US-12619956-B2 - Predictive care of customer premises equipments

Abstract

Various example embodiments for supporting predictive care of customer premises equipments are presented herein. Various example embodiments for supporting predictive care of customer premises equipments may be configured to support machine learning predictive care of customer premises equipments based on application of various machine learning capabilities.

Inventors

• GINO DION

Assignees

• NOKIA SOLUTIONS AND NETWORKS OY

Dates

Publication Date

20260505

Application Date

20230130

Claims (18)

1. 1 . An apparatus, comprising: at least one processor; and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to: obtain, for a set of customer premises equipments, time-based customer premises equipment operating statistics data indicative of operation of the set of customer premises equipments; obtain, for the set of customer premises equipments, historical customer care data identifying a set of customer care events triggered for the set of customer premises equipments; learn, based on application of machine learning to the historical customer care data and the time-based customer premises equipment operating statistics data, a learned threshold for a condition experienced by at least a portion of the customer premises equipments, wherein the learned threshold for the condition represents a state of the condition at which a customer care action would be triggered for addressing the condition; determine, based on the learned threshold for the condition, an operating statistic threshold for the condition; and define, for the condition experienced by at least a portion of the customer premises equipments, a predictive signature configured to predict that the condition will be experienced by the customer premises equipments within a time frame, wherein the predictive signature includes an indication of the condition, the operating statistic threshold for the condition, and an action to be performed to address the condition; determine, for a given customer premises equipment based on an operating statistic of the given customer premises equipment, that the predictive signature is detected for the given customer premises equipment, wherein the predictive signature includes a memory leak predictive signature in which the condition is a memory leak condition, the operating statistic threshold relates to a free memory statistic, and the action to be performed includes a reboot operation; and initiate, for the given customer premises equipment based on detection of the predictive signature for the given customer premises equipment, the action to be performed to address the condition.
2. 2 . The apparatus of claim 1 , wherein the time-based customer premises equipment operating statistics data includes at least one of a set of time-based customer premises equipment health statistics associated with ones of the customer premises equipment or a set of time-based customer premises equipment performance statistics associated with ones of the customer premises equipment.
3. 3 . The apparatus of claim 1 , wherein the time-based customer premises equipment operating statistics data includes at least one of a set of time-based operating statistics curves or a set of time-based operating statistics traces.
4. 4 . The apparatus of claim 1 , wherein, to obtain the time-based customer premises equipment operating statistics data, the instructions, when executed by the at least one processor, cause the apparatus to: obtain, for the set of customer premises equipments, a set of operating statistics measured by the customer premises equipments and a respective set of time stamps indicative of times at which the respective operating statistics were measured by the respective customer premises equipments; and associate the time stamps with the operating statistics to form the time-based customer premises equipment operating statistics data.
5. 5 . The apparatus of claim 4 , wherein the set of operating statistics measured by the customer premises equipments includes at least one of a set of customer premises equipment health statistics associated with ones of the customer premises equipment or a set of customer premises equipment performance statistics associated with ones of the customer premises equipment.
6. 6 . The apparatus of claim 4 , wherein, for at least one of the customer premises equipments, the set of operating statistics measured by the respective customer premises equipment includes at least one of at least one a wide area network statistic associated with a wide area network supporting the respective customer premises equipment, a radio access network statistic associated with a radio access network supporting the respective customer premises equipment, a WiFi statistic associated with a WiFi network supporting the respective customer premises equipment, an operating system statistic associated with operation of an operating system of the respective customer premises equipment, an application statistic associated with an application running on the respective customer premises equipment, or an environmental statistic associated with the respective customer premises equipment.
7. 7 . The apparatus of claim 4 , wherein the set of operating statistics measured by the customer premises equipments is received from at least one of the set of customer premises equipments or a customer premises equipment management system configured to provide management functions for the set of customer premises equipments.
8. 8 . The apparatus of claim 1 , wherein the time-based customer premises equipment operating statistics data is obtained from a time-series database configured to store a set of operating statistics measured by the customer premises equipments and a respective set of time stamps indicative of times at which the respective operating statistics were measured by the respective customer premises equipments.
9. 9 . The apparatus of claim 1 , wherein the historical customer care data includes at least one of customer care ticket data for a set of customer care tickets opened for the set of customer premises equipments, customer care workflow data for a set of customer care workflow operations performed for the set of customer premises equipments, or customer care dispatch data for a set of customer care dispatch events performed for the set of customer premises equipments.
10. 10 . The apparatus of claim 1 , wherein the historical customer care data is obtained from a service provider workflow management platform.
11. 11 . The apparatus of claim 1 , wherein the operating statistic threshold for the condition is determined, based on the learned threshold, in a manner tending to decrease a probability that the condition is experienced at a customer premises equipment and results in a failure of the customer premises equipment.
12. 12 . The apparatus of claim 1 , wherein the predictive signature includes a predictive index based on a probability value indicative of a probability that a customer will initiate a customer care contact for the condition and a time frame value associated with the probability value where the time frame value provides an indication of a time frame within which the customer care contact may occur.
13. 13 . The apparatus of claim 12 , wherein the instructions, when executed by the at least one processor, cause the apparatus at least to: obtain, for the set of customer premises equipments, new time-based customer premises equipment operating statistics data indicative of operation of the set of customer premises equipments; obtain, for the set of customer premises equipments, new historical customer care data identifying a new set of customer care events triggered for the set of customer premises equipments; and adapting, based on processing of at least one of the new time-based customer premises equipment operating statistics data or the new historical customer care data using a machine learning process, the predictive index of the predictive signature.
