DETAILED ACTION
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
The amendment filed on 12/17/2025 has been received and considered by the examiner. Claims 1, 7, 17, and 20-21 were amended, and all uncancelled claims remain pending.
The 112(b) rejection for claim 7 is withdrawn because the amendment corrected it to depend on claim 1.
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Response to Arguments
Applicant’s arguments with respect to claim(s) 1, 3-5, 7-11, 13-14, 16-17, and 19-21 have been considered but are moot because the new ground of rejection does not rely on the combination of references applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
Claim Rejections - 35 USC § 103
The text of those sections of Title 35, U.S. Code not included in this action can be found in a prior Office action.
Claim(s) 1, 3, 5, 8, 13, 16-17, and 19-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Wang et al. (US 2020/0304381 A1, hereinafter “Wang”) in view of Ottersten et al. (US 2021/0345134 A1, hereinafter “Ottersten”) and further in view of Fu et al. (US 2018/0310203 A1, hereinafter “Fu”).
As to Claim 1:
Wang describes a method to predict device performance based on a machine learning analysis of network conditions.
Specifically, Wang teaches:
Predicting one or more cellular network performance parameters associated with user equipment (UE) within a three-dimensional (3D) space having one or more cellular nodes, including a 5G cellular node
(“Implementations of new broadband cellular networks (e.g., 3GPP 5G networks) and other wireless networks (e.g., IEEE 802.11ax) are raising user expectations for increased network speeds.... The network evaluation service described herein may provide a solution to ‘learn’ about parameters of essentially every packet size of each application on each device in the network.... In one instance, the learning result can be used to predict the application performance on each device” (Wang, 0001, 0014).
Here, “predict the application performance” maps to “predicting one or more cellular network performance parameters”,
“on each device” maps to “associated with user equipment (UE)”, and
“5G networks” maps to “a three-dimensional (3D) space having one or more cellular nodes, including a 5G cellular node”).
Determining, for each of a plurality of UEs within the 3D space, values associated with one or more UE-side features that are specific to each UE
(“[T]raffic evaluation function 322 [in Fig. 3-4C] may provide output 450 including a current condition of the network 502. The current condition of the network 502 may include, for each application in each client device 150, whether there is a degrading (e.g., data rate, latency, jitter, etc.)” (Wang, 0053).
Here, “provide” maps to “determining”,
“in each client device 150” maps to “for each of a plurality of UEs within the 3D space”,
“current conditions of the network” for “each client device” maps to “values associated with one or more UE-side features that are specific to each UE”).
Predicting UE-specific values of the one or more cellular network performance parameters for each UE as a function of the values associated with the one or more UE-side features of each respective UE to generate predicted UE-specific values
(“The network evaluation service described herein may provide a solution to ‘learn’ about parameters of essentially every packet size of each application on each device in the network.... In one instance, the learning result can be used to predict the application performance on each device” (Wang, 0014). Also, Fig. 4B in Wang shows real-time prediction of UE packet parameters.
Here, “predict the application performance on each device” maps to “predicting UE-specific values of the one or more cellular network performance parameters for each UE ... to generate predicted UE-specific values”,
“can be used” maps to “as a function of”, and
element 412 in Fig. 4B, “live traffic”, of “each device” maps to “values associated with the one or more UE-side features of each respective UE”).
Predicting UE-specific values of the one or more cellular network performance parameters includes predicting first UE-specific values of the one or more cellular performance parameters that are specific for the first UE ... to generate predicted UE-specific values for the first UE and predicting second UE-specific values of the one or more cellular network performance parameters that are specific for the second UE ... to generate predicted UE-specific values for the second UE
(“The network evaluation service described herein may provide a solution to ‘learn’ about parameters of essentially every packet size of each application on each device in the network.... In one instance, the learning result can be used to predict the application performance on each device in different network conditions” (Wang, 0014).
Here, “predict” maps to “predicting”, and
“application performance on each device” maps to “UE-specific values of the one or more cellular network performance parameters includes predicting first UE-specific values of the one or more cellular performance parameters that are specific for the first UE ... to generate predicted UE-specific values for the first UE and predicting second UE-specific values of the one or more cellular network performance parameters that are specific for a second UE ... to generate predicted UE-specific values for the second UE”).
Predicting cellular data throughput and one or more of signal strength and level of carrier aggregation based on inputting the values associated with the one or more UE-side features that are specific to each UE to a machine learning model
(“The current condition of the network 502 may include, for each application in each client device ... whether there is a degrading (e.g., data rate, latency, jitter, etc.) ... Network prediction module 510 may learn from the current conditions 502 ... Fig. 6 is an exemplary table which may illustrate prediction output 506.... Entries 630 for field 620 may include one or more predicted performance parameters, such as data speed, latency, jitter, etc” (Wang, 0053-0055).
Here, “predicted performance” which includes “data speed” maps to “predicting cellular data throughput”,
“includes” maps to “include”, and
“throughput” maps to “cellular data throughput” from the list of “one or more of cellular data throughput, signal strength and level of carrier aggregation”,
“jitter” maps to “signal strength” from the list of “one or more of signal strength and level of carrier aggregation” because jitter is a measure of noise which is inversely proportional to relative signal strength, so measuring jitter is measuring the inverse of signal strength,
“learn from” maps to “based on inputting ... to a machine learning model”, and
“current condition of the network 502” for “each client device” maps to “the values associated with the one or more UE-side features that are specific to each UE”).
The output of the machine learning module are the predicted UE-specific values including the predicted UE-specific values for the first UE and the predicted UE-specific values for the second UE
(“The current condition of the network 502 may include, for each application in each client device ... whether there is a degrading (e.g., data rate, latency, jitter, etc.) ... Network prediction module 510 may learn from the current conditions 502 ... Fig. 6 is an exemplary table which may illustrate prediction output 506.... Entries 630 for field 620 may include one or more predicted performance parameters, such as data speed, latency, jitter, etc” (Wang, 0053-0055).
Here, “prediction output” maps to “the output of the machine learning module”,
element 620 in Fig. 6, “predicted performance of each client device”, maps to “the predicted UE-specific values including the predicted UE-specific values for the first UE and the predicted UE-specific values for the second UE”).
The predicted UE-specific values for the first UE include predicted cellular data throughput and one or more of signal strength and level or carrier aggregation specific for the first UE
(“The current condition of the network 502 may include, for each application in each client device ... whether there is a degrading (e.g., data rate, latency, jitter, etc.) ... Network prediction module 510 may learn from the current conditions 502 ... Fig. 6 is an exemplary table which may illustrate prediction output 506.... Entries 630 for field 620 may include one or more predicted performance parameters, such as data speed, latency, jitter, etc” (Wang, 0053-0055).
