# US 12,604,208 B2 · Learning method, wireless quality estimation method, learning device, wireless quality estimation device, and program > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,604,208 B2 - Document ID: US12604208B2 - Status: Granted - Kind: B2 - Application number: 18/245,962 - Filed: 2020-11-06 - Published: 2023-11-30 - Granted: 2026-04-14 - Earliest claimed: 2020-11-06 - Assignees: NTT, Inc. - Applicants: NTT, Inc. - Inventors: Keisuke Wakao; Kenichi Kawamura; Takatsune Moriyama - Web page: https://patentgrep.com/patent/US12604208B2 ## CPC - H04W 16/18: Network planning tools - G01R 29/10: Radiation diagrams of antennas - G06N 3/0499: Feedforward networks - G06N 3/08: Learning methods - G06N 3/09: Supervised learning - G06N 3/096: Transfer learning - H04W 24/08: Testing, {supervising or monitoring} using real traffic - H04W: WIRELESS COMMUNICATION NETWORKS (broadcast communication H04H; communication systems using wireless links for non-selective communication, e.g. wireless extensions H04M1/72) - G01R: MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES (indicating correct tuning of resonant circuits H03J3/12) - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS ## Abstract A learning method in which a learning device performs learning on a model for estimating wireless quality by machine learning, and the learning method includes a learning step of generating a pre-trained model by performing the learning on the model by using measurement data of wireless quality obtained by a plurality of terminals, and a retraining step of retraining the pre-trained model by using measurement data satisfying a predetermined condition. ## Claims ### Claim 1 A learning method for a learning device that performs learning on a model for estimating wireless quality by machine learning, the learning method comprising: generating a pre-trained model by performing the learning on the model by using measurement data of wireless quality obtained by a plurality of terminals; determining whether measurement data satisfy a predetermined condition, said condition is satisfied if the measurement data is acquired from a terminal including a wireless communication chip having a same model number as a model number of a wireless communication chip that is included in a terminal that is a target of wireless quality estimation, and retraining the pre-trained model by using the measurement data satisfying the predetermined condition, wherein the retraining includes extracting measurement data of a specific terminal that is a target of the wireless quality estimation or measurement data of a terminal close to the specific terminal from the measurement data of the plurality of terminals, and performing the retraining by using the extracted measurement data, wherein the measurement data used for retraining is extracted based on terminal information and position information, and wherein the terminal information includes at least one of a device model, a wireless communication standard, antenna performance, or a model number of the wireless communication chip such that a terminal from which the measurement data is acquired is close to the terminal that is the target of wireless quality estimation. ### Claim 2 (depends on 1) The learning method according to claim 1, wherein the model is for a neural network, and the retraining includes designating a range of neurons to be retrained in the pre-trained model. ### Claim 3 (depends on 1) A wireless quality estimation method executed by a wireless quality estimation device, comprising: estimating a location of a terminal that is a target of wireless quality estimation; and estimating wireless quality at an estimated position of the terminal that is the target of the wireless quality estimation by using the model retrained by the learning method according to claim 1. ### Claim 4 A learning device that performs learning on a model for estimating wireless quality by machine learning, the learning device comprising: a processor; and a memory storing program instructions that cause the processor to: generate a pre-trained model by performing the learning on the model by using measurement data of wireless quality obtained by a plurality of terminals; determine whether the measurement data satisfy a predetermined condition, said condition is satisfied if the measurement data is acquired from a terminal including a wireless communication chip having a same model number as a model number of a wireless communication chip that is included in a terminal that is a target of wireless quality estimation, and retrain the pre-trained model by using measurement data satisfying the predetermined condition, wherein the processor is further caused to extract measurement data of a specific terminal that is a target of the wireless quality estimation or measurement data of a terminal close to the specific terminal from the measurement data of the plurality of terminals, and perform the retraining by using the extracted measurement data, wherein the measurement data used for retraining is extracted based on terminal information and position information, and wherein the terminal information includes at least one of a device model, a wireless communication standard, antenna performance, or a model number of the wireless communication chip such that a terminal from which the measurement data is acquired is close to the terminal that is the target of wireless quality estimation. ### Claim 5 A wireless quality estimation device, comprising: a processor; and a memory storing program instructions that cause the processor to: estimate a position of a terminal that is a target of wireless quality estimation; and estimate wireless quality at an estimated position of the terminal that is the target of the wireless quality estimation by using a model generated by determining whether measurement data satisfy a predetermined condition, said condition is satisfied if the measurement data is acquired from a terminal including a wireless communication chip having a same model number as a model number of a wireless communication chip that is included in a terminal that is a target of wireless quality estimation and retraining, by using the measurement data satisfying the predetermined condition, a pre-trained model obtained by performing machine learning by using measurement data of wireless quality obtained by a plurality of terminals, wherein the retraining includes extracting measurement data of a specific terminal that is a target of the wireless quality estimation or measurement data of a terminal close to the specific terminal from the measurement data of the plurality of terminals, and performing the retraining by using the extracted measurement data, wherein the measurement data used for retraining is extracted based on terminal information and position information, and wherein the terminal information includes at least one of a device model, a wireless communication standard, antenna performance, or a model number of the wireless communication chip such that a terminal from which the measurement data is acquired is close to the terminal that is the target of wireless quality estimation. ### Claim 6 (depends on 1) A non-transitory computer-readable recording medium having stored therein a program for causing a computer to perform the learning method according to claim 1. ### Claim 7 (depends on 3) A non-transitory computer-readable recording medium having stored therein a program for causing a computer to perform the wireless quality estimation method according to claim 3. ### Claim 8 (depends on 2) The learning method according to claim 2, wherein the retraining includes updating only three intermediate layers located immediately before an output layer of the neural network included in the pre-trained model.