# US 11,909,594 B2 · Purging IoT devices in a cellular network > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 11,909,594 B2 - Document ID: US11909594B2 - Status: Granted - Kind: B2 - Application number: 16/383,379 - Filed: 2019-04-12 - Published: 2020-10-15 - Granted: 2024-02-20 - Earliest claimed: 2019-04-12 - Assignees: T-Mobile USA, Inc. - Applicants: T-Mobile USA, Inc. - Inventors: Muhammad Tawhidur Rahman - Web page: https://patentgrep.com/patent/US11909594B2 ## CPC - H04L 41/145: {involving simulating, designing, planning or modelling of a network} - G06F 16/2379: {Updates performed during online database operations; commit processing} - G06N 20/10: using kernel methods, e.g. support vector machines [SVM] - H04L 41/142: using statistical or mathematical methods - H04L 67/12: specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks - H04L 67/30: Profiles - H04W 4/70: Services for machine-to-machine communication [M2M] or machine type communication [MTC] - G06N 3/088: Non-supervised learning, e.g. competitive learning - G06N 20/00: Machine learning - H04W 8/22: Processing or transfer of terminal data, e.g. status or physical capabilities - H04W 84/042: {Public Land Mobile systems, e.g. cellular systems} - H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M) - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS - H04W: WIRELESS COMMUNICATION NETWORKS (broadcast communication H04H; communication systems using wireless links for non-selective communication, e.g. wireless extensions H04M1/72) ## Abstract Processes and systems for identifying and removing non-functional Internet-of-Things (IoT) devices from a cellular network are discussed herein. In one example, a probability distribution function of a data profile transmitted by a type of IoT devices is generated. The probability distribution function may be chosen based on a known distribution such as a Gaussian, Bessel, or linear regressive model. The probability distribution function may be empirically generated by applying data received from IoT devices to a machine learning model using supervised or unsupervised learning. After generating the model representing the expected data profile of data received from a type or class IoT devices, data transmitted by the IoT devices may be applied to the model to determine whether the IoT devices are functional or non-functional. Non-functional IoT devices may be removed from the cellular network. ## Claims ### Claim 1 A method for removing a non-functioning Internet-of-Things (IoT) device from a cellular network, comprising: receiving, as training data, first data from first one or more IoT devices, the training data being collected over a time interval, including, after receiving a first portion of the training data: determining, during the time interval, characteristics of the first portion of the training data collected over the time interval; and determining, during the time interval, a length of the time interval over which the training data is collected based at least in part on the characteristics of the first portion of the training data, wherein at least a second portion of the training data is received during a remainder of the time interval after the determining of the length of the time interval based at least in part on the characteristics of the first portion of the training data; generating, based at least in part on the training data, a model representing an expected data profile to be transmitted to the cellular network from second one or more IoT devices; receiving, as received data, second data from an IoT device; comparing the received data from the IoT device to the model representing the expected data profile to be transmitted to identify the loT device as the non-functioning IoT device; and removing, based at least in part on comparing the received data to the model representing the expected data profile to be transmitted, the non-functioning IoT device from the cellular network. ### Claim 2 (depends on 1) The method of claim 1, wherein the model represents a plurality of classes of IoT devices, and wherein the method further comprises: determining a class of the plurality of classes associated with the IoT device; and wherein removing the non-functioning loT device from the cellular network is further based in part on the class associated with the loT device. ### Claim 3 (depends on 1) The method of claim 1, wherein generating the model includes: processing the received data in a time domain or in a frequency domain, wherein processing the received data in a time domain includes processing one or more amplitudes of the received data and processing in the frequency domain includes processing a frequency content of the received data. ### Claim 4 (depends on 1) The method of claim 1, wherein generating the model includes: receiving additional training data from the first one or more IoT devices over a fixed time interval; and updating the model based in part on the additional training data. ### Claim 5 (depends on 1) The method of claim 1, wherein generating the model includes: receiving additional training data from a non-functioning IoT device; and updating the model based in part on the additional training data, wherein the model represents a data profile transmitted by a non-functioning IoT device. ### Claim 6 (depends on 1) The method of claim 1, wherein generating includes: receiving additional data from the first one or more IoT devices; and updating the model based on the additional data. ### Claim 7 (depends on 1) The method of claim 1, further comprising: receiving an indication of a statistical representation of data expected to be transmitted by the IoT device; and wherein generating the model includes configuring the model based on the indication of the statistical representation of the data expected to be transmitted by the IoT device. ### Claim 8 (depends on 