Here, “predicted performance of each client device” in Fig. 6 maps to “the predicted UE-specific values for the first UE”,
“such as” maps to “include”,
“data speed” maps to “predicted cellular data strength”,
“jitter” maps to “signal strength and level” from the list of “signal strength and level or carrier aggregation”, and
“of each client device” maps to “specific for the first UE”).
The predicted UE-specific values for the second UE include predicted cellular data throughput and one or more of signal strength and level or carrier aggregation specific for the second UE
(“The current condition of the network 502 may include, for each application in each client device ... whether there is a degrading (e.g., data rate, latency, jitter, etc.) ... Network prediction module 510 may learn from the current conditions 502 ... Fig. 6 is an exemplary table which may illustrate prediction output 506.... Entries 630 for field 620 may include one or more predicted performance parameters, such as data speed, latency, jitter, etc” (Wang, 0053-0055).
Here, “predicted performance of each client device” in Fig. 6 maps to “the predicted UE-specific values for the second UE”,
“such as” maps to “include”,
“data speed” maps to “predicted cellular data strength”,
“jitter” maps to “signal strength and level” from the list of “signal strength and level or carrier aggregation”, and
“of each client device” maps to “specific for the second UE”).
The machine learning module has been trained using truth data associated with the one or more UE-side features
(“As shown in Fig. 4A, training module 335 may be a machine-learning-based function located in the cloud. In one implementation, training module 335 may receive historical training data 402 from benchmark performance database 330 or another data source” (Wang, 0049).
Here, “machine-learning-based function” maps to “the machine learning module”,
“receive historical training data” maps to “has been trained using truth data associated with the one or more UE-side features”).
Wang does not explicitly disclose:
The UE-side features include mobility features, the mobility features including direction and speed of movement relative to the 3D space that are specific for each UE
The UE-side features include first mobility features including direction and speed of movement relative to the 3D space that are specific to a first UE and second mobility features including direction and speed of movement relative to the 3D space that are specific to a second UE
Predicting first UE-specific values ... based on the first mobility features to generate predicted UE-specific values for the first UE
Predicting second UE-specific values ... based on the second mobility features to generate predicted UE-specific values for the second UE
The predicted UE-specific values ... based on the first mobility features
The predicted UE-specific values for the second UE ... based on the second mobility features
However, Ottersten does describe a machine learning model that can predict the optimal configuration of a network node based on the features of user devices it serves.
Specifically, Ottersten teaches:
The UE-side features include mobility features, the mobility features including direction and speed of movement relative to the 3D space that are specific for each UE
(“The input data may comprise one or more input parameters such as received signal strength, angle of arrival, measured or estimated UE speed, target block or bit error rates, just to give some examples” (Ottersten, 0103).
Here, “input data” maps to “the UE-side features”,
“angle of arrival” and “UE speed” map to “mobility features ... that are specific for each UE”,
“angle of arrival” maps to “the mobility features including direction”, and
“UE speed” maps to “speed of movement relative to the 3D space”).
The UE-side features include first mobility features including direction and speed of movement relative to the 3D space that are specific to a first UE and second mobility features including direction and speed of movement relative to the 3D space that are specific to a second UE
(“The input data may comprise one or more input parameters such as received signal strength, angle of arrival, measured or estimated UE speed, target block or bit error rates, just to give some examples” (Ottersten, 0103).
Here, “[t]he input data may include” maps to “the UE-side features include”,
“UE speed” maps to “first mobility features including direction and speed of movement relative to the 3D space that are specific to a first UE”, and
“UE speed” maps to “second mobility features including direction and speed of movement relative to the 3D space that are specific to a second UE” because MPEP 2144 VI B lists “duplication of parts” as an example of an obvious modification).
Predicting first UE-specific values ... based on the first mobility features to generate predicted UE-specific values for the first UE
(“The input data may comprise one or more input parameters such as received signal strength, angle of arrival, measured or estimated UE speed, target block or bit error rates, just to give some examples” (Ottersten, 0103).
Here, “input data may comprise ... measured or estimated UE speed” maps to “predicting first UE-specific values ... based on the first mobility features to generate predicted UE-specific values for the first UE”).
Predicting second UE-specific values ... based on the second mobility features to generate predicted UE-specific values for the second UE
(“The input data may comprise one or more input parameters such as received signal strength, angle of arrival, measured or estimated UE speed, target block or bit error rates, just to give some examples” (Ottersten, 0103).
Here, “input data may comprise ... measured or estimated UE speed” maps to “predicting second UE-specific values ... based on the first mobility features to generate predicted UE-specific values for the second UE”).
The predicted UE-specific values for the first UE ... based on the first mobility features
(“The input data may comprise one or more input parameters such as received signal strength, angle of arrival, measured or estimated UE speed, target block or bit error rates, just to give some examples” (Ottersten, 0103).
Here, “measured or estimated UE speed” maps to “the predicted UE-specific values for the first UE”, and
“the input data may comprise ... measured or estimated UE speed” maps to “based on the first mobility features”).
The predicted UE-specific values for the second UE ... based on the second mobility features
(“The input data may comprise one or more input parameters such as received signal strength, angle of arrival, measured or estimated UE speed, target block or bit error rates, just to give some examples” (Ottersten, 0103).
Here, “measured or estimated UE speed” maps to “the predicted UE-specific values for the second UE”, and
“the input data may comprise ... measured or estimated UE speed” maps to “based on the first mobility features”).
Thus, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to use the mobility features described in Ottersten as the inputs for the machine learning model in Wang. UE mobility features can help to predict future throughput and noise, just as past throughput and noise measurements can.
The combination of Wang and Ottersten does not explicitly disclose:
Predicting UE-specific values for the first UE ... independent of the second mobility features
Predicting UE-specific values for the second UE ... independent of the first mobility features
However, Fu does describe a method to predict traffic throughput for individual wireless devices.
Specifically, Fu teaches:
Predicting UE-specific values for the first UE ... independent of the second mobility features
(“[T]he traffic throughput prediction module 200b [in Fig. 6] receives ... previous traffic throughput for the wireless device as inputs ... [M]essages from the traffic throughput prediction module 200b can include predicted traffic throughput for the wireless device 140” (Fu, 0078, 0080).
Here, “predicted traffic throughput for the wireless device” in Fig. 6 maps to “predicting UE-specific values for the first UE”, and
“previous traffic throughput for the wireless device as inputs” maps to “independent of the second mobility features”).
