# US 10,484,881 B2 · Optimization of cellular network architecture based on device type-specific traffic dynamics > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 10,484,881 B2 - Document ID: US10484881B2 - Status: Granted - Kind: B2 - Application number: 15/633,339 - Filed: 2017-06-26 - Published: 2017-10-12 - Granted: 2019-11-19 - Earliest claimed: 2010-12-15 - Assignees: AT&T Intellectual Property I, L.P.; Board of Trustees of Michigan State University - Applicants: AT&T Intellectual Property I, L.P.; Board of Trustees of Michigan State University - Inventors: Jia Wang; Lusheng Ji; Xiang-Yang Alexander Liu; Muhammad Zubair Shafiq - Web page: https://patentgrep.com/patent/US10484881B2 ## CPC - H04W 16/22: Traffic simulation tools or models - H04B 17/3913: {Predictive models, e.g. based on neural network models} - H04W 16/04: Traffic adaptive resource partitioning - H04W 24/08: Testing, {supervising or monitoring} using real traffic - H04W 72/0473: {the resource being transmission power} - H04W: WIRELESS COMMUNICATION NETWORKS (broadcast communication H04H; communication systems using wireless links for non-selective communication, e.g. wireless extensions H04M1/72) - H04B: TRANSMISSION ## Abstract A method, a computer readable medium and an apparatus for optimizing a cellular network architecture are disclosed. For example, the method obtains network traffic data for a plurality of different endpoint device types, wherein the network traffic data comprises network traffic data for each of the plurality of different endpoint device types, and predicts a future traffic pattern for one of the plurality of different endpoint device types based on the network traffic data. The method then adjusts a parameter of the cellular network architecture in response to the future traffic pattern predicted for the one of the plurality of different endpoint device types. ## Claims ### Claim 1 A method for adjusting a parameter of a cellular network, the method comprising: obtaining, by a processor in communication with a non-transitory computer-readable medium, network traffic data for a plurality of different endpoint device types, wherein the network traffic data is derived from voice calls of the cellular network, wherein the network traffic data comprises current network traffic data and a training data set comprising past network traffic data, wherein the network traffic data comprises network traffic data for each of the plurality of different endpoint device types, wherein each of the plurality of different endpoint device types comprises a different physical endpoint device type, wherein the network traffic data is stored in a plurality of records, wherein each record of the plurality of records contains a device type identifier that identifies an endpoint device type of the plurality of different endpoint device types; segregating, by the processor, at least one record in the plurality of records based on a corresponding device type identifier that identifies one of the plurality of different endpoint device types; predicting, by the processor, a future traffic pattern for the one of the plurality of different endpoint device types based on the at least one record in the plurality of records that contain the corresponding device type identifier that identifies the one of the plurality of different endpoint device types, wherein the predicting comprises: creating a model for the network traffic data that is specific to the one of the plurality of different endpoint device types, wherein the model indicates a degree of correlation between a present network traffic state represented by the current network traffic data and a number of past network traffic states represented by the training data set; and determining, using the model, that a geographic concentration of endpoint devices of the one of the plurality of different endpoint device types is expected at a first base station of the cellular network; and adjusting, by the processor, the parameter of the cellular network in response to the future traffic pattern predicted for the one of the plurality of different endpoint device types. ### Claim 2 (depends on 1) The method of claim 1, wherein the predicting further comprises: creating a Zipf-like distribution model for the network traffic data that is specific to the one of the plurality of different endpoint device types, wherein the Zipf-like distribution model segregates the network traffic data according to a number of application types. ### Claim 3 (depends on 1) The method of claim 1, further comprising: updating the training data set based on the current network traffic data. ### Claim 4 (depends on 1) The method of claim 1, wherein the adjusting the parameter comprises: adding a mobile base station to the cellular network proximate to the first base station. ### Claim 5 (depends on 1) The method of claim 1, wherein the adjusting the parameter comprises: decreasing a transmit power in a second base station proximate to the first base station. ### Claim 6 (depends on 1) The method of claim 1, wherein the adjusting the parameter comprises: increasing a transmit power of the first base station. ### Claim 7 (depends on 1) The method of claim 1, wherein the processor is a processor of an application server. ### Claim 8 (depends on 1) The method of claim 1, wherein the network traffic data comprises a type allocation code identifying a particular endpoint device type. ### Claim 9 A non-transitory computer readable medium storing instructions which, when executed by a processor, cause the processor to perform operations for adjusting a parameter of a cellular network, the operations comprising: obtaining network traffic data for a plurality of different endpoint device types, wherein the network traffic data is derived from voice calls of the cellular network, wherein the network traffic data comprises current network traffic data and a training data set comprising past network traffic data, wherein the network traffic data comprises network traffic data for each of the plurality of different endpoint device types, wherein each of the plurality of different