1) The method of claim 1, wherein the model represents a distribution of data that is indicative of a Gaussian distribution, a Bessel distribution, a linear distribution, or a stochastic distribution. ### Claim 9 (depends on 1) The method of claim 1, wherein the model is generated based on a machine learning algorithm. ### Claim 10 (depends on 1) The method of claim 1, wherein the model is generated based on supervised learning or unsupervised learning. ### Claim 11 (depends on 1) The method of claim 1, wherein the non-functioning IoT device (1) transmits data with a signal to noise ratio below a threshold, (2) transmits corrupted data, or (3) fails to transmit data within a period of time. ### Claim 12 (depends on 1) The method of claim 1, wherein removing includes removing a permission associated with the IoT device, the permission allowing the IoT device to access the cellular network. ### Claim 13 (depends on 1) The method of claim 1, wherein removing includes removing an identifier of the IoT device from a record in a database, the database identifying an IoT device is capable of accessing a cellular network. ### Claim 14 A system for removing non-functioning Internet-of-Things (IoT) devices from a cellular network, the system comprising: one or more processors; and a memory storing computer instructions that when executed cause the one or more processors to perform acts comprising: receiving, as training data, first data from first one or more IoT devices, the training data being collected over a time interval, including, after receiving a first portion of the training data: determining, during the time interval, characteristics of the first portion of the training data collected over the time interval; and determining, during the time interval, a length of the time interval over which the training data is collected based at least in part on the characteristics of the first portion of the training data, wherein at least a second portion of the training data is received during a remainder of the time interval after the determining of the length of the time interval based at least in part on the characteristics of the first portion of the training data; generating, based at least in part on the training data, a model representing an expected data profile to be transmitted to the cellular network by second one or more IoT devices; receiving, as received data, second data from an IoT device; comparing the received data from the IoT device to the model representing the expected data profile to be transmitted to identify the IoT device as a non-functioning IoT device; and removing, based at least in part on comparing the received data to the model representing the expected data profile to be transmitted, the non-functioning IoT device from the cellular network. ### Claim 15 (depends on 14) The system of claim 14, wherein the model represents the expected data profile of a class of IoT devices. ### Claim 16 (depends on 14) The system of claim 14, the acts further comprising: receiving the received data from a type of IoT devices; and processing the received data in a time domain or in a frequency domain, wherein processing the received data in a time domain includes processing one or more amplitudes of the received data and processing in the frequency domain includes processing a frequency content of the received data. ### Claim 17 (depends on 14) The system of claim 14, the acts further comprising: receiving additional training data from the first one or more IoT devices; and updating the model based on the additional training data. ### Claim 18 (depends on 14) The system of claim 14, wherein the model represents a distribution of data that is indicative of a Gaussian distribution, a Bessel distribution, a linear distribution, or a stochastic distribution. ### Claim 19 A non-transitory computer-readable media for removing non-functioning Internet-of-Things (loT) devices from a cellular network, the non-transitory computer-readable media including instructions that when executed by a processor, cause a node in the cellular network to perform acts comprising: receiving, as training data, first data from first one or more IoT devices, the training data being collected over a time interval, including, after receiving a first portion of the training data: determining, during the time interval, characteristics of the first portion of the training data collected over the time interval; and determining, during the time interval, a length of the time interval over which the training data is collected based at least in part on the characteristics of the first portion of the training data, wherein at least a second portion of the training data is received during a remainder of the time interval after the determining of the length of the time interval based at least in part on the characteristics of the first portion of the training data; generating, based at least in part on the training data, a model representing an expected data profile to be transmitted to the cellular network by second one or more IoT devices; receiving, as received data, second data from an IoT device; comparing the received data from the IoT device to the model representing the expected data profile to be transmitted to identify the IoT device as a non-functioning IoT device; and removing, based at least in part on comparing the received data to the model representing the expected data profile to be transmitted, the non-functioning IoT device from the cellular network. ### Claim 20 (depends on 19) The non-transitory computer-readable media of claim 19, wherein the model represents an (1) an indication of a statistical representation of data expected to be transmitted by the IoT device, (2) a Gaussian distribution, (3) a Bessel distribution, (4) a linear regression distribution, (5) a stochastic distribution, or (6) a machine learning algorithm.