Predicting UE-specific values for the second UE ... independent of the first mobility features
(“[T]he traffic throughput prediction module 200b [in Fig. 6] receives ... previous traffic throughput for the wireless device as inputs ... [M]essages from the traffic throughput prediction module 200b can include predicted traffic throughput for the wireless device 140” (Fu, 0078, 0080).
Here, “predicted traffic throughput for the wireless device” in Fig. 6 maps to “predicting UE-specific values for the second UE”, and
“previous traffic throughput for the wireless device as inputs” maps to “independent of the first mobility features”).
Thus, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate Fu’s practice of using device-specific input data into Wang’s method for predicting wireless device throughput. The past throughput values for a device would be an obvious candidate input to predict that device’s throughput in the future.
As to Claim 3:
Wang teaches:
Cellular data throughput is one or more of downlink throughput and uplink throughput
(“Entries 630 for field 620 [in Fig. 6] may include ... data speed” (Wang).
Here, “data speed” maps to “cellular data throughput is ... downlink throughput and uplink throughput” from the list of “cellular data throughput is one or more of downlink throughput and uplink throughput”).
As to Claim 5:
Wang teaches:
Predicting cellular data performance across the UEs based on the predicted values of the one or more cellular network performance parameters
(“The network evaluation service described herein may provide a solution to ‘learn’ about parameters of essentially every packet size of each application on each device in the network.... In one instance, the learning result can be used to predict the application performance on each device in different network conditions” (Wang, 0014).
Here, “predict the application performance on each device” maps to “predicting cellular data performance across the UEs”,
“can be used” maps to “based on”, and
“the learning result” maps to “the one or more cellular network performance parameters”).
As to Claim 8:
Wang teaches:
The UE-side features further include connection-based features
(“[D]ata speed, jitter, and latency can be considered three basic features of each packet size traversing a network” (Wang, 0013).
Here, “features of each packet size traversing a network” maps to “the UE-side features”, and
“data speed, jitter and latency” map to “connection-based features”).
As to Claim 13:
Wang teaches:
The UE-side features further include connection-based features, the connection-based features including past values of cellular data throughput associated with the UE
(“[A]pplications running on client device 150 may have traffic characteristics, such as data speed, latency, and jitter for each packet size and direction, that can be identified and characterized by the network evaluation service” (Wang, 0023).
Here, “traffic characteristics” maps to “the UE-side features”,
“such as” maps to “include”, and
“data speed” maps to “connection-based features, the connection-based features including past values of cellular data throughput associated with the UE”).
As to Claim 16:
Wang teaches:
The UEs include one or more of cellular telephones, computer tablets, and computers
(“Client device 150 may be a mobile device. For example, a client device 150 may be implemented as a smartphone, a tablet device, a netbook, a computer (e.g., a laptop, a palmtop, etc.), or another type of mobile device” (Wang, 0023).
Here, “Client device 150” maps to “the UEs”,
“a smartphone” maps to “cellular telephones”,
“a tablet device” maps to “computer tablets”, and
“a computer” maps to “computers”).
As to Claim 17:
Wang teaches:
One or more cellular nodes, including one or more 5G panels
(“Implementations of new broadband cellular networks (e.g., 3GPP 5G networks) and other wireless networks (e.g., IEEE 802.11ax) are raising user expectations for increased network speeds” (Wang, 0001).
Here, “5G networks” map to “one or more cellular nodes, including one or more 5G panels”).
A computing system connected to the cellular nodes, the computing system including a machine learning module
(Fig. 4B in Wang shows a machine learning module predicting network performance.
Here, element 400 in Fig. 4B maps to “a computing system connected to the cellular nodes”, and
“Trained Model 410” maps to “the computing system including a machine learning module”).
Determine, for each of a plurality of UEs within the 3D space, values associated with one or more UE-side features that are specific to each UE
(“[T]raffic evaluation function 322 [in Fig. 3-4C] may provide output 450 including a current condition of the network 502. The current condition of the network 502 may include, for each application in each client device 150, whether there is a degrading (e.g., data rate, latency, jitter, etc.)” (Wang, 0053).
Here, “provide” maps to “determine”,
“in each client device 150” maps to “for each of a plurality of UEs within the 3D space”,
“current conditions of the network” for “each client device” maps to “values associated with one or more UE-side features that are specific to each UE”).
Predict UE-specific values of the one or more cellular network performance parameters for each UE as a function of the values associated with the one or more UE-side features of each respective UE to generate predicted UE-specific values
(“The network evaluation service described herein may provide a solution to ‘learn’ about parameters of essentially every packet size of each application on each device in the network.... In one instance, the learning result can be used to predict the application performance on each device” (Wang, 0014). Also, Fig. 4B in Wang shows real-time prediction of UE packet parameters.
Here, “predict the application performance on each device” maps to “predict UE-specific values of the one or more cellular network performance parameters for each UE ... to generate predicted UE-specific values”,
“can be used” maps to “as a function of”, and
element 412 in Fig. 4B, “live traffic”, of “each device” maps to “values associated with the one or more UE-side features of each respective UE”).
To predict UE-specific values of the one or more cellular network performance parameters, the computing system is configured to predict first UE-specific values of the one or more cellular performance parameters that are specific for the first UE ... to generate predicted UE-specific values for the first UE and predicting second UE-specific values of the one or more cellular network performance parameters that are specific for the second UE ... to generate predicted UE-specific values for the second UE
(“The network evaluation service described herein may provide a solution to ‘learn’ about parameters of essentially every packet size of each application on each device in the network.... In one instance, the learning result can be used to predict the application performance on each device in different network conditions” (Wang, 0014).
Here, “predict” maps to “to predict”, and
“application performance on each device” maps to “UE-specific values of the one or more cellular network performance parameters”,
“predict” maps to “the computing system is configured to predict”,
“application performance on each device” maps to “first UE-specific values of the one or more cellular performance parameters that are specific for the first UE ... to generate predicted UE-specific values for the first UE and predicting second UE-specific values of the one or more cellular network performance parameters that are specific for a second UE ... to generate predicted UE-specific values for the second UE”).
Predict cellular data throughput and one or more of signal strength and level of carrier aggregation based on inputting the values associated with the one or more UE-side features that are specific to each UE to a machine learning model
(“The current condition of the network 502 may include, for each application in each client device ... whether there is a degrading (e.g., data rate, latency, jitter, etc.) ... Network prediction module 510 may learn from the current conditions 502 ... Fig. 6 is an exemplary table which may illustrate prediction output 506.... Entries 630 for field 620 may include one or more predicted performance parameters, such as data speed, latency, jitter, etc” (Wang, 0053-0055).
