kit packets. The backend system is performed using a second communication device of the edge device.In some embodiments, the edge device's second communication device is a satellite terminal device configured to send sensor kit packets to a satellite that routes the sensor kit to a public network.

[0105] In some embodiments, this method further includes capturing sensor data by multiple sensors and transmitting the sensor data to an edge device via a self-configured sensor kit network. In some of these embodiments, transmitting sensor data via a self-configured sensor kit network includes each sensor of the multiple sensors directly transmitting an instance of the sensor data to the edge device using a short-range communication protocol. The configuration of the sensor kit network is a star network. In some embodiments, this method further includes initiating the configuration of the self-configured sensor kit network by an edge processing system.

[0106] In this embodiment, the self-configured sensor kit network is a mesh network, and each of the multiple sensors includes a communication device. Some methods of these embodiments further include establishing a communication channel with at least one other sensor among the multiple sensors by the communication device of each sensor among the multiple sensors. This method involves at least one of the multiple sensors receiving instances of sensor data from one or more other sensors among the multiple sensors, and at least one of the multiple sensors routing the received instances of sensor data toward an edge device.

[0107] In some embodiments, the self-configured sensor kit network is a hierarchical network, and the sensor kit includes one or more collection devices. Some methods of these embodiments further include: reporting packets from one or more sensors among multiple sensors by at least one of the multiple collection devices; and routing the reported packets to edge devices by at least one of the multiple collection devices. In some embodiments, each collection device is located in a different section of the pipeline configuration and collects sensor data from sensors of multiple sensors located in their respective rooms.

[0108] In some embodiments, the method further includes storing instances of sensor data captured by multiple sensors of the sensor kit in one or more storage devices of the edge device. In this embodiment, the edge device further includes one or more storage devices that store a model data store that stores one or more machine learning models, each configured to predict or classify the state of the agricultural environment and / or agricultural components. The configuration is based on a set of functions derived from instances of sensor data captured by one or more of the multiple sensors.

[0109] In some embodiments, performing one or more edge operations includes the edge processing system generating a feature vector based on one or more instances of sensor data received from one or more sensors of a plurality of sensors; the edge processing system inputting the feature vector into a machine learning model to obtain predictions or classifications related to the state of the agricultural environment or a particular component of the agricultural environment, and confidence levels corresponding to the predictions or classifications; and, based on the predictions or classifications, the edge processing system selectively encoding one or more instances of the sensor data before sending them to a backend system. Some of these embodiments include the edge processing system compressing one or more instances of the sensor data using a lossy codec in response to obtaining one or more predictions or classifications related to the agricultural environment and the conditions of each component of the agricultural environment, which collectively indicate that there are likely no problems related to the agricultural environment or any component of the agricultural environment. In some embodiments, compressing one or more instances of the sensor data using a lossy codec includes: the edge processing system normalizing one or more instances of the sensor data to their respective pixel values; and the edge processing system encoding each pixel value into a media content frame. The edge processing system compresses blocks of media content frames using a lossy codec to obtain compressed blocks, where the lossy codec is a video codec, and the compressed blocks include the media content frames and one or more other media content frames, including normalized pixel values ​​of other instances of sensor data. In some embodiments, the backend system receives the compressed blocks in one or more sensor kit packets and determines the sensor data collected by the sensor kit by decompressing the compressed blocks using a lossy codec.

[0110] In some embodiments, selective encoding of one or more instances of sensor data includes the edge processing system compressing one or more instances of sensor data using a lossless codec in response to obtaining predictions or classifications related to a particular state. A component or agricultural environment indicates that there is a high probability of problems related to that particular component or agricultural environment. In this embodiment, encoding one or more instances of sensor data includes the edge processing system refraining from compressing one or more instances of sensor data in response to obtaining predictions or classifications related to a particular component or agricultural state. A setting indicates that there is a high probability of problems related to that particular component or agricultural environment. In some embodiments, selective encoding of one or more instances of sensor data includes the edge processing system selecting a stream of sensor data instances for uncompressed transmission.

[0111] In some embodiments, performing one or more edge operations includes the edge processing system generating a feature vector based on one or more instances of sensor data received from one or more sensors of a plurality of sensors; the edge processing system inputting the feature vector into a machine learning model to obtain predictions or classifications related to the state of the agricultural environment or a particular component of the agricultural environment, and confidence levels corresponding to the predictions or classifications; and the edge processing system selectively storing one or more instances of sensor data in the storage of one or more storage devices based on the predictions or classifications. In some of these embodiments, selectively storing one or more instances of sensor data includes the edge processing system storing one or more instances of sensor data that have expired in response to obtaining one or more predictions or classifications in a storage device. Storing one or more instances of sensor data related to the state of the agricultural environment and each component of the agricultural environment collectively indicates that there is a high probability that there are no problems related to the agricultural environment or any component of the agricultural environment. Expiration is performed so that one or more instances of sensor data are deleted from the storage device according to their expiration. In some embodiments, selectively storing one or more instances of sensor data includes the edge processing system storing one or more instances of sensor data indefinitely in a storage device in response to obtaining state-related predictions or classifications. A particular component or agricultural environment indicates that there is a high probability of a problem related to that particular component or agricultural environment. In some embodiments, the plurality of sensors include a first set of sensors of a first sensor type selected from the group including speed sensors, weight sensors, and camera sensors, and a second set of sensors of a second sensor type selected from the group including light sensors, humidity sensors, temperature sensors, carbon dioxide sensors, and fan sensors.

[0112] According to some embodiments of the present disclosure, a sensor kit configured to monitor an industrial manufacturing setup is disclosed. The sensor kit includes an edge device and a plurality of sensors that capture sensor data and transmit the sensor data over a self-configured 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. At least one of the plurality of sensors includes a sensing component that captures sensor measurements and outputs instances of sensor data, a processing unit that generates report packets based on one or more instances of sensor data and outputs the report packets, and a communication device configured to receive the report packets from the processing unit and transmit the report packets over the self-configured sensor kit network based on a first communication protocol. Each report packet includes routing data and one or more instances of sensor data. The plurality of sensors includes two or more sensor types selected from the group of metal penetration sensors, concrete penetration sensors, vibration sensors, light sensors, strain sensors, rust sensors, biological sensors, temperature sensors, chemical sensors, valve integrity sensors, rotational speed sensors, vibration sensors, flow sensors, cavitation sensors, pressure sensors, weight sensors, and camera sensors. The edge device includes a communication system having a first communication device that receives report packets from multiple sensors via a self-configured sensor kit network. A second communication device transmits sensor kit packets to a backend system via a public network. The edge device further includes a processing system having one or more processors that execute computer executable instructions causing the processing system to: receive report packets from the communication system; perform one or more edge operations on instances of sensor data in the report packets; and generate sensor kit packets based on instances of sensor data, where each sensor kit packet contains at least one instance of sensor data.The sensor kit packet is then output to the communication system, which in turn sends the report packet to the backend system via the public network.

[0113] In some embodiments, the sensor kit further includes a gateway device configured to receive sensor kit packets from an edge device via a wired communication link and transmit the sensor kit packets to a backend system via a public network on behalf of the edge device. In some of these embodiments, the gateway device includes a satellite terminal device configured to transmit the sensor kit packets to a satellite that routes the sensor kit to a public network. Alternatively, in some embodiments, the gateway device includes a cellular chipset pre-configured to transmit the sensor kit packets to a cell phone tower of a pre-selected cellular provider.

[0114] In some embodiments, the edge device's second communication device is a satellite terminal device configured to send sensor kit packets to a satellite that routes the sensor kit to a public network.

[0115] In this embodiment, the edge device further includes one or more storage devices that store a sensor data store that stores instances of sensor data captured by multiple sensors of the sensor kit.

[0116] In some embodiments, the edge device further includes one or more storage devices (memory devices) that store a model data store, each of which is trained to predict or classify the state of an industrial manufacturing environment and / or industrial components of an industrial manufacturing environment based on a set of features derived from instances of sensor data captured by one or more of the multiple sensors. In some embodiments, performing one or more edge operations includes the steps of: generating a feature vector based on one or more instances of sensor data received from one or more of the multiple sensors; inputting the feature vector into a trained model to obtain a prediction or classification related to the state of an industrial manufacturing setup or a particular industrial component of an industrial manufacturing setup, and a confidence level corresponding to the prediction or classification; and selectively encoding one or more instances of sensor data based on the state or prediction before sending them to a backend system. In some of these embodiments, selectively encoding one or more instances of sensor data includes compressing one or more instances of sensor data using a lossy codec in response to obtaining one or more predictions or classifications relating to the state of the industrial manufacturing setup and each of the industrial components of the industrial manufacturing setup, which collectively indicate that there are likely to be no problems relating to the industrial manufacturing setup and any industrial component of the industrial manufacturing setup. In some embodiments, compressing one or more instances of sensor data using a lossy codec includes the steps of normalizing the one or more instances of sensor data to their respective pixel values, encoding each pixel value into a video frame, and compressing a block of video frames using a 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 containing the normalized pixel values ​​of other instances of sensor data.In embodiments, selective encoding one or more instances of sensor data includes compressing one or more instances of sensor data using a lossless codec in response to obtaining predictions or classifications related to the state of a particular industrial component or industrial manufacturing environment that indicate a high probability of a problem related to that particular industrial component or industrial manufacturing environment. In embodiments, selective encoding one or more instances of sensor data includes refraining from compressing one or more instances of sensor data in response to obtaining predictions or classifications related to the state of a particular industrial component or industrial manufacturing environment that indicate a high probability of a problem related to that particular industrial component or industrial manufacturing setup. In embodiments, performing one or more edge operations includes generating a feature vector based on one or more instances of sensor data received from one or more sensors among a plurality of sensors; inputting the feature vector into a machine learning model to obtain predictions or classifications related to the state of a particular industrial component or industrial manufacturing environment of an industrial manufacturing environment, and confidence levels corresponding to the predictions or classifications; and selectively storing one or more instances of sensor data in a storage device of an edge device based on the predictions or classifications. In embodiments, selectively storing one or more instances of sensor data includes storing one or more instances of sensor data on a storage device with an expiration date, such that in response to obtaining one or more predictions or classifications that collectively indicate that there are likely no problems related to the industrial manufacturing setup and the industrial components of the industrial manufacturing setup, one or more instances of sensor data are deleted from the storage device according to the expiration date.In the embodiment, selectively storing one or more instances of sensor data includes storing one or more instances of sensor data indefinitely in a storage device in response to obtaining predictions or classifications related to the state of a particular industrial component or industrial manufacturing setup that indicate a high probability of a problem related to that particular industrial component or industrial manufacturing setup.

[0117] In some embodiments, the self-configured sensor kit network is a star network in which each sensor of a plurality of sensors directly transmits each instance of sensor data to an edge device using a short-range communication protocol. In some of these embodiments, a computer-executable instruction further causes one or more processors in the edge device to initiate the configuration of the self-configured sensor kit network.

[0118] In some embodiments, the self-configuring sensor kit network is a mesh network in which the communication device of each of the multiple sensors is configured to establish a communication channel with at least one other sensor among the multiple sensors, and at least one sensor among the multiple sensors is configured to receive instances of sensor data from one or more other sensors among the multiple sensors and route the received instances of sensor data toward edge devices. In some of these embodiments, a computer executable instruction further causes one or more processors in the edge devices to initiate the configuration of the self-configuring sensor kit network, and the multiple sensors form a mesh network in response to the edge devices initiating the configuration of the self-configuring sensor kit network.

[0119] In some embodiments, the self-configured sensor kit network is a hierarchical network. In some of these embodiments, the sensor kit further includes one or more collection devices configured to receive report packets from one or more of a plurality of sensors and route the report packets to edge devices. In some embodiments, each collection device is installed in a separate room of an industrial production site and collects sensor data from a plurality of sensors located in each room.

[0120] According to some embodiments of the present disclosure, a sensor kit configured to monitor an underwater industrial environment is disclosed. The sensor kit includes an edge device and a plurality of sensors that acquire sensor data and transmit the sensor data via a self-configured 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. At least one of the plurality of sensors includes a sensing component that acquires sensor measurements and outputs instances of sensor data, a processing unit that generates a report packet containing routing data and one or more instances of sensor data based on one or more instances of the sensor data and outputs the report packet, and a communication device configured to receive the report packet from the processing unit and transmit the report packet to the edge device via the self-configured sensor kit network according to a first communication protocol. The plurality of sensors include two or more sensor types selected from the group including infrared sensors, sonar sensors, LIDAR sensors, water permeability sensors, light sensors, strain sensors, rust sensors, biosensors, temperature sensors, chemical sensors, valve integrity sensors, vibration sensors, flow sensors, cavitation sensors, pressure sensors, weight sensors, and camera sensors. The edge device includes a communication system comprising a first communication device that receives report packets from multiple sensors via a self-configured sensor kit network, and a second communication device that transmits sensor kit packets to a backend system via a public network.The edge device receives a report packet from the communication system, performs one or more edge operations on the instances of sensor data in the report packet, generates a sensor kit packet containing at least one instance of sensor data based on the instances of sensor data, outputs the sensor kit packet to the communication system, and the communication system further includes a processing system having one or more processors that execute computer-executable instructions, the communication system sending the report packet to the backend system over a public network.

[0121] In some embodiments, the sensor kit further includes a gateway device configured to receive sensor kit packets from an edge device via a wired communication link and, on behalf of the edge device, transmit the sensor kit packets to a backend system via a public network. In some of these embodiments, the gateway device includes a satellite terminal device configured to transmit the sensor kit packets to a satellite that routes the sensor kit to a public network. Alternatively, in some embodiments, the gateway device includes a cellular chipset pre-configured to transmit the sensor kit packets to a cellular tower of a pre-selected cellular provider.

[0122] In some embodiments, the edge device's second communication device is a satellite terminal device configured to transmit sensor kit packets to a satellite that routes the sensor kit to the public network.

[0123] In the embodiment, the edge device further includes one or more storage devices that store a sensor data store that stores instances of sensor data captured by multiple sensors of the sensor kit.

[0124] In embodiments, the edge device further includes one or more storage devices that store a model data store that stores one or more machine learning models, each trained to predict or classify the state of an underwater industrial environment and / or industrial components of an underwater industrial environment based on a set of features derived from instances of sensor data captured by one or more of the multiple sensors. In some embodiments, performing one or more edge operations includes: generating a feature vector based on one or more instances of sensor data received from one or more of the multiple sensors; inputting the feature vector into a machine learning model to obtain a prediction or classification related to the state of an underwater industrial environment or a particular industrial component of an underwater industrial environment, and a confidence level corresponding to the prediction or classification; and selectively encoding one or more instances of sensor data before sending them to a backend system based on the state or prediction. In embodiments, selectively encoding one or more instances of sensor data includes compressing one or more instances of sensor data using a lossy codec in response to obtaining one or more predictions or classifications related to the state of the underwater industrial environment and each industrial component of the underwater industrial environment, which collectively indicate that there is a high probability that there are no problems related to any of the industrial components of the underwater industrial environment and the underwater industrial setting. In one embodiment, compressing one or more instances of sensor data using a lossy codec includes normalizing one or more instances of sensor data to their respective pixel values, encoding each pixel value into a video frame, and compressing blocks of video frames using a lossy codec, wherein the lossy codec is a video codec, and the blocks of video frames include a video frame and one or more other video frames containing the normalized pixel values ​​of other instances of sensor data.In embodiments, selective encoding one or more instances of sensor data includes compressing one or more instances of sensor data using a lossless codec in response to obtaining a prediction or classification that indicates a high probability of a problem related to a particular industrial component or underwater industrial environment. In embodiments, selective encoding one or more instances of sensor data includes refraining from compressing one or more instances of sensor data in response to obtaining a prediction or classification that indicates a high probability of a problem related to a particular industrial component or underwater industrial environment. In embodiments, performing one or more edge operations includes generating a feature vector based on one or more instances of sensor data received from one or more sensors among a plurality of sensors; inputting the feature vector into a machine learning model to obtain a prediction or classification related to the state of an underwater industrial environment or a particular industrial component of an underwater industrial environment, and a confidence level corresponding to the prediction or classification; and selectively storing one or more instances of sensor data in a storage device of an edge device based on the prediction or classification. In the embodiment, selectively storing one or more instances of sensor data includes storing one or more instances of sensor data on a storage device with an expiration date in response to obtaining one or more predictions or classifications regarding the state of each industrial component of the underwater industrial environment and underwater industrial setting, which collectively indicate that there is a high probability that there are no problems related to any industrial component of the underwater industrial environment and underwater industrial setting, and that one or more instances of sensor data are purged from the storage device according to the expiration date.In the embodiment, selective storage of one or more instances of sensor data includes storing one or more instances of sensor data indefinitely in a storage device in response to obtaining predictions or classifications related to the state of a particular industrial component or underwater industrial setting that indicate a high probability of a problem related to that particular industrial component or underwater industrial setting.

[0125] In some embodiments, the self-configured sensor kit network is a star network in which each sensor of a plurality of sensors directly transmits each instance of sensor data to an edge device using a short-range communication protocol. In some of these embodiments, a computer-executable instruction further causes one or more processors in the edge device to initiate the configuration of the self-configured sensor kit network.