14. 14 . An apparatus, comprising: at least one processor; and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to: maintain a predictive signature configured to indicate that a condition will be experienced within a time frame, wherein the predictive signature includes an indication of the condition, an operating statistic threshold for the condition, and an action to be performed to address the condition, wherein the operating statistic threshold is based on a learned threshold for the condition determined based on a set of customer premises equipments, wherein the learned threshold for the condition is based on application of machine learning to historical customer care data for the set of customer premises equipments and time-based customer premises equipment operating statistics data for the set of customer premises equipments, wherein the learned threshold for the condition represents a state of the condition at which a customer care action would be triggered for addressing the condition; obtain, for a given customer premises equipment, an operating statistic of the given customer premises equipment; determine, based on comparison of the operating statistic of the given customer premises equipment and the operating statistic threshold of the predictive signature, that the predictive signature is detected for the given customer premises equipment, wherein the predictive signature includes at least one of: a memory leak predictive signature in which the condition is a memory leak condition, the operating statistic threshold relates to a free memory statistic, and the action to be taken includes a reboot operation; a wireless signal interference predictive signature in which the condition is a wireless interference condition, the operating statistic threshold relates to at least one of a packets dropped statistics or a packets errored statistic, and the action to be taken includes a wireless channel switching operation; or a temperature anomaly predictive signature in which the condition is a temperature anomaly condition, the operating statistic threshold relates to a customer premises equipment temperature statistic, and the action to be taken includes at least one of sending a message to a customer or dispatching a technician; and initiate, for the given customer premises equipment based on the predictive signature, the action to be performed to address the condition.
15. 15 . The apparatus of claim 14 , wherein, to initiate the action to be performed to address the condition, the instructions, when executed by the at least one processor, cause the apparatus at least to: identify, from the predictive signature based on detection of the predictive signature for the given customer premises equipment, the action to be performed to address the condition; and send, toward the given customer premises equipment, a message including an indication of the action to be performed to address the condition.
16. 16 . The apparatus of claim 14 , wherein the action to be performed to address the condition includes at least one of a reconfiguration of the given customer premises equipment, a reset of the given customer premises equipment, or a reboot of the given customer premises equipment.
17. 17 . An apparatus, comprising: at least one processor; and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to: obtain, for a set of customer premises equipments, time-based customer premises equipment operating statistics data indicative of operation of the set of customer premises equipments; obtain, for the set of customer premises equipments, historical customer care data identifying a set of customer care events triggered for the set of customer premises equipments; learn, based on application of machine learning to the historical customer care data and the time-based customer premises equipment operating statistics data, a learned threshold for a condition experienced by at least a portion of the customer premises equipments, wherein the learned threshold for the condition represents a state of the condition at which a customer care action would be triggered for addressing the condition; determine, based on the learned threshold for the condition, an operating statistic threshold for the condition; and define, for the condition experienced by at least a portion of the customer premises equipments, a predictive signature configured to predict that the condition will be experienced by the customer premises equipments within a time frame, wherein the predictive signature includes an indication of the condition, the operating statistic threshold for the condition, and an action to be performed to address the condition; determine, for a given customer premises equipment based on an operating statistic of the given customer premises equipment, that the predictive signature is detected for the given customer premises equipment, wherein the predictive signature includes a wireless signal interference predictive signature in which the condition is a wireless interference condition, the operating statistic threshold relates to at least one of a packets dropped statistics or a packets errored statistic, and the action to be taken includes a wireless channel switching operation; and initiate, for the given customer premises equipment based on detection of the predictive signature for the given customer premises equipment, the action to be performed to address the condition.
18. 18 . An apparatus, comprising: at least one processor; and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to: obtain, for a set of customer premises equipments, time-based customer premises equipment operating statistics data indicative of operation of the set of customer premises equipments; obtain, for the set of customer premises equipments, historical customer care data identifying a set of customer care events triggered for the set of customer premises equipments; learn, based on application of machine learning to the historical customer care data and the time-based customer premises equipment operating statistics data, a learned threshold for a condition experienced by at least a portion of the customer premises equipments, wherein the learned threshold for the condition represents a state of the condition at which a customer care action would be triggered for addressing the condition; determine, based on the learned threshold for the condition, an operating statistic threshold for the condition; and define, for the condition experienced by at least a portion of the customer premises equipments, a predictive signature configured to predict that the condition will be experienced by the customer premises equipments within a time frame, wherein the predictive signature includes an indication of the condition, the operating statistic threshold for the condition, and an action to be performed to address the condition; determine, for a given customer premises equipment based on an operating statistic of the given customer premises equipment, that the predictive signature is detected for the given customer premises equipment, wherein the predictive signature includes a temperature anomaly predictive signature in which the condition is a temperature anomaly condition, the operating statistic threshold relates to a customer premises equipment temperature statistic, and the action to be taken includes at least one of sending a message to a customer or dispatching a technician; and initiate, for the given customer premises equipment based on detection of the predictive signature for the given customer premises equipment, the action to be performed to address the condition.