Here, “predicted performance” which includes “data speed” maps to “predict cellular data throughput”,
“includes” maps to “include”, and
“throughput” maps to “cellular data throughput” from the list of “one or more of cellular data throughput, signal strength and level of carrier aggregation”,
“jitter” maps to “signal strength” from the list of “one or more of signal strength and level of carrier aggregation” because jitter is a measure of noise which is inversely proportional to relative signal strength, so measuring jitter is measuring the inverse of signal strength,
“learn from” maps to “based on inputting ... to a machine learning model”, and
“current condition of the network 502” for “each client device” maps to “the values associated with the one or more UE-side features that are specific to each UE”).
The output of the machine learning module are the predicted UE-specific values including the predicted UE-specific values for the first UE and the predicted UE-specific values for the second UE
(“The current condition of the network 502 may include, for each application in each client device ... whether there is a degrading (e.g., data rate, latency, jitter, etc.) ... Network prediction module 510 may learn from the current conditions 502 ... Fig. 6 is an exemplary table which may illustrate prediction output 506.... Entries 630 for field 620 may include one or more predicted performance parameters, such as data speed, latency, jitter, etc” (Wang, 0053-0055).
Here, “prediction output” maps to “the output of the machine learning module”,
element 620 in Fig. 6, “predicted performance of each client device”, maps to “the predicted UE-specific values including the predicted UE-specific values for the first UE and the predicted UE-specific values for the second UE”).
The predicted UE-specific values for the first UE include predicted cellular data throughput and one or more of signal strength and level or carrier aggregation specific for the first UE
(“The current condition of the network 502 may include, for each application in each client device ... whether there is a degrading (e.g., data rate, latency, jitter, etc.) ... Network prediction module 510 may learn from the current conditions 502 ... Fig. 6 is an exemplary table which may illustrate prediction output 506.... Entries 630 for field 620 may include one or more predicted performance parameters, such as data speed, latency, jitter, etc” (Wang, 0053-0055).
Here, “predicted performance of each client device” in Fig. 6 maps to “the predicted UE-specific values for the first UE”,
“such as” maps to “include”,
“data speed” maps to “predicted cellular data strength”,
“jitter” maps to “signal strength and level” from the list of “signal strength and level or carrier aggregation”, and
“of each client device” maps to “specific for the first UE”).
The predicted UE-specific values for the second UE include predicted cellular data throughput and one or more of signal strength and level or carrier aggregation specific for the second UE
(“The current condition of the network 502 may include, for each application in each client device ... whether there is a degrading (e.g., data rate, latency, jitter, etc.) ... Network prediction module 510 may learn from the current conditions 502 ... Fig. 6 is an exemplary table which may illustrate prediction output 506.... Entries 630 for field 620 may include one or more predicted performance parameters, such as data speed, latency, jitter, etc” (Wang, 0053-0055).
Here, “predicted performance of each client device” in Fig. 6 maps to “the predicted UE-specific values for the second UE”,
“such as” maps to “include”,
“data speed” maps to “predicted cellular data strength”,
“jitter” maps to “signal strength and level” from the list of “signal strength and level or carrier aggregation”, and
“of each client device” maps to “specific for the second UE”).
The machine learning module has been trained using truth data associated with the one or more UE-side features
(“As shown in Fig. 4A, training module 335 may be a machine-learning-based function located in the cloud. In one implementation, training module 335 may receive historical training data 402 from benchmark performance database 330 or another data source” (Wang, 0049).
Here, “machine-learning-based function” maps to “the machine learning module”,
“receive historical training data” maps to “has been trained using truth data associated with the one or more UE-side features”).
Wang does not explicitly disclose:
The UE-side features include mobility features, the mobility features including direction and speed of movement relative to the 3D space that are specific for each UE
The UE-side features include first mobility features including direction and speed of movement relative to the 3D space that are specific to a first UE and second mobility features including direction and speed of movement relative to the 3D space that are specific to a second UE
Predict first UE-specific values ... based on the first mobility features to generate predicted UE-specific values for the first UE
Predict second UE-specific values ... based on the first mobility features to generate predicted UE-specific values for the second UE
The predicted UE-specific values ... based on the first mobility features
The predicted UE-specific values for the second UE ... based on the second mobility features
However, Ottersten does teach:
The UE-side features include mobility features, the mobility features including direction and speed of movement relative to the 3D space that are specific for each UE
(“The input data may comprise one or more input parameters such as received signal strength, angle of arrival, measured or estimated UE speed, target block or bit error rates, just to give some examples” (Ottersten, 0103).
Here, “input data” maps to “the UE-side features”,
“angle of arrival” and “UE speed” map to “mobility features ... that are specific for each UE”,
“angle of arrival” maps to “the mobility features including direction”, and
“UE speed” maps to “speed of movement relative to the 3D space”).
The UE-side features include first mobility features including direction and speed of movement relative to the 3D space that are specific to a first UE and second mobility features including direction and speed of movement relative to the 3D space that are specific to a second UE
(“The input data may comprise one or more input parameters such as received signal strength, angle of arrival, measured or estimated UE speed, target block or bit error rates, just to give some examples” (Ottersten, 0103).
Here, “[t]he input data may include” maps to “the UE-side features include”,
“UE speed” maps to “first mobility features including direction and speed of movement relative to the 3D space that are specific to a first UE”, and
“UE speed” maps to “second mobility features including direction and speed of movement relative to the 3D space that are specific to a second UE” because MPEP 2144 VI B lists “duplication of parts” as an example of an obvious modification).
Predict first UE-specific values ... based on the first mobility features to generate predicted UE-specific values for the first UE
(“The input data may comprise one or more input parameters such as received signal strength, angle of arrival, measured or estimated UE speed, target block or bit error rates, just to give some examples” (Ottersten, 0103).
Here, “input data may comprise ... measured or estimated UE speed” maps to “predict first UE-specific values ... based on the first mobility features to generate predicted UE-specific values for the first UE”).
Predict second UE-specific values ... based on the second mobility features to generate predicted UE-specific values for the second UE
(“The input data may comprise one or more input parameters such as received signal strength, angle of arrival, measured or estimated UE speed, target block or bit error rates, just to give some examples” (Ottersten, 0103).
Here, “input data may comprise ... measured or estimated UE speed” maps to “predict second UE-specific values ... based on the first mobility features to generate predicted UE-specific values for the second UE”).