[0126] In some embodiments, the self-configuring sensor kit network is a mesh network in which the communication device of each of the multiple sensors is configured to establish a communication channel with at least one other sensor among the multiple sensors, and at least one sensor among the multiple sensors is configured to receive instances of sensor data from one or more other sensors among the multiple sensors and route the received instances of sensor data toward edge devices. In some of these embodiments, a computer executable instruction further causes one or more processors in the edge devices to initiate the configuration of the self-configuring sensor kit network, and the multiple sensors form a mesh network in response to the edge devices initiating the configuration of the self-configuring sensor kit network.

[0127] In some embodiments, the self-configured sensor kit network is a hierarchical network. In some of these embodiments, the sensor kit further includes one or more collection devices configured to receive report packets from one or more of a plurality of sensors and route the report packets to edge devices. In some of these embodiments, each collection device is installed in a different section of the underwater industrial environment and collects sensor data from a plurality of sensors located in each section.

[0128] According to some embodiments of this disclosure, a system for monitoring industrial settings is disclosed. The system includes a set of sensor kits, each having a set of sensors that are registered with the respective industrial environment and configured to monitor the physical characteristics of the industrial environment. The system also includes a set of communication gateways for communicating instances of sensor values ​​from the sensor kits to a backend system. The backend system is configured to process instances of sensor values ​​and monitor the industrial environment, and upon receiving registration data for the sensor kits to the industrial environment, the backend system automatically configures and displays a dashboard for the owner or operator of the industrial environment. The dashboard provides monitoring information based on instances of sensor values ​​for the industrial environment.

[0129] In some embodiments, sensor kit registration includes an interface for specifying the type of entity or industry setting to monitor. In some of these embodiments, the backend system configures a dashboard based on the registered entity or industry setting type. In some embodiments, the backend system includes analytical functions configured based on the type of entity or industry setting. In some embodiments, the backend system includes machine learning functions configured based on the type of entity or industry setting.

[0130] In one embodiment, the communication gateway is configured to provide a virtual container for instances of sensor values ​​so that only the registered owner or operator of the industrial setting can access the sensor values.

[0131] In this embodiment, when registering a sensor kit to an industrial setting, the user can select a set or parameters for monitoring, and a set of services and functions of the backend system is automatically provisioned based on the selected parameters.

[0132] In the embodiment, at least one of the sensor kit, communication gateway, and backend system includes an edge computing system for automatically calculating metrics for industrial settings based on multiple instances of sensor values ​​from the set of sensor kits.

[0133] In some embodiments, the sensor kit is a self-configurable sensor kit network. In some embodiments, the sensor kit network is a star network in which each sensor of a plurality of sensors directly transmits its respective instance of sensor data to a communication gateway using a short-range communication protocol. In some embodiments, a computer-executable instruction causes one or more processors in the communication gateway device to initiate the configuration of the self-configurable sensor kit network. In some embodiments, the self-configurable sensor kit network is a mesh network in which the communication device of each sensor of a plurality of sensors is configured to establish a communication channel with at least one other sensor of the plurality of sensors, and at least one sensor of the plurality of sensors is configured to receive instances of sensor data from one or more other sensors of the plurality of sensors and route the received instances of sensor data to the communication gateway. In some embodiments, a computer-executable instruction further causes one or more processors in the communication gateway to initiate the configuration of the self-configurable sensor kit network, and in response to the communication gateway initiating the configuration of the self-configurable sensor kit network, the plurality of sensors form a mesh network. In some embodiments, the self-configurable sensor kit network is a hierarchical network.

[0134] According to some embodiments of this disclosure, a system for monitoring industrial environments is disclosed. The system includes a set of sensor kits, each having a set of sensors registered with an industrial environment and configured to monitor the physical characteristics of the industrial environment; and a communication gateway for processing instances of sensor values ​​from the sensor kits and communicating instances of sensor values ​​to a backend system that monitors the industrial environment, the backend system, upon receiving registration data for the sensor kits to the industrial environment, automatically configures and populates a dashboard for the owner or operator of the industrial environment, the dashboard providing monitoring information based on instances of sensor values ​​to the industrial environment. In some embodiments, the registration of the sensor kits includes an interface for specifying the type of entity or industrial setting to monitor. In embodiments, the backend system configures the dashboard based on the type of registered entity or industrial setting. In some embodiments, the backend system includes analytical functions configured based on the type of entity or industrial setting. In embodiments, the backend system includes machine learning functions configured based on the type of entity or industrial setting.

[0135] In some embodiments, the communication gateway is configured to provide a virtual container for instances of sensor values ​​so that only the registered owner or operator of the industrial setting can access the sensor values. In embodiments, when registering a sensor kit to an industrial environment, the user can select a set of parameters for monitoring, and a set of services and capabilities of the backend system is automatically provisioned based on the selected parameters. In some embodiments, at least one of the sensor kit, communication gateway, and backend system includes an edge computing system for automatically calculating metrics of the industrial environment based on multiple instances of sensor values ​​from the set of sensor kits.

[0136] In some embodiments, the sensor kit is a self-configurable sensor kit network. In some of these embodiments, the sensor kit network is a star network such that each sensor of a plurality of sensors directly transmits its respective instance of sensor data to a communication gateway using a short-range communication protocol. In embodiments, a computer-executable instruction causes one or more processors in the communication gateway device to initiate the configuration of the self-configurable sensor kit network.

[0137] In some embodiments, the self-configured sensor kit network is a mesh network in which the communication device of each of a plurality of sensors is configured to establish a communication channel with at least one other sensor among the plurality of sensors, and at least one sensor among the plurality of sensors is configured to receive instances of sensor data from one or more other sensors among the plurality of sensors and route the received instances of sensor data toward a communication gateway. In some of these embodiments, a computer executable instruction further causes one or more processors of the communication gateway to initiate the configuration of the self-configured sensor kit network, and the plurality of sensors form a mesh network in response to the communication gateway initiating the configuration of the self-configured sensor kit network. In some embodiments, the self-configured sensor kit network is a hierarchical network.

[0138] According to some embodiments of the present disclosure, a method for monitoring multiple industrial settings using a set of sensor kits, a set of communication gateways, and a backend system is disclosed. The method includes the steps of: registering each sensor kit of the multiple sensor kits with each of the multiple industrial settings; configuring each sensor kit of the multiple sensor kits to monitor the physical characteristics of each industrial setting to which the sensor kit is registered; having each communication gateway in the set of communication gateways transmit instances of sensor data from each sensor kit of the multiple sensor kits to the backend system; having the backend system process the instances of sensor data received from each sensor kit of the multiple sensor kits; having the backend system automatically configure and populate a dashboard for the owner or operator of each industrial setting when it receives registration data for the sensor kits of the multiple sensor kits; and having the dashboard provide monitoring information based on instances of sensor data for each industrial environment.

[0139] In some embodiments, the step of registering each sensor kit includes providing an interface for specifying the type of entity or industrial environment to monitor. In some of these embodiments, the step of configuring each sensor kit to monitor the physical characteristics of each industrial setting includes the backend system configuring a dashboard based on the type of registered entity or industrial setting. In some embodiments, the backend system includes analytical functions configured based on the type of entity in the industrial setting. In embodiments, the backend system includes machine learning equipment configured based on the type of entity or industrial setting.

[0140] In some embodiments, the method further includes the step of providing a virtual container for instances of sensor data by each of a plurality of communication gateways so that only the registered owner or operator of the respective industrial setting can access the sensor data. In embodiments, when registering a sensor kit to an industrial setting, the user can select a set of parameters for monitoring. In some embodiments, the method further includes the step of having a backend system automatically provision a set of services and capabilities of the backend system based on the selected parameters. In embodiments, at least one of the sensor kits among the plurality of sensor kits, the communication gateways among the plurality of communication gateways, and the backend system includes an edge computing system for automatically calculating metrics for the industrial setting based on multiple instances of sensor data from the set of sensor kits.

[0141] In some embodiments, at least one of the multiple sensor kits is a self-configured sensor kit network comprising multiple sensors. In some of these embodiments, the method further includes the steps of capturing sensor data by the multiple sensors and transmitting the sensor data to an edge device via the self-configured sensor kit network by the multiple sensors. In some embodiments, the step of transmitting sensor data via the self-configured sensor kit network includes each sensor of the multiple sensors directly transmitting an instance of the sensor data to the edge device using a short-range communication protocol, and the self-configured sensor kit network is a star network. In some embodiments, the method further includes the step of initiating the configuration of the self-configured sensor kit network by an edge processing system.

[0142] In some embodiments, the self-configured sensor kit network is a mesh network, and each of the multiple sensors includes a communication device. In some of these embodiments, the method further includes: establishing a communication channel with at least one other sensor of the multiple sensors by the communication device of each sensor of the multiple sensors; receiving instances of sensor data from one or more other sensors of the multiple sensors by at least one of the multiple sensors; and routing the received instances of sensor data to an edge device by at least one of the multiple sensors.

[0143] In some embodiments, the self-configured sensor kit network is a hierarchical network, and the sensor kit includes one or more acquisition devices. In some embodiments, the multiple sensors include a first set of sensors of a first sensor type and a second set of sensors of a second sensor type.

[0144] According to some embodiments of the present disclosure, a sensor kit configured for monitoring an industrial environment is disclosed. The sensor kit includes an edge device, a plurality of sensors including one or more sensors of a first sensor type and one or more sensors of a second sensor type that acquire sensor data and transmit the sensor data over a self-configured sensor kit network, a processing unit that generates and outputs a report packet including routing data and one or more instances of sensor data based on one or more instances of sensor data, and a communication device configured to receive the report packet from the processing unit and transmit the report packet to the edge device over the self-configured sensor kit network according to a first communication protocol, wherein at least one of the plurality of sensors includes a sensing component that acquires sensor measurements and outputs instances of sensor data. The edge device includes a communication system having a first communication device that receives report packets from multiple sensors via a self-configured sensor kit network, a second communication device that transmits sensor kit packets to a backend system via a public network, and a processing system having one or more processors that execute computer-executable instructions causing the processing system to receive report packets from the communication device and generate a data block based on sensor data obtained from the report packets, wherein the data block includes (i) a block header that defines the address of the data block, and (ii) a block body that defines sensor data and the parent address of another data block to which the data block is linked, and transmits the data block to one or more node computing devices that collectively store a distributed ledger composed of multiple data blocks.

[0145] In some embodiments, data block generation includes generating a hash value of the block body. In embodiments, data block generation includes encrypting the block body.

[0146] In some embodiments, the distributed ledger includes one or more conditions related to collected sensor data, and a smart contract that defines one or more actions to be initiated by the smart contract when one or more conditions are met. In some embodiments, the smart contract receives a data block from the sensor kit and determines whether one or more conditions are met based on at least the sensor data stored in the data block. In embodiments, the smart contract corresponds to an insurance company. In some embodiments, the actions defined in the smart contract trigger the transfer of funds to an account related to the operator associated with the sensor kit in response to the meeting of one or more conditions. In embodiments, one or more conditions include a first condition that determines whether the sensor kit has reported a sufficient amount of sensor data, and a second condition that determines whether the reported sensor data indicates that the industrial setting is functioning correctly.

[0147] In some embodiments, the smart contract corresponds to a regulatory body. In some of these embodiments, an action defined in the smart contract triggers the issuance of tokens to an operator associated with a sensor kit in response to the fulfillment of one or more conditions. In embodiments, one or more conditions include a first condition requiring a certain amount of reported sensor data reported by the sensor kit, and a second condition requiring that the reported sensor data comply with reporting rules.

[0148] In some embodiments, the edge device is one of the node computing devices.

[0149] According to some embodiments of this disclosure, a method for monitoring an industrial environment is disclosed using a sensor kit having a plurality of sensors and an edge device including a processing system. The method includes the steps of: the processing system receiving a report packet from each of one or more of the plurality of sensors, each of which includes routing data and one or more instances of sensor data; and the processing system generating a data block based on the sensor data obtained from the report packet, wherein the data block includes (i) a block header defining the address of the data block, and (ii) a block body defining sensor data and the parent address of another data block to which the data block is linked, and the processing system sending the data block to one or more node computing devices that collectively store a distributed ledger composed of a plurality of data blocks. In some embodiments, the step of generating the data block includes the processing system generating a hash value of the block body. In embodiments, the generation of the data block includes the processing system encrypting the block body.

[0150] In some embodiments, the distributed ledger includes one or more conditions related to collected sensor data, and a smart contract that defines one or more actions to be initiated by the smart contract in response to the fulfillment of one or more conditions. In some of these embodiments, the smart contract receives a data block from the sensor kit and determines whether one or more conditions are met based on at least the sensor data stored in the data block. In some embodiments, the smart contract corresponds to an insurance company. In embodiments, the actions defined in the smart contract trigger the transfer of funds to an account related to the operator associated with the sensor kit in response to the fulfillment of one or more conditions. In some embodiments, the one or more conditions include a first condition that determines whether the sensor kit has reported a sufficient amount of sensor data, and a second condition that determines whether the reported sensor data indicates that the industrial setting is functioning correctly.

[0151] In some embodiments, the smart contract corresponds to a regulatory body. In some of these embodiments, the actions defined in the smart contract trigger the issuance of tokens to the operator associated with the sensor kit in response to the fulfillment of one or more conditions.

[0152] In some embodiments, one or more conditions include a first condition requiring a certain amount of reported sensor data to be reported by the sensor kit, and a second condition requiring that the reported sensor data comply with reporting rules.

[0153] In some embodiments, the edge device is one of the node computing devices.

[0154] In some embodiments, the plurality of sensors includes a first set of sensors of a first sensor type and a second set of sensors of a second sensor type.

[0155] According to some embodiments of the present disclosure, a system is disclosed. The system includes a backend system including one or more servers configured to deploy a smart contract to a distributed ledger that defines one or more conditions relating to collected sensor data and one or more actions initiated by the smart contract when the one or more conditions are met, on behalf of a user; and a sensor kit including a plurality of sensors, including one or more sensors of a first sensor type and one or more sensors of a second sensor type, configured to monitor an industrial environment, an edge device, and a plurality of sensors that acquire sensor data and transmit the sensor data over a self-configured sensor kit network, wherein at least one of the plurality of sensors is a sensing component that acquires sensor measurements and outputs instances of sensor data; a processing unit that generates and outputs a report packet including routing data and instances of one or more sensor data based on one or more instances of sensor data; and a communication device configured to receive the report packet from the processing unit and transmit the report packet over the self-configured sensor kit network to the edge device according to a first communication protocol.The edge device includes a communication system having a first communication device that receives report packets from multiple sensors via a self-configured sensor kit network, and a second communication device that transmits sensor kit packets to a backend system via a public network, and a processing system having one or more processors that execute computer-executable instructions causing the processing system to receive report packets from the communication device and generate a data block based on sensor data obtained from the report packets, wherein the data block includes (i) a block header that defines the address of the data block, and (ii) a block body that defines sensor data and the parent address of another data block to which the data block is linked, and transmits the data block to one or more node computing devices that collectively store a distributed ledger composed of multiple data blocks.

[0156] In some embodiments, data block generation includes generating a hash value for the block body. In some embodiments, data block generation includes encrypting the block body.

[0157] In some embodiments, the smart contract receives a data block from the sensor kit and determines whether one or more conditions are met based on at least the sensor data stored in the data block. In some of these embodiments, the smart contract corresponds to an insurance company. In some embodiments, an action defined in the smart contract triggers a transfer of funds to an account associated with the operator associated with the sensor kit in response to the meeting of one or more conditions. In embodiments, one or more conditions include a first condition determining whether the sensor kit has reported a sufficient amount of sensor data, and a second condition determining whether the reported sensor data indicates that the industrial setting is functioning correctly. In some embodiments, the smart contract corresponds to a regulatory body. In embodiments, an action defined in the smart contract triggers the issuance of tokens to the operator associated with the sensor kit in response to the meeting of one or more conditions. In some embodiments, one or more conditions include determining whether the sensor kit has reported the required amount of sensor data as defined by the regulation.

[0158] In some embodiments, the edge device is one of the node computing devices.

[0159] According to some embodiments of the present disclosure, a method is disclosed for monitoring an industrial environment using a sensor kit that communicates with a backend system, wherein the sensor kit comprises a plurality of sensors and edge devices. The method comprises the steps of: the backend system deploying a smart contract to a distributed ledger on behalf of a user, which defines one or more conditions relating to collected sensor data and one or more actions initiated by the smart contract in response to the fulfillment of the one or more conditions; an edge processing system of an edge device receiving a report packet from each of one or more sensors of the plurality of sensors, which includes routing data and one or more instances of sensor data; and the edge processing system generating a data block based on the sensor data obtained from the report packet, wherein the data block includes (i) a block header defining the address of the data block, and (ii) a block body defining sensor data and the parent address of another data block to which the data block is linked, and transmitting the data block to one or more node computing devices that collectively store a distributed ledger composed of a plurality of data blocks.

[0160] In some embodiments, the step of generating a data block includes generating a hash value of the block body by an edge processing system. In embodiments, the step of generating a data block includes encrypting the block body by an edge processing system.

[0161] In some embodiments, a distributed ledger receives data blocks from a sensor kit and determines whether one or more conditions of a smart contract are met based on at least the sensor data stored in the data blocks. In some of these embodiments, the smart contract corresponds to an insurance company. In embodiments, an action defined in the smart contract triggers a transfer of funds to an account associated with the operator associated with the sensor kit in response to the meeting of one or more conditions. In some embodiments, one or more conditions include a first condition that determines whether the sensor kit has reported a sufficient amount of sensor data, and a second condition that determines whether the reported sensor data indicates that the industrial setting is functioning correctly.

