Predictive care for customer premises equipment

Abstract

To provide support for customer premises equipment in communication systems.SOLUTION: Various example embodiments for supporting predictive care of customer premises equipment are presented herein. Various example embodiments for supporting predictive care of customer premises equipment may be configured to support machine learning predictive care of customer premises equipment based on application of various machine learning capabilities.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] Various exemplary embodiments generally relate to communication systems, and more specifically, but not limited to, providing support for customer premise equipment in a communication system.

Background Art

[0002] In communication networks, various communication technologies may be used to support various types of communications.

Summary of the Invention

[0003] In at least some exemplary 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 to at least obtain time-based customer premise equipment operation statistics data indicative of the operation of a set of customer premise equipment for the set of customer premise equipment; obtain historical customer care data identifying a set of customer care events triggered for the set of customer premise equipment for the set of customer premise equipment; learn a learned threshold for a condition experienced by at least a portion of the customer premise equipment based on application of machine learning to the historical customer care data and the time-based customer premise equipment operation statistics data; and define a prediction signature configured to predict that the condition will be experienced within a time frame for a condition experienced by at least a portion of the customer premise equipment, the prediction signature including an indication of the condition, an operation statistics threshold for the condition determined based on the learned threshold, and an action to be executed to address the condition.

[0004] In at least some exemplary embodiments, the time-based customer premise equipment operation statistics data includes at least one of a set of time-based customer premise equipment health statistics related to one of the customer premise equipment or a set of time-based customer premise equipment performance statistics related to one of the customer premise equipment.

[0005] In at least some exemplary embodiments, time-based customer premises equipment operation statistics data includes at least one of a set of time-based operation statistics curves or a set of time-based operation statistics traces.

[0006] In at least some exemplary embodiments, in order to obtain time-based customer premises equipment operation statistics data, an instruction, when executed by at least one processor, causes the device to obtain, for a set of customer premises equipment, a set of operation statistics measured by the customer premises equipment and a set of timestamps indicating the time each operation statistic was measured by each customer premises equipment, associate the timestamps with the operation statistics, and form time-based customer premises equipment operation statistics data.

[0007] In at least some exemplary embodiments, the set of operational statistics measured by customer premises equipment includes at least one of the following: a set of customer premises equipment health statistics related to one of the customer premises equipment, or a set of customer premises equipment performance statistics related to one of the customer premises equipment.

[0008] In at least some exemplary embodiments, for at least one of the customer premises equipment, the set of operational statistics measured by each customer premises equipment may include at least one of the following: wide area network statistics related to the wide area network supporting each customer premises equipment; wireless access network statistics related to the wireless access network supporting each customer premises equipment; WiFi statistics related to the WiFi network supporting each customer premises equipment; operating system statistics related to the operation of the operating system of each customer premises equipment; application statistics related to the applications running on each customer premises equipment; or environmental statistics related to each customer premises equipment.

[0009] In at least some exemplary embodiments, a set of operational statistics measured by customer premises equipment is received from at least one of the customer premises equipment or a customer premises equipment management system configured to provide management functions for the customer premises equipment.

[0010] In at least some exemplary embodiments, time-based customer premises equipment operation statistics data is obtained from a time-series database configured to store a set of operation statistics measured by customer premises equipment and a set of timestamps indicating the time each operation statistic was measured by each customer premises equipment.

[0011] In at least some exemplary embodiments, the historical customer care data includes at least one of the following: customer care ticket data for a set of customer care tickets opened for a set of customer premises equipment; customer care workflow data for a set of customer care workflow operations performed for a set of customer premises equipment; or customer care dispatch data for a set of customer care dispatch events performed for a set of customer premises equipment.

[0012] In at least some exemplary embodiments, historical customer care data is retrieved from a service provider workflow management platform. In at least some exemplary embodiments, learned thresholds for conditions represent the conditional states that trigger customer care actions to address those conditions.

[0013] In at least some exemplary embodiments, the behavioral statistical threshold for a condition is determined based on learned thresholds in a way that tends to reduce the probability that the condition is experienced in customer premises equipment and results in a failure of the customer premises equipment.

[0014] In at least some exemplary embodiments, the predictive signature includes at least one memory leak predictive signature where the condition is a memory leak condition, a memory leak predictive signature where the operational statistics threshold is related to free memory statistics, an action including a reboot operation, and a radio signal interference predictive signature where the condition is a radio interference condition, the operational statistics threshold being related to at least one of packet loss statistics and packet error statistics, and the action to be taken includes a radio channel switching operation, or a temperature anomaly predictive signature where the condition is a temperature anomaly condition, the operational statistics threshold being related to customer premises equipment temperature statistics, and the action to be taken including at least one of sending a message to the customer or dispatching a technician.

[0015] In at least some exemplary embodiments, an instruction, when executed by at least one processor, causes the device to determine, at least based on the operational statistics of a given customer premises device, that a predictive signature has been detected for a given customer premises device, the predictive signature being configured to indicate that a certain condition is experienced by the customer premises within a time frame, the predictive signature including an indication of the condition, an operational statistics threshold for the condition, and addressing the condition and, based on the predictive signature, initiating an action to be performed on the given customer premises device to address the condition.

[0016] In at least some exemplary embodiments, a non-temporary computer-readable medium, when executed by the device, includes computer program instructions causing the device to perform the steps of: obtaining time-based customer premises equipment behavior statistics data indicating the behavior of a set of customer premises equipment; obtaining historical customer care data identifying a set of customer care events triggered on a set of customer premises equipment; learning learned thresholds for conditions experienced by at least a portion of the customer premises equipment based on the application of machine learning to the historical customer care data and the time-based customer premises equipment behavior statistics data; and defining a predictive signature for a condition experienced by at least a portion of the customer premises equipment, configured to predict that the condition will be experienced within a time frame, wherein the predictive signature includes a description of the condition, a behavior statistics threshold for the condition determined based on the learned thresholds, and an action to be taken to address the condition.

[0017] In at least some exemplary embodiments, time-based customer premises equipment operation statistics data includes at least one of a set of time-based customer premises equipment health statistics related to one of the customer premises equipment, or a set of time-based customer premises equipment performance statistics related to one of the customer premises equipment.

[0018] In at least some exemplary embodiments, time-based customer premises equipment operation statistics data includes at least one of a set of time-based operation statistics curves or a set of time-based operation statistics traces.

[0019] In at least some exemplary embodiments, to obtain time-based customer premises equipment operation statistics, a computer program instruction, when executed by the device, causes the device to obtain for a set of customer premises equipment a set of operation statistics measured by the customer premises equipment and a set of timestamps indicating the time each operation statistic was measured by each customer premises equipment, and associates the timestamps with the operation statistics to form time-based customer premises equipment operation statistics.

[0020] In at least some exemplary embodiments, the set of operational statistics measured by customer premises equipment includes at least one of the following: a set of customer premises equipment health statistics related to one of the customer premises equipment, or a set of customer premises equipment performance statistics related to one of the customer premises equipment.

[0021] In at least some exemplary embodiments, for at least one of the customer premises equipment, the set of operational statistics measured by each customer premises equipment may include at least one of the following: wide area network statistics related to the wide area network supporting each customer premises equipment; wireless access network statistics related to the wireless access network supporting each customer premises equipment; WiFi statistics related to the WiFi network supporting each customer premises equipment; operating system statistics related to the operation of the operating system of each customer premises equipment; application statistics related to the applications running on each customer premises equipment; or environmental statistics related to each customer premises equipment.

[0022] In at least some exemplary embodiments, a set of operational statistics measured by customer premises equipment is received from at least one of a set of customer premises equipment or a customer premises equipment management system configured to provide management functions for a set of customer premises equipment.

[0023] In at least some exemplary embodiments, time-based customer premises equipment operation statistics data is obtained from a time-series database configured to store a set of operation statistics measured by customer premises equipment and a set of timestamps indicating the time each operation statistic was measured by each customer premises equipment.

[0024] In at least some exemplary embodiments, historical customer care data includes at least one of the following: customer care ticket data for a set of customer care tickets opened for a set of customer premises equipment; customer care workflow data for a set of customer care workflow actions performed for a set of customer premises equipment; or customer care dispatch data for a set of customer care dispatch events performed for a set of customer premises equipment. In at least some exemplary embodiments, historical customer care data is retrieved from a service provider workflow management platform.

[0025] In at least some exemplary embodiments, the learned threshold for a condition represents the conditional state that triggers a customer care action to address the condition. In at least some exemplary embodiments, the behavioral statistics threshold for a condition is determined based on the learned threshold in a way that tends to reduce the probability that the condition is experienced in customer premises equipment and results in a failure of the customer premises equipment.

[0026] In at least some exemplary embodiments, the predictive signature includes at least one of a memory leak predictive signature where the condition is a memory leak condition, a memory leak predictive signature where the operational statistics threshold is a free memory statistic, an action including a reboot operation, and a radio signal interference predictive signature where the condition is a radio interference condition, the operational statistics threshold being at least one of packet loss statistics and packet error statistics, and the action to be taken including a radio channel switching operation, or a temperature anomaly predictive signature where the condition is a temperature anomaly condition, the operational statistics threshold being a customer premises equipment temperature statistic, and the action to be taken including at least one of sending a message to the customer or dispatching a technician.

[0027] In at least some exemplary embodiments, a computer program instruction, when executed by the device, causes the device to determine, at least based on the operational statistics of a given customer premises device, that a predictive signature has been detected for a given customer premises device, the predictive signature being configured to indicate that a certain condition will be experienced by the customer premises within a time frame, the predictive signature including an indication of the condition, an operational statistics threshold for the condition, and an action to be performed to address the condition and to initiate an action to be performed on a given customer premises device based on the predictive signature to address the condition.

[0028] In at least some exemplary embodiments, the method includes, for a set of customer premise equipment, obtaining time-based customer premise equipment operation statistics data indicative of the operation of the set of customer premise equipment; obtaining historical customer care data for the set of customer premise equipment identifying a set of customer care events triggered for the set of customer premise equipment; learning a learned threshold for a condition experienced by at least a portion of the customer premise equipment based on an application of machine learning to the historical customer care data and the time-based customer premise equipment operation statistics data; and defining a predictive signature configured to predict that a condition experienced by at least a portion of the customer premise equipment will be experienced within a time frame, the predictive signature including an indication of the condition, an operational statistics threshold for the condition determined based on the learned threshold, and an action to be performed to address the condition.