endpoint device types comprises a different physical endpoint device type, wherein the network traffic data is stored in a plurality of records, wherein each record of the plurality of records contains a device type identifier that identifies an endpoint device type of the plurality of different endpoint device types; segregating, by the processor, at least one record in the plurality of records based on a corresponding device type identifier that identifies one of the plurality of different endpoint device types; predicting a future traffic pattern for the one of the plurality of different endpoint device types based on the at least one record in the plurality of records that contain the corresponding device type identifier that identifies the one of the plurality of different endpoint device types, wherein the predicting comprises: creating a model for the network traffic data that is specific to the one of the plurality of different endpoint device types, wherein the model indicates a degree of correlation between a present network traffic state represented by the current network traffic data and a number of past network traffic states represented by the training data set; and determining, using the model, that a geographic concentration of endpoint devices of the one of the plurality of different endpoint device types is expected at a first base station of the cellular network; and adjusting the parameter of the cellular network in response to the future traffic pattern predicted for the one of the plurality of different endpoint device types. ### Claim 10 (depends on 9) The non-transitory computer readable medium of claim 9, wherein the predicting further comprises: creating a Zipf-like distribution model for the network traffic data that is specific to the one of the plurality of different endpoint device types, wherein the Zipf-like distribution segregates the network traffic data according to a number of application types. ### Claim 11 (depends on 9) The non-transitory computer readable medium of claim 9, the operations further comprising, updating the training data set based on the current network traffic data. ### Claim 12 (depends on 9) The non-transitory computer readable medium of claim 9, wherein the adjusting the parameter comprises: adding a mobile base station to the cellular network proximate to the first base station. ### Claim 13 (depends on 9) The non-transitory computer readable medium of claim 9, wherein the adjusting the parameter comprises: decreasing a transmit power in a second base station proximate to the first base station. ### Claim 14 (depends on 9) The non-transitory computer readable medium of claim 9, wherein the adjusting the parameter comprises: increasing a transmit power of the first base station. ### Claim 15 (depends on 9) The non-transitory computer readable medium of claim 9, wherein the processor is a processor of an application server. ### Claim 16 (depends on 9) The non-transitory computer readable medium of claim 9, wherein the network traffic data comprises a type allocation code identifying a particular endpoint device type. ### Claim 17 An apparatus for adjusting a parameter of a cellular network, the apparatus comprising: a processor; and a non-transitory computer-readable medium storing instructions which, when executed by the processor, cause the processor to perform operations, the operations comprising: obtaining network traffic data for a plurality of different endpoint device types, wherein the network traffic data is derived from voice calls of the cellular network, wherein the network traffic data comprises current network traffic data and a training data set comprising past network traffic data, wherein the network traffic data comprises network traffic data for each of the plurality of different endpoint device types, wherein each of the plurality of different endpoint device types comprises a different physical endpoint device type, wherein the network traffic data is stored in a plurality of records, wherein each record of the plurality of records contains a device type identifier that identifies an endpoint device type of the plurality of different endpoint device types; segregating at least one record in the plurality of records based on a corresponding device type identifier that identifies one of the plurality of different endpoint device types; predicting a future traffic pattern for the one of the plurality of different endpoint device types based on the at least one record in the plurality of records that contain the corresponding device type identifier that identifies the one of the plurality of different endpoint device types, wherein the predicting comprises: creating a model for the network traffic data that is specific to the one of the plurality of different endpoint device types, wherein the model indicates a degree of correlation between a present network traffic state represented by the current network traffic data and a number of past network traffic states represented by the training data set; and determining, using the model, that a geographic concentration of endpoint devices of the one of the plurality of different endpoint device types is expected at a first base station of the cellular network; and adjusting the parameter of the cellular network in response to the future traffic pattern predicted for the one of the plurality of different endpoint device types. ### Claim 18 (depends on 17) The apparatus of claim 17, wherein the predicting further comprises: creating a Zipf-like distribution model for the network traffic data that is specific to the one of the plurality of different endpoint device types, wherein the Zipf-like distribution model segregates the network traffic data according to a number of application types. ### Claim 19 (depends on 17) The apparatus of claim 17, the operations further comprising, updating the training data set based on the current network traffic data. ### Claim 20 (depends on 17) The apparatus of claim 17, wherein the adjusting the parameter comprises: adding a mobile base station to the cellular network proximate to the first base station.