Autonomous situation awareness with ambient and reflexive context and anomaly-driven predicates

A computer-implemented method, comprising:

• accessing a stream of multivariate observations of (i) system variables about a present operational state of an asset, and (ii) command variables about a present configuration of controls of the asset;
• supplementing the multivariate observations with estimates of ambient variables that are estimated from the multivariate observations by a first machine learning model, wherein the first machine learning model is trained to estimate values for the ambient variables that are consistent with sensed values for the ambient variables sensed during a reference operation of the asset;
• determining anomaly statuses of the system variables based on sequential analysis of residuals between the system variables and estimates of the system variables that are estimated from the supplemented multivariate observations by a second machine learning model, wherein the second machine learning model is trained to estimate values for the system variables that are consistent with the reference operation of the asset;
• evaluating the anomaly statuses of the system variables with one or more predicates associated with controls of the asset to select one command variable for adjustment out of the command variables;
• generating a suggested value for the selected command variable based on present and future values of the supplemented multivariate observations using a third machine learning model, wherein the third machine learning model is trained to predict what the future values will be after an interval that are consistent with the reference operation of the asset;
• presenting the selected command variable and the suggested value for human-in-the-loop confirmation;
• and
• automatically adjust the selected command variable to match the suggested value to cause the asset to avoid a hazard.

A system, comprising:

• at least one processor connected to at least one memory;
• a sensor system that is configured to sense physical phenomena associated with an asset and deliver a stream of multivariate observations of (i) system variables about a present operational status of the asset, and (ii) command variables about a present configuration of controls of the asset;
• a human-in-the-loop control interface configured to (a) emphasize a selected control of an asset with respect to other controls, and (b) present recommended values for the selected control of the asset;
• one or more non-transitory computer-readable media that include stored thereon instructions that, when executed by the at least one processor accessing the memory cause the system to:
• supplement the multivariate observations with machine learning estimates of ambient variables based on the system variables and command variables;
• determine anomaly statuses of the system variables based on sequential analysis of residuals between the system variables and machine learning estimates of expected values for the system variables that are estimated from the supplemented multivariate observations;
• evaluate the anomaly statuses with predicates associated with the controls of the asset to select one of the command variables to be adjusted;
• generate a suggested value for the selected command variable based on machine learning predictions of future values for the system variables following an interval;
• in the human-in-the-loop control interface,
• emphasize the selected command variable in relation to other command variables, and
• display the suggested value for the selected command variable;
• and
• automatically adjust the selected command variable to match the suggested value to cause the asset to avoid a hazard.