[0162] In some embodiments, the smart contract corresponds to a regulatory body. In some of these embodiments, the actions defined in the smart contract trigger the issuance of tokens to an operator associated with the sensor kit in response to the fulfillment of one or more conditions. In some embodiments, one or more conditions include determining whether the sensor kit has reported the required amount of sensor data as defined by the regulation. In embodiments, the edge device is one of the node computing devices. In some embodiments, the backend system is one of the node computing devices. In embodiments, the plurality of sensors includes a first set of sensors of a first sensor type and a second set of sensors of a second sensor type.

[0163] A more complete understanding of this disclosure will be derived from the following description and accompanying drawings, as well as the claims. [Brief explanation of the drawing]

[0164] The accompanying drawings, included to provide a better understanding of this disclosure, illustrate embodiments(s) of this disclosure and, together with the description, serve to illustrate the principles of this disclosure.

[0165] [Figure 1] This is a schematic diagram illustrating an example of a sensor kit deployed in an industrial environment according to some embodiments of the present disclosure.

[0166] [Figure 2A] This is a schematic diagram illustrating an example of a sensor kit network having a star network topology according to some embodiments of the present disclosure.

[0167] [Figure 2B] This is a schematic diagram showing an example of a sensor kit network having a mesh network topology according to some embodiments of the present disclosure.

[0168] [Figure 2C] This is a schematic diagram illustrating an example of a sensor kit network having a hierarchical network topology according to some embodiments of the present disclosure.

[0169] [Figure 3A] This is a schematic diagram showing an example of a sensor according to some embodiments of the present disclosure.

[0170] [Figure 3B] This is a schematic diagram showing an example of a report packet schema according to some embodiments of the present disclosure.

[0171] [Figure 4] This is a schematic diagram showing an example of an edge device for a sensor kit according to some embodiments of the present disclosure.

[0172] [Figure 5] This schematic diagram illustrates an example of a backend system that receives sensor data from a sensor kit deployed in an industrial environment, according to some embodiments of the present disclosure.

[0173] [Figure 6]This flowchart shows an example of a set of operations for encoding sensor data captured by a sensor kit, according to some embodiments of the present disclosure.

[0174] [Figure 7] This flowchart shows an example of a set of operations for decoding sensor data provided to a backend system by a sensor kit, according to some embodiments of the present disclosure.

[0175] [Figure 8] This flowchart shows an example of a set of operations for encoding sensor data captured by a sensor kit using a media codec, according to some embodiments of the present disclosure.

[0176] [Figure 9] This flowchart shows an example of a set of operations for decoding sensor data provided by a sensor kit to a backend system using a media codec, according to some embodiments of the present disclosure.

[0177] [Figure 10] This flowchart shows an example of a set of operations for determining a transmission strategy and / or storage strategy for sensor data collected by a sensor kit, according to some embodiments of the present disclosure.

[0178] [Figure 11] This is a schematic diagram illustrating different configurations of a sensor kit according to several embodiments of the present disclosure. [Figure 12] This is a schematic diagram illustrating different configurations of a sensor kit according to several embodiments of the present disclosure. [Figure 13] This is a schematic diagram illustrating different configurations of a sensor kit according to several embodiments of the present disclosure. [Figure 14]This is a schematic diagram illustrating different configurations of a sensor kit according to several embodiments of the present disclosure. [Figure 15] This is a schematic diagram illustrating different configurations of a sensor kit according to several embodiments of the present disclosure.

[0179] [Figure 16] This flowchart shows an example of a set of operations for monitoring an industrial environment using an automatically configured backend system, according to some embodiments of the present disclosure.

[0180] [Figure 17] This is a plan view of a manufacturing facility showing exemplary embodiments of sensor kits including edge devices according to some embodiments of the present disclosure.

[0181] [Figure 18] This is a plan view of a surface portion of an underwater industrial facility showing exemplary embodiments of a sensor kit including an edge device according to some embodiments of the present disclosure.

[0182] [Figure 19] This is a plan view of an indoor agricultural facility showing an exemplary embodiment of a sensor kit including an edge device, according to some embodiments of the present disclosure. [Modes for carrying out the invention]

[0183] Various configurations of sensor kits are disclosed. A sensor kit may be a purpose-configured system containing sensors for monitoring a specific type of industrial environment, and the sensors may be provided in an optionally unified kit along with other devices, systems, and components, such as those providing communication, processing, and intelligence functions. In embodiments, the owner or operator of an industrial environment may purchase or otherwise obtain a sensor kit. During the purchasing process, the owner or operator, or a user associated with the industrial environment, may provide or indicate one or more characteristics of the industrial environment (e.g., the type of environment, the location of the environment, the size of the environment, whether the environment is indoors or outdoors, the components and / or types of components being monitored, the number of each component and / or type of component being monitored, etc.). In embodiments, a sensor kit may be pre-configured based on the characteristics and requirements of the industrial operator or owner. A sensor kit may be pre-configured so that the owner or operator can install the sensor kit in a "plug-and-play" manner, thereby eliminating the need for the owner or operator to configure a sensor kit network through which the devices in the sensor kit communicate.

[0184] Figure 1 - Sensor kit environment

[0185] Figure 1 is a schematic diagram showing an industrial environment 120 in which a sensor kit 100 is installed. In embodiments, the sensor kit 100 may refer to a fully deployable, purpose-configured industrial IoT system provided as a unified kit, ready for deployment to the industrial environment 120 by a consumer entity (e.g., the owner or operator of the industrial environment 120). In embodiments, the sensor kit 100 enables the owner or operator to install and deploy the sensor kit without performing any configuration (e.g., setting user permissions, setting passwords, and / or setting notification and / or display priorities). The term "sensor kit" 100 may refer to a collection of devices installed in an industrial environment 120 (e.g., a factory, mine, oil field, oil pipeline, refinery, commercial kitchen, industrial park, storage facility, construction site, etc.). The collection of devices comprising the sensor kit 100 includes one or more sets of Internet of Things (IoT) sensors 102 and one or more sets of edge devices 104. For the purposes of this discussion, references to "sensors" or "sensor devices" should be understood to mean IoT sensors unless otherwise specified.

[0186] In an embodiment, the sensor kit 100 includes a set of IoT sensors 102 configured to be deployed in, on, or around industrial components, types of industrial components (e.g., turbines, generators, fans, pumps, valves, assembly lines, pipes or pipelines, food inspection lines, server racks, etc.), industrial environments 120, and / or types of industrial environments 120 (e.g., indoors, outdoors, manufacturing, mining, drilling, resource extraction, underground, underwater, etc.), and a set of edge devices capable of processing sensor inputs and providing network-based communication. In an embodiment, the edge device 104 may include or communicate with a local data processing system (e.g., a device configured to compress sensor data, filter sensor data, analyze sensor data, issue notifications based on sensor data, etc.) that can provide local outputs such as signals and analysis results which are the result of local processing. In embodiments, the edge device 104 includes, or is capable of communicating with, a communication system (e.g., a Wi-Fi chipset, a cellular chipset, a satellite transceiver, a cognitive radio, one or more Bluetooth® chips and / or other networking devices) that can communicate data (e.g., raw and / or processed sensor data, notifications, command instructions, etc.) inside and outside the industrial environment. In embodiments, the communication system is configured to operate independently of the primary data or communication network of the industrial environment 120. In embodiments, the communication system includes security features and commands that maintain complete physical and data isolation from the primary data or communication network of the industrial environment 120. For example, in embodiments, a Bluetooth-enabled edge device may be configured to allow pairing only with pre-registered components of the kit, rather than pairing with other Bluetooth-enabled devices within the industrial environment 120.

[0187] In an embodiment, the IoT sensor 102 is a sensor device configured to collect sensor data and communicate the sensor data to another device using at least one communication protocol. In an embodiment, the IoT sensor 102 is configured to be deployed in, on, or around a defined type of industrial entity. The term industrial entity may refer to any object that may be monitored in the industrial environment 120. In an embodiment, the industrial entity may include industrial components (e.g., turbines, generators, fans, pumps, valves, assembly lines, pipes or conduits, food inspection lines, server racks, etc.). In an embodiment, the industrial entity may include living organisms associated with the industrial environment 120 (e.g., humans working in the industrial environment 120, or livestock monitored in the industrial environment 120). Depending on the intended use, configuration, or purpose of the sensor kit 100, the configuration and form factor of the IoT sensor 102 will vary. Examples of different types of sensors include vibration sensors, inertial sensors, temperature sensors, humidity sensors, motion sensors, LIDAR sensors, smoke / fire sensors, current sensors, pressure sensors, pH sensors, light sensors, radiation sensors, etc.

[0188] In an embodiment, the edge device 104 may be a computing device configured to receive sensor data from one or more IoT sensors 102 and to perform one or more edge-related processes associated with the sensor data. Edge-related processes can refer to processes performed on the edge device 104 to store sensor data, reduce communication network bandwidth, and / or reduce the computing resources required by the backend system. Examples of edge processes may include data filtering, signal filtering, data processing, compression, encoding, quick prediction, quick notification, and emergency alerts.

[0189] In some embodiments, the sensor kit 100 is pre-configured so that devices within the sensor kit 100 (e.g., sensor 102, edge device 104, collection device, gateway, etc.) communicate with each other via the sensor kit network without the user having to configure the sensor kit network. The sensor kit network may refer to a closed communication network established among various devices of the sensor kit, utilizing two or more different communication protocols and / or communication media to enable communication of data between devices and to broader communication networks such as public communication networks 190 (e.g., the Internet, satellite networks, and / or one or more cellular networks). For example, some devices in the sensor kit network may communicate using the Bluetooth communication protocol, while other devices may communicate with each other using the Near Field Communication Protocol, Zigbee® protocol, and / or Wi-Fi communication protocol. In some implementations, the sensor kit 100 may be configured to establish a mesh network having various devices acting as routing nodes within the sensor kit network. For example, sensor 102 may be configured to collect data and transmit the collected data to edge device 104 via the sensor kit network, or it may be configured to receive and route data packets from other sensors 102 in the sensor kit network toward edge device 104.

[0190] In embodiments, the sensor kit network may include additional types of devices. In embodiments, the sensor kit 100 may include one or more collection devices (not shown in Figure 1) that function as routing nodes in the sensor network, so that the collection devices become part of a mesh network. In embodiments, the sensor kit 100 may include a gateway device (not shown in Figure 1) that enables communication with a wider network, so that the gateway device communicates with the edge device 104 via a wired or wireless communication medium in an industrial environment 120 (e.g., inside a factory with very thick concrete walls) where the edge device 104 would be prevented from communicating with the public communication network 190. Embodiments of the sensor kit 100 may include additional devices without departing from the scope of this disclosure.

[0191] In one embodiment, the sensor kit 100 is configured to communicate with a backend system 150 via a communication network such as a public communication network 190. In another embodiment, the backend system 150 is configured to receive sensor data from the sensor kit 100 and to perform one or more backend operations on the received sensor data. Examples of backend operations include storing the sensor data in a database, performing analytical tasks on the sensor data, providing the user with the analytical results and / or visualization of the sensor data via a portal and / or dashboard, training one or more machine learning models using the sensor data, determining predictions and / or classifications related to the operation of the industrial environment 120 and / or industrial devices in the industrial environment 120 based on the sensor data, controlling aspects and / or industrial devices in the industrial environment 120 based on the predictions and / or classifications, and issuing notifications to the user via a portal and / or dashboard based on the predictions and / or classifications.

[0192] In some embodiments, it is understood that the sensor kit 100 may provide additional types of data to the backend system 150. For example, the sensor kit 100 may provide diagnostic data indicating detected problems (e.g., malfunction, low battery level, etc.) or potential problems of the sensor 102 or other devices within the sensor kit 100.

[0193] In some embodiments, the sensor kit 100 is configured to self-monitor for a faulty component (e.g., a faulty sensor 102) and report the faulty component to the operator. For example, in some embodiments, the edge device 104 may be configured to detect a failure of sensor 102 based on the absence of reports from the sensor, the absence of a response to a request (e.g., "ping"), and / or unreliable data (e.g., data regularly deviating from expected sensor readings). In some embodiments, the edge device 104 can maintain a sensor kit network map showing the location of each device in the sensor kit network and provide the user with the approximate location and / or identifier of the faulty sensor.

[0194] In some embodiments, the sensor kit 100 may be implemented to allow post-installation configuration. Post-installation configuration may refer to updating the sensor kit 100 by adding devices and / or services to the sensor kit 100 after it has been installed. In some of these embodiments, a user of the system (e.g., an operator of an industrial environment 120) may subscribe to or purchase certain edge “services.” For example, the sensor kit 100 may be configured to run certain programs installed on one or more devices of the sensor kit 100 only if the user has a valid subscription or ownership right to access programmatically supported edge services. If the user no longer has a valid subscription and / or ownership right, the sensor kit 100 may suspend the execution of those programs. For example, the user may subscribe to unsubscribe from AI-based edge services, mesh networking capabilities, self-monitoring services, compression services, facility notifications, etc.

[0195] In some embodiments, a user can add a new sensor 102 to the post-installation sensor kit in a plug-and-play manner. In some of these embodiments, the edge device 104 and sensor 102 (or other devices added to the sensor kit 100) may include their respective short-range communication capabilities (e.g., a near-field communication (NFC) chip, an RFID chip, a Bluetooth chip, a Wi-Fi adapter, etc.). In these embodiments, sensor 102 may include persistent storage for identification data (e.g., a sensor identifier value) and other data that will be used to add sensor 102 to the sensor kit 100 (e.g., the type of industrial device, supported communication protocols, etc.). In some embodiments, a user can initiate a post-installation addition to the sensor kit 100 by pressing a button on the edge device 104 and / or by bringing sensor 102 near the edge device 104. In some embodiments, in response to the user initiating a post-installation addition to the sensor kit, the edge device 104 may emit a signal (e.g., a radio frequency). The edge device 104 may emit a signal, for example, as a result of a human user pressing a button, or at predetermined time intervals. The emitted signal may trigger a sensor 102 that is close enough to receive the signal, which may transmit the sensor ID of sensor 102 and any other appropriate configuration data (e.g., device type, communication protocol, etc.). In response to sensor 102 transmitting its configuration data (e.g., sensor ID and other relevant configuration data) to the edge device 104, the edge device 104 may add sensor 102 to the sensor kit 102. Adding sensor 102 to the sensor kit 104 may include updating a data store or manifest stored in the edge device 104 that identifies the devices of the sensor kit 100 and their associated data.Non-limiting examples of data that may be stored in the manifest associated with each sensor 102 include the communication protocol used by the sensor 102 to communicate with the edge device 104 (or intermediate device), the type of sensor data provided by the sensor 102 (e.g., vibration sensor data, temperature data, humidity data, etc.), the model used to analyze the sensor data from the sensor 102 (e.g., model identifier), and the alarm limits associated with the sensor 102.

[0196] In embodiments, the sensor kit 100 (e.g., edge device 104) may be configured to update the distributed ledger 162 with sensor data captured by the sensor kit 100. In embodiments, the distributed ledger 162 is a blockchain or any other suitable distributed ledger 162. The distributed ledger 162 may be a public ledger or a private ledger. A private ledger reduces the power consumption requirements for maintaining the distributed ledger 162, while a public ledger consumes more power but provides stronger security. In embodiments, the distributed ledger 162 may be distributed among multiple node computing devices 160. The node computing devices 160 may be any suitable computing device, including physical servers, virtual servers, personal computing devices, etc. In some embodiments, the node computing devices 160 are authorized (e.g., via a consensus mechanism) before the node computing devices 160 join the distributed ledger. In some embodiments, the distributed ledger 162 may be stored privately. For example, the distributed ledger may be stored among a pre-approved set of node computing devices so that the distributed ledger 162 cannot be accessed by unauthorized devices. In some embodiments, the node computing devices 160 are edge devices 104 of the sensor kit 102 and other sensor kits 102.

[0197] In an embodiment, the distributed ledger 162 consists of a set of linked data structures (e.g., blocks, data records, etc.) such that the linked data structures form a non-periodic graph. For convenience of explanation, the data structures will be referred to as blocks. In an embodiment, each block may include a header containing the block's unique ID, a body containing the data stored in the block, and a pointer. In an embodiment, the pointer is the block ID of the block's parent block, the parent block being a block created before the block was written to. The data stored in each block may be sensor data captured by each sensor kit 100. Depending on the implementation, the type and amount of sensor data stored in each body of the block may differ. For example, a block may store a set of sensor measurements from one or more types of sensors 102 of the sensor kit 100 captured over a period of time (e.g., sensor data 102 captured from all sensors 102 of the sensor kit 100 over a period of time of one hour or one day), and associated metadata (e.g., the sensor identifier for each sensor measurement, and the timestamp for each sensor measurement or group of sensor measurements). In some embodiments, the block may store sensor measurements that have been determined to be abnormal (e.g., sensor measurements outside the standard deviation of expected sensor measurements or deltas exceeding a threshold), and / or sensor measurements indicating a problem or potential problem, as well as associated metadata (e.g., the sensor ID for each sensor measurement and a timestamp for each sensor measurement or group of sensor measurements). In some embodiments, the sensor data stored in the block may be compressed and / or encoded sensor data such that the edge device 104 compresses / encodes the sensor data into a more compact format. In embodiments, the edge device 104 may generate a hash of the body such that the contents of the body (e.g., the block ID of the parent block and the sensor data) are hashed and cannot be altered without changing the hash value.In one embodiment, the edge device 104 may encrypt the content within the block to prevent unauthorized devices from reading the content.