[0029] In at least some exemplary embodiments, the time-based customer premise equipment operation statistics data includes at least one of a set of time-based customer premise equipment health statistics associated with one of the customer premise equipment or a set of time-based customer premise equipment performance statistics associated with one of the customer premise equipment.

[0030] In at least some exemplary embodiments, the time-based customer premise equipment operation statistics data includes at least one of a set of time-based operation statistics curves or a set of time-based operation statistics traces.

[0031] In at least some exemplary embodiments, obtaining the time-based customer premise equipment operation statistics data includes, for the set of customer premise equipment, obtaining a set of operation statistics measured by the customer premise equipment and a set of respective timestamps indicating the times at which the respective operation statistics were measured by the respective customer premise equipment, associating the timestamps with the operation statistics, and forming the time-based customer premise equipment operation statistics data.

[0032] In at least some exemplary embodiments, the set of operational statistics measured by customer premises equipment includes at least one of the following: a set of customer premises equipment health statistics related to one of the customer premises equipment, or a set of customer premises equipment performance statistics related to one of the customer premises equipment.

[0033] In at least some exemplary embodiments, for at least one of the customer premises equipment, the set of operational statistics measured by each customer premises equipment includes at least one of the following: wide area network statistics related to the wide area network supporting each customer premises equipment; wireless access network statistics related to the wireless access network supporting each customer premises equipment; WiFi statistics related to the WiFi network supporting each customer premises equipment; operating system statistics related to the operation of the operating system of each customer premises equipment; application statistics related to the applications running on each customer premises equipment; or environmental statistics related to each customer premises equipment.

[0034] In at least some exemplary embodiments, a set of operational statistics measured by customer premises equipment is received from at least one of the customer premises equipment or a customer premises equipment management system configured to provide management functions for the customer premises equipment.

[0035] In at least some exemplary embodiments, time-based customer premises equipment operation statistics data is obtained from a time-series database configured to store a set of operation statistics measured by customer premises equipment and a set of timestamps indicating the time each operation statistic was measured by each customer premises equipment.

[0036] In at least some exemplary embodiments, the historical customer care data includes at least one of the following: customer care ticket data for a set of customer care tickets opened for a set of customer premises equipment; customer care workflow data for a set of customer care workflow operations performed for a set of customer premises equipment; or customer care dispatch data for a set of customer care dispatch events performed for a set of customer premises equipment.

[0037] In at least some exemplary embodiments, historical customer care data is retrieved from a service provider workflow management platform. In at least some exemplary embodiments, learned thresholds for conditions represent the conditional states that trigger customer care actions to address those conditions.

[0038] In at least some exemplary embodiments, the behavioral statistical threshold for a condition is determined based on learned thresholds in a way that tends to reduce the probability that the condition is experienced in customer premises equipment and results in a failure of the customer premises equipment.

[0039] In at least some exemplary embodiments, the predictive signature includes at least one of a memory leak predictive signature where the condition is a memory leak condition, a memory leak predictive signature where the operational statistics threshold is a free memory statistic, an action including a reboot operation, and a radio signal interference predictive signature where the condition is a radio interference condition; the operational statistics threshold is for at least one of packet loss statistics and packet error statistics, and the action to be taken includes a radio channel switching operation, or a temperature anomaly predictive signature where the condition is a temperature anomaly condition; the operational statistics threshold is for customer premises equipment temperature statistics, and the action to be taken includes at least one of sending a message to the customer or dispatching a technician.

[0040] In at least some exemplary embodiments, the method includes the step of determining, based on operational statistics of a given customer premises device, that a predictive signature has been detected for a given customer premises device, wherein the predictive signature is configured to indicate that a certain condition will be experienced by the customer premises within a time frame, and the predictive signature includes a representation of the condition, an operational statistics threshold for the condition, an action to be taken to address the condition, and an action to initiate the action to be taken to address the condition for the given customer premises device based on the predictive signature.

[0041] In at least some exemplary embodiments, the apparatus includes means for acquiring time-based customer premises equipment behavior statistics data indicating the operation of a set of customer premises equipment; means for acquiring historical customer care data identifying a set of customer care events triggered for a set of customer premises equipment; means for learning learned thresholds for conditions experienced by at least a portion of the customer premises equipment based on the application of machine learning to the historical customer care data and the time-based customer premises equipment behavior statistics data; and means for defining predictive signatures configured to predict that a condition will be experienced within a time frame, wherein the predictive signatures include an indication of the condition, a behavior statistics threshold for the condition determined based on the learned thresholds, and an action to be taken to address the condition.

[0042] In at least some exemplary embodiments, time-based customer premises equipment operation statistics data includes at least one of a set of time-based customer premises equipment health statistics related to one of the customer premises equipment, or a set of time-based customer premises equipment performance statistics related to one of the customer premises equipment.

[0043] In at least some exemplary embodiments, time-based customer premises equipment operation statistics data includes at least one of a set of time-based operation statistics curves or a set of time-based operation statistics traces.

[0044] In at least some exemplary embodiments, means for obtaining time-based customer premises equipment operation statistics data include, for a set of customer premises equipment, means for obtaining a set of operation statistics measured by the customer premises equipment, and for each set of timestamps indicating the time at which each operation statistic was measured by each customer premises equipment, and means for associating the timestamps with the operation statistics to form time-based customer premises equipment operation statistics data.

[0045] In at least some exemplary embodiments, the set of operational statistics measured by customer premises equipment includes at least one of the following: a set of customer premises equipment health statistics related to one of the customer premises equipment, or a set of customer premises equipment performance statistics related to one of the customer premises equipment.

[0046] In at least some exemplary embodiments, for at least one of the customer premises equipment, the set of operational statistics measured by each customer premises equipment includes at least one of the following: wide area network statistics related to the wide area network supporting each customer premises equipment; wireless access network statistics related to the wireless access network supporting each customer premises equipment; WiFi statistics related to the WiFi network supporting each customer premises equipment; operating system statistics related to the operation of the operating system of each customer premises equipment; application statistics related to the applications running on each customer premises equipment; or environmental statistics related to each customer premises equipment.

[0047] In at least some exemplary embodiments, a set of operational statistics measured by customer premises equipment is received from at least one of the customer premises equipment or a customer premises equipment management system configured to provide management functions for the customer premises equipment.

[0048] In at least some exemplary embodiments, time-based customer premises equipment operation statistics data is obtained from a time-series database configured to store a set of operation statistics measured by customer premises equipment and a set of timestamps indicating the time each operation statistic was measured by each customer premises equipment.

[0049] In at least some exemplary embodiments, the historical customer care data includes at least one of the following: customer care ticket data for a set of customer care tickets opened for a set of customer premises equipment; customer care workflow data for a set of customer care workflow operations performed for a set of customer premises equipment; or customer care dispatch data for a set of customer care dispatch events performed for a set of customer premises equipment.

[0050] In at least some exemplary embodiments, historical customer care data is retrieved from a service provider workflow management platform. In at least some exemplary embodiments, learned thresholds for conditions represent the conditional states that trigger a customer care action to address the condition. In at least some exemplary embodiments, behavioral statistics thresholds for conditions are determined based on learned thresholds in a way that reduces the probability that the condition is experienced in customer premises equipment and results in a failure of the customer premises equipment.

[0051] In at least some exemplary embodiments, the predictive signature includes at least one of a memory leak predictive signature where the condition is a memory leak condition, a memory leak predictive signature where the operational statistics threshold is a free memory statistic, an action including a reboot operation, and a radio signal interference predictive signature where the condition is a radio interference condition, the operational statistics threshold being at least one of packet loss statistics and packet error statistics, and the action to be taken including a radio channel switching operation, or a temperature anomaly predictive signature where the condition is a temperature anomaly condition, the operational statistics threshold being a customer premises equipment temperature statistic, and the action to be taken including at least one of sending a message to the customer or dispatching a technician.

[0052] In at least some exemplary embodiments, the apparatus includes means for determining, based on operational statistics of a given customer premises equipment, that a predictive signature has been detected for a given customer premises equipment, wherein the predictive signature is configured to indicate that a certain condition is experienced by the customer premises within a time frame, and the predictive signature includes a representation of the condition, an operational statistics threshold for the condition, and an action to be taken to address the condition, and the apparatus includes means for initiating an action to be taken for a given customer premises equipment to address the condition based on the predictive signature.

[0053] In at least some exemplary embodiments, the device includes at least one processor and at least one memory for storing instructions, which, when executed by the at least one processor, causes the device to determine, based on operational statistics of a given customer premises device, that a predictive signature has been detected for a given customer premises device, the predictive signature being configured to indicate that a certain condition is experienced by the customer premises within a time frame, the predictive signature including an indication of the condition, an operational statistics threshold for the condition, and an action to be performed to address the condition and to initiate an action to be performed on a given customer premises device based on the predictive signature to address the condition.

[0054] In at least some exemplary embodiments, in order to determine that a predictive signature has been detected for a given customer premises device, the instruction is executed by at least one processor, causing the device to obtain operational statistics for at least a given customer premises device, and to determine, based on the operational statistics for the given customer premises device and the operational statistics threshold for the predictive signature, that the predictive signature has been detected for the given customer premises device.

[0055] In at least some exemplary embodiments, the predictive signature includes at least one of a memory leak predictive signature where the condition is a memory leak condition, a memory leak predictive signature where the operational statistics threshold is a free memory statistic, an action including a reboot operation, and a radio signal interference predictive signature where the condition is a radio interference condition; the operational statistics threshold is with respect to at least one of packet loss statistics and packet error statistics, and the action to be taken includes a radio channel switching operation, or a temperature anomaly predictive signature where the condition is a temperature anomaly condition; the operational statistics threshold is with respect to customer premises equipment temperature statistics, and the action to be taken includes at least one of sending a message to the customer or dispatching a technician.

[0056] In at least some exemplary embodiments, to initiate an action to be performed to address a condition, an instruction, when executed by at least one processor, causes the device to identify an action to be performed to address a condition from a predictive signature based on the detection of a predictive signature for a given customer premises equipment, and to send to the given customer premises equipment, the message includes instructions for the action to be performed to address the condition. In at least some exemplary embodiments, the action to be performed to address the condition includes at least one of reconfiguring the given customer premises equipment, resetting the given customer premises equipment, or restarting the given customer premises equipment.

[0057] In at least some exemplary embodiments, the non-transient computer-readable medium includes, when executed by the device, computer program instructions causing the device to determine, based on the operational statistics of a given customer premises equipment, that a predictive signature has been detected for a given customer premises equipment, wherein the predictive signature is configured to indicate that a certain condition is experienced by the customer premises within a time frame, and the predictive signature includes an indication of the condition, an operational statistics threshold for the condition, and an action to be performed to address the condition and to initiate an action to be performed on a given customer premises equipment based on the predictive signature to address the condition.

