Orchestrator reporting of probability of downtime from machine learning process
Disclosed embodiments relate to reporting Electronic Control Unit (ECU) errors or faults to a remote monitoring server. Operations may include receiving operational data from a plurality of ECUs in the vehicle, the operational data being indicative of a plurality of runtime attributes of the plurality of ECUs; generating, through a machine learning process, a statistical model of the operational data; receiving live, runtime updates from the plurality of ECUs in the communications network of the vehicle; identifying an ECU error associated with an ECU in the communications network of the vehicle, the ECU error being determined by a comparison of the live, runtime updates with the statistical model of the operational data to identify at least one deviation from the operational data; and wirelessly sending a report to the remote monitoring server based on the live, runtime updates, the report identifying the ECU and the identified ECU error.
1. A non-transitory computer-readable medium including instructions that, when executed by at least one processor, cause the at least one processor to perform error-monitoring operations for reporting and remediating controller errors, comprising:
receiving operational data from a plurality of controllers, the operational data being indicative of a plurality of runtime attributes associated with the plurality of controllers, the plurality of runtime attributes comprising at least one of: a memory activity or a processing activity;
generating a statistical model of the operational data received from the plurality of controllers based on a predetermined set of controller functionalities, wherein the statistical model is based on time-based analysis of attributes associated with one or more of the plurality of controllers, the attributes comprising at least one of: a central processing unit operation, information stored in a memory component, or information accessed from a memory component;
receiving live, runtime updates from the plurality of controllers;
identifying, based on a comparison of the live, runtime updates to the statistical model, an error of a monitored controller; and
generating a report based on the identified error, the report identifying the monitored controller and the identified error.
2. The non-transitory computer-readable medium of claim 1 , wherein the statistical model is configured to apply a machine learning function to determine if at least one of the plurality of controllers is operating within a predetermined envelope of operational attributes.
3. The non-transitory computer-readable medium of claim 1 , wherein the error-monitoring operations further comprise causing the monitored controller to adjust from executing a first version of controller software to a second version of controller software in response to the identified error.
4. The non-transitory computer-readable medium of claim 3 , wherein causing the monitored controller to adjust from executing the first version of controller software to the second version of controller software comprises implementing a delta file on the monitored controller.
5. The non-transitory computer-readable medium of claim 3 , wherein the second version of controller software was previously deployed to the monitored controller.
6. The non-transitory computer-readable medium of claim 1 , wherein the identification of the error of the monitored controller is based on a detected stack overflow in the monitored controller.
7. The non-transitory computer-readable medium of claim 1 , wherein the error-monitoring operations further comprise activating a safe mode of the monitored controller in response to the identified error.
8. The non-transitory computer-readable medium of claim 1 , wherein the statistical model is based on an execution frequency or execution path of software.
9. The non-transitory computer-readable medium of claim 1 , wherein the identified error includes a drift of an application execution profile.
10. The non-transitory computer-readable medium of claim 1 , wherein the error-monitoring operations further comprise determining whether the identified error is detrimental to the monitored controller.
11. The non-transitory computer-readable medium of claim 10 , wherein the error-monitoring operations further comprise performing a control action in response to the determination that the identified error is detrimental to the monitored controller, the control action including at least one of:
issuing a prompt to adjust the monitored controller from executing a first version of controller software to a second version of controller software;
sending an alert associated with the monitored controller; or
blocking an instruction sent from the monitored controller.
12. The non-transitory computer-readable medium of claim 1 , wherein the monitored controller is one of the plurality of controllers.
13. The non-transitory computer-readable medium of claim 1 , wherein at least part of the plurality of controllers are in separate devices across a wide area communications network.
14. A computer-implemented method for reporting and remediating controller errors, comprising:
receiving operational data from a plurality of controllers, the operational data being indicative of a plurality of runtime attributes associated with the plurality of controllers, the plurality of runtime attributes comprising at least one of: a memory activity or a processing activity;
generating a statistical model of the operational data received from the plurality of controllers based on a predetermined set of controller functionalities, wherein the statistical model is based on time-based analysis of attributes associated with one or more of the plurality of controllers, the attributes comprising at least one of: a central processing unit operation, information stored in a memory component, or information accessed from a memory component;
receiving live, runtime updates from the plurality of controllers;
identifying, based on a comparison of the live, runtime updates to the statistical model, an error of a monitored controller; and
generating a report based on the identified error, the report identifying the monitored controller and the identified error.
15. The computer-implemented method of claim 14 , wherein the statistical model is configured to apply a machine learning function to determine if at least one of the plurality of controllers is operating within a predetermined envelope of operational attributes.
16. The computer-implemented method of claim 14 , further comprising causing the monitored controller to adjust from executing a first version of controller software to a second version of controller software in response to the identified error.
17. The computer-implemented method of claim 16 , wherein causing the monitored controller to adjust from executing the first version of controller software to the second version of controller software comprises implementing a delta file on the monitored controller.
18. The computer-implemented method of claim 16 , wherein the second version of controller software was previously deployed to the monitored controller.
19. The computer-implemented method of claim 14 , wherein the monitored controller is one of the plurality of controllers.
20. The computer-implemented method of claim 14 , wherein at least part of the plurality of controllers are in separate devices across a wide area communications network.