The predicted UE-specific values for the first UE ... based on the first mobility features
(“The input data may comprise one or more input parameters such as received signal strength, angle of arrival, measured or estimated UE speed, target block or bit error rates, just to give some examples” (Ottersten, 0103).
Here, “measured or estimated UE speed” maps to “the predicted UE-specific values for the first UE”, and
“the input data may comprise ... measured or estimated UE speed” maps to “based on the first mobility features”).
The predicted UE-specific values for the second UE ... based on the second mobility features
(“The input data may comprise one or more input parameters such as received signal strength, angle of arrival, measured or estimated UE speed, target block or bit error rates, just to give some examples” (Ottersten, 0103).
Here, “measured or estimated UE speed” maps to “the predicted UE-specific values for the second UE”, and
“the input data may comprise ... measured or estimated UE speed” maps to “based on the first mobility features”).
Thus, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to use the mobility features described in Ottersten as the inputs for the machine learning model in Wang. UE mobility features can help to predict future throughput and noise, just as past throughput and noise measurements can.
The combination of Wang and Ottersten does not explicitly disclose:
Predict UE-specific values for the first UE ... independent of the second mobility features
Predict UE-specific values for the second UE ... independent of the first mobility features
However, Fu does teach:
Predict UE-specific values for the first UE ... independent of the second mobility features
(“[T]he traffic throughput prediction module 200b [in Fig. 6] receives ... previous traffic throughput for the wireless device as inputs ... [M]essages from the traffic throughput prediction module 200b can include predicted traffic throughput for the wireless device 140” (Fu, 0078, 0080).
Here, “predicted traffic throughput for the wireless device” in Fig. 6 maps to “predict UE-specific values for the first UE”, and
“previous traffic throughput for the wireless device as inputs” maps to “independent of the second mobility features”).
Predict UE-specific values for the second UE ... independent of the first mobility features
(“[T]he traffic throughput prediction module 200b [in Fig. 6] receives ... previous traffic throughput for the wireless device as inputs ... [M]essages from the traffic throughput prediction module 200b can include predicted traffic throughput for the wireless device 140” (Fu, 0078, 0080).
Here, “predicted traffic throughput for the wireless device” in Fig. 6 maps to “predict UE-specific values for the second UE”, and
“previous traffic throughput for the wireless device as inputs” maps to “independent of the first mobility features”).
Thus, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate Fu’s practice of using device-specific input data into Wang’s method for predicting wireless device throughput. The past throughput values for a device would be an obvious candidate input to predict that device’s throughput in the future.
As to Claim 19:
Wang teaches:
The computing system comprises a laptop, a server, or a cloud computing platform
(“[T]he UE 110 is implemented as a smartphone but may be implemented as any suitable computing or electronic device, such as a ... laptop computer” (Wang, 0040).
Here, any suitable computing or electronic device, such as a ... laptop computer” maps to “a laptop, a server, or a cloud computing platform”).
As to Claim 20:
Wang teaches:
A non-transitory, computer-readable medium comprising executable instructions, which when executed by processing circuitry, cause a computing device
(“Some operations of the example methods may be described in the general context of executable instructions stored on computer-readable storage memory that is local and/or remote to a computer processing system, and implementations can include software applications, programs, functions, and the like” (Wang, 0140).
Here, “computer-readable storage memory” maps to “a non-transitory, computer-readable medium”,
“stored on” maps to “comprising”,
“executable instructions” map to “executable instructions, which when executed ... cause a computing device”, and
“a computer processing system” maps to “processing circuitry”).
Determine, for each of a plurality of UEs within the 3D space, values associated with one or more UE-side features that are specific to each UE
(“[T]raffic evaluation function 322 [in Fig. 3-4C] may provide output 450 including a current condition of the network 502. The current condition of the network 502 may include, for each application in each client device 150, whether there is a degrading (e.g., data rate, latency, jitter, etc.)” (Wang, 0053).
Here, “provide” maps to “determine”,
“in each client device 150” maps to “for each of a plurality of UEs within the 3D space”,
“current conditions of the network” for “each client device” maps to “values associated with one or more UE-side features that are specific to each UE”).
Predict UE-specific values of the one or more cellular network performance parameters for each UE as a function of the values associated with the one or more UE-side features of each respective UE to generate predicted UE-specific values
(“The network evaluation service described herein may provide a solution to ‘learn’ about parameters of essentially every packet size of each application on each device in the network.... In one instance, the learning result can be used to predict the application performance on each device” (Wang, 0014). Also, Fig. 4B in Wang shows real-time prediction of UE packet parameters.
Here, “predict the application performance on each device” maps to “predict UE-specific values of the one or more cellular network performance parameters for each UE ... to generate predicted UE-specific values”,
“can be used” maps to “as a function of”, and
element 412 in Fig. 4B, “live traffic”, of “each device” maps to “values associated with the one or more UE-side features of each respective UE”).
Predicting UE-specific values of the oen or more cellular network performance parameters includes predicting first UE-specific values of the one or more cellular performance parameters that are specific for the first UE ... to generate predicted UE-specific values for the first UE and predicting second UE-specific values of the one or more cellular network performance parameters that are specific for the second UE ... to generate predicted UE-specific values for the second UE
(“The network evaluation service described herein may provide a solution to ‘learn’ about parameters of essentially every packet size of each application on each device in the network.... In one instance, the learning result can be used to predict the application performance on each device in different network conditions” (Wang, 0014).
Here, “predict” maps to “predicting”, and
“application performance on each device” maps to “UE-specific values of the one or more cellular network performance parameters”,
“predict” maps to “predicting”,
“application performance on each device” maps to “first UE-specific values of the one or more cellular performance parameters that are specific for the first UE ... to generate predicted UE-specific values for the first UE and predicting second UE-specific values of the one or more cellular network performance parameters that are specific for a second UE ... to generate predicted UE-specific values for the second UE”).
Predicting cellular data throughput and one or more of signal strength and level of carrier aggregation based on inputting the values associated with the one or more UE-side features that are specific to each UE to a machine learning model
(“The current condition of the network 502 may include, for each application in each client device ... whether there is a degrading (e.g., data rate, latency, jitter, etc.) ... Network prediction module 510 may learn from the current conditions 502 ... Fig. 6 is an exemplary table which may illustrate prediction output 506.... Entries 630 for field 620 may include one or more predicted performance parameters, such as data speed, latency, jitter, etc” (Wang, 0053-0055).