[0058] In at least some exemplary embodiments, in order to determine that a predictive signature has been detected for a given customer premises device, a computer program instruction, when executed by the device, causes the device to at least obtain operational statistics for the given customer premises device, and, based on the operational statistics for the given customer premises device and the operational statistics threshold for the predictive signature, determine that the predictive signature has been detected for the given customer premises device.

[0059] In at least some exemplary embodiments, the predictive signature includes at least one memory leak predictive signature where the condition is a memory leak condition, a memory leak predictive signature where the operational statistics threshold is related to free memory statistics, an action including a reboot operation, and a radio signal interference predictive signature where the condition is a radio interference condition, the operational statistics threshold being related to at least one of packet loss statistics and packet error statistics, and the action to be taken includes a radio channel switching operation, or a temperature anomaly predictive signature where the condition is a temperature anomaly condition, the operational statistics threshold being related to customer premises equipment temperature statistics, and the action to be taken including at least one of sending a message to the customer or dispatching a technician.

[0060] In at least some exemplary embodiments, to initiate an action to be performed to address a condition, a computer program instruction, when executed by the device, causes the device to identify an action to be performed to address a condition from a predictive signature based on the detection of a predictive signature for a given customer premises device, and to send to the given customer premises device a message containing instructions for the action to be performed to address the condition.

[0061] In at least some exemplary embodiments, the actions taken to address the condition include at least one of reconfiguring a given customer premises device, resetting a given customer premises device, or restarting a given customer premises device.

[0062] In at least some exemplary embodiments, the method includes the step of determining, based on operational statistics of a given customer premises equipment, that a predictive signature has been detected for a given customer premises equipment, wherein the predictive signature is configured to indicate that a certain condition will be experienced by the customer premises within a time frame, and the predictive signature includes an indication of the condition, an operational statistics threshold for the condition, an action to be taken to address the condition, and initiating the action to be taken to address the condition for the given customer premises equipment based on the predictive signature.

[0063] In at least some exemplary embodiments, determining that a predictive signature has been detected for a given customer premises device includes obtaining operational statistics for the given customer premises device and determining that a predictive signature has been detected for the given customer premises device based on the operational statistics for the given customer premises device and the operational statistics threshold for the predictive signature.

[0064] In at least some exemplary embodiments, the predictive signature includes at least one memory leak predictive signature where the condition is a memory leak condition, a memory leak predictive signature where the operational statistics threshold is related to free memory statistics, an action including a reboot operation, and a radio signal interference predictive signature where the condition is a radio interference condition, the operational statistics threshold being related to at least one of packet loss statistics and packet error statistics, and the action to be taken includes a radio channel switching operation, or a temperature anomaly predictive signature where the condition is a temperature anomaly condition, the operational statistics threshold being related to customer premises equipment temperature statistics, and the action to be taken including at least one of sending a message to the customer or dispatching a technician.

[0065] In at least some exemplary embodiments, initiating an action to be performed to address a condition includes identifying an action to be performed to address a condition from a predictive signature based on the detection of a predictive signature for a given customer premises device, and sending a message to the given customer premises device containing instructions for the action to be performed to address the condition.

[0066] In at least some exemplary embodiments, the actions taken to address the condition include at least one of reconfiguring a given customer premises device, resetting a given customer premises device, or restarting a given customer premises device.

[0067] In at least some exemplary embodiments, the apparatus includes means for determining, based on operational statistics of a given customer premises equipment, that a predictive signature has been detected for a given customer premises equipment, wherein the predictive signature is configured to indicate that a certain condition is experienced by the customer premises within a time frame, and the predictive signature includes an indication of the condition, an operational statistics threshold for the condition, an action to be performed to address the condition, and means for initiating the action to be performed to address the condition for a given customer premises equipment based on the predictive signature.

[0068] In at least some exemplary embodiments, means for determining whether a predictive signature has been detected for a given customer premises equipment includes means for obtaining operational statistics for a given customer premises equipment, and means for determining whether a predictive signature has been detected for a given customer premises equipment based on the operational statistics for a given customer premises equipment and an operational statistics threshold for the predictive signature.

[0069] In at least some exemplary embodiments, the predictive signature includes at least one of a memory leak predictive signature where the condition is a memory leak condition, a memory leak predictive signature where the operational statistics threshold is a free memory statistic, an action including a reboot operation, and a radio signal interference predictive signature where the condition is a radio interference condition, the operational statistics threshold being at least one of packet loss statistics and packet error statistics, and the action to be taken including a radio channel switching operation, or a temperature anomaly predictive signature where the condition is a temperature anomaly condition, the operational statistics threshold being a customer premises equipment temperature statistic, and the action to be taken including at least one of sending a message to the customer or dispatching a technician.

[0070] In at least some exemplary embodiments, means for initiating an action to be performed to address a condition include means for identifying an action to be performed to address a condition from a predictive signature based on the detection of a predictive signature for a given customer premises device, and means for sending a message containing instructions for the action to be performed to address the condition to the given customer premises device.

[0071] In at least some exemplary embodiments, the actions taken to address the condition include at least one of reconfiguring a given customer premises device, resetting a given customer premises device, or restarting a given customer premises device. [Brief explanation of the drawing]

[0072] The teachings in this specification can be easily understood by considering the following detailed description in conjunction with the accompanying drawings.

[0073] [Figure 1] This figure shows an exemplary embodiment of a communication system configured to support predictive care of customer premises equipment. [Figure 2] This figure shows an exemplary embodiment of a method for supporting predictive care of customer premises equipment. [Figure 3] This figure shows an exemplary embodiment of a method for generating predictive signatures for use in supporting predictive care of customer premises equipment. [Figure 4] This figure shows an exemplary embodiment of a method for generating predictive signatures for use in supporting predictive care of customer premises equipment. [Figure 5] This figure shows an exemplary embodiment of a method for using predictive signatures to support predictive care of customer premises equipment. [Figure 6] This figure shows an exemplary embodiment of a method for supporting predictive care of customer premises equipment. [Figure 7] This figure shows an exemplary embodiment of a method for supporting predictive care of customer premises equipment. [Figure 8] This figure shows an exemplary embodiment of a computer suitable for use in performing the various functions presented herein.

[0074] To facilitate understanding, the same reference numbers are used in this specification whenever possible to indicate identical elements common to various figures. [Modes for carrying out the invention]

[0075] Various exemplary embodiments for supporting predictive care of customer premises equipment are presented herein. These various exemplary embodiments for supporting predictive care of customer premises equipment may be configured to support machine learning-based predictive care of customer premises equipment based on the application of various types of machine learning capabilities to various types of data related to supporting service assurance functions for customer premises equipment.

[0076] Various exemplary embodiments for supporting predictive care for customer premises equipment may be configured to apply machine learning to the service assurance concept. Various exemplary embodiments for supporting predictive care for customer premises equipment may be configured to support service assurance for customer premises equipment based on the use of machine learning and / or artificial intelligence in the context of a feedback loop that validates assertions related to service assurance for customer premises equipment. Various exemplary embodiments for supporting predictive care for customer premises equipment may be configured to provide a feedback loop that validates assertions related to service assurance for customer premises equipment by supplying the service provider's historical customer care data (e.g., actual historical records of customer care workflows, data (e.g., from customers calling the self-service portal, from customers using the self-service portal, or equivalent), workflow management data, trouble ticket dispatch data, etc.) to the machine learning process, so that a unique mapping of traditional statistics can be achieved and a fully adaptive threshold point that predicts when conditions that may result in a customer care call or dispatch are about to occur (and thus predicts when a customer care call or dispatch is about to occur).

[0077] Various exemplary embodiments for supporting predictive care of customer premises equipment (CME) can be configured to support CEM service assurance based on the use of machine learning and / or artificial intelligence within the context of a feedback loop that validates assertions related to CEM service assurance. From the end customer's perspective, the platform can predict when CEM issues will occur and take proactive steps to prevent CEM issues from occurring in the first place. Based on the use of a feedback loop that uses customer care data (e.g., call, ticket, and / or dispatch data) and applying machine learning to a combination of customer care data and CEM data points at the exact time when the issue actually occurred, the platform overcomes the complexity of tracking rapidly moving targets of thresholds for conditions experienced by CEM, enabling prediction of conditions in CEM that may lead to problems in CEM.

[0078] Various exemplary embodiments for supporting predictive care of customer premises equipment may be configured to support truly predictive customer premises equipment service assurance by applying machine learning techniques to a combination of behavioral statistics associated with customer premises equipment operation and historical customer care data associated with customer premises equipment operation, thereby enabling highly accurate prediction of when service interruptions may occur and enabling proactive actions to be taken before service interruptions actually occur. Various exemplary embodiments for supporting predictive care of customer premises equipment may be configured to combine the analysis of operational data (e.g., customer premises equipment statistics, network statistics, or similar, and various combinations thereof) with historical customer care workflow data (e.g., tickets, workflows, and / or dispatch data) to provide predictive indicators of customer care issues (e.g., occurrence of customer care issues on customer premises equipment, and / or reporting of customer care issues by customers) within a specific time frame (e.g., within the next 24 hours, 48 ​​hours, one week, one month, etc.) with a specific level of confidence (e.g., at least 70% confidence, at least 75% confidence, at least 85% confidence, etc.).

[0079] Various exemplary embodiments for supporting predictive care for customer premises equipment (CAMEs) can be configured to provide a framework that automatically learns various threshold points, which, when crossed or violated, trigger a customer care call, ticket, or dispatch. These thresholds are validated in a complete feedback loop across a wide range of metrics and data points from the CAMEs (e.g., Wide Area Network (WAN) statistics related to the WAN supporting the CAMEs, Radio Access Network (RAN) statistics related to the RAN supporting the CAMEs, WiFi statistics related to the WiFi network supporting the CAMEs, Operating System (OS) statistics related to the operation of the CAMEs' OS, Application statistics related to applications running on the CAMEs, Environmental statistics related to the CAMEs, and various combinations thereof). After learning the thresholds, the framework can then apply various predictive and proactive automated actions to correct problems and prevent triggering customer care calls, tickets, or dispatches. This has a significant impact on reducing the service provider's operating expenses (OPEX) costs while increasing customer satisfaction.