[0198] As described above, the distributed ledger 162 may be used for different purposes. In some embodiments, the distributed ledger 162 may further include one or more smart contracts. A smart contract is a self-executing digital contract. A smart contract may include code (e.g., executable instructions) that defines one or more conditions that trigger one or more actions. A smart contract may be written by a developer in a scripting language (e.g., JavaScript), an object-code language (e.g., Java), or a compiled language (e.g., C++ or C). The written smart contract may be encoded into a block and deployed to the distributed ledger 162. In embodiments, the backend system 150 is configured to receive smart contracts from users and write the smart contracts to their respective distributed ledgers 162. In embodiments, the address of the smart contract (e.g., the block ID of the block containing the smart contract) may be provided to one or more parties to the smart contract so that each party can invoke the smart contract using the address. In some embodiments, a smart contract may include an API that enables the parties to provide data (e.g., block addresses) and / or send data (e.g., instructions to transfer funds to an account).

[0199] In an exemplary implementation, the insurance company may allow the owner and / or operator, who is insured of the industrial environment 120, to agree to share sensor data with the insurance company to prove that the equipment in the facility is functioning properly, and in return, the insurance company may issue a rebate or refund to the owner and / or operator if the owner and / or operator complies with the agreement with the insurance company. Compliance with the contract may be verified electronically by participant nodes of the distributed ledger and / or sensor kit 100 via a smart contract. In an embodiment, the insurance company may deploy a smart contract that triggers the issuance of a rebate or refund of a portion of the premium if the sensor kit 100 provides the insurance company via the distributed ledger with sufficient sensor data indicating that the facility is functioning properly (for example, by adding the smart contract to the distributed ledger 162). In some of these embodiments, the smart contract may include a first condition requiring that a certain amount of sensor data be reported from the facility, and a second condition that each instance of the sensor data is equal to a value (e.g., no classified or predicted problems) or a range of values ​​(e.g., all sensor measurements are within a predefined range of values). In some embodiments, the action taken in response to one or more conditions being met may be to deposit funds (e.g., via wire transfer or cryptocurrency) into an account. In this example, the edge device 104 may write blocks containing the sensor data to a distributed ledger. The edge device 104 may also provide the smart contract with the addresses of these blocks (e.g., using the smart contract's API). Once the smart contract has verified the first and second conditions of the contract, the smart contract may initiate the transfer of funds from the insurer's account to the insured's account.

[0200] In another example, a regulatory body (e.g., a state, local, or federal regulatory body) may require facility operators to report sensor data to ensure compliance with one or more regulations. For example, a regulatory body may regulate food testing facilities, pharmaceutical manufacturing facilities (e.g., manufacturing facility 1700), indoor farming facilities (e.g., indoor farming facility 1800), offshore oil drilling facilities (e.g., underwater industrial facility 1900), etc. In an embodiment, the regulatory body may deploy a smart contract configured to receive and verify sensor data from the industrial environment 120, and in response to verifying the sensor data, issue a compliance token (or certificate) to the facility owner's account. In some of these embodiments, the smart contract may include a condition requiring that a certain amount of sensor data be reported from the facility, and a second condition requiring that the sensor data comply with reporting rules. In this example, the edge device 104 may write blocks containing the sensor data to the distributed ledger 162. The edge device 104 may also provide the smart contract with the addresses of these blocks (e.g., using the smart contract's API). Once the smart contract has verified the first and second conditions of the contract, the smart contract may generate a token indicating compliance by the facility operator and initiate the transfer of funds to an account associated with the facility (e.g., a digital wallet).

[0201] The distributed ledger 162 can be adapted to additional or alternative applications without departing from the scope of this disclosure.

[0202] Figures 2A, 2B, and 2C - Components and Networking

[0203] Figures 2A, 2B, and 2C show examples of the configuration of the sensor kit network 200. Depending on the sensor kit 100 and the industrial environment 120 in which the sensor kit 100 is installed, the sensor kit network 200 can communicate in different ways.

[0204] Figure 2A shows an example of a sensor kit network 200A, which is a star network. In these embodiments, the sensor 102 communicates directly with the edge device 104. In these embodiments, the communication protocol(s) used by the sensor device 102 and the edge device 104 to communicate depends on one or more of the physical area of ​​the sensor kit network 102, the available power supply, and the type of sensor 102 in the sensor kit 100. For example, in a setting where the area to be monitored is relatively small and the sensor 102 cannot be connected to a power supply, the sensor 102 may be manufactured as a Bluetooth Low Energy (BLE) microchip that communicates using the Bluetooth Low Energy protocol (e.g., the Bluetooth 5 protocol maintained by the Bluetooth Special Interest Group). In another example, in a relatively small area where many sensors 102 are deployed, the sensor 102 may be manufactured as a Wi-Fi microchip that communicates using the IEEE 802.11 protocol. In the embodiment of Figure 2A, the sensor 102 may be configured to perform unidirectional or bidirectional communication. In embodiments where the edge device 104 does not need to communicate data and / or commands to the sensor 102, the sensor 102 may be configured for one-way communication. In embodiments where the edge device 104 communicates data and / or commands to the sensor 102, the sensor 102 may be configured as a transceiver for two-way communication. The star network may be configured as a device having other suitable communication devices without departing from the scope of this disclosure.

[0205] Figure 2B shows an exemplary sensor kit network 200B, which is a mesh network in which nodes (e.g., sensor 102) are directly, dynamically, and / or non-hierarchically connected to each other to cooperate in order to efficiently route data to and from edge devices 104. In some embodiments, the devices in the mesh network (e.g., sensor 102, edge device 104, and / or any other devices in sensor kit network 200B) may be configured to self-organize and self-configure the mesh network so that sensor 102 and / or edge device 104 determine which devices will route data on behalf of other devices and / or redundancy for transmission in the event of a routing node (e.g., sensor 102) failure. In embodiments, the sensor kit 100 may be configured to implement the mesh network in an industrial environment 120 where the area to be monitored is relatively large (e.g., a radius greater than 100 meters from edge devices 104), and / or where the sensors 102 in the sensor kit 100 are intended to be installed in close proximity to each other. In the latter scenario, the power consumption of each individual sensor 102 can be reduced because the distance that each individual sensor 102 needs to transmit is relatively small compared to that of sensors 102 in a star network. In embodiments, the sensor 102 may be manufactured from a Zigbee microchip, a Digi XBee® microchip, a Bluetooth Low Energy microchip, and / or any other suitable communication device configured to participate in a mesh network.

[0206] Figure 2C shows an example of a sensor kit network 200C, which is a hierarchical network. In these embodiments, the sensor kit 100 includes a set of acquisition devices 206. The acquisition devices 206 may refer to non-sensor devices that receive sensor data from the sensor device 104 and route the sensor data to the edge device 104, either directly or via another acquisition device 206. In embodiments, the hierarchical network may refer to a network topography in which one or more intermediate devices (e.g., acquisition devices 206) route data from one or more respective peripheral devices (e.g., sensor device 102) to a central device (e.g., edge device 104). The hierarchical network may include wired and / or wireless connections. In embodiments, the sensor device 102 may be configured to communicate with the acquisition device 206 via any suitable communication device (e.g., a Bluetooth Low Energy microchip, a Wi-Fi microchip, a Zigbee microchip, etc.). In one embodiment, the hierarchical sensor kit network may be implemented in an industrial environment 120 where a power supply is available to power the data acquisition device 206 and / or where the sensors 102 are likely to be too far apart to support a reliable mesh network.

[0207] Examples in Figures 2A to 2C are provided for different topologies of the sensor kit network. These examples are not intended to limit the types of sensor kit networks 200 that can be formed by the sensor kits 100. Furthermore, the sensor kit network 200 may be configured as a hybrid of a star network, a hierarchical network, and / or a mesh network, depending on the industrial environment 120 in which each sensor kit 200 is deployed.

[0208] Figures 3A, 3B, 4, and 5 - Example configuration of sensors, edge devices, and backend systems

[0209] Figure 3A shows an exemplary IoT sensor 102 (or sensor) according to an embodiment of the present disclosure. Embodiments of the IoT sensor 102 may include, but are not limited to, one or more sensing components 302, one or more storage devices 304, one or more power supplies 306, one or more communication devices 308, and a processing device 310. In an embodiment, the processing device 310 may perform an edge reporting module 312.

[0210] Sensor 102 includes at least one sensing component 302. The sensing component 302 may be any digital, analog, chemical, and / or mechanical component that outputs raw sensor data to the processing device 310. It is understood that different types of sensors 102 may be manufactured with different types of sensing components. In embodiments, the sensing component 302 of an inertial sensor may include one or more accelerometers and / or one or more gyroscopes. In embodiments, the sensing component 302 of a temperature sensor may include one or more thermistors or other temperature sensing mechanisms. In embodiments, the sensing component 302 of a heat flux sensor may include, for example, a thin-film sensor, a surface-mount sensor, a polymer-based sensor, a chemical sensor, etc. In embodiments, the sensing component 302 of a motion sensor may include a LIDAR device, a radar device, a sonar device, etc. In embodiments, the sensing component 302 of the occupancy sensor may include a surface being monitored for occupancy, a pressure-actuated switch embedded beneath the surface of the occupancy sensor, and / or a piezoelectric element integrated into the surface of the occupancy sensor, such that an electrical signal is generated when an object occupies the surface being monitored for occupancy. In embodiments, the sensing component 302 of the humidity sensor may include a capacitive element (e.g., a metal oxide between electrodes) that outputs an electrical capacitance value corresponding to ambient humidity; a resistive element comprising a salt medium having electrodes on two sides of the medium, thereby whose variable resistance measured at the electrodes corresponds to ambient humidity; and / or a thermal element including a first thermal sensor that outputs the temperature of a dry medium (e.g., dry nitrogen) and a second thermal sensor that outputs the ambient temperature of the sensor's environment, such that the humidity is determined based on the change between the temperature of the dry medium and the ambient temperature, i.e., delta. In embodiments, the sensing component 302 of the vibration sensor may include an accelerometer component, a position sensing component, a torque sensing component, and the like. It should be understood that the list of sensor types and their sensing components is provided as examples.Additional or alternative types of sensors and sensing components may be integrated into sensor 102 without departing from the scope of this disclosure. Furthermore, in some embodiments, sensor 102 of sensor kit 100 may include audio, visual, or audio / visual sensors in addition to non-audio / visual sensors 102 (i.e., sensors that do not capture video or audio). In these embodiments, sensing component 392 may include a camera and / or one or more microphones. In some embodiments, the microphones may be directional microphones such that the direction of a sound source can be determined.

[0211] The storage device 304 may be any suitable medium for storing data to be transmitted to the edge device 104. In embodiments, the storage device 304 may be a persistent storage medium such as a flash memory device. In embodiments, the storage device 304 may be a transient storage medium such as a random access memory device. In embodiments, the storage device 304 may be a circuit configured to store charge, thereby the magnitude of the charge stored by the component indicates a sensed value or incremental count. In these embodiments, this type of storage device 304 may be used when power availability and size are concerns and / or when the sensor data is count-based (e.g., number of detected events). It is understood that any other suitable storage device 304 may be used. In embodiments, the storage device 304 may include a cache 314, which is configured to store sensor data that has not yet been reported to the edge device 104. In these embodiments, the edge reporting module 312 may clear the cache 314 after the sensor data stored in the cache 314 has been transmitted to the edge device 104.

[0212] The power supply 306 is any suitable component that supplies power to other components of the sensor 102, including the sensing component 302, the storage device 304, the communication device 306, and / or the processing device 308. In embodiments, the power supply 306 includes a wired connection to an external power source (e.g., alternating current from a power outlet, or direct current from a battery or photovoltaic power source). In embodiments, the power supply 306 may include a power inverter that converts alternating current to direct current (or vice versa). In embodiments, the power supply 306 may include an integrated power source such as a rechargeable lithium-ion battery or a solar element. In embodiments, the power supply 306 may include a self-generating element such as a piezoelectric element. In these embodiments, the piezoelectric element can output a voltage when sufficient mechanical stress or force is applied to the element. This voltage may be stored in a capacitor or used to power the sensing element 302. In embodiments, the power supply may include an antenna (e.g., a receiver or transceiver) that receives radio frequencies that energize the sensor 102. In these embodiments, the radio frequency may "wake up" the sensor 102 and trigger an action by the sensor 102, such as performing a sensor measurement and / or reporting the sensor data to the edge device 104. The power supply 306 may also include additional or alternative components.

[0213] In some embodiments, the communication device 308 is a device that enables wired or wireless communication with another device in the sensor kit network 200. In most sensor kit configurations 100, the sensor 102 is configured to communicate wirelessly. In these embodiments, the communication device 308 may include a transmitter or transceiver that transmits data to other devices in the sensor kit network 200. Furthermore, in some of these embodiments, the communication device 308 having a transceiver may receive data from other devices in the sensor kit network 200. In wireless embodiments, the transceiver may be integrated into a chip configured to perform communication using the respective communication protocol. In some embodiments, the communication device 308 may be a Zigbee microchip, a Digi XBee microchip, a Bluetooth microchip, a Bluetooth Low Energy microchip, a Wi-Fi microchip, or another suitable short-range communication microchip. In embodiments where the sensor kit 200 supports a mesh network, the communication device 308 may be a microchip implementing a communication protocol that supports mesh networking (e.g., ZigBee PRO mesh networking protocol, Bluetooth Mesh, 802.11a / b / g / n / ac, etc.). In these embodiments, the communication device 308 may be configured to establish a mesh network and handle the routing of data packets received from other devices according to the communication protocol implemented by the communication device 308. In some embodiments, the sensor 102 may consist of two or more communication devices 308. In these embodiments, the sensor 102 may be added to the configuration of different sensor kits 100 and / or allow for flexible configuration of the sensor kit 102 depending on the industrial environment 120.

[0214] In an embodiment, the processing device 310 may be a microprocessor. The microprocessor may include a memory (e.g., read-only memory (ROM)) for storing computer-executable instructions and one or more processors for executing the computer-executable instructions. In an embodiment, the processing device 310 executes the edge reporting module 312. In an embodiment, the edge reporting module 312 is configured to send data to the edge device 104. Depending on the configuration of the sensor kit network 200 and the location of the sensor 102 relative to the edge device 104, the edge reporting module 312 may send the data (e.g., sensor data) directly to the edge device 104 or to an intermediate device (e.g., a collection device 206 or another sensor device 102) that routes the data toward the edge device 104. In an embodiment, the edge reporting module 312 retrieves raw sensor data from the sensing component 302 or from the storage device 304 and packets the raw sensor data into reporting packets 320.

[0215] Figure 3B shows exemplary reporting packets 320 according to several embodiments of the present disclosure. In some of these embodiments, the edge reporting module 312 may take a reporting packet template as input to obtain the reporting packet 320. In embodiments, the reporting packet 320 may include a first field 322 indicating the sensor ID of sensor 102 and a second field 326 indicating sensor data. Furthermore, the reporting packet 320 may include additional fields such as a routing data field 324 indicating the destination of the packet (e.g., the address or identifier of edge device 104), a timestamp field 328 indicating a timestamp, and / or a checksum field 330 indicating a checksum (e.g., a hash value of the contents of the reporting packet). The reporting packet may include additional or alternative fields (e.g., error codes) without departing from the scope of the present disclosure.

[0216] Returning to Figure 3A, in this embodiment, the edge reporting module 312 may generate a report packet 320 for each instance of sensor data. Alternatively, the edge reporting module 312 may generate a report packet 320 containing a batch of sensor data (e.g., the previous N sensor readings, or all sensor readings held in the sensor 102 cache 314 since the cache 314 was last purged). Once the report packet 320 is generated, the edge reporting module 312 may output the report packet 320 to the communication device 308, which then transmits the report packet 320 to the edge device 104 (either directly or via one or more intermediate devices). The edge reporting module 312 may generate and transmit the report packet 320 continuously or triggered at predetermined intervals (e.g., every second, every minute, every hour) or when triggered (e.g., when powered on or when receiving a command from the edge device 104).

[0217] In an embodiment, the edge reporting module 312 instructs the sensing component 302 to capture sensor data. In an embodiment, the edge reporting module 312 may instruct the sensing component 302 to capture sensor data at predetermined intervals. For example, the edge reporting module 312 may instruct the sensing component 302 to capture sensor data every second, every minute, or every hour. In an embodiment, the edge reporting module 312 may instruct the sensing component 302 to capture sensor data when the power supply 306 is energized. For example, the power supply 306 may be energized by a radio frequency or when a pressure switch is activated and closes the circuit. In an embodiment, the edge reporting module 312 may instruct the sensing component 302 to capture sensor data in response to receiving a command to report sensor data from an edge device 104 or a human user (for example, in response to a user pressing a button).