Here, “predicted performance” which includes “data speed” maps to “predicting cellular data throughput”,
“includes” maps to “include”, and
“throughput” maps to “cellular data throughput” from the list of “one or more of cellular data throughput, signal strength and level of carrier aggregation”,
“jitter” maps to “signal strength” from the list of “one or more of signal strength and level of carrier aggregation” because jitter is a measure of noise which is inversely proportional to relative signal strength, so measuring jitter is measuring the inverse of signal strength,
“learn from” maps to “based on inputting ... to a machine learning model”, and
“current condition of the network 502” for “each client device” maps to “the values associated with the one or more UE-side features that are specific to each UE”).
The output of the machine learning module are the predicted UE-specific values including the predicted UE-specific values for the first UE and the predicted UE-specific values for the second UE
(“The current condition of the network 502 may include, for each application in each client device ... whether there is a degrading (e.g., data rate, latency, jitter, etc.) ... Network prediction module 510 may learn from the current conditions 502 ... Fig. 6 is an exemplary table which may illustrate prediction output 506.... Entries 630 for field 620 may include one or more predicted performance parameters, such as data speed, latency, jitter, etc” (Wang, 0053-0055).
Here, “prediction output” maps to “the output of the machine learning module”,
element 620 in Fig. 6, “predicted performance of each client device”, maps to “the predicted UE-specific values including the predicted UE-specific values for the first UE and the predicted UE-specific values for the second UE”).
The predicted UE-specific values for the first UE include predicted cellular data throughput and one or more of signal strength and level or carrier aggregation specific for the first UE
(“The current condition of the network 502 may include, for each application in each client device ... whether there is a degrading (e.g., data rate, latency, jitter, etc.) ... Network prediction module 510 may learn from the current conditions 502 ... Fig. 6 is an exemplary table which may illustrate prediction output 506.... Entries 630 for field 620 may include one or more predicted performance parameters, such as data speed, latency, jitter, etc” (Wang, 0053-0055).
Here, “predicted performance of each client device” in Fig. 6 maps to “the predicted UE-specific values for the first UE”,
“such as” maps to “include”,
“data speed” maps to “predicted cellular data strength”,
“jitter” maps to “signal strength and level” from the list of “signal strength and level or carrier aggregation”, and
“of each client device” maps to “specific for the first UE”).
The predicted UE-specific values for the second UE include predicted cellular data throughput and one or more of signal strength and level or carrier aggregation specific for the second UE
(“The current condition of the network 502 may include, for each application in each client device ... whether there is a degrading (e.g., data rate, latency, jitter, etc.) ... Network prediction module 510 may learn from the current conditions 502 ... Fig. 6 is an exemplary table which may illustrate prediction output 506.... Entries 630 for field 620 may include one or more predicted performance parameters, such as data speed, latency, jitter, etc” (Wang, 0053-0055).
Here, “predicted performance of each client device” in Fig. 6 maps to “the predicted UE-specific values for the second UE”,
“such as” maps to “include”,
“data speed” maps to “predicted cellular data strength”,
“jitter” maps to “signal strength and level” from the list of “signal strength and level or carrier aggregation”, and
“of each client device” maps to “specific for the second UE”).
The machine learning module has been trained using truth data associated with the one or more UE-side features
(“As shown in Fig. 4A, training module 335 may be a machine-learning-based function located in the cloud. In one implementation, training module 335 may receive historical training data 402 from benchmark performance database 330 or another data source” (Wang, 0049).
Here, “machine-learning-based function” maps to “the machine learning module”,
“receive historical training data” maps to “has been trained using truth data associated with the one or more UE-side features”).
Wang does not explicitly disclose:
The UE-side features include mobility features, the mobility features including direction and speed of movement relative to the 3D space that are specific for each UE
The UE-side features include first mobility features including direction and speed of movement relative to the 3D space that are specific to a first UE and second mobility features including direction and speed of movement relative to the 3D space that are specific to a second UE
Predicting first UE-specific values for the first UE ... based on the first mobility features to generate predicted UE-specific values for the first UE
Predicting second UE-specific values for the second UE ... based on the second mobility features to generate predicted UE-specific values for the second UE
The predicted UE-specific values ... based on the first mobility features
The predicted UE-specific values for the second UE ... based on the second mobility features
However, Ottersten does teach:
The UE-side features include mobility features, the mobility features including direction and speed of movement relative to the 3D space that are specific for each UE
(“The input data may comprise one or more input parameters such as received signal strength, angle of arrival, measured or estimated UE speed, target block or bit error rates, just to give some examples” (Ottersten, 0103).
Here, “input data” maps to “the UE-side features”,
“angle of arrival” and “UE speed” map to “mobility features ... that are specific for each UE”,
“angle of arrival” maps to “the mobility features including direction”, and
“UE speed” maps to “speed of movement relative to the 3D space”).
The UE-side features include first mobility features including direction and speed of movement relative to the 3D space that are specific to a first UE and second mobility features including direction and speed of movement relative to the 3D space that are specific to a second UE
(“The input data may comprise one or more input parameters such as received signal strength, angle of arrival, measured or estimated UE speed, target block or bit error rates, just to give some examples” (Ottersten, 0103).
Here, “[t]he input data may include” maps to “the UE-side features include”,
“UE speed” maps to “first mobility features including direction and speed of movement relative to the 3D space that are specific to a first UE”, and
“UE speed” maps to “second mobility features including direction and speed of movement relative to the 3D space that are specific to a second UE” because MPEP 2144 VI B lists “duplication of parts” as an example of an obvious modification).
Predicting first UE-specific values ... based on the first mobility features to generate predicted UE-specific values for the first UE
(“The input data may comprise one or more input parameters such as received signal strength, angle of arrival, measured or estimated UE speed, target block or bit error rates, just to give some examples” (Ottersten, 0103).
Here, “input data may comprise ... measured or estimated UE speed” maps to “predicting first UE-specific values ... based on the first mobility features to generate predicted UE-specific values for the first UE”).
Predicting second UE-specific values ... based on the second mobility features to generate predicted UE-specific values for the second UE
(“The input data may comprise one or more input parameters such as received signal strength, angle of arrival, measured or estimated UE speed, target block or bit error rates, just to give some examples” (Ottersten, 0103).
Here, “input data may comprise ... measured or estimated UE speed” maps to “predicting second UE-specific values ... based on the first mobility features to generate predicted UE-specific values for the second UE”).
The predicted UE-specific values for the first UE ... based on the first mobility features
(“The input data may comprise one or more input parameters such as received signal strength, angle of arrival, measured or estimated UE speed, target block or bit error rates, just to give some examples” (Ottersten, 0103).