[0080] Various exemplary embodiments for supporting predictive care of customer premises equipment (CME) can be used to calculate threshold points for various dimensions of CEM health and performance data using historical CEM data (e.g., CEM care ticket, workflow, and dispatch data from a service provider's CEM care management platform that shows when and why an end user experienced a problem significant enough to trigger a care call, ticket, or dispatch) and CEM operation statistics data (e.g., in the form of CEM operation data curves generated based on the storage of CEM operation statistics data in a time-series database) to detect CEM conditions before relevant problems occur, and therefore before care calls, tickets, and / or dispatches that might otherwise occur occur. The threshold points for various dimensions of CEM health and performance data that can be used to detect CEM conditions before relevant problems occur can be calculated based on the use of machine learning to cross-reference historical CEM data with time-based CEM operation statistics data. Next, using the obtained threshold points for various dimensions of customer premises equipment health and performance data, predictive signatures can be defined that can be used to detect various types of conditions in customer premises equipment before related problems occur, and therefore before care calls, tickets, and / or dispatches that might otherwise occur.

[0081] Various exemplary embodiments for supporting predictive care of customer premises equipment may be configured to support predictive, proactive care rather than reactive care (e.g., Figure 6), where problems are resolved only after they have actually occurred by accurately predicting impending problems before they occur and taking proactive steps to prevent them from occurring. Various exemplary embodiments for supporting predictive care of customer premises equipment may be configured to support predictive care of customer premises equipment in highly dynamic environments (e.g., changing conditions on customer premises equipment, changes in software on customer premises equipment, changes in network conditions in the network supporting customer premises equipment, and various combinations thereof) without requiring constant intervention (e.g., planning, tuning, etc.) which generally result in significantly large false positives, complexity, and costs.

[0082] These and various other exemplary embodiments for supporting predictive care of customer premises equipment, as well as the advantages or potential advantages of these exemplary embodiments for supporting predictive care of customer premises equipment, can be further understood by referring to the various figures discussed further below.

[0083] Figure 1 shows an exemplary embodiment of a communication system configured to support predictive care of customer premises equipment.

[0084] The communication system 100 includes a set of customer premises 110-1 to 110-N (collectively, customer premises 110) which include customer premises equipment 111-1 to 111-N (collectively, customer premises equipment 111), a communication network 120 which supports communication of customer premises equipment 111, and a customer premises equipment management system 130 which is configured to support the management of customer premises equipment 111. The customer care system 140 may be configured to support customer care for customer premises equipment 111, and the customer premises equipment predictive care system 150 may be configured to support various functions for providing predictive care for customer premises equipment 111. It will be understood that the communication system 100 may include various other elements which have been omitted for clarity.

[0085] The customer premises 110 represents a customer location where the customer premises equipment 111 is deployed. For example, the customer premises 110 may include a home, a business, etc. The customer premises equipment 111 may be a device that acts as an interface between the customer premises 110 and a communication network 120 that supports communication for the customer premises 110. For example, the customer premises equipment may include an optical network terminal (ONT), a residential gateway, a fixed wireless access gateway, etc. The customer premises equipment 111 may be configured to support communication for various other customer devices that may be located in the customer premises 110 (e.g., smartphones, computers, set-top boxes, smart TVs, game systems, smart appliances, Internet of Things (IoT) devices, etc., and various combinations thereof), which have been omitted for clarity.

[0086] The customer premises equipment 111 may be configured to support various functions that can be used to support customer care (and therefore service assurance) of the customer premises equipment 111, including supporting the collection and reporting of operational statistics related to the operation of the customer premises equipment 111 (e.g., customer premises equipment health statistics, customer premises equipment performance statistics, and various combinations thereof), and supporting actions (e.g., reconfiguration, reset, restart, etc.) requested by management systems (e.g., customer premises equipment management system 130, customer premises equipment management system 140, and / or customer premises equipment predictive care system 150) within the context of supporting customer care for the customer premises equipment 111, as well as various combinations thereof.

[0087] As described above, the customer premises equipment 111 may be configured to support customer care by supporting the collection and reporting of operational statistics related to the operation of the customer premises equipment 111. For example, operational statistics collected and reported by the customer premises equipment 111 may include wide area network statistics related to the wide area network supporting the customer premises equipment 111, wireless access network statistics related to the wireless access network supporting the customer premises equipment 111, WiFi statistics related to the WiFi network supporting the customer premises equipment 111, operating system statistics related to the operation of the operating system of the customer premises equipment 111, application statistics related to applications running on the customer premises equipment 111, environmental statistics related to the customer premises equipment 111, and at least one of various combinations thereof. It will be understood that the customer premises equipment 111 may be configured to support the collection and reporting of various other types of operational statistics related to the operation of the customer premises equipment 111.

[0088] As described above, the customer premises equipment 111 may be configured to support customer care by supporting actions requested by a management system (e.g., customer premises equipment management system 130, customer care system 140, and / or customer premises equipment predictive care system 150) under conditions that support customer care for the customer premises equipment 111. For example, actions requested by a management system and performed by the customer premises equipment 111 may include reconfiguration actions (e.g., downloading and installing updated operating settings used by the customer premises equipment 111 to handle customer traffic, downloading and installing updated software (e.g., operating system software, protocol stack software, etc.) on the customer premises equipment 111), triggering the customer premises equipment 111 to switch to different communication channels (e.g., different optical wavelengths, different WiFi channels, etc.), as well as various combinations thereof, resetting the customer premises equipment 111, rebooting the customer premises equipment 111, and various combinations thereof. It will be understood that the customer premises equipment may be configured to support various other actions requested by a management system under conditions that support customer care for the customer premises equipment 111.

[0089] The communication network 120 may be configured to support communication of customer premises equipment 111 within the customer premises 110, as well as communication of various elements involved in the management and customer care of the customer premises equipment 111 within the customer premises 110. The communication network 120 may be based on various types of communication technologies, depending on the type of customer premises equipment 111 supported. For example, the communication network 120 may support wired access by the customer premises equipment 111 (e.g., optical networks and networking in fiber-to-the-home (FTTH) networks, cable-based networks and networking, and various combinations thereof), wireless access by the customer premises equipment 111 (e.g., cellular-based access, WiFi-based access, and various combinations thereof), and various combinations thereof. The communication network 120 may include various communication elements (omitted for clarity) that can support communication of customer premises equipment 111 within the customer premises 110, as well as communication of various elements involved in the management and customer care of customer premises equipment 111 within the customer premises 110, such as wired communication devices (e.g., routers, switches, gateways, etc.), wireless communication devices (e.g., wireless access network (RAN) devices, wireless access points (WAPs), core wireless network elements, etc.), communication and communication management support elements (e.g., policy controllers, policy and billing rule functions, etc.), and various combinations thereof. It will be understood that the communication network 120 may support various other communication functions.

[0090] The customer premises equipment management system 130 may be configured to support the management of customer premises equipment 111. The customer premises equipment management system 130 may be configured to support the configuration of customer premises equipment 111, monitoring of customer premises equipment 111, reconfiguration of customer premises equipment 111, control over the operation of customer premises equipment 111, and various combinations thereof. The customer premises equipment management system 130 may be configured to control the reporting of operational statistics measured by customer premises equipment 111 (e.g., periodically polling customer premises equipment 111 for various operational statistics, configuring customer premises equipment 111 to periodically measure and report various operational statistics, requesting customer premises equipment 111 for various operational statistics under various conditions, and various combinations thereof). The customer premises equipment management system 130 may be configured to receive and store customer premises equipment operation statistics data measured and provided by the customer premises equipment 111, where such customer premises equipment operation statistics data may include raw operation statistics data received from the customer premises equipment 111, processed operation statistics data generated based on processing of the raw operation statistics data received from the customer premises equipment 111, and various combinations thereof.

[0091] The customer premises equipment management system 130 may be configured to support predictive care functions to support predictive care of customer premises equipment 111. The customer premises equipment management system 130 may be configured to acquire customer premises equipment operation statistics data measured and provided by customer premises equipment 111 for use by the customer premises equipment predictive care system 150 when providing predictive care to customer premises equipment 111. The customer premises equipment management system 130 may be configured to generate and maintain time-based customer premises equipment operation statistics data 131 for use by the customer premises equipment predictive care system 150 when providing predictive care to customer premises equipment 111. For example, the customer premises equipment management system 130 may generate time-based customer premises equipment operation statistics data 131 by acquiring operation statistics measured by customer premises equipment 111 and a timestamp indicating the time when the operation statistics were measured by customer premises equipment 111. The timestamp is then associated with the operation statistics to form the time-based customer premises equipment operation statistics data 131. The customer premises equipment management system 130 may be configured to support predictive care actions initiated by the customer premises equipment predictive care system 150 when providing predictive care to the customer premises equipment 111. For example, the customer premises equipment management system 130 may be configured to receive requests for action execution on the customer premises equipment 111 (e.g., requests from the customer premises equipment predictive care system 150 for reconfiguration, reset, or restart of the customer premises equipment 111) and to send messages to the customer premises equipment 111 to trigger the execution of actions on the customer premises equipment 111 (e.g., based on one or more messages sent to the customer premises equipment 111). It will be understood that the customer premises equipment management system 130 may be configured to support a variety of other predictive care functions to support predictive care for the customer premises equipment 111.

[0092] The time-based customer premises equipment operation statistics data 131 may include various types of customer premises operation statistics data, may be maintained in various forms, and may be various combinations thereof. The time-based customer premises equipment operation statistics data 131 may be maintained in a time-series database (TSDB), from which time-based operation statistics curves and / or time-based operation statistics traces can be generated. The time-based customer premises equipment operation statistics data 131 may include such time-based operation statistics curves and / or time-based operation statistics traces. It should be understood that the time-based customer premises equipment operation statistics data 131 may include various other types of customer premises operation statistics data, may be maintained in various other forms, and may be various combinations thereof.

[0093] The customer care system 140 may be configured to support customer care for the customer premises equipment 111. The customer care system 140 may be configured to support various customer care activities for the customer premises equipment 111, including supporting customer care contact from customers of the customer premises equipment 111 (e.g., phone calls to report problems, online submissions to report problems, and various combinations thereof), customer care ticketing related to providing customer care to customers of the customer premises equipment 111 (e.g., opening tickets, resolving tickets, and various combinations thereof), customer care technician management (e.g., supporting remote activities by customer care technicians working to provide customer care to customers of the customer premises equipment 111, supporting the dispatch of customer care technicians to the customer premises 110 of customers of the customer premises equipment 111, and various combinations thereof), and various combinations thereof.