Description

TECHNICAL FIELD Various example embodiments relate generally to communication systems and, more particularly but not exclusively, to providing support for customer premises equipment in communication systems. BACKGROUND In communication networks, various communications technologies may be used to support various types of communications. SUMMARY In at least some example embodiments, an apparatus includes at least one processor and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus at least to obtain, for a set of customer premises equipments, time-based customer premises equipment operating statistics data indicative of operation of the set of customer premises equipments, obtain, for the set of customer premises equipments, historical customer care data identifying a set of customer care events triggered for the set of customer premises equipments, learn, based on application of machine learning to the historical customer care data and the time-based customer premises equipment operating statistics data, a learned threshold for a condition experienced by at least a portion of the customer premises equipments, and define, for the condition experienced by at least a portion of the customer premises equipments, a predictive signature configured to predict that the condition will be experienced within a time frame, wherein the predictive signature includes an indication of the condition, an operating statistic threshold for the condition that is determined based on the learned threshold, and an action to be performed to address the condition. In at least some example embodiments, the time-based customer premises equipment operating statistics data includes at least one of a set of time-based customer premises equipment health statistics associated with ones of the customer premises equipment or a set of time-based customer premises equipment performance statistics associated with ones of the customer premises equipment. In at least some example embodiments, the time-based customer premises equipment operating statistics data includes at least one of a set of time-based operating statistics curves or a set of time-based operating statistics traces. In at least some example embodiments, to obtain the time-based customer premises equipment operating statistics data, the instructions, when executed by the at least one processor, cause the apparatus to obtain, for the set of customer premises equipments, a set of operating statistics measured by the customer premises equipments and a respective set of time stamps indicative of times at which the respective operating statistics were measured by the respective customer premises equipments and associate the time stamps with the operating statistics to form the time-based customer premises equipment operating statistics data. In at least some example embodiments, the set of operating statistics measured by the customer premises equipments includes at least one of a set of customer premises equipment health statistics associated with ones of the customer premises equipment or a set of customer premises equipment performance statistics associated with ones of the customer premises equipment. In at least some example embodiments, for at least one of the customer premises equipments, the set of operating statistics measured by the respective customer premises equipment includes at least one of at least one a wide area network statistic associated with a wide area network supporting the respective customer premises equipment, a radio access network statistic associated with a radio access network supporting the respective customer premises equipment, a WiFi statistic associated with a WiFi network supporting the respective customer premises equipment, an operating system statistic associated with operation of an operating system of the respective customer premises equipment, an application statistic associated with an application running on the respective customer premises equipment, or an environmental statistic associated with the respective customer premises equipment. In at least some example embodiments, the set of operating statistics measured by the customer premises equipments is received from at least one of the set of customer premises equipments or a customer premises equipment management system configured to provide management functions for the set of customer premises equipments. In at least some example embodiments, the time-based customer premises equipment operating statistics data is obtained from a time-series database configured to store a set of operating statistics measured by the customer premises equipments and a respective set of time stamps indicative of times at which the respective operating statistics were measured by the respective customer premises equipments. In at least some example embodiments, the historical customer care data includes at least one of customer care ticket data for