[0218] In embodiments, the sensor 102 includes a housing (not shown). The sensor housing may have any suitable form factor. In embodiments where the sensor 102 is used outdoors, the sensor may have a housing that is waterproof and / or resistant to extreme cold and / or extreme heat. In embodiments, the housing may have a suitable coupling mechanism for detachably coupling to an industrial component.

[0219] The foregoing is an example of sensor 102. Sensor 102 may have additional or alternative components without departing from the scope of this disclosure.

[0220] Figure 4 shows an example of an edge device 104. In this embodiment, the edge device 104 may include a storage system 402, a communication system 404, and a processing system 406. The edge device 104 may also include additional components, such as a power supply and a user interface, which are not shown.

[0221] The storage system 402 includes one or more storage devices. The storage devices may include persistent storage media (e.g., flash memory drives, hard disk drives) and / or transient storage devices (e.g., RAM). The storage system 402 may store one or more datastores. The datastores may include one or more databases, tables, indexes, records, file systems, folders and / or files. In the illustrated embodiment, the storage device stores a configuration datastore 410, a sensor datastore 412, and a model datastore 414. The storage system 402 may store additional or alternative datastores without departing from the scope of the disclosure.

[0222] In some embodiments, the configuration data store 410 stores data relating to the configuration of the sensor kit 100, including the devices of the sensor kit 100. In some embodiments, the configuration data store 410 may maintain a set of device records. A device record may indicate a device identifier that uniquely identifies the devices of the sensor kit 100. A device record may further indicate the type of device (e.g., sensor, collection device, gateway device, etc.). In embodiments where the network path from each device to the edge device 104 remains constant, the device record may also indicate the network path of the device to the edge device 104 (e.g., any intermediate devices in the device's network path). Furthermore, if the device record corresponds to a sensor 102, the device record may indicate the type of sensor (e.g., sensor type identifier) ​​and / or the type of data provided by the sensor 102.

[0223] In one embodiment, the configuration data store 410 may maintain a set of sensor type records, each corresponding to a different type of sensor 102 within the sensor kit 100. The sensor type record may indicate a type identifier that identifies the type of sensor and / or the type of sensor data provided by the sensor. In another embodiment, the sensor type record may further indicate relevant information related to the sensor data, such as the maximum or minimum value of the sensor data and the error code output by the sensor 102 of the sensor type.

[0224] In some embodiments, the configuration data store 410 may maintain a map of the sensor kit network 200. The map of the sensor kit network 200 may show the network topology of the sensor kit network 200, including the network paths of the collection of devices within the sensor kit 100. In some embodiments, the map may also include the physical locations of the sensors. The physical location of sensor 102 may be defined as the room or area in which sensor 102 is located, the specific industrial component that sensor 102 is monitoring, a set of relative coordinates of edge device 104 (e.g., x, y, z coordinates relative to edge device 104, or the angle and distance of sensor 102 relative to edge device 104), the estimated longitude and latitude of sensor 102, or other suitable format for relative or absolute positioning and / or measurement.

[0225] In some embodiments, the sensor data store 412 stores sensor data collected from the sensor 102 of the sensor kit 100. In some embodiments, the sensor data store 412 maintains sensor data collected over a period of time. In some of these embodiments, the sensor data store 412 may be a cache that stores sensor data until it is reported to and backed up by the backend system 150. In these embodiments, the cache may be cleared when the sensor data is reported to the backend system 150. In some embodiments, the sensor data store 412 stores all sensor data collected by the sensor kit 412. In these embodiments, the sensor data store 412 may provide a backup over time of all sensor data collected by the sensor kit 100, thereby ensuring that the owner of the sensor kit 100 retains ownership of the data.

[0226] In embodiments, the model data store 414 stores machine learning models. Machine learning models may include any suitable type of model, including neural networks, deep neural networks, recurrent neural networks, Bayesian neural networks, regression-based models, decision trees, predictor trees, classification trees, hidden Markov models, and / or other suitable types of models. Machine learning models may be trained on training data, which may be expert-created data, historical data, and / or outcome-based data. Outcome-based data may be data collected after a prediction or classification has been made, indicating whether the prediction or classification was correct or incorrect and the resulting outcome. A training data instance may refer to a unit of training data containing a set of features and labels. In embodiments, the labels on a training data instance may indicate the state of an industrial component or industrial environment 120 at a given time. Examples of states vary considerably depending on the industrial environment 120 and the conditions under which the machine learning model is trained to predict or classify. Examples of labels in a manufacturing facility include, but are not limited to, no problems detected, mechanical failure of a component, electrical failure of a component, chemical leak detected, etc. Examples of labels in mining facilities include, but are not limited to, no problem detected, oxygen deficiency, presence of toxic gases, and structural component failure. Examples of labels in oil and gas facilities (e.g., oil fields, gas fields, refineries, pipelines) include, but are not limited to, no problem detected, mechanical failure of components (e.g., a faulty valve or O-ring), and leaks. Examples of labels in indoor agricultural facilities include, but are not limited to, no problem detected, plants dying, plants wilting, plants turning a specific color (e.g., brown, purple, orange, or yellow), and mold being found. In these examples, there are certain features that may be related to the condition and some features that may have little to do with the condition.Through a machine learning process (which may run on the backend system 150 or another system), the model is trained to make predictions or classifications based on a set of features. Thus, the set of features of the training data instance may include sensor data that is temporarily close to the time when the state of the industrial component or industrial environment 120 occurred (e.g., a label associated with the industrial component or industrial environment 120).

[0227] In some embodiments, the machine learning model may include a predictive model used to predict potential problems related to the monitored industrial component. In some of these embodiments, the machine learning model may be trained on training data (expert-generated data and / or historical data) corresponding to one or more states associated with a particular component. In some of these embodiments, the training dataset may include sensor data corresponding to scenarios where maintenance or some intervention was subsequently required, and sensor data corresponding to scenarios where maintenance or some intervention was ultimately not required. In these exemplary embodiments, the machine learning model may be used to determine a prediction of one or more potential problems that may occur with respect to one or more monitored industrial components and / or the monitored industrial environment 120.

[0228] In some embodiments, the machine learning model may include a classification model that classifies the state of the monitored industrial component and / or industrial environment 120. In some of these embodiments, the machine learning model may be trained on training data (e.g., expert-generated data and / or historical data) corresponding to one or more states associated with a particular component. In some of these embodiments, the training dataset may include sensor data corresponding to scenarios in which each industrial component and / or each industrial environment 120 was operating under normal conditions, and sensor data in which each industrial component and / or each industrial environment 120 was operating under abnormal conditions. In training data instances where abnormal conditions occurred, the training data instance may include a label indicating the type of abnormal condition. For example, a training data instance corresponding to an indoor agricultural facility that was determined to have excessively high humidity compared to ideal growing conditions may include a label indicating that the facility's humidity was too high.

[0229] In embodiments, the communication system 404 includes two or more communication devices, including at least one internal communication device that communicates with the sensor kit network 200 and at least one external communication device that communicates with a public communication network (e.g., the Internet) directly or via a gateway device. The at least one internal communication device may include a Bluetooth chip, a Zigbee chip, an XBee chip, a Wi-Fi chip, and the like. The selection of the internal communication device may depend on the environment of the industrial environment 120 and its impact on the sensors 102 installed therein (e.g., whether the sensors 102 have a reliable power supply, whether the sensors 102 are close to each other but spaced apart, whether the sensors 102 need to transmit through walls, etc.). The external communication device may perform wired or wireless communication. In embodiments, the external communication device may include a cellular chipset (e.g., a 4G or 5G chipset), an Ethernet card, a satellite communication card, or other suitable communication device. The external communication device(s) of the edge device 104 may be selected based on the environment of the industrial environment 120 (e.g., indoor versus outdoor, thick walls that obstruct wireless communication versus thin walls that enable wireless communication, located near a cell phone tower versus located in a remote area), and the preferences of the operator of the industrial environment 120 (e.g., the operator allows the edge device 104 to access the private network of the industrial environment 120, or the operator does not allow the edge device 104 to access the private network of the industrial environment 120).

[0230] In embodiments, the processing system 406 may include one or more memory devices (e.g., ROM and / or RAM) for storing computer-executable instructions, and one or more processors for executing computer-executable instructions. The processing system 406 may execute one or more of the data processing module 420, encoding module 422, instant AI module 424, notification module 426, configuration module 428, and distributed ledger module 430. The processing system 406 may execute additional or alternative modules without departing from the scope of this disclosure. Furthermore, the modules described herein may include submodules that perform one or more functions of each module.

[0231] In embodiments, the data processing module 420 receives sensor data from the sensor kit network 200 and performs one or more data processing operations on the received sensor data. In embodiments, the data processing module 420 receives report packets 320 containing sensor data. In some of these embodiments, the data processing module 420 may filter out duplicate data records (for example, filtering out one of two report packets 320 received from two separate sensors monitoring the same component for redundancy). The data processing module 420 may additionally or alternatively filter and / or flag report packets 320 containing sensor data that is clearly erroneous (for example, a sensor whose type is outside a given acceptable range, or a sensor containing an error code). In embodiments, the data processing module 420 may store and / or index the sensor data in a sensor data store.

[0232] In an embodiment, the data processing module 420 may aggregate sensor data received from the sensors 102 of the sensor kit 100 or a subset thereof over a period of time and transmit the sensor data to the backend system 150. When transmitting the sensor data to the backend system 150, the data processing module 420 may generate a sensor kit report packet containing one or more instances of the sensor data. The sensor data in the sensor kit report packet may be compressed or not. In an embodiment, the sensor kit report packet may indicate a sensor kit identifier that identifies the source of the data packet to the backend system 150. In an embodiment, when the data processing module 420 receives sensor data from the sensor 102, it may transmit the sensor data at predetermined intervals (e.g., every second, every minute, every hour, every day) or in response to a trigger condition (e.g., a prediction or classification that there is a problem with an industrial component or industrial environment 120 based on the received sensor data). In some embodiments, the sensor data may be encoded / compressed so that sensor data collected from multiple sensors 102 and / or over a period of time is transmitted more efficiently. In embodiments, the data processing module 420 may utilize the quick-decision AI module 424 to determine whether the industrial components of the industrial environment 120 and / or the industrial environment 120 itself are likely to be in a normal state. If the quick-decision AI module 424 determines with high certainty that the industrial components and / or the industrial environment 120 are in a normal state, the data processing module 420 may delay or skip sending the sensor data used for classification to the backend system 150. In addition or alternatively, if the quick-decision AI module 424 determines with high certainty that the industrial components and / or the industrial environment 120 are in a normal state, the data processing module 420 may compress the sensor data, and may compress it at a higher rate. The data processing module 420 may perform additional or alternative functions without departing from the scope of this disclosure.

[0233] In embodiments, the encoding module 422 may receive sensor data and encode, compress, and / or encrypt the sensor data. The encoding module 422 may employ other techniques to compress the sensor data. In embodiments, the encoding module 422 may employ horizontal or compression techniques to compress the sensor data. For example, the encoding module 422 may use the Lempel-Zev-Welch algorithm or a variation thereof. In some embodiments, the encoding module 522 may represent the sensor data in its original integer or "count format" and with the associated calibration coefficients and offsets at the time of acquisition. In these embodiments, the coefficients and offsets may be combined at the time of acquisition, when the exact signal path is known, so that one floating-point coefficient and one integer offset are stored for each channel.

[0234] In some embodiments, the encoding module 422 may employ one or more codecs to compress the sensor data. The codecs may be proprietary codecs and / or publicly available codecs. In some embodiments, the encoding module 422 may compress the sensor data using a media compression codec (e.g., a video compression codec). For example, the encoding module 422 may normalize the sensor data to values ​​that fit within the range and format of a media frame (e.g., normalize the sensor data to acceptable pixel values ​​for inclusion in a video frame) and embed the normalized sensor data into the media frame. The encoding module 422 may embed the normalized sensor data collected from the sensor 102 of the sensor kit 100 into the media frame according to a predefined mapping (e.g., a mapping of each sensor 102 to one or more pixels in the media frame). The encoding module 422 may thus generate a set of consecutive media frames and compress the media frames using a media codec (such as an H.264 / MPEG-4 codec, an H.265 / MPEG-H codec, an H.263 / MPEG-4 codec, or a proprietary codec) to obtain an encoding of the sensor data. The encoding module 422 then sends the sensor data encoding to a backend system, which may decompress and recalculate the sensor data based on normalized values. In these embodiments, the codec used for compression and the mapping to the sensor pixels may be selected to reduce lossiness or to increase the compression ratio. Furthermore, the aforementioned techniques may be applied to sensor data that tends to be more static and less varied between samples, and / or sensor data collected from different sensors that tend to vary less when sampled simultaneously. The encoding module 422 may employ additional or alternative encoding / compression techniques without departing from the scope of this disclosure.

[0235] In an embodiment, the instant AI module 424 may utilize a limited set of machine learning models to generate predictions and / or classifications of the state of the monitored industrial component and / or the state of the monitored industrial environment 120. In an embodiment, the instant AI module 424 may receive a set of features (e.g., one or more sensor data values) and request a specific type of prediction or classification based on them. In an embodiment, the instant AI module 424 may leverage a machine learning model corresponding to the requested prediction or classification. The instant AI module 424 may generate a feature vector based on the received features, such that the feature vector includes one or more sensor data values ​​obtained from one or more sensors 102 of the sensor kit 100. The instant AI module 424 may feed the feature vector to a machine learning model. The machine learning model may output the prediction or classification and the confidence level of the prediction or classification. In an embodiment, the instant AI module 424 may output the prediction or classification to the data processing module 420 (or another module that requested the prediction or classification). For example, in one embodiment, the data processing module 420 may use a classification that the industrial component and / or industrial environment 120 is in a normal state to delay or skip the transmission of sensor data and / or compress the sensor data. In another embodiment, the data processing module 420 may use a prediction or classification that the industrial component and / or industrial environment 120 is likely to encounter a failure to send uncompressed sensor data to the backend system 150, which may further analyze the sensor data and / or notify a human user of the potential problem.

[0236] In some embodiments, the notification module 426 may provide a notification or alarm to the user based on sensor data. In some of these embodiments, the notification module 426 may apply a set of rules that trigger a notification or alarm when certain conditions are met. A state may define a sensor data value that strongly correlates with an undesirable (e.g., emergency) state. Upon receiving sensor data from the data processing module 420, the notification module 426 may apply one or more rules to the sensor data. If the conditions for triggering an alarm or notification are met, the notification module 426 may issue an alarm or notification to a human user. The manner in which the alarm or notification is provided to the human user (e.g., by providing it to a user device or by triggering an audible alarm) may be predefined or, in some embodiments, defined by an operator of the industrial environment 120.

[0237] In an embodiment, the configuration module 428 constitutes the sensor kit network 200. In an embodiment, the configuration module 428 may send configuration requests to other devices of the sensor kit 100 when the sensor 102, edge device 104, and any other devices installed in the industrial environment 120 are present. In some of these embodiments, the sensor 102 and / or other devices may establish a mesh network or a hierarchical network in response to the configuration request. In an embodiment, the sensor 102 and other devices in the sensor kit network may respond to the configuration request. In an embodiment, the configuration module 428 may generate a device record corresponding to the responding device based on the device ID of those devices and any additional data provided in response to the configuration request.

[0238] In some embodiments, the configuration module 428 adds a new device to the sensor kit 100. In these embodiments, the configuration module 428 adds a new sensor 102 to the installed sensor kit 100 in a plug-and-play manner. In some of these embodiments, the communication devices 404, 308 of the edge device 104 and sensor 102 (or other devices added to the sensor kit 100) may include short-range communication capabilities (e.g., Near Field Communication (NFC) chips). In these embodiments, sensor 102 may include persistent storage for identification data (e.g., sensor ID value) and other data that will be used to add the sensor to the sensor kit (e.g., device type, supported communication protocol, etc.). In response to a user initiating an added device to the sensor kit 100 after installation (e.g., the user pressing a button on the edge device 104 and / or bringing sensor 102 near the edge device 104), the configuration module 428 may cause the communication system 404 to transmit a signal (e.g., radio frequency). The transmitted signal may trigger a sensor 102 that is close enough to receive the signal to transmit its sensor ID and any other appropriate configuration data (e.g., device type, communication protocol, etc.). In response to the sensor 102 transmitting its configuration data (sensor ID and other relevant configuration data) to the edge device 104, the configuration module 428 may add the new sensor 102 to the sensor kit 102. In one embodiment, adding the sensor 102 to the sensor kit 104 may include generating a new device record corresponding to the new sensor 102 based on the sensor ID, which updates the configuration data store 410 with the new device record. The configuration module 428 may add the new sensor 102 to the sensor kit 100 in any other appropriate way.

[0239] In the embodiment, the edge device 104 may include a distributed ledger module 430. In the embodiment, the distributed ledger module 430 may be configured to update the distributed ledger 162 with sensor data captured by the sensor kit 100. In the embodiment, the distributed ledger may be distributed among multiple node computing devices 160. As described above, in the embodiment, the distributed ledger 162 consists of a set of linked data structures (e.g., blocks, data records, etc.). For convenience of explanation, we will refer to the data structures as blocks.