[0094] The customer care system 140 may be configured to support predictive care functions to support predictive care for customer premises equipment 111. The customer care system 140 may be configured to generate customer care data within conditions that support customer care for customer premises equipment 111, and to maintain customer care data associated with customer premises equipment 111, or similar data, as well as various combinations thereof. The customer care data may include customer care data contained in customer care tickets opened for customer premises equipment 111, customer care data extracted from customer care tickets opened for customer premises equipment 111, as well as various combinations thereof. The customer care system 140 may be configured to maintain historical customer care data 141 used by the customer premises equipment predictive care system 150 when providing predictive care for customer premises equipment 111. It will be understood that the customer care system 140 may be configured to support various other predictive care functions to support predictive care for customer premises equipment 111.

[0095] The historical customer care data 141 may include various types of customer care data. This may include raw historical customer care data (e.g., original customer care records, customer care data sourced from original customer care records), pre-processed historical customer care data (e.g., pre-processed customer care records, pre-processed customer care data sourced from original customer care records before processing), and various combinations thereof. For each customer care event with associated customer care records, the historical customer care data 141 may include various types of information that may be supplied from customer care records created based on the customer care event. For example, historical customer care data that can be obtained from customer care records created based on customer care events may include event information related to the customer care event (e.g., date and time of customer care contact by the customer, indication of whether a technician was dispatched to the customer's facility, and various combinations thereof), problem information related to the problem that caused the customer care event (e.g., problem information related to the problem that caused the customer care event (e.g., indication of the type of problem, details of the problem related to the problem, and various other combinations), and test information related to tests conducted to evaluate the problem that caused the customer care event (e.g., type of test conducted, test results of the test conducted, and their results). It may also include condition information related to the conditions that caused the problem that led to the customer care event (e.g., display of the type of condition, details of the conditions related to the condition, and various other combinations), resolution information (e.g., display of how the problem or potential problem that triggered the customer's customer care contact was resolved, display of how the conditions that caused the problem or potential problem that triggered the customer's customer care contact were resolved, and various combinations thereof). It will be understood that the historical customer care data 141 may also include various other types of data that can be obtained in the course of providing customer care to the customer premises equipment 111.

[0096] The customer premises equipment predictive care system 150 can be configured to support various functions for providing predictive care to customer premises equipment 111 in the customer premises 110. The customer premises equipment predictive care system 150 can be configured to acquire and learn time-based customer premises equipment behavior statistics data that shows the behavior of a set of customer premises equipment, and historical customer care data that identifies a set of customer care events triggered for a set of customer premises equipment. Based on the application of machine learning to the historical customer care data and time-based customer premises equipment behavior statistics data, learned thresholds for conditions experienced by at least a portion of the customer premises equipment define predictive signatures configured to predict that conditions will be experienced within a time frame for conditions experienced by at least a portion of the customer premises equipment. A predictive signature includes a condition instruction, a behavioral statistics threshold for the condition determined based on learned thresholds, and an action to be taken to address the condition. Based on the behavioral statistics of a given customer premises equipment, it is determined and initiated that a predictive signature has been detected for a given customer premises equipment. For a given customer premises equipment based on the predictive signature, the action should be taken to address the condition.

[0097] The customer premises equipment predictive care system 150 may be configured to acquire time-based customer premises equipment operation statistics 131 for customer premises equipment 111, which indicates the operation of a set of customer premises equipment 111. The customer premises equipment predictive care system 150 may acquire the time-based customer premises equipment operation statistics 131 locally (for example, if the time-based customer premises equipment operation statistics 131 is already provided to the customer premises equipment predictive care system 150), from the customer premises equipment management system 130, and from various combinations thereof. While primarily presented, it will be understood that the time-based customer premises equipment operation statistics 131 are generated by and acquired from the customer premises equipment management system 130, exemplary embodiments will be presented. In at least some exemplary embodiments, time-based customer premises equipment operation statistics data 131 may be generated by the customer premises equipment predictive care system 150 (for example, by obtaining a set of operation statistics measured by the customer premises equipment 111 and a set of timestamps indicating the time each operation statistic was measured by the customer premises equipment 111, and associating the timestamps with the operation statistics to form time-based customer premises equipment operation statistics data 131).

[0098] The customer premises equipment predictive care system 150 may be configured to acquire historical customer care data 141 for customer premises equipment 111, identifying a set of customer care events triggered for customer premises equipment 111. The customer premises equipment predictive care system 150 may acquire the historical customer care data 141 locally (e.g., if the historical customer care data 141 has already been provided to the customer premises equipment predictive care system 150), from the customer care system 140, and from various combinations thereof. While the present invention is primarily presented with respect to exemplary embodiments in which the historical customer care data 141 is generated by and acquired from the customer care system 140, it will be understood that in at least some exemplary embodiments, the historical customer care data 141 may be generated by the customer premises equipment predictive care system 150 (e.g., for customer premises equipment 111, by acquiring a set of customer care records generated based on customer care communications between the service provider and the customer of the customer premises equipment 111, and processing the customer care records to provide the historical customer care data 141).

[0099] The customer premises equipment predictive care system 150 may be configured to learn learned thresholds for conditions experienced by at least some of the customer premises equipment, based on the application of machine learning to historical customer care data 141 and time-based customer premises equipment operation statistics data 131. The use of machine learning to learn learned thresholds for a particular type of condition may involve using machine learning to match a timestamp of the time when the customer premises equipment 111 experienced that particular type of condition (as determined from historical customer care data 141) with a timestamp of the time when operation statistics were measured by the customer premises equipment 111 (as determined from time-based customer premises equipment operation statistics data 131), so associating the occurrence of that particular type of condition with the operation conditions in the customer premises equipment 111 when an instance of that particular type of condition occurred. The results of the matching can be further processed based on machine learning to determine the trend of the customer premises equipment 111's operation statistics when the condition occurred, and the trend of the customer premises equipment 111's operation statistics when the condition occurred can be used to determine learned thresholds for a particular type of condition. Since learned thresholds can be used to evaluate the operational statistics of customer premises equipment 111 to determine whether a condition is predictable for customer premises equipment 111, it will be understood that learned thresholds for conditions can also be considered operational statistics thresholds.

[0100] The customer premises equipment predictive care system 150 can be configured to define predictive signatures 151 that the customer premises equipment 111 will use to predict problems with the customer premises equipment 111 before problems actually occur, based on the conditions that the customer premises equipment 111 will experience. A predictive signature 151 defined for a customer premises equipment condition (or state) may be configured to represent the customer premises equipment condition (or state) with a given health or performance statistical threshold and a set of one or more automated actions to be taken to resolve the customer premises equipment condition before problems actually occur as a result of the customer premises equipment condition. A predictive signature 151 defined for a customer premises equipment condition may also be configured to predict that the condition will be experienced within a time frame, and the predictive signature 151 includes a description of the condition, an action statistical threshold for the condition determined based on learned thresholds, and an action to be taken to address the condition. It will be understood that various types of predictive signatures 151 can be configured and defined to predict the occurrence of various types of conditions before the occurrence of problems (e.g., errors, failures, etc.) resulting from such conditions.

[0101] For example, a memory leak prediction signature may be created and applied to predict and prevent the occurrence of memory leak conditions on a CPE. A memory leak prediction signature may be defined as follows: (1) Condition: Memory leak, (2) Condition: The FreeMemoryStatus counter on the CPE is less than or equal to 5% of the total available memory on the CPE, and (3) Action: Reboot the CPE (e.g., immediately, or scheduled during inactive time).

[0102] For example, a WiFi interference prediction signature can be created and applied to predict and prevent the occurrence of WiFi interference conditions at a CPE. A WiFi interference prediction signature may be defined as follows: (1) Condition: WiFi interference, (2) Condition: Wireless interface PacketsDropped and PacketsErrored, and (3) Action: Initiate a WiFi channel scan for new and best WiFi channel selection.

[0103] For example, an optical temperature anomaly prediction signature may be created and applied to predict and prevent the occurrence of optical temperature anomaly conditions in CPE. An optical temperature anomaly prediction signature may be defined as follows: (1) Condition: Optical temperature anomaly, (2) Condition: ONT Transiver Temperature, and (3) Action: Initiate a predictive technician dispatch to resolve the issue before potential damage or failure of the equipment occurs.

[0104] The customer premises equipment predictive care system 150 may be configured to maintain predictive signatures 151 for use in predicting problems with customer premises equipment 111 before problems actually occur. The customer premises equipment predictive care system 150 may be configured to monitor customer premises equipment 111 based on the predictive signatures 151 to predict the occurrence of conditions that are expected to cause problems with customer premises equipment 111 and to initiate preventive corrective actions to prevent problems from occurring with customer premises equipment 111. This improves the customer experience by preventing the occurrence of problems that would negatively impact the customer experience of customers of customer premises equipment 111 and reduces the service provider's costs by eliminating the need for costly customer care contacts between customers of customer premises equipment 111 and the service provider's customer care resources. The application of predictive signatures 151 to support predictive care for specific customer premises equipment 111 will be discussed further below.

[0105] The customer premises equipment predictive care system 150 may be configured to generate predictive indices for conditions as output of applying machine learning to time-based customer premises equipment operation statistics data 131 and historical customer care data 141 for conditions in which predictive signatures are generated. The predictive index may be based on two values: (1) a probability value indicating the probability that a customer will initiate a customer care contact for a condition (e.g., expressed as a percentage or other appropriate value), and (2) a time frame value associated with the probability value, where the time frame value indicates a time frame in which a customer care contact may occur (may occur with a probability defined by the probability value). For example, for a particular condition, there may be a predictive index indicating a 75% probability that the particular condition will cause an issue resulting in a customer care contact within the next 48 hours. Similarly, for example, for a particular condition, there may be a predictive index indicating an 83% probability that the particular condition will cause an issue resulting in a customer care contact within the next 72 hours. Similarly, for example, for a particular condition, there may be a predictive index indicating a 97% probability that the particular condition will cause an issue resulting in a customer care contact within the next 24 hours. It will be understood that multiple such predictive indices may be defined for several conditions. It will be understood that such predictive metrics may be used to support predictive care of customer premises equipment. It will be understood that such predictive metrics may form part of predictive signatures, may be used to support predictive care of customer premises equipment, may be maintained separately from predictive signatures, but may be used together with predictive signatures to support predictive care of customer premises equipment, as well as in various combinations thereof. It will be understood that predictive metrics may be adapted over time by machine learning processes (e.g., as new time-based customer premises equipment operation statistics data 131 and / or historical customer care data 141 are acquired and processed based on machine learning, etc.), as well as in various combinations thereof.