Here, “measured or estimated UE speed” maps to “the predicted UE-specific values for the first UE”, and
“the input data may comprise ... measured or estimated UE speed” maps to “based on the first mobility features”).
The predicted UE-specific values for the second UE ... based on the second mobility features
(“The input data may comprise one or more input parameters such as received signal strength, angle of arrival, measured or estimated UE speed, target block or bit error rates, just to give some examples” (Ottersten, 0103).
Here, “measured or estimated UE speed” maps to “the predicted UE-specific values for the second UE”, and
“the input data may comprise ... measured or estimated UE speed” maps to “based on the first mobility features”).
Thus, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to use the mobility features described in Ottersten as the inputs for the machine learning model in Wang. UE mobility features can help to predict future throughput and noise, just as past throughput and noise measurements can.
The combination of Wang and Ottersten does not explicitly disclose:
Predict UE-specific values for the first UE ... independent of the second mobility features
Predict UE-specific values for the second UE ... independent of the first mobility features
However, Fu does teach:
Predict UE-specific values for the first UE ... independent of the second mobility features
(“[T]he traffic throughput prediction module 200b [in Fig. 6] receives ... previous traffic throughput for the wireless device as inputs ... [M]essages from the traffic throughput prediction module 200b can include predicted traffic throughput for the wireless device 140” (Fu, 0078, 0080).
Here, “predicted traffic throughput for the wireless device” in Fig. 6 maps to “predict UE-specific values for the first UE”, and
“previous traffic throughput for the wireless device as inputs” maps to “independent of the second mobility features”).
Predict UE-specific values for the second UE ... independent of the first mobility features
(“[T]he traffic throughput prediction module 200b [in Fig. 6] receives ... previous traffic throughput for the wireless device as inputs ... [M]essages from the traffic throughput prediction module 200b can include predicted traffic throughput for the wireless device 140” (Fu, 0078, 0080).
Here, “predicted traffic throughput for the wireless device” in Fig. 6 maps to “predict UE-specific values for the second UE”, and
“previous traffic throughput for the wireless device as inputs” maps to “independent of the first mobility features”).
Thus, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate Fu’s practice of using device-specific input data into Wang’s method for predicting wireless device throughput. The past throughput values for a device would be an obvious candidate input to predict that device’s throughput in the future.
Claim(s) 4 is/are rejected under 35 U.S.C. 103 as being unpatentable over Wang (US 2020/003) in view of Ottersten (US 2021/0345134 A1) and Fu (US 2018/0310203 A1) and further in view of Hu et al. (US 2023/0362057 A1, hereinafter “Hu”).
As to Claim 4:
Wang teaches:
The predicted values of the one or more cellular network performance parameters
(“The network evaluation service described herein may provide a solution to ‘learn’ about parameters of essentially every packet size of each application on each device in the network.... In one instance, the learning result can be used to predict the application performance on each device in different network conditions” (Wang, 0014).
Here, “predict the application performance on each device” maps to “the predicted values of the one or more cellular network performance parameters”).
Optimize the cellular data throughput
(“The current condition of the network 502 may include, for each application in each client device ... whether there is a degrading (e.g., data rate, latency, jitter, etc.) ... Network prediction module 510 may learn from the current conditions 502 ... Fig. 6 is an exemplary table which may illustrate prediction output 506.... Entries 630 for field 620 may include one or more predicted performance parameters, such as data speed, latency, jitter, etc” (Wang, 0053-0055).
Here, “predicted performance” maps to “optimize”, and
“data speed” maps to “the cellular data throughput”).
The combination of Wang, Ottersten, and Fu does not explicitly disclose:
Selecting network channels based on ... parameters ... to additionally optimize one or more of quality of service, handoff, and data prefetch across the UEs
However, Hu does describe a method for allocating resources for network slices.
Specifically, Hu teaches:
Selecting network channels based on ... parameters ... to additionally optimize one or more of quality of service, handoff, and data prefetch across the UEs
(“[I]f the selected data transport channels also comply with the security requirements, the latency requirements and the QoS requirements for the TN NSS to be created, the TN NSSFM may map the TN NSS to at least one of the selected data transport channels” (Hu, 0082).
Here, “map ... to at least one of the selected data transport channels” maps to “selecting network channels”,
“comply with the ... requirements” maps to “based on ... parameters”, and
“the QoS requirements” maps to “to additionally optimize ... quality of service” from the list of “to additionally optimize one or more of quality of service, handoff, and data prefetch across the UEs”).
Thus, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to select network frequencies to optimize QoS, as described in Hu, into Wang’s method for optimizing network parameters. QoS is an important aspect of network performance, and selecting unused network channels is an obvious path to improving it.
Claim(s) 7 is/are rejected under 35 U.S.C. 103 as being unpatentable over Wang (US 2020/0304381 A1) in view of Ottersten (US 2021/0345134 A1) and Fu (US 2018/0310203 A1) and further in view of Bellamkonda et al. (US 11,012,872 B1, hereinafter “Bellamkonda”).
As to Claim 7:
The combination of Wang, Ottersten, and Fu does not explicitly disclose:
The direction and speed of movement relative to the 3D space for at least one UE indicates that the at least one UE is not moving within the 3D space
However, Bellamkonda does describe a method for allocating slice resources.
Specifically, Bellamkonda teaches:
The direction and speed of movement relative to the 3D space for at least one UE indicates that the at least one UE is not moving within the 3D space
(“[T]he polymorphic algorithms may include ... mobility algorithms (e.g., stationary)” (Bellamkonda col. 6, line 67; col. 7, lines 1-3).
Here, “stationary” maps to “the direction and speed of movement relative to the 3D space for at least one UE indicates that the at least one UE is not moving within the 3D space”).
Thus, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate Bellamkonda’s practice of accounting for stationary mobility into Wang’s method for optimizing network performance. A stationary device is one possible mobility scenario, so it would have been obvious to account for it when using mobility parameters.
Claim(s) 9-11 is/are rejected under 35 U.S.C. 103 as being unpatentable over Wang (US 2020/0304381 A1) in view of Ottersten (US 2021/0345134 A1) and Fu (US 2018/0310203 A1) and further in view of Svendsen et al. (US 2023/0299835 A1, hereinafter “Svendsen”).
As to Claim 9:
The combination of Wang, Ottersten, and Fu does not explicitly disclose:
The UE-side features further include tower-based features
The tower-based features are selected from features including distance between panel and UE, UE-panel positional angle, and UE-panel mobility angle
However, Svendsen does describe a method for optimizing the distance between two panels on a UE to achieve antenna isolation.