[0240] As described, each block may include a header containing the block's unique ID and a body containing the data stored in the block and a pointer to its parent block. In an embodiment, the pointer within a block is the block ID of its parent block. The data stored in each block may be sensor data captured by each sensor kit 100. Depending on the embodiment, the type and amount of sensor data stored in each body of the block may differ. For example, a block may store a set of sensor measurements from one or more types of sensors 102 in the sensor kit 100 captured over a period of time (e.g., sensor data 102 captured from all sensors 102 in the sensor kit 100 over a period of time of one hour or one day), and associated metadata (e.g., the sensor ID of each sensor measurement and a timestamp of each sensor measurement or group of sensor measurements). In some embodiments, a block may store sensor measurements that have been determined to be abnormal (e.g., sensor measurements outside the standard deviation of expected sensor measurements or deltas exceeding a threshold), and / or sensor measurements indicating a problem or potential problem, as well as associated metadata (e.g., the sensor ID for each sensor measurement and a timestamp for each sensor measurement or group of sensor measurements). In some embodiments, the sensor data stored in the block may be compressed and / or encoded sensor data such that the encoding module 422 compresses / encodes the sensor data into a more compact format. In embodiments, the distributed ledger module 430 may generate a hash of the body such that the contents of the body (e.g., the block ID of the parent block and the sensor data) are hashed and cannot be altered unless the hash value is changed. In embodiments, the distributed ledger module 430 may encrypt the contents within the block to prevent the contents from being read by unauthorized devices.

[0241] In this embodiment, the distributed ledger module 430 generates blocks in response to trigger events. Examples of trigger events may include a predetermined time interval (e.g., every minute, every hour, every day), when a potential problem is classified or predicted, or when one or more sensor readings are outside an acceptable threshold. In response to a trigger event, the distributed ledger module 430 may generate blocks based on sensor data to be reported. Depending on the configuration of the server kit 100 and the intended use of the distributed ledger 162, the amount and type of data included in the blocks may vary. For example, in a manufacturing or resource extraction setting such as a manufacturing facility 1700 or an underwater industrial environment 1800, the distributed ledger 162 may be used to demonstrate functional machinery and / or to predict the need for maintenance. In this example, the distributed ledger module 430 may be accessed by an insurer for setting insurance premiums and / or issuing refunds. Therefore, in this example, the distributed ledger module 430 may include any sensor readings (and associated metadata) that are outside the acceptable threshold or instance in which a problem is classified or predicted. In another example, the distributed ledger may be accessible to regulatory bodies to ensure that the facility operates in accordance with one or more regulations. In these embodiments, the distributed ledger module 430 may store one or more sets of sensor measurements (and associated metadata) in blocks so that the sensor measurements can be analyzed by regulatory bodies. In some of these embodiments, the sensor measurements may be compressed to store more sensor data in a single block. In response to generating a block, the distributed ledger module 430 may send the block to one or more node computing devices 160. Once the block is validated (e.g., using a consensus mechanism), each node computing device 160 may update the distributed ledger 162 with the new block.

[0242] As discussed, in some embodiments, the distributed ledger may further include smart contracts. The written smart contracts may be encoded into blocks and deployed to the distributed ledger 162. The address of the smart contract (e.g., the block ID of the block containing the smart contract) may be provided to one or more parties to the smart contract so that each party can invoke the smart contract using the address. In some of these embodiments, the address of the smart contract may be provided to the distributed ledger module 430 so that the distributed ledger module 430 can report items to the smart contract. In some embodiments, the distributed ledger module 430 may leverage the smart contract's API to report items to the smart contract.

[0243] In the implementation example described above, the insurance company may use a smart contract to enable the owner and / or operator of the insured property to certify that the equipment within the property is functioning properly. In some embodiments, the smart contract may trigger the issuance of a rebate or refund of a portion of the premium if the owner and / or operator of the property provides sufficient sensor data demonstrating that the property is functioning correctly. In some of these embodiments, the smart contract may include a first condition requiring that a certain amount of sensor data be reported from the property, and a second condition that each instance of the sensor data is equal to a value (e.g., no classified or predicted problems) or a range of values ​​(e.g., all sensor measurements are within a predefined range of values). In some embodiments, the action may be to deposit funds (e.g., via wire transfer or cryptocurrency) into an account in response to the first and second conditions being met. In this example, the distributed ledger module 430 may write a block containing the sensor data to the distributed ledger 162. The distributed ledger module 430 may also provide the addresses of these blocks to the smart contract (for example, via the smart contract's API). Once the smart contract has confirmed the first and second conditions of the contract, it may initiate the transfer of funds from the insurer's account to the insured's account.

[0244] In another example described above, a regulatory body (e.g., a state, local, or federal regulatory body) may use a smart contract to monitor a facility (e.g., a food testing facility, a pharmaceutical manufacturing facility, an indoor farm facility, an offshore oil drilling facility, etc.) based on reported sensor data and ensure compliance with one or more regulations. In embodiments, the smart contract may be configured to receive and verify sensor data from the facility (e.g., via the smart contract's API) and, in response to verifying the sensor data, issue a compliance token (or certificate) to the facility owner's account. In some of these embodiments, the smart contract may include a first condition requiring a certain amount of sensor data to be reported from the facility, and a second condition requiring that the sensor data comply with reporting rules. In this example, the distributed ledger module 430 may write blocks containing the sensor data to the distributed ledger. The sensor kit 100 may also provide the smart contract with addresses for these blocks (e.g., using the smart contract's API). Once the smart contract has confirmed the first and second conditions of the contract, the smart contract may generate a token indicating compliance by the facility operator and initiate the transfer of funds to an account associated with the facility (e.g., a digital wallet).

[0245] Figure 5 shows an exemplary backend system 150 according to several embodiments of the present disclosure. In embodiments, the backend system 150 may be implemented as a cloud service running on one or more physical server devices. In embodiments, the backend system 150 may include a storage system 502, a communication system 504, and a processing system 506. The backend system 150 may include additional components not shown.

[0246] The storage system 502 includes one or more storage devices. The storage devices may include persistent storage media (e.g., flash memory drives, hard disk drives) and / or transient storage devices (e.g., RAM). The storage system 502 may store one or more datastores. The datastores may include one or more databases, tables, indexes, records, file systems, folders and / or files. In the illustrated embodiment, the storage system 502 stores a sensor kit datastore 510 and a model datastore 512. The storage system 502 may store additional or alternative datastores without departing from the scope of the disclosure.

[0247] In some embodiments, the sensor kit data store 510 stores data associated with each sensor kit 100. In some embodiments, the sensor kit data store 510 may store sensor kit data corresponding to each installed sensor kit 100. In some embodiments, the sensor kit data may indicate devices within the sensor kit 100, including each sensor 102 (e.g., sensor ID) within the sensor kit 100. In some embodiments, the sensor kit data may indicate sensor data captured by the sensor kit 100. In some of these embodiments, the sensor kit data may identify each instance of sensor data captured by the sensor kit 100, and for each instance of sensor data, the sensor kit data may indicate the sensor 102 that captured the sensor data, and in some embodiments, it may indicate a timestamp corresponding to the sensor data.

[0248] In embodiments, the model data store 512 stores machine learning models trained by the AI ​​system 524 based on training data. The machine learning models may include predictive models and classification models. In embodiments, the training data used to train a particular model includes data collected from one or more sensor kits 100 that monitor the same type of industrial environment 120. The training data may additionally or alternatively include historical data and / or expert-generated data. In embodiments, each machine learning model may be associated with a respective type of industrial environment 120. In some of these embodiments, the AI ​​system 524 may periodically update the machine learning models for the types of industrial environments 120 based on sensor data collected from the sensor kits 100 that monitor those types of industrial environments 120 and the results obtained from those industrial environments 120. In embodiments, the machine learning models for the types of industrial environments 120 may be provided to the edge devices 104 of the sensor kits 100 that monitor that type of industrial environment 120.

[0249] In an embodiment, the communication system 504 includes one or more communication devices, including at least one external communication device that communicates with a public communication network (e.g., the Internet). The external communication device may perform wired or wireless communication. In an embodiment, the external communication device may include a cellular chipset (e.g., a 4G or 5G chipset), an Ethernet card and / or a Wi-Fi card, or other suitable communication device.

[0250] In embodiments, the processing system 506 may include one or more memory devices (e.g., ROM and / or RAM) for storing computer-executable instructions, and one or more processors for executing computer-executable instructions. The processors may run in parallel or in a distributed manner. The processors may be located on the same physical server device or on different server devices. The processing system 506 may execute one or more of the decoding module 520, data processing module 522, AI module 524, notification module 526, analysis module 528, control module 530, dashboard module 532, configuration module 534, and distributed ledger management module 536. The processing system 406 may execute additional or alternative modules without departing from the scope of this disclosure. Furthermore, the modules described herein may include submodules that perform one or more functions of each module.

[0251] In some embodiments, the sensor kit 100 may transmit an encoded sensor kit packet containing sensor data to the backend system 150. In these embodiments, the decoding module 520 may receive the encoded sensor data from the edge device 104 and decode, decode, and / or decompress the encoded sensor kit packet to obtain the sensor data and metadata associated with the received sensor data (e.g., the sensor kit ID and one or more sensor IDs of the sensor that captured the sensor data). The decoding module 520 may output the sensor data and other metadata to the data processing module 522.

[0252] In some embodiments, the data processing module 522 may process sensor data received from the sensor kit 100. In some embodiments, the data processing module 522 may receive sensor data and store the sensor data in the sensor kit data store 510 in relation to the sensor kit 100 that provided the sensor data. In some embodiments, the data processing system 522 may provide AI-related requests to the AI ​​module 524. In these embodiments, the data processing system 522 may extract relevant sensor data instances from the received sensor data and provide the extracted sensor data instances to the AI ​​module 524 in a request indicating the type of request (e.g., what type of prediction or classification) and the sensor data to be used. If a potential problem is predicted or classified, the data processing module 522 may execute a workflow related to the potential problem. The workflow may define how the potential problem is handled. For example, the workflow may indicate that a notification should be sent to a human user, that corrective action should be initiated, and / or other appropriate action. The data processing module 522 may perform additional or alternative processing tasks without departing from the scope of this disclosure.

[0253] In embodiments, the AI ​​module 524 trains a machine learning model used to make predictions or classifications. The machine learning model may include any suitable type of model, including neural networks, deep neural networks, recurrent neural networks, Bayesian neural networks, regression-based models, decision trees, predictor trees, classification trees, hidden Markov models, and / or any other suitable type of model. The AI ​​module 524 may train the machine learning model on a training dataset. The training dataset may include expert-generated data, historical data, and / or outcome-based data. Outcome-based data may be data collected after a prediction or classification has been made, indicating whether the prediction or classification was correct or incorrect, and / or the realized outcome. A training data instance may refer to a unit of training data containing a set of features and labels. In embodiments, the labels of a training data instance may indicate the state of an industrial component or industrial environment 120 at a given time. Examples of states may vary considerably depending on the industrial environment 120 and the conditions under which the machine learning model is trained to predict or classify. Examples of labels in manufacturing facilities include, but are not limited to, no problems detected, mechanical failure of a component, electrical failure of a component, or detection of a chemical leak. Examples of labels in mining facilities include, but are not limited to, no problems detected, oxygen deficiency, presence of toxic gases, or failure of structural components. Examples of labels in oil and gas facilities (e.g., oil fields, gas fields, refineries, pipelines) include, but are not limited to, no problems detected, mechanical failure of a component (e.g., a faulty valve or O-ring), or leaks. Examples of labels in indoor agricultural facilities include, but are not limited to, no problems detected, plants dying, plants wilting, plants turning a specific color (e.g., brown, purple, orange, or yellow), or mold being found.In each of these examples, there are certain features that may be related to the state and several features that may be little related to the state. In embodiments, the AI ​​module 524 may enhance the machine learning model as more sensor data and results related to the machine learning model are received. In embodiments, the machine learning models may be stored in the model data store 512. Each model may be stored with a model identifier, which may indicate (e.g., mapped) the type of industrial environment 120 the model performs in, the type of prediction or classification the model performs in, and the features the model receives. In some embodiments, one or more machine learning models (and subsequent updates) may be pushed to each sensor kit 100 so that the edge devices 104 of each sensor kit 100 may use one or more machine learning models to perform prediction and / or classification without relying on the backend system 150.

[0254] In an embodiment, the AI ​​module 524 receives a request for prediction and / or classification and determines a prediction and / or classification based on the request. In an embodiment, the request may indicate the type of prediction or classification being requested and may include a set of features for making the prediction or classification. In response to the request, the AI ​​module 524 may select a machine learning model to utilize based on the type of prediction or classification being requested, thereby the selected model receiving a specific set of features. The AI ​​module 524 may then generate a feature vector containing one or more instances of sensor data and supply the feature vector to the selected model. In response to the feature vector, the selected model may output the prediction or classification and the confidence level of the prediction or classification (e.g., a confidence score). The AI ​​module 524 may output the prediction or classification and the confidence level therein to the module that provided the request.

[0255] In an embodiment, the notification module 526 may issue a notification to the user and / or the respective industrial environment 120 when a problem is detected in the respective setting. In an embodiment, the notification may be sent to the user's user device indicating the nature of the problem. The notification module 526 may implement an API (e.g., a REST API) so that the user's user device associated with the industrial environment 120 may request a notification from the backend system 150. In response to the request, the notification module 526 may provide any notification to the user device if any. In an embodiment, the notification may also be sent to a device located in the industrial environment 120 so that the device may, in response to the industrial environment 120, generate an alarm in the industrial environment 120.

[0256] In an embodiment, the analysis module 528 may perform analysis-related tasks on sensor data collected by the backend system 150 and stored in the sensor kit data store 510. In an embodiment, the analysis tasks may be performed on sensor data received from individual sensor kits. Furthermore, or alternatively, the analysis tasks may be performed on sensor data. Examples of analysis tasks may be performed on sensor data obtained from various sensor kits 100 monitoring different industrial environments 120. Examples of analysis tasks may include energy utilization analysis, quality analysis, process optimization analysis, financial analysis, predictive analysis, yield optimization analysis, failure prediction analysis, scenario planning analysis, and many others.

[0257] In an embodiment, the control module 530 may control one or more aspects of the industrial environment 120 based on decisions made by the AI ​​system 524. In an embodiment, the control module 530 may be configured to provide commands to devices or systems in the industrial environment 120 to take corrective action in response to the detection of a particular problem. For example, the control module 530 may issue a command to a manufacturing facility to stop an assembly line in response to a determination that a critical component on the assembly line is likely or likely to be faulty. In another example, the control module 530 may issue a command to an agricultural facility to activate a dehumidifier in response to a determination that the humidity level inside the facility is too high. In yet another example, the control module 530 may issue a command to close a valve in an oil pipeline in response to a determination that a component in an oil pipeline downstream of the valve is likely or likely to be faulty. For a particular industrial environment 120, the control module 530 may perform corrective actions defined by a human user associated with the industrial environment 120, in which case the human user can define what conditions trigger the corrective actions.

[0258] In embodiments, the dashboard module 532 presents the dashboard to a human user via a user device 140 associated with the human user. In embodiments, the dashboard provides a graphical user interface that enables a display related to the sensor kit 100 (e.g., an employee in an industrial environment 120) associated with the human user. In these embodiments, the dashboard module 532 may retrieve and display raw sensor data provided by the sensor kit, analytical data related to the sensor data provided by the sensor kit 100, predictions or classifications made by the backend system 150 based on the sensor data, etc.

[0259] In an embodiment, the dashboard module 532 allows a human user to configure aspects of the sensor kit 100. In an embodiment, the dashboard module 532 may present a graphical user interface that allows a human user to configure one or more aspects of the associated sensor kit 100. In an embodiment, the dashboard may allow a user to configure alarm limits for one or more sensor types and / or conditions. For example, a user can define a temperature value at which a notification is sent to a human user. In another example, a user can define a set of conditions that trigger an alarm when predicted by the AI ​​module and / or edge device. In an embodiment, the dashboard allows a user to define which users receive notifications when an alarm is triggered. In an embodiment, the dashboard may allow a user to subscribe to additional functions of the backend system 150 and / or edge device 104.

[0260] In one embodiment, the dashboard may allow a user to add one or more subscriptions to the sensor kit 100. These subscriptions may include access to backend services and / or edge services. The user may select services to add to the sensor kit 100 and provide payment information for payment of these services. Upon verification of payment information, the backend system 150 may provide the sensor kit 100 with access to those functions. Examples of services that can be subscribed to include analytics services, AI services, notification services, and the like. The dashboard may allow the user to perform additional or alternative configurations.

[0261] In the embodiments, the configuration module 534 maintains the configuration of each sensor kit 100. Initially, when the new sensor kit 100 is deployed to the industrial environment 120, the configuration module 534 may update the sensor kit data store 510 with the device ID of each device in the newly installed sensor kit 100. Once the sensor kit data store 510 has been updated to reflect the newly installed sensor kit 100, the backend system 150 may begin storing sensor data from the sensor kit 100. In the embodiments, a new sensor 102 may be added to each sensor kit 100. In these embodiments, the edge device 104 may provide an add request to the backend system 150 when attempting to add a device to the sensor kit 100. In the embodiments, the request may indicate the sensor ID of the new sensor. In response to the request, the configuration module 534 may add the sensor ID of the new sensor to the sensor kit data of the sensor kit 100 in the request within the sensor kit data store 510.