[0106] The customer premises equipment predictive care system 150 determines, based on operational statistics of a given customer premises equipment 111, that a predictive signature 151 has been detected for the given customer premises equipment 111, and can be configured to initiate an action defined by the detected predictive signature 151 for the given customer premises equipment 111 based on the detected predictive signature 151, which proactively prevents problems that may arise from the conditions defined by the detected predictive signature. In this way, predictive care is provided to the given customer premises equipment 111 to improve the customer experience of the customer of the given customer premises equipment 111 and to reduce the cost for the service provider to service the given customer premises equipment 111.

[0107] The customer premises equipment predictive care system 150 may be configured to determine, based on the operational statistics of a given customer premises equipment 111, that a predictive signature 151 has been detected for that customer premises equipment 111.

[0108] The customer premises equipment predictive care system 150 may be configured to obtain operational statistics for a given customer premises equipment 111 in various ways for use in determining whether any of the predictive signatures 151 hit the given customer premises equipment 111. For example, the customer premises equipment predictive care system 150 may receive operational statistics for a given customer premises equipment 111 directly from the customer premises equipment 111, or from the customer premises equipment management system 130 (for example), and the customer premises equipment management system may monitor the customer premises equipment 111 to obtain operational statistics from the customer premises equipment 111. For example, the customer premises equipment predictive care system 150 may receive operational statistics for a given customer premises equipment 111 periodically, in response to one or more conditions, or equivalent, and in various combinations thereof. It will be understood that the customer premises equipment predictive care system 150 may be configured to obtain operational statistics for a given customer premises equipment 111 in various other ways.

[0109] The customer premises equipment predictive care system 150 may be configured to search for predictive signatures 151 based on the operational statistics of a given customer premises equipment 111 to determine whether any of the conditions defined by any of the predictive signatures 151 are currently being experienced by the given customer premises equipment 111. The customer premises equipment predictive care system 150 can search for predictive signatures 151 by evaluating the condition field of the predictive signatures 151 (which includes learned operational statistics thresholds on which the operational statistics of the customer premises equipment 111 can be compared, and which are also referred to herein as the "Stat" field in the example predictive signatures provided above) based on the operational statistics of a given customer premises equipment 111 to determine whether a given customer premises equipment 111 is experiencing any of the conditions of the predictive signatures 151. The search for predictive signatures 151 based on the operational statistics of a given customer premises equipment 111 to determine whether any of the conditions defined by any of the predictive signatures 151 are currently being experienced by a given customer premises equipment 111 can be further understood by considering specific embodiments of the predictive signatures described above.

[0110] For example, the operational statistics of a given customer premises equipment 111 include a FreeMemoryStatus counter value indicating that the free memory of the given customer premises equipment 111 is 4% of the total available memory of the given customer premises equipment 111, and a “memory leak” condition is predicted for the given customer premises equipment 111. For example, a “memory leak” prediction signature is identified as being associated with the given customer premises equipment 111 based on a comparison between a 4% FreeMemoryStatus counter value and a 5% FreeMemoryStatus counter threshold, as defined in the “condition” portion of the “memory leak” prediction signature. As will be discussed further below, the action associated with the “memory leak” prediction signature may then be initiated on the given customer premises equipment 111 to resolve the condition and prevent any resulting problems (e.g., service interruption, failure, or equivalent) that would otherwise be detected by the customer and reported to the service provider by the customer at the customer care contact.

[0111] For example, the operational statistics of a given customer premises equipment 111 include wireless interface PacketsDropped and PacketsErrored values ​​indicating WiFi interference at the given customer premises equipment 111, and the “WiFi interference” condition is predicted for the given customer premises equipment 111. For example, a “WiFi interference” prediction signature is identified as being associated with a given customer premises equipment 111 based on a comparison of the wireless interface PacketsDropped and PacketsErrored values ​​from the operational statistics of the given customer premises equipment 111 with the wireless interface PacketsDropped and PacketsErrored thresholds, respectively, defined in the “Stat” portion of the “WiFi interference” prediction signature. As will be discussed further below, the associated action of the “WiFi interference” prediction signature may then be initiated for the given customer premises equipment 111 to resolve the condition and prevent any resulting problems (e.g., service interruption, failure, or equivalent) that would otherwise be detected by the customer and reported to the service provider by the customer in a customer care contact.

[0112] For example, the operational statistics of a given customer premises equipment 111 include an ONT TransiverTemperature value indicating an optical temperature anomaly in the given customer premises equipment 111, and the “optical temperature anomaly” condition is predicted for the given customer premises equipment 111. For example, an “optical temperature anomaly” prediction signature is identified as being associated with the given customer premises equipment 111 based on a comparison between the ONT TransiverTemperature value from the operational statistics of the given customer premises equipment 111 and the ONT TransiverTemperature threshold defined in the “Stat” portion of the “optical temperature anomaly” prediction signature. As will be discussed further below, the associated action of the “optical temperature anomaly” prediction signature may then be initiated on the given customer premises equipment 111 to resolve the condition and prevent any resulting problems (e.g., service interruption, failure, or equivalent) that would otherwise be detected by the customer and reported to the service provider by the customer at the customer care contact.

[0113] While primarily presented, this method relates to the detection of a single predictive signature 151 associated with a given customer premises equipment 111. However, it will be understood that multiple predictive signatures 151 may be detected for a given customer premises equipment 111 based on its operational statistics.

[0114] The customer premises equipment predictive care system 150 may be configured to initiate an action defined by a detected predictive signature 151 for a given customer premises equipment 111, based on a predictive signature 151 detected for the given customer premises equipment 111 based on operational statistics of the given customer premises equipment 111. The actions defined by the detected predictive signature 151 may include reconfiguration actions (e.g., one or more of the following) configured to cause the given customer premises equipment 111 to be reconfigured, such as port or port configuration(s), interface or interface configuration(s), memory or memory configuration(s), OS or OS configuration(s), channel in use, reset actions(s), and various combinations thereof(s), reset actions (e.g., an action to reset the given customer premises equipment 111 and / or one or more elements or components of the given customer premises equipment 111), restart actions (e.g., an action to restart the given customer premises equipment 111), and various combinations thereof. The detected predictive signature 151 may include, or alternatively, actions to trigger customer care contact between the given customer premises equipment 111 and the customer (e.g., a call from a customer care technician to the customer, dispatching a customer care technician to the customer premises 110 for the given customer premises equipment 111, and various combinations thereof.

[0115] The customer premises equipment predictive care system 150 can initiate actions defined by the detected predictive signature 151 in various ways. The customer premises equipment predictive care system 150 may initiate actions defined by the detected predictive signature 151 by sending one or more messages configured to cause a given customer premises equipment 111 to perform the actions defined by the detected predictive signature 151 (for example, on behalf of the given customer premises equipment 111, by the given customer premises equipment 111, or equivalents, and various combinations). For example, the customer premises equipment predictive care system 150 may send one or more action messages to the given customer premises equipment 111 to cause one or more actions to be performed by the given customer premises equipment 111. The customer premises equipment predictive care system 150 may send to 40 to cause the customer care system 140 to perform one or more actions on behalf of a given customer premises equipment 111, or may send one or more action messages to the customer premises equipment management system 130 to cause the customer care system 140 to initiate one or more messages to a given customer premises equipment 111, to cause the customer premises equipment management system 130 to perform one or more actions on behalf of a given customer premises equipment 111, and / or to cause the customer premises equipment management system 13 to initiate one or more messages, and it will be understood that the actions defined by the detected predictive signature 151 can be initiated in a variety of other ways.

[0116] For example, a “memory leak” predictive signature is identified as being associated with a given customer premises equipment 111 based on the operational statistics of the customer premises equipment 111 and the “status” portion of the “memory leak” predictive signature. The associated action of the “memory leak” predictive signature may then be initiated on the given customer premises equipment 111 to resolve the condition and prevent any resulting problems (e.g., service interruption, failure, or equivalent) that would otherwise be detected by the customer and reported by the customer to the customer care contact service provider. As described above, the action indicated for the “memory leak” predictive signature includes restarting the given customer premises equipment 111, and therefore the customer premises equipment predictive care system 150 sends a message to the given customer premises equipment 111 with instructions to be performed when the given customer premises equipment 111 is restarted.

[0117] For example, a “WiFi interference” prediction signature is identified as being associated with a given customer premises equipment 111 based on the operational statistics of the customer premises equipment 111 and the “Stat” portion of the “WiFi interference” prediction signature. The associated action for the “WiFi interference” prediction signature may then be initiated on the given customer premises equipment 111 to resolve the condition and prevent any resulting problems (e.g., service interruption, failure, etc.) that would otherwise be detected by the customer and reported by the customer to the customer care contact service provider. As described above, the action indicated for the “WiFi interference” prediction signature includes initiating a WiFi channel scan for a new best WiFi channel selection. Therefore, the customer premises equipment prediction care system 150 sends a message to the given customer premises equipment 111 with a command to perform a WiFi channel scan on the given customer premises equipment 111 so that the given customer premises equipment 111 switches to a better WiFi channel.

[0118] For example, a “optical temperature anomaly” predictive signature is identified as being associated with a given customer premises equipment 111 based on the operational statistics of the customer premises equipment 111 and the “Stat” portion of the “optical temperature anomaly” predictive signature. The associated action for the “optical temperature anomaly” predictive signature may then be initiated for the given customer premises equipment 111 to resolve the condition and prevent any resulting problems (e.g., service interruption, failure, or equivalent) that would otherwise be detected by the customer and reported by the customer to the customer care contact service provider. As described above, the action indicated for the “optical temperature anomaly” predictive signature includes initiating a predictive technician dispatch to resolve the issue before any possible equipment damage or failure occurs, and therefore the customer premises equipment predictive care system 150 sends a message to the customer care system 140 to schedule a technician dispatch to the customer premises 110 associated with the given customer premises equipment 111.

[0119] The customer premises equipment predictive care system 150 may be configured to support a customer premises equipment "centralized care" mode in which, under certain conditions, a specific customer premises equipment may be automatically placed into a highly monitored condition (e.g., a highly monitored condition with more frequent collection and analysis of operating parameters from the customer premises equipment). This may be applied to customer premises equipment for a variety of reasons. For example, the "centralized care" mode may be applied to customer premises equipment associated with a new customer (e.g., ensuring a "first-time-right" setup by defaulting new customers to an accelerated data collection and analysis phase with less resource and storage concentration). For example, the "centralized care" mode may be applied to customer premises equipment based on the detection of a predictive signature of the customer premises equipment (e.g., the customer premises equipment is automatically placed into a detailed data collection and analysis phase for 48 hours, and if the associated conditions are not automatically clarified, the customer premises is prompted to resolve through automated suboptimal actions or referenced tiered support).

