Unified analytics engine

A method for proactively detecting equipment failures at satellite ground stations, the method comprising:

• obtaining, for each of a plurality of satellite passes, power consumption data representing an amount of power consumed by a digitizer during a respective satellite pass, wherein:
• the digitizer is located at a satellite ground station that is configured to receive data from a satellite during each respective satellite pass,
• the power consumption data comprises time series data indicating a cumulative amount of power consumed by the digitizer at each of a plurality of times during the respective satellite pass, and
• the plurality of satellite passes occur of over a first time period;
• computing, for each of the plurality of satellite passes, a mean power consumption value representing the amount of power consumed by the digitizer during the respective satellite pass;
• generating time series data comprising at least (i) a timestamp of a time that the respective satellite pass began and (ii) the mean power consumption value computed for the respective satellite pass;
• determining, using a changepoint detection model, a quantity of changepoints detected within the time series data and times associated with each of the detected changepoints;
• providing at least one of the quantity of changepoints or the times associated with each of the detected changepoints to a machine learning model configured to determine whether the digitizer experienced at least one of a power failure event or a power disruption event;
• obtaining, from the machine learning model, based on the at least one of the quantity of changepoints or the times associated with each of the detected changepoints, an indication that the digitizer experienced at least one of the power failure event or the power disruption event;
• retrieving, from a client database, a quality of service (QoS) score of a client indicating how well satellite constellation access programs previously provided to the client complied with one or more satellite constellation access targets specified by the client;
• updating, based on the indication obtained from the machine learning model, the QoS score to obtain an updated QoS score for the client;
• and
• generating, using a scheduling system, based on the updated QoS score for the client, a satellite constellation access program for the client, wherein the satellite constellation access program indicates a time interval that the client is authorized to access the satellite ground station to receive data from the satellite during a subsequent satellite pass.

A non-transitory computer readable storage medium storing instructions for proactively detecting equipment failures at satellite ground stations that, when executed by one or more processors, cause the one or more processors to:

• obtain, for each of a plurality of satellite passes, power consumption data representing an amount of power consumed by a digitizer during a respective satellite pass, wherein:
• the digitizer is located at a satellite ground station that is configured to receive data from a satellite during each respective satellite pass,
• the power consumption data comprises time series data indicating a cumulative amount of power consumed by the digitizer at each of a plurality of times during the respective satellite pass, and
• the plurality of satellite passes occur of over a first time period;
• compute, for each of the plurality of satellite passes, a mean power consumption value representing the amount of power consumed by the digitizer during the respective satellite pass;
• generate time series data comprising at least (i) a timestamp of a time that the respective satellite pass began and (ii) the mean power consumption value computed for the respective satellite pass;
• determine, using a changepoint detection model, a quantity of changepoints detected within the time series data and times associated with each of the detected changepoints;
• provide at least one of the quantity of changepoints or the times associated with each of the detected changepoints to a machine learning model configured to determine whether the digitizer experienced at least one of a power failure event or a power disruption event;
• obtain, from the machine learning model, based on the at least one of the quantity of changepoints or the times associated with each of the detected changepoints, an indication that the digitizer experienced at least one of the power failure event or the power disruption event;
• retrieve, from a client database, a quality of service (QoS) score of a client indicating how well satellite constellation access programs previously provided to the client complied with one or more satellite constellation access targets specified by the client;
• update, based on the indication obtained from the machine learning model, the QoS score to obtain an updated QoS score for the client;
• and
• generate, using a scheduling system, based on the updated QoS score for the client, a satellite constellation access program for the client, wherein the satellite constellation access program indicates a time interval that the client is authorized to access the satellite ground station to receive data from the satellite during a subsequent satellite pass.