[0262] In some embodiments, the backend system 150 includes a distributed ledger management module 536. In some of these embodiments, the distributed ledger management module 536 enables a user to update and / or configure the distributed ledger. In some of these embodiments, the distributed ledger management module 536 enables a user to define or upload a smart contract. As discussed, a smart contract may include one or more conditions to be validated by the smart contract and one or more actions to be triggered when the conditions are validated. In some embodiments, a user may provide the distributed ledger management module 536 with one or more conditions to be validated via a user interface. In some of these embodiments, a user may provide code that defines the conditions (e.g., JavaScript code, Java code, C code, C++ code, etc.). A user may also provide actions to be performed in response to certain conditions being met. In response to the smart contract being uploaded / created, the distributed ledger management module 536 may deploy the smart contract. In some embodiments, the distributed ledger management module 536 may generate a block containing a smart contract. The block may include a header defining the address of the block and a body containing the address to the previous block and the smart contract. In some embodiments, the distributed ledger management module 536 may determine a hash value based on the body of the block and / or encrypt the block. The distributed ledger management module 536 may send the block to one or more node computing devices 160, which then update the distributed ledger with the block containing the smart contract. The distributed ledger management module 536 may further provide the address of the block to one or more parties who may have access to the smart contract. The distributed ledger management module 536 may perform additional or alternative functions without departing from the scope of this disclosure.

[0263] The backend system 150 may include additional or alternative components, data stores, and / or modules that are not discussed.

[0264] Figures 6-9 - Exemplary methods for encoding and / or decoding sensor data

[0265] Figure 6 shows an exemplary set of operations for method 600 for compressing sensor data obtained by sensor kit 100. In an embodiment, method 600 may be performed by the edge device 104 of sensor kit 100.

[0266] In 610, the edge device 104 receives sensor data from one or more sensors 102 of the sensor kit 100 via the sensor kit network 200. In the embodiment, the sensor data from each sensor 102 may be received in a report packet. Each report packet may include the device identifier of the sensor 102 that generated the report packet and one or more instances of the sensor data captured by the sensor 102. The report packet may include additional data such as a timestamp or other metadata.

[0267] In 612, the edge device 104 processes the sensor data. In an embodiment, the edge device 104 may dedup any duplicate reporting packets. In an embodiment, the edge device 104 may filter out sensor data that is clearly incorrect (e.g., out of acceptable range). In an embodiment, the edge device 104 may aggregate sensor data obtained from multiple sensors 102. In an embodiment, the edge device 104 may perform one or more AI-related tasks, such as determining a prediction or classification related to the state of one or more industrial components in the industrial environment 120. In some of these embodiments, the decision to compress the sensor data may depend on whether the edge device 104 determines that there is a potential problem with the industrial component. For example, the edge device 104 may compress the sensor data if there is no predicted or classified problem. In other embodiments, the edge device 104 may compress any sensor data being sent to the backend system, or a specific type of sensor data (e.g., sensor data obtained from a temperature sensor).

[0268] In 614, the edge device 104 may compress the sensor data. The edge device 104 may employ any suitable compression technique for compressing the sensor data. For example, the edge device 104 may employ a vertical or horizontal compression technique. The edge device 104 may consist of a codec for compressing the sensor data. The codec may be a proprietary codec or an "off-the-shelf" codec.

[0269] In 616, the edge device 104 may transmit the compressed sensor data to the backend system 150. In an embodiment, the edge device 104 may generate a sensor kit packet containing the compressed data. The sensor kit packet may specify the source of the sensor kit packet (e.g., a sensor kit ID or an edge device ID) and may include additional metadata (e.g., a timestamp). In an embodiment, the edge device 104 may encrypt the sensor kit packet before transmitting it to the backend system 150. In an embodiment, the edge device 104 transmits the sensor kit packet directly to the backend system 150 (e.g., via a cellular connection, network connection, or satellite uplink). In another embodiment, the edge device 104 transmits the sensor kit packet to the backend system 150 via a gateway device, and the gateway device transmits the sensor kit packet directly to the backend system 150 (e.g., via a cellular connection or satellite uplink).

[0270] Figure 7 shows an example of the operation of method 700 for processing compressed sensor data received from sensor kit 100. In this embodiment, method 700 is performed by backend system 150.

[0271] In 710, the backend system 150 receives compressed sensor data from the sensor kit. In this embodiment, the compressed sensor data may be received in a sensor kit packet.

[0272] In step 712, the backend system 150 decompresses the received sensor data. In one embodiment, the backend system may use a codec to decompress the received sensor data. Before decompressing the received sensor data, the backend system 150 may decode the sensor kit packet containing the compressed sensor data.

[0273] In 714, the backend system 150 performs one or more backend operations on the decompressed sensor data. These backend operations may include storing the data, filtering the data, performing AI-related tasks on the sensor data, issuing one or more notifications in relation to the results of the AI-related tasks, performing one or more analysis-related tasks, and controlling industrial components of the industrial environment 120.

[0274] Figure 8 shows an example of the operation of method 800 for streaming sensor data from sensor kit 100 to backend system 150. In embodiments, method 800 may be performed by edge device 104 of sensor kit 100.

[0275] In 810, the edge device 104 receives sensor data from one or more sensors 102 of the sensor kit 100 via the sensor kit network 200. In an embodiment, the sensor data from each sensor 102 may be received in a report packet. Each report packet may include the device identifier of the sensor 102 that generated the report packet and one or more instances of the sensor data captured by the sensor 102. The report packet may include additional data such as a timestamp or other metadata. In an embodiment, the edge device 104 may process the sensor data. For example, the edge device 104 may dedup any duplicate report packets and / or filter out sensor data that is clearly incorrect (e.g., out of acceptable range). In an embodiment, the edge device 104 may aggregate the sensor data obtained from multiple sensors 102.

[0276] In 812, the edge device 104 may normalize and / or convert the sensor data to a media frame-compliant format. In embodiments, the edge device 104 may normalize and / or convert each instance of sensor data to a value compliant with the constraints of the media frame that will contain the sensor data. For example, in embodiments where the media frame is a video frame, the edge device 104 may normalize and / or convert the instances of sensor data to an acceptable pixel frame. The edge device 104 may employ one or more mapping and / or normalization functions to convert and / or normalize the sensor data.

[0277] In 814, the edge device 104 may generate blocks of media frames based on the transformed and / or normalized sensor data. For example, in an embodiment where the media frame is a video frame, the edge device 104 may assign each instance of the transformed and / or normalized sensor data to each pixel of the video frame. The manner in which the edge device 104 assigns instances of the transformed and / or normalized sensor data to each pixel may be defined by a mapping that associates each sensor with each pixel value. In an embodiment, the mapping may be defined to minimize the variance between the values ​​of adjacent pixels. In an embodiment, the edge device 104 may generate a series of time-sequenced media frames such that each consecutive media frame corresponds to a subsequent set of sensor data instances.

[0278] In 816, the edge device 104 may encode blocks of media frames. In embodiments, the edge device 104 may employ an encoder of a media codec (e.g., a video codec) to compress blocks of media frames. The codec may be a proprietary codec or a "off-the-shelf" codec. For example, the media codec may be an H.264 / MPEG-4 codec, an H.265 / MPEG-H codec, an H.263 / MPEG-4 codec, a proprietary codec, etc. The codec receives blocks of media frames and generates encoded media blocks based on them.

[0279] In 818, the edge device 104 may transmit the encoded media blocks to the backend system 150. In an embodiment, the edge device 104 may stream the encoded media blocks to the backend system 150. Each encoded block may specify the source of the block (e.g., sensor kit ID or edge device ID) and may include additional metadata (e.g., timestamp and / or block identifier). In an embodiment, the edge device 104 may encrypt the encoded media blocks before transmitting them to the backend system 150. The edge device 104 may transmit the encoded media blocks to the backend system 150 directly (e.g., via a cellular connection, network connection, or satellite uplink) or via a gateway device, which transmits the encoded media blocks directly to the backend system 150 (e.g., via a cellular connection or satellite uplink).

[0280] Edge device 104 may continue to execute the aforementioned method 800 to deliver a stream of live sensor data from the sensor kit. The aforementioned method 900 may be executed in a setting where many sensors are deployed in the environment and the sensors are sampled frequently or continuously. In this way, the bandwidth required to provide sensor data to the backend system is reduced.

[0281] FIG. 9 shows an example of the operation of method 900 for capturing a sensor data stream from edge device 104. In an embodiment, method 900 is executed by a backend system.

[0282] At 910, backend system 150 receives an encoded media block from the sensor kit. Backend system 150 may receive an encoded media block as part of a sensor data stream.

[0283] At 912, backend system 150 uses a decoder corresponding to the codec used to encode the media block to decode the encoded block to obtain a set of consecutive media frames. As described with respect to the encoding operation, the codec may be a proprietary codec or an "off-the-shelf" codec. For example, the media codec may be an H.264 / MPEG-4 codec, an H.265 / MPEG-H codec, an H.263 / MPEG-4 codec, a proprietary codec, and the like. The codec receives the encoded blocks of the media frame, decodes the encoded blocks, and obtains a series of consecutive media frames.

[0284] At 914, the backend system 150 reconstructs sensor data based on the media frame. In an embodiment, the backend system 150 determines normalized and / or transformed sensor values embedded in each respective media frame. For example, in an embodiment where the media frame is a video frame, the backend system 150 may determine pixel values for each pixel within the media frame. The pixel values may correspond to each respective sensor 102 of the sensor kit 100, and the values may represent a normalized and / or transformed instance of the sensor data. In an embodiment, the backend system 150 may reconstruct the sensor data by reversing the normalization and / or transformation of the pixel values. In an embodiment, the backend system 150 may utilize an inverse transformation and / or inverse normalization function to obtain each reconstructed sensor data instance.

[0285] At 918, the backend system 150 performs one or more backend operations based on the reconstructed sensor data. Backend operations may include storing data, filtering data, performing AI-related tasks on the sensor data, issuing one or more notifications related to the results of the AI-related tasks, performing one or more analysis-related tasks, controlling industrial components of the industrial environment 120, and the like.

[0286] FIG. 10 - Exemplary Method for Determining a Transmission Plan

[0287] Figure 10 shows a set of operations of method 1000 for determining a transmission plan and / or storage plan for sensor data collected by sensor kit 100 based on sensor data. The transmission plan may define how the sensor data is transmitted (if any) to a backend system. For example, the sensor data may be compressed using an aggressive lossy codec, compressed using a lossless codec, or transmitted uncompressed. The storage plan may define how the sensor data is stored in edge device 104. For example, the sensor data may be stored permanently (or until a human deletes the sensor data), stored for a certain period (e.g., one year), or discarded. Method 1000 may be performed by edge device 104. Method 1000 may be performed to reduce the network bandwidth consumed by sensor kit 100 and / or to reduce storage constraints in edge device 104.

[0288] In 1010, the edge device 104 receives sensor data from the sensor 102 of the sensor kit 100. The data may be received continuously or intermittently. In one embodiment, the sensor 102 may push sensor data to the edge device 104, and / or the edge device 104 may periodically request sensor data 102 from the sensor 102. In one embodiment, the edge device 104 may process the sensor data upon receipt, including deduplication of the sensor data.

[0289] In some embodiments, the edge device 104 may be configured to perform one or more AI-related tasks before transmission over the satellite uplink. In some of these embodiments, the edge device 104 may be configured to determine, based on sensor data and one or more machine learning models, whether there are likely to be any problems related to any of the components and / or the industrial environment 120.

[0290] In 1012, the edge device 104 may generate one or more feature vectors based on sensor data. The feature vectors may include sensor data from a single sensor 102, a subset of sensors 102, or all sensors 102 of the sensor kit 100. In scenarios where a single sensor or a subset of sensors 102 is included in the feature vectors, a machine learning model may be trained to identify one or more problems related to the industrial component or industrial environment 120, but this may not be sufficient to fully determine that the entire environment is likely to be problem-free / safe. Furthermore, or alternatively, the feature vectors may correspond to a single snapshot in time (e.g., all sensor data in the feature vectors correspond to the same sampling event) or to a period of time (sensor data samples from the most recent sampling event and sensor data samples from previous sampling events). In embodiments where the feature vectors define sensor data from a single snapshot, the machine learning model may be trained to identify potential problems without temporal context. In embodiments where the feature vectors define sensor data over a period of time, the machine learning model may be trained to identify potential problems in the context of what the sensor(s) 102 had previously reported. In these embodiments, the edge device 104 may maintain a cache of sensor data sampled over a predetermined period of time (e.g., previous time, previous day, previous N days) such that the cache is cleared in a first-in, first-out manner. In these embodiments, the edge device 104 may retrieve previous sensor data samples from the cache for use in generating a feature vector having data samples over a certain period of time.

[0291] In 1014, the edge device 104 may input one or more feature vectors to one or more machine learning models. Each model may output predictions or classifications related to the industrial components and / or industrial environment 120, as well as confidence scores related to the predictions or classifications.

[0292] In 1016, the edge device 104 may determine a transmission plan and / or storage plan based on the output of a machine learning model. In some embodiments, the edge device 104 may make decisions relating to the manner in which the sensor data is transmitted to the backend system 150. In some embodiments, the edge device 104 may make decisions relating to the manner in which the sensor data is transmitted to the backend system 150 and / or stored in the edge device. In some of these embodiments, the edge device 104 may compress the sensor data if there are no likely problems across the entire industrial environment 120 and its individual components. For example, if the machine learning model predicts that there are likely no problems and classifies with high confidence (e.g., confidence score greater than 0.98) that there are currently no problems, the edge device 104 may compress the sensor data. Alternatively, in a scenario where the machine learning model predicts that there are likely no problems and classifies with high confidence that there are currently no problems, the edge device 104 may refrain from transmitting the sensor data but store it in the edge device 104 for a predetermined period (e.g., a one-year expiration period). In scenarios where a machine learning model predicts potential problems or classifies current problems, the edge device 104 may transmit sensor data uncompressed or using a lossless compression codec. Furthermore, or alternatively, in scenarios where a machine learning model predicts potential problems or classifies current problems, the edge device 104 may store indefinitely the sensor data used to make the prediction or classification, as well as data collected before and / or after the state was predicted or classified.

[0293] Figures 11-15 - Exemplary sensor kit configuration

[0294] Figure 11 shows an example configuration of a sensor kit 1100 according to several embodiments of the present disclosure. In the illustrated example, the sensor kit 1100 is configured to communicate with a communication network 180 via an uplink 1108 to a satellite 1110. In an embodiment, the sensor kit 1100 of Figure 11 is configured for use in a remote industrial environment 120 where cellular coverage is unreliable or nonexistent. In an embodiment, the sensor kit 1100 may be installed in natural resource extraction, natural resource transport systems, power generation facilities, etc. For example, the sensor kit 1100 may be deployed in oil or natural gas fields, offshore oil drilling rigs, mines, oil or gas pipelines, solar thermal power plants, wind power plants, hydroelectric power plants, etc.

[0295] In the example shown in Figure 11, the server kit 1100 includes an edge device 104 and a set of sensors 102. The sensors 102 may include various types of sensors 102 that differ depending on the industrial environment 120. In the illustrated example, the sensors 102 communicate with the edge device 104 via a mesh network. In these embodiments, the sensors 102 may communicate sensor data to nearby sensors 102 so that the sensor data is propagated to the edge device 104 located in a remote / peripheral area of ​​the industrial environment 120. Although a mesh network is shown, the sensor kit 1100 in Figure 11 may include alternative network topologies such as a hierarchical topology (e.g., some or all of the sensors 102 communicate with the edge device 104 via their respective data collectors) or a star topology (e.g., the sensors 102 communicate directly with the edge devices).

[0296] In the embodiment shown in Figure 11, the edge device 104 includes a satellite terminal equipped with a directional antenna for communicating with a satellite. The satellite terminal may be pre-configured to communicate with an Earth-synchronous satellite or a low Earth orbit satellite. The edge device 104 may receive sensor data from a sensor kit network established by the sensor kit 1100. The edge device 104 may then transmit the sensor data to the backend system 150 via satellite 1110.

[0297] In the embodiments, the configuration of the server kit 1100 is suitable for industrial environments 120 that cover remote areas where external power sources are not abundant. In the embodiments, the sensor kit 1100 may include an external power source such as a battery, rechargeable battery, generator, and / or solar panel. In these embodiments, the external power source may be deployed to power the sensor 102, the edge device 104, and any other devices in the sensor kit 1100.

[0298] In these embodiments, the configuration of the server kit 1100 is suitable for an outdoor industrial environment 120. In these embodiments, the sensor 102, the edge device 104, and other devices of the sensor kit 100 (e.g., a data collection device) may be configured in a weatherproof housing. In these embodiments, the sensor kit 1100 may be deployed in an outdoor environment.