[0120] It will be understood that the customer premises equipment predictive care system 150 may be configured to support various other predictive care capabilities to support predictive care of customer premises equipment.

[0121] Although primarily presented as a standalone system, the various functions of the customer premises equipment management system 130 and the customer care system 140 are understood, and the customer premises equipment predictive care system 150 can be combined in various ways (for example, the customer premises equipment management system 130 and the customer care system 140 can be implemented together, the customer care system 140 and the customer premises equipment predictive care system 150 can be implemented together, or similar), distributed in various other ways, or similar, and various combinations thereof can be made.

[0122] Figure 2 shows an exemplary embodiment of a method for supporting predictive care of customer premises equipment. Although presented herein primarily as being performed sequentially, it will be understood that at least some of the functions of Method 200 may be performed simultaneously or in a different order than that presented in Figure 2.

[0123] In block 201, method 200 is initiated.

[0124] In block 210, a predictive signature is defined that predicts that a condition experienced by a set of customer premises equipment will be experienced within a time frame, based on the use of machine learning to process historical customer care data and time-based customer premises equipment operation statistics data associated with the set of customer premises equipment, the predictive signature including an indication of the condition, an operation statistics threshold for the condition, and an action to be taken to address the condition. It will be understood that in at least some exemplary embodiments, block 210 may be provided using the method or a part thereof as shown in Figures 3 and 4.

[0125] In block 220, based on the determination that a predictive signature has been detected for a given customer premises device based on the operational statistics of that device, an action to be taken to address the condition is initiated. In at least some exemplary embodiments, it will be understood that block 220 may be provided using the method or a portion thereof shown in Figure 4.

[0126] Method 200 terminates in block 299.

[0127] It will be understood that the various predictive care capabilities illustrated and described within the context of the communication system 100 in Figure 1 may also be incorporated into the method 200 in Figure 2, and / or alternatively.

[0128] Figures 3 and 4 show exemplary embodiments of a method for generating predictive signatures for use in supporting predictive care of customer premises equipment. It will be understood that in at least some exemplary embodiments, the method 300 (or a portion thereof) of Figures 3 and 4 can be used to provide the block 210 of Figure 2. Although presented herein primarily as being performed sequentially, it will be understood that at least some of the functions of method 300 may be performed simultaneously or in a different order than those presented in Figures 3 and 4.

[0129] In block 301, method 300 is initiated.

[0130] In block 310, time-based customer premises equipment operation statistics data are obtained for the customer premises equipment set, indicating the operation of the customer premises equipment set.

[0131] In block 320, historical customer care data is obtained that identifies the set of customer care events triggered for a set of customer premises equipment.

[0132] In block 330, learned thresholds for conditions experienced by at least some of the customer premises equipment are learned based on the application of machine learning to historical customer care data and time-based customer premises equipment operation statistics data.

[0133] In block 340, for conditions experienced by at least some of the customer's premises equipment, a predictive signature is defined that is configured to predict whether the condition will be experienced within a time frame, the predictive signature including a description of the condition, an action statistics threshold for the condition determined based on learned thresholds, and an action to be taken to address the condition.

[0134] Method 300 terminates in block 399.

[0135] It will be understood that the various predictive care capabilities illustrated and described within the context of the communication system 100 in Figure 1 may also be incorporated into the method 300 in Figures 3 and 4, and / or alternatively.

[0136] Figure 5 depicts an exemplary embodiment of a method for using predictive signatures to support predictive care of customer premises equipment. It will be understood that in at least some exemplary embodiments, the method 400 (or a portion thereof) of Figure 5 can be used to provide block 220 of Figure 2. Although presented herein primarily as being performed sequentially, it will be understood that at least some of the functions of method 400 may be performed simultaneously or in a different order than that presented in Figure 5.

[0137] In block 401, method 400 is initiated.

[0138] In block 410, based on the operational statistics of a given customer premises equipment, it is determined that a predictive signature has been detected for the given customer premises equipment, the predictive signature being configured to indicate that a certain condition will be experienced by the customer premises within a time frame, and the predictive signature includes an indication of the condition, an operational statistics threshold for the condition, and an action to be taken to address the condition.

[0139] In block 420, based on the predictive signature, an action is initiated for a given customer premises device to address the condition.

[0140] Method 400 terminates in block 499.

[0141] It will be understood that the various predictive care capabilities illustrated and described within the context of the communication system 100 in Figure 1 may also be incorporated into method 400 in Figure 5, and / or alternatively.

[0142] Figures 6 and 7 illustrate exemplary embodiments of a method for supporting predictive care of customer premises equipment. While presented herein primarily as being performed sequentially, it will be understood that at least some of the functions of Method 500 may be performed simultaneously or in a different order than those shown in Figures 6 and 7.

[0143] In block 501, method 500 is initiated.

[0144] In block 510, time-based customer premises equipment operation statistics data are obtained for the customer premises equipment set, indicating the operation of the customer premises equipment set.

[0145] In block 520, historical customer care data is obtained that identifies the set of customer care events triggered for a set of customer premises equipment.

[0146] In block 530, learned thresholds for conditions experienced by at least some of the customer premises equipment are learned based on the application of machine learning to historical customer care data and time-based customer premises equipment operation statistics data.

[0147] In block 540, for conditions experienced by at least some of the customer's premises equipment, a predictive signature is defined that is configured to predict whether the condition will be experienced within a time frame, the predictive signature including a description of the condition, an action statistics threshold for the condition determined based on learned thresholds, and an action to be taken to address the condition.

[0148] In block 550, based on the operational statistics of a given customer premises device, it is determined that a predictive signature has been detected for that customer premises device.

[0149] In block 560, based on the predictive signature, an action is initiated for a given customer premises device to address the condition.

[0150] Method 500 terminates in block 599.

[0151] It will be understood that the various predictive care capabilities illustrated and described within the context of the communication system 100 in Figure 1 may also be incorporated into the method 500 in Figures 6 and 7, and / or alternatively.

[0152] Various exemplary embodiments for supporting predictive care of customer premises equipment may offer various advantages or potential advantages. For example, various exemplary embodiments for supporting predictive care of customer premises equipment may be configured to support proactive, predictive care rather than reactive care by accurately predicting imminent problems before they occur and taking proactive steps to prevent them from occurring. For example, various exemplary embodiments for supporting predictive care of customer premises equipment may be configured to provide predictive care that prevents problems before they occur, thereby eliminating the need to initiate corrective action after a problem has occurred (e.g., based on the static use of a “counter” to detect the problem), and thus improving the quality of the customer experience and reducing the service provider’s OPEX by preventing unnecessary customer care contacts (e.g., phone calls and dispatches). For example, various exemplary embodiments for supporting predictive care of customer premises equipment may be configured to eliminate the need for static use of data points for reactive care of customer premises equipment (e.g., thresholds are statically configured by expensive domain knowledge) by supporting predictive care of customer premises equipment based on continuously learning and adapting thresholds of data points to support the prediction of problems before problems occur (e.g., evolving with customer premises equipment, networks, etc.). For example, various exemplary embodiments for supporting predictive care of customer premises equipment may be configured to provide predictive metrics for customer care problems within a specific time frame (e.g., within the next 24 hours, 48 ​​hours, week, month, or equivalent) so that preventive measures can be taken for problems before problems actually occur.

[0153] Various exemplary embodiments for supporting predictive care of customer premises equipment may be configured to enable the identification of signatures of customer premises equipment that indicate a tendency to experience conditions that could cause problems (e.g., errors, failures, or other issues that could negatively impact customer-facing services), thereby enabling service providers to proactively initiate actions to prevent or mitigate conditions that could cause problems if not addressed, thereby improving the customer's end-user experience (e.g., by preventing errors or failures before they occur) and reducing service provider expenses (e.g., reducing OPEX by avoiding the need to dispatch technicians to customer premises, reducing CAPEX by preventing issues that could result in damage to equipment such as customer premises equipment, and various combinations thereof).

[0154] For example, various exemplary embodiments for supporting predictive care of customer premises equipment may be configured to support predictive care of customer premises equipment in a highly dynamic environment (e.g., environment), where changes to conditions on customer premises equipment, changes to software on customer premises equipment, changes to network conditions in the network supporting customer premises equipment, or similar, as well as various combinations thereof, may exist without requiring constant intervention (e.g., planning, tuning, etc.), which typically results in significantly greater false positives, complexity, and cost.

[0155] For example, various exemplary embodiments for supporting predictive care of customer premises equipment may be configured to support a customer premises equipment "centralized care" mode in which, under certain conditions, certain customer premises equipment may be automatically placed under high-level monitoring conditions (e.g., conditions in which operational parameters from customer premises equipment are collected and analyzed more frequently). For example, various exemplary embodiments for supporting predictive care of customer premises equipment may be configured to eliminate the need for conventional thresholding and monitoring of counters and statistics, and the use of conventional alarms based on such thresholding and monitoring (e.g., Simple Network Management Protocol (SNMP) alarms, Network Configuration Protocol (NETCONF) alarms, or equivalents), which are generally not truly effective or adaptable, generally require significant intervention and support, all of which lead to increased OPEX by the service provider. Various exemplary embodiments for supporting predictive care of customer premises equipment may offer various other advantages or potential advantages.

[0156] Figure 8 shows an exemplary embodiment of a computer suitable for use in performing the various functions presented herein.

[0157] The computer 600 includes a processor 602 (e.g., a central processing unit (CPU), a processor, a processor having a set of processor cores, a processor core of a processor, etc.) and memory 604 (e.g., random access memory (RAM), read-only memory (ROM), etc.). In at least some exemplary embodiments, the computer 600 may include at least one processor and at least one memory that stores instructions, when executed by at least one processor, causing the computer to perform various functions presented herein.

[0158] Computer 600 may also include a cooperative element 605. The cooperative element 605 may be a hardware device. The cooperative element 605 may be a process that is loaded into memory 604 and executed by processor 602 to implement various functions presented herein (in this case, for example, the cooperative element 605 (including associated data structures) may be stored on a non-temporary computer-readable medium such as a storage device or other suitable type of storage element (e.g., a magnetic drive, an optical drive, etc.)).

[0159] Computer 600 may also include one or more input / output devices 606. The input / output devices 606 may include one or more of the following: user input devices (e.g., keyboard, keypad, mouse, microphone, camera, etc.), user output devices (e.g., display, speaker, etc.), one or more network communication devices or elements (e.g., input port, output port, receiver, transmitter, transceiver, etc.), one or more storage devices (e.g., tape drive, floppy drive, hard disk drive, compact disk drive, etc.), and various combinations thereof.

[0160] It will be understood that computer 600 may represent a general architecture and functionality suitable for implementing the functional elements described herein, parts of the functional elements described herein, and various combinations thereof. For example, computer 600 may provide a general architecture and functionality suitable for implementing one or more of the elements presented herein.

[0161] It will be understood that at least some of the functions presented herein may be implemented in software (for example, through a software implementation on one or more processors) to run on a general-purpose computer (for example, through execution on one or more processors) in order to provide a dedicated computer, and / or in hardware (for example, using a general-purpose computer, one or more application-specific integrated circuits, and / or any other hardware equivalents).

[0162] It will be understood that at least some of the functions presented herein can be implemented in hardware, for example, as circuits that work with a processor to perform various functions. Parts of the functions / elements described herein may be implemented as computer program products, and computer instructions, when processed by the computer, adapt the computer's operation such that the methods and / or techniques described herein are invoked or otherwise provided. Instructions for invoking various methods may be stored in a fixed or removable medium (e.g., a non-temporary computer-readable medium), transmitted via a data stream in a broadcast medium or other signal carrier medium, and / or stored in memory within a computing device operating according to the instructions.

[0163] As used herein, the term “non-transient” should be understood to refer to limitations of the medium itself (i.e., tangible rather than signal-based), as opposed to limitations of data storage persistence (e.g., RAM vs. ROM).

[0164] As used herein, “at least one of the <list of two or more elements>” and “at least one of the following” are understood to mean, where “list of two or more elements>” and similar phrases mean at least one of the elements, or at least two or more of the elements, or at least all of the elements, when the list of two or more elements is joined by “and” or “or”.

[0165] As used herein, the term “or” will be understood to mean a non-exclusive “or” unless otherwise indicated (e.g., the use of “other” or “or alternative”).

[0166] Although various embodiments incorporating the teachings presented herein are shown and described in detail herein, those skilled in the art will find that many other various embodiments incorporating these teachings can still be readily devised.

Claims

1. The device comprises at least one processor and at least one memory for storing instructions, and when an instruction is executed by the at least one processor, the device provides at least one The steps include obtaining time-series customer premises equipment operation statistics data showing the operation of a set of customer premises equipment, The steps include obtaining historical customer care data for the set of customer premises equipment, indicating when and why the end user experienced a problem, and identifying a set of customer care events triggered for the set of customer premises equipment, The steps include learning learned thresholds for conditions experienced by at least some of the customer premises equipment, based on the application of machine learning to the historical customer care data and the time-series customer premises equipment operation statistics data, An apparatus characterized by causing the apparatus to perform the steps of defining a predictive signature configured to predict that a condition experienced by at least a portion of the customer's premises equipment will be experienced within a time frame, wherein the predictive signature includes the condition, an operational statistical threshold for the condition determined based on the learned threshold, and an action to be taken to address the condition.

2. The apparatus according to claim 1, characterized in that the time-series customer premises equipment operation statistics data includes at least one of a set of time-series customer premises equipment health statistics related to one of the customer premises equipment, or a set of time-series customer premises equipment performance statistics related to one of the customer premises equipment.

3. The apparatus according to claim 1, characterized in that the time-series customer premises equipment operation statistics data includes at least one of a set of time-series operation statistics curves or a set of time-series operation statistics traces.

4. In order to obtain the aforementioned time-series customer premises equipment operation statistics data, the instruction, when executed by the at least one processor, is directed to the device. The steps include obtaining, for each set of customer premises equipment, a set of operational statistics measured by the customer premises equipment and a set of timestamps indicating the time each operational statistic was measured by each of the customer premises equipment, The apparatus according to claim 1, characterized in that it performs the step of associating the timestamp with the operation statistics to form the time-series customer premises equipment operation statistics data.

5. The apparatus according to claim 4, characterized in that the set of operational statistics measured by the customer premises equipment includes at least one of a set of customer premises equipment health statistics related to one of the customer premises equipment, or a set of customer premises equipment performance statistics related to one of the customer premises equipment.

6. The apparatus according to claim 4, wherein for at least one of the customer premises equipment, the set of operational statistics measured by each customer premises equipment includes at least one of the following: wide area network statistics related to the wide area network supporting each customer premises equipment; wireless access network statistics related to the wireless access network supporting each customer premises equipment; Wi-Fi statistics related to the Wi-Fi network supporting each customer premises equipment; operating system statistics related to the operation of the operating system of each customer premises equipment; application statistics related to the application running on each customer premises equipment; or environmental statistics related to each customer premises equipment.

7. The apparatus according to claim 4, wherein the set of operational statistics measured by the customer premises equipment is received from at least one of the set of customer premises equipment or a customer premises equipment management system configured to provide management functions for the set of customer premises equipment.

8. The apparatus according to claim 1, characterized in that the time-series customer premises equipment operation statistics data is obtained from a time-series database configured to store a set of operation statistics measured by the customer premises equipment and a set of timestamps indicating the time when each of the operation statistics was measured by each of the customer premises equipment.

9. The apparatus according to claim 1, wherein the historical customer care data includes at least one of the following: customer care ticket data for a set of customer care tickets opened for the set of customer premises equipment; customer care workflow data for a set of customer care workflow operations performed for the set of customer premises equipment; or customer care dispatch data for a set of customer care dispatch events performed for the set of customer premises equipment.

10. The apparatus according to claim 1, characterized in that the aforementioned historical customer care data is obtained from a service provider workflow management platform.

11. The apparatus according to claim 1, wherein the learned threshold for the condition represents the state of the condition that triggers a customer care action to address the condition.

12. The aforementioned prediction signature is, The aforementioned condition is a memory leak condition, the aforementioned operational statistics threshold is with respect to free memory statistics, and the aforementioned action to be taken includes a restart operation, along with a memory leak prediction signature. The aforementioned conditions are radio signal interference conditions, the operational statistics threshold relates to at least one of packet drop statistics or packet error statistics, and the action to be taken includes a radio interference prediction signature, which includes a wireless channel switching operation. The apparatus according to claim 1, characterized in that the condition is a temperature anomaly condition, the operational statistics threshold is related to customer premises equipment temperature statistics, and the action to be taken includes at least one of sending a message to the customer or dispatching a technician, and includes a temperature anomaly prediction signature.

13. When the instruction is executed by the at least one processor, the device will: A step of determining that a predictive signature has been detected for a predetermined customer premises device, based on the operational statistics of the predetermined customer premises device, wherein the predictive signature is configured to indicate that a certain condition is experienced by the customer premises within a time frame, and the predictive signature includes an indication of the condition, an operational statistics threshold for the condition, and an action to be performed to address the condition. The apparatus according to claim 1, characterized in that it causes the predetermined customer premises equipment to perform the step of initiating an action to be performed in order to address the conditions based on the predictive signature.

14. A method performed by an apparatus comprising at least one processor and at least one memory for storing instructions, wherein the method is: The steps include obtaining time-series customer premises equipment operation statistics data showing the operation of a set of customer premises equipment, The steps include obtaining historical customer care data for the set of customer premises equipment, indicating when and why the end user experienced a problem, and identifying a set of customer care events triggered for the set of customer premises equipment, The steps include learning learned thresholds for conditions experienced by at least some of the customer premises equipment, based on the application of machine learning to the historical customer care data and the time-series customer premises equipment operation statistics data, A method comprising the steps of defining a predictive signature configured to predict that a condition experienced by at least a portion of the customer's premises equipment will be experienced within a time frame, wherein the predictive signature includes the condition, an operational statistical threshold for the condition determined based on the learned threshold, and an action to be taken to address the condition.

15. The device comprises at least one processor and at least one memory for storing instructions, and when an instruction is executed by the at least one processor, the device provides at least one Steps include detecting, with respect to a given customer premises equipment, that the operational statistics of the given customer premises equipment have exceeded a threshold for a condition in a predictive signature for the given customer premises equipment, wherein the threshold is a learned threshold for the condition experienced by at least a portion of the customer premises equipment, based on the application of machine learning to historical customer care data and time-series customer premises equipment operational statistics data, the historical customer care data indicates when and why an end user experienced a problem and identifies a set of customer care events triggered for the set of customer premises equipment, the predictive signature is configured to indicate that the condition will be experienced by the customer premises within a time frame, and the predictive signature includes the condition, an operational statistics threshold for the condition determined based on the learned threshold, and an action to be taken to address the condition. An apparatus characterized by causing the device to perform the step of initiating an action to be performed on a predetermined customer premises device to address the conditions, based on the predictive signature.

16. In order to determine that the predictive signature has been detected for the predetermined customer premises equipment, the instruction, when executed by the at least one processor, sends at least the following to the device: The steps include: obtaining the operation statistics of the predetermined customer premises equipment; The apparatus according to claim 15, characterized in that it performs the step of determining that the predictive signature has been detected for the predetermined customer premises equipment based on the operation statistics of the predetermined customer premises equipment and the operation statistics threshold of the predictive signature.

17. The aforementioned prediction signature is, The aforementioned conditions are memory leak conditions, the operational statistics threshold is related to free memory statistics, and the action to be performed includes a restart operation, along with a memory leak prediction signature. The aforementioned conditions are radio interference conditions, the operational statistics threshold relates to at least one of packet drop statistics or packet error statistics, and the Action to be performed includes a radio channel switching operation, along with a radio signal interference prediction signature. The apparatus according to 15, characterized in that the above condition is a temperature anomaly condition, the operation statistics threshold is related to customer premises equipment temperature statistics, and the action to be performed includes at least one of sending a message to the customer or dispatching a technician, and includes a temperature anomaly prediction signature.

18. To initiate an action to be taken to address the aforementioned conditions, the instruction, once executed by the at least one processor, causes the device to: The steps include: identifying from the predictive signatures, based on the detection of the predictive signatures for the predetermined customer premises equipment, an action to be taken to address the conditions; The apparatus according to claim 15, characterized in that it performs the step of sending a message to a predetermined customer premises device, which includes instructions for an action to be taken to address the conditions.

19. The apparatus according to claim 15, characterized in that the actions performed to address the conditions include at least one of the following: reconfiguring the specified customer premises equipment, resetting the specified customer premises equipment, or restarting the specified customer premises equipment.

Images