[0299] In embodiments, the edge device 104 may be configured to perform one or more AI-related tasks before transmission over the satellite uplink. In some of these embodiments, the edge device 104 may be configured to determine, based on sensor data and one or more machine-learned models, whether it is likely that there are no problems related to any of the components and / or the industrial environment 120. In embodiments, the edge device 104 may receive sensor data from various sensors and generate one or more feature vectors based on it. The feature vectors may include sensor data from a single sensor 102, a subset of sensors 102, or all of the sensors 102 in the sensor kit 1100. In scenarios where a single sensor or a subset of sensors 102 is included in the feature vectors, the machine-learned models may be trained to identify one or more problems related to the industrial components or the industrial environment 120, but may not be sufficient to fully determine that the entire environment is likely to be problem-free / safe. Furthermore, or alternatively, the feature vectors may correspond to a single snapshot in time (e.g., all sensor data in the feature vectors correspond to the same sampling event) or to a period of time (sensor data samples from the most recent sampling event and sensor data samples from previous sampling events). In embodiments where the feature vector defines sensor data from a single snapshot, the machine learning model may be trained to identify potential problems without temporal context. In embodiments where the feature vector defines sensor data over a period of time, the machine learning model may be trained to identify potential problems in the context of what the sensor(s) 102 previously reported. In these embodiments, the edge device 104 may maintain a cache of sensor data sampled over a predetermined time period (e.g., previous hour, previous day, previous N days) such that the cache is cleared in a first-in, first-out manner.In these embodiments, the edge device 104 may retrieve previous sensor data samples from a cache for use in generating a feature vector having data samples over a certain period of time.

[0300] In some embodiments, the edge device 104 may supply one or more feature vectors to one or more machine-learned models. Each model may output predictions or classifications related to the industrial component and / or industrial environment 120, as well as confidence scores related to the predictions or classifications. In some embodiments, the edge device 104 may make decisions related to how the sensor data is transmitted to and / or stored in the backend system 150. For example, in some embodiments, the edge device 104 may compress the sensor data based on the prediction or classification. In some of these embodiments, the edge device 104 may compress the sensor data if there are no likely problems across the entire industrial environment 120 and its individual components. For example, if a machine-learned model predicts that there are likely no problems and classifies that there are currently no problems with high confidence (e.g., confidence score greater than 0.98), the edge device 104 may compress the sensor data. Alternatively, in scenarios where a machine learning model predicts that there is a high probability of no problem and classifies with high confidence that there is currently no problem, the edge device 104 may refrain from transmitting the sensor data but may store it on the edge device 104 for a predetermined period (e.g., one year). In scenarios where a machine learning model predicts potential problems or classifies current problems, the edge device 104 may transmit the sensor data either uncompressed or using a lossless compression codec. In this way, since the sensor data is mostly compressed or not transmitted at all, the amount of bandwidth transmitted over the satellite uplink may be reduced.

[0301] In embodiments, the edge device 104 may apply one or more rules to determine whether a trigger condition exists. In embodiments, one or more rules may be tailored to identify a potential hazard and / or emergency. In these embodiments, the edge device 104 may trigger one or more notifications or alarms when a trigger condition exists. Furthermore, or alternatively, the edge device 104 may transmit sensor data without compression when a trigger condition exists.

[0302] Figure 12 shows an example configuration of a sensor kit 1200 according to several embodiments of the present disclosure. In the illustrated example, the sensor kit 1200 is configured to include a gateway device 1206 that communicates with a communication network 180 via an uplink 1108 to a satellite 1110. In an embodiment, the sensor kit 1200 of Figure 12 is configured for use in an industrial environment 120 located in a remote area where cellular coverage is unreliable or nonexistent, and the edge device 104 is located in a location where physical transmission to the satellite is unreliable or impossible. In an embodiment, the sensor kit 1100 may be installed in an underground or underwater facility, or a facility with very thick walls. For example, the sensor kit 1100 may be deployed in an underground mine, an underwater oil or gas pipeline, an underwater hydroelectric power plant, etc.

[0303] In the example shown in Figure 12, the server kit 1200 includes an edge device 104, a set of sensors 102, and a gateway device 1206. In an embodiment, the gateway device 1206 is a communication device including a satellite terminal having a directional antenna for communicating with a satellite. The satellite terminal may be pre-configured to communicate with a geosynchronous satellite or a low-Earth orbit satellite. In an embodiment, the gateway device 1206 may communicate with the edge device 104 via a wired communication link 1208 (e.g., Ethernet). The edge device 104 may receive sensor data from a sensor kit network established by the sensor kit 1200. The edge device 104 may then transmit the sensor data to the gateway device 1206 via the wired communication link 1208. The gateway device 1206 may then communicate the sensor data to the backend system 150 via a satellite uplink 1108.

[0304] The sensor 102 may include various types of sensors 102 that differ depending on the industrial environment 120. In the illustrated example, the sensor 102 communicates with edge devices 104 via a mesh network. In these embodiments, the sensor 102 may communicate sensor data to nearby sensors 102 so that the sensor data is propagated to edge devices 104 located in remote / peripheral areas of the industrial environment 120. Although a mesh network is shown, the sensor kit 1200 in Figure 12 may include alternative network topologies such as a hierarchical topology (e.g., some or all of the sensors 102 communicate with edge devices 104 via their respective data acquisition devices) or a star topology (e.g., the sensors 102 communicate directly with the edge devices).

[0305] In an embodiment, the configuration of the server kit 1200 is suitable for an industrial environment 120 that covers remote areas where external power sources are not abundant. In an embodiment, the sensor kit 1200 may include an external power source such as a battery, a rechargeable battery, a generator, and / or a solar panel. In these embodiments, the external power source may be deployed to supply power to the sensor 102, the edge device 104, and any other devices within the sensor kit 1200.

[0306] In an embodiment, the configuration of the server kit 1200 is suitable for an industrial environment 120 underground or underwater. In an embodiment, the sensor 102, the edge device 104, and other devices of the sensor kit 100 (e.g., the collection device) may be configured with a waterproof housing or other airtight housing (to prevent dust from entering the edge device 104 and / or the sensor device 102). Further, since the gateway device 1208 is likely to be located outdoors, the gateway device 1208 may include a weather-resistant housing.

[0307] In embodiments, the edge device 104 may be configured to perform one or more AI-related tasks before transmission over the satellite uplink. In some of these embodiments, the edge device 104 may be configured to determine, based on sensor data and one or more machine-learned models, whether it is likely that there are no problems related to any of the components and / or the industrial environment 120. In embodiments, the edge device 104 may receive sensor data from various sensors and generate one or more feature vectors based on it. The feature vectors may include sensor data from a single sensor 102, a subset of sensors 102, or all of the sensors 102 in the sensor kit 1200. In scenarios where a single sensor or a subset of sensors 102 is included in the feature vectors, the machine-learned models may be trained to identify one or more problems related to the industrial components or the industrial environment 120, but may not be sufficient to definitively determine that the entire environment is likely to be problem-free / safe. Furthermore, or alternatively, the feature vectors may correspond to a single snapshot in time (e.g., all sensor data in the feature vectors correspond to the same sampling event) or to a period of time (sensor data samples from the most recent sampling event and sensor data samples from previous sampling events). In embodiments where the feature vector defines sensor data from a single snapshot, the machine learning model may be trained to identify potential problems without temporal context. In embodiments where the feature vector defines sensor data over a period of time, the machine learning model may be trained to identify potential problems in the context of what the sensor(s) 102 previously reported. In these embodiments, the edge device 104 may maintain a cache of sensor data sampled over a predetermined time period (e.g., previous hour, previous day, previous N days) such that the cache is cleared in a first-in, first-out manner.In these embodiments, the edge device 104 may retrieve previous sensor data samples from a cache for use in generating a feature vector having data samples over a certain period of time.

[0308] In some embodiments, the edge device 104 may supply one or more feature vectors to one or more machine-learned models. Each model may output predictions or classifications related to the industrial component and / or industrial environment 120, as well as confidence scores related to the predictions or classifications. In some embodiments, the edge device 104 may make decisions related to how the sensor data is transmitted to and / or stored in the backend system 150. For example, in some embodiments, the edge device 104 may compress the sensor data based on the prediction or classification. In some of these embodiments, the edge device 104 may compress the sensor data if there are no likely problems across the entire industrial environment 120 and its individual components. For example, if a machine-learned model predicts that there are likely no problems and classifies that there are currently no problems with high confidence (e.g., confidence score greater than 0.98), the edge device 104 may compress the sensor data. Alternatively, in scenarios where a machine learning model predicts that there is a high probability of no problem and classifies with high confidence that there is currently no problem, the edge device 104 may refrain from transmitting the sensor data but may store it on the edge device 104 for a predetermined period (e.g., one year). In scenarios where a machine learning model predicts potential problems or classifies current problems, the edge device 104 may transmit the sensor data either uncompressed or using a lossless compression codec. In this way, since the sensor data is mostly compressed or not transmitted at all, the amount of bandwidth transmitted over the satellite uplink may be reduced.

[0309] In embodiments, the edge device 104 may apply one or more rules to determine whether a trigger condition exists. In embodiments, one or more rules may be tailored to identify a potential hazard and / or emergency. In these embodiments, the edge device 104 may trigger one or more notifications or alarms when a trigger condition exists. Furthermore or alternatively, the edge device 104 may transmit sensor data (via the gateway device 1206) without compression when a trigger condition exists.

[0310] Figure 13 shows example configurations of a server kit 1300 according to several embodiments of the present disclosure. In the example in Figure 13, the server kit 1300 includes an edge device 104, a set of sensors, and a set of data collection devices. In the embodiment, the configuration of the server kit 1300 is suitable for an industrial environment 120 that covers a large area and has abundant power, but where the industrial operator does not wish to connect the sensor kit 1400 to the private network of the industrial environment 120. In the embodiment, the edge device 104 includes a cellular communication device (e.g., a 4G LTE chipset or a 5G LTE chipset) with a transceiver that communicates with a cellular tower 1310. The cellular communication may be pre-configured to communicate with a cellular data provider. For example, in the embodiment, the edge device 104 may include a SIM card registered with a cellular provider having a cellular tower 1310 close to the industrial environment 120. The edge device 104 may receive sensor data from a sensor kit network established by the sensor kit 1400. The edge device 104 may process the sensor data and then transmit it to the backend system 150 via the cellular tower 1310.

[0311] The sensor 102 may include various types of sensors 102 depending on the industrial environment 120. In the illustrated example, the sensor 102 communicates with the edge device 104 via a hierarchical network. In these embodiments, the sensor 102 may also communicate sensor data to a data acquisition device 206, which may communicate sensor data to the edge device 104 via a wired or wireless communication link. The hierarchical network may be deployed in locations where the area to be monitored is rather large (e.g., 40,000 square feet or more) and power supply is abundant, such as factories, power plants, food inspection facilities, and indoor cultivation facilities. Although a hierarchical network is shown, the sensor kit 1300 in Figure 13 may include alternative network topologies such as a mesh topology or a star topology (e.g., the sensor 102 communicates directly with the edge device).

[0312] In embodiments, the edge device 104 may be configured to perform one or more AI-related tasks before transmission over the satellite uplink. In some of these embodiments, the edge device 104 may be configured to determine, based on sensor data and one or more machine-trained models, whether it is likely that there are no problems related to any of the components and / or the industrial environment 120. In embodiments, the edge device 104 may receive sensor data from various sensors and generate one or more feature vectors based on it. The feature vectors may include sensor data from a single sensor 102, a subset of sensors 102, or all of the sensors 102 in the sensor kit 1300. In scenarios where a single sensor or a subset of sensors 102 is included in the feature vectors, the machine-trained models may be trained to identify one or more problems related to the industrial components or the industrial environment 120, but may not be sufficient to definitively determine that the entire environment is likely to be problem-free / safe. Furthermore, or alternatively, the feature vectors may correspond to a single snapshot in time (e.g., all sensor data in the feature vectors correspond to the same sampling event) or to a period of time (sensor data samples from the most recent sampling event and sensor data samples from previous sampling events). In embodiments where the feature vector defines sensor data from a single snapshot, the machine learning model may be trained to identify potential problems without temporal context. In embodiments where the feature vector defines sensor data over a period of time, the machine learning model may be trained to identify potential problems using the context of what the sensor(s) 102 had previously reported. In these embodiments, the edge device 104 may maintain a cache of sensor data sampled over a predetermined time period (e.g., previous hour, previous day, previous N days) such that the cache is cleared in a first-in, first-out manner.In these embodiments, the edge device 104 may retrieve previous sensor data samples from a cache for use in generating a feature vector having data samples over a certain period of time.

[0313] In some embodiments, the edge device 104 may supply one or more feature vectors to one or more machine-learned models. Each model may output predictions or classifications related to the industrial component and / or industrial environment 120, as well as confidence scores related to the predictions or classifications. In some embodiments, the edge device 104 may make decisions related to how the sensor data is transmitted to and / or stored in the backend system 150. For example, in some embodiments, the edge device 104 may compress the sensor data based on the prediction or classification. In some of these embodiments, the edge device 104 may compress the sensor data if there are no likely problems across the entire industrial environment 120 and its individual components. For example, if a machine-learned model predicts that there are likely no problems and classifies that there are currently no problems with high confidence (e.g., confidence score greater than 0.98), the edge device 104 may compress the sensor data. Alternatively, in scenarios where a machine learning model predicts that there is likely no problem and classifies with high confidence that there is currently no problem, the edge device 104 may refrain from transmitting the sensor data but may store it for a predetermined period (e.g., one year). In scenarios where a machine learning model predicts a potential problem or classifies a current problem, the edge device 104 may transmit the sensor data either uncompressed or using a lossless compression codec. In this way, since the sensor data is mostly compressed or not transmitted at all, the amount of bandwidth transmitted over the cellular tower may be reduced.

[0314] In embodiments, the edge device 104 may apply one or more rules to determine whether a trigger condition exists. In embodiments, one or more rules may be tailored to identify a potential hazard and / or emergency. In these embodiments, the edge device 104 may trigger one or more notifications or alarms when a trigger condition exists. Furthermore, or alternatively, the edge device 104 may transmit sensor data without compression when a trigger condition exists.

[0315] Figure 14 shows an example configuration of a server kit 1400 according to several embodiments of the present disclosure. In the example of Figure 14, the server kit 1400 includes an edge device 104, a pair of sensors 102, a pair of data collection devices 206, and a gateway device 1406. In embodiments, the configuration of the server kit 1400 is suitable for an industrial environment 120 that covers a large area and has abundant power, but where the industrial operator does not want to connect the sensor kit 1400 to the private network of the industrial environment 120, and where wireless communication (e.g., cellular communication) is unreliable or impossible due to the walls of the industrial environment 120. In embodiments, the gateway device 1406 is a cellular network gateway device that includes a cellular communication device (e.g., a 4G, 5G chipset) with transceivers that communicate with a cellular tower 1310. The cellular communication may be pre-configured to communicate with a cellular data provi...

Claims

1. A method for monitoring multiple industrial environments using multiple sensor kits, multiple communication gateways, and a backend system, The backend system performs the steps of registering each of the multiple sensor kits with each of the multiple industrial environments, The backend system is configured to monitor the physical characteristics of each of the multiple sensor kits, each of which is registered in the industrial environment. The steps include: each of the multiple communication gateways transmits an instance of sensor data from each of the multiple sensor kits to the backend system; The backend system performs the steps of processing instances of sensor data received from each of the multiple sensor kits, Upon receiving registration data for the multiple sensor kits via the aforementioned backend system, the system automatically configures and inputs a dashboard for the owner or operator of each industrial environment. The dashboard includes the step of providing monitoring information based on instances of sensor data for each industrial environment, At least one of the aforementioned multiple sensor kits is a self-configurable sensor kit network including multiple sensors, The steps include acquiring sensor data with the aforementioned multiple sensors, The steps include the plurality of sensors transmitting sensor data to an edge device via the self-configured sensor kit network, The self-configured sensor kit network is a mesh network, and each of the plurality of sensors includes a communication device. The steps include establishing a communication channel between the communication device of each of the plurality of sensors and at least one other sensor of the plurality of sensors, The steps include: receiving an instance of sensor data from one or more other sensors among the plurality of sensors using at least one of the plurality of sensors; A method further comprising the step of routing an instance of the received sensor data toward the edge device by at least one of the plurality of sensors.

2. The method according to claim 1, characterized in that the step of registering each of the aforementioned sensor kits includes providing an interface for specifying the type of entity or industrial environment to be monitored.

3. The method according to claim 2, wherein the step of configuring each sensor kit to monitor the physical characteristics of each industrial environment includes configuring a dashboard by the backend system based on the type of registered entity or industrial environment.

4. The method according to claim 2, characterized in that the backend system includes an analytical function configured based on the type of entity in the industrial environment.

5. The method according to claim 2, characterized in that the backend system includes machine learning equipment configured based on the type of entity or the industrial environment.

6. The method according to claim 1, further comprising providing a virtual container for an instance of sensor data by each of the plurality of communication gateways, so that only the registered owner or operator of the respective industrial environment can access the sensor data.

7. The method according to claim 1, characterized in that when registering a sensor kit in an industrial environment, the backend system allows the user to select a set of parameters for monitoring.

8. The method according to 7, further comprising the backend system automatically provisioning a set of services and capabilities of the backend system based on selected parameters.

9. The method according to claim 1, wherein at least one of the sensor kits among the plurality of sensor kits, the communication gateway among the plurality of communication gateways, and the backend system includes an edge computing system for automatically calculating metrics for an industrial environment based on multiple instances of sensor data from the plurality of sensor kits.

10. The method according to claim 1, further comprising initiating the configuration of the self-configured sensor kit network by at least one of the plurality of communication gateways.

11. The method according to claim 1, characterized in that the plurality of sensors include a first set of sensors of a first sensor type and a second set of sensors of a second sensor type.

Images