Specifically, Svendsen teaches:
The UE-side features further include tower-based features
(“In some exemplary embodiments, the UE panel-to-panel distance may be large enough for panel-to-panel interaction to not be considered as a near-field coupling at mmWave frequencies” (Svendsen, 0062).
Here, “the UE panel-to-panel distance” maps to “the UE-side features further include tower-based features” because this is an example of a “tower-based feature” given later in this claim).
The tower-based features are selected from features including distance between panel and UE, UE-panel positional angle, and UE-panel mobility angle
(“In some exemplary embodiments, the UE panel-to-panel distance may be large enough for panel-to-panel interaction to not be considered as a near-field coupling at mmWave frequencies” (Svendsen, 0062).
Here, “UE panel-to-panel distance” maps to “the tower based features are selected from ... distance between panel and UE” from the list of “the tower based features are selected from features including distance between panel and UE, UE-panel positional angle and UE-panel mobility angle”).
Thus, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to use the UE to panel distance taught in Svendsen to as an input criteria in the machine learning model described in Wang. UE to panel distance can decisively impact network performance, making the former a logical feature to consider when predicting the latter.
As to Claim 10:
The combination of Wang, Ottersten, and Fu does not explicitly disclose:
The UE-side features further include tower-based features
However, Svendsen does teach:
The UE-side features further include tower-based features
(“In some exemplary embodiments, the UE panel-to-panel distance may be large enough for panel-to-panel interaction to not be considered as a near-field coupling at mmWave frequencies” (Svendsen, 0062).
Here, “the UE panel-to-panel distance” maps to “the UE-side features further include tower-based features” because this is an example of a “tower-based feature” given in Claim 11).
Thus, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to use the UE to panel distance taught in Svendsen to as an input criteria in the machine learning model described in Wang. UE to panel distance can decisively impact network performance, making the former a logical feature to consider when predicting the latter.
As to Claim 11:
The combination of Wang, Ottersten, and Fu does not explicitly disclose:
The tower-based features are selected from features including distance between panel and UE, UE-panel positional angle and UE-panel mobility angle
However, Svendson does teach:
The tower-based features are selected from features including distance between panel and UE, UE-panel positional angle and UE-panel mobility angle
(“In some exemplary embodiments, the UE panel-to-panel distance may be large enough for panel-to-panel interaction to not be considered as a near-field coupling at mmWave frequencies” (Svendsen, 0062).
Here, “UE panel-to-panel distance” maps to “the tower based features are selected from ... distance between panel and UE” from the list of “the tower based features are selected from features including distance between panel and UE, UE-panel positional angle and UE-panel mobility angle”).
Thus, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to use the UE to panel distance taught in Svendsen to as an input criteria in the machine learning model described in Wang. UE to panel distance can decisively impact network performance, making the former a logical feature to consider when predicting the latter.
Claim(s) 14 is/are rejected under 35 U.S.C. 103 as being unpatentable over Wang (US 2020/0304381 A1) in view of Ottersten (US 2021/0345134 A1) and Fu (US 2018/0310203 A1) and further in view of Tayamon et al. (US 2024/0205956 A1, hereinafter “Tayamon”).
As to Claim 14:
The combination of Wang, Ottersten, and Fu does not explicitly disclose:
The UE-side features include factors, associated with each respective UE, that attenuate 5G signals
However, Tayamon does describe a machine learning model for selecting the format of a PUCCH (Physical Uplink Control Channel).
Specifically, Tayamon teaches:
The UE-side features include factors, associated with each respective UE, that attenuate 5G signals
(“Examples of UE information are as follows: ... Signal attenuation measurements between the user device and one or more network nodes. This may include measurements of pathloss, fading, shadowing over one or multiple communication frequencies” (Tayamon, 0065, 0068).
Here, “UE information” maps to “UE-side feature”, and
“measurements of pathloss, fading, shadowing” maps to “factors, associated with each respective UE, that attenuate 5G signals”).
Thus, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to combine the input data indicating signal attenuation taught in Tayamon with the model taught in Wang which predicts network performance. Signal attenuation can severely degrade network performance, so it makes sense to incorporate it as an input in a machine learning model that predicts performance.
Claim(s) 21 is/are rejected under 35 U.S.C. 103 as being unpatentable over Wang (US 2020/0304381 A1) in view of Guha et al. (US 10,993,140 B1) and further in view of Ottersten (US 2021/0345134 A1) and Fu (US 2018/0310203 A1).
As to Claim 21:
Wang teaches:
Predicting cellular performance for user equipment (UE) within a three-dimensional (3D) space having a plurality of cellular nodes
(“Implementations of new broadband cellular networks (e.g., 3GPP 5G networks) and other wireless networks (e.g., IEEE 802.11ax) are raising user expectations for increased network speeds.... The network evaluation service described herein may provide a solution to ‘learn’ about parameters of essentially every packet size of each application on each device in the network.... In one instance, the learning result can be used to predict the application performance on each device” (Wang, 0001, 0014).
Here, “predict the application performance” maps to “predicting cellular performance”,
“on each device” maps to “user equipment (UE)”, and
“5G networks” maps to “a three-dimensional (3D) space having a plurality of cellular nodes”).
Determining, for each of a plurality of UEs within the 3D space, values associated with one or more UE-side features that are specific for each UE
(“Client devices 150 [in Fig. 1] may store applications or ‘apps’) that receive and/or generate traffic (e.g., packets) ... Referring to Fig. 4B, at traffic pattern recognition function 320, each segment of live traffic 412 will be input to traffic summary extraction 420 to generate a data rate, jitter, and latency value for each category of packets” (Wang, 0023, 0050).
Here, “live traffic 412 will be input” for “client devices 150” maps to “determining ... values associated with one or more UE-side features that are specific for each UE”, and
“client devices 150” map to “a plurality of UEs within the 3D space”).
Estimating ... cellular network performance for each UE as a function of the values
(“The network evaluation service described herein may provide a solution to ‘learn’ about parameters of essentially every packet size of each application on each device in the network.... In one instance, the learning result can be used to predict the application performance on each device” (Wang, 0014). Also, Fig. 4B in Wang shows real-time prediction of UE packet parameters.
Here, “predict the application performance on each device in the network” maps to “estimating ... cellular network performance for each UE”, and
“the learning result can be used” maps to “as a function of the values”).
Estimating cellular network performance includes applying the values to a machine learning module trained using truth data associated with the one or more UE-side features
(“The network evaluation service described herein may provide a solution to ‘learn’ about parameters of essentially every packet size of each application on each device in the network.... In one instance, the learning result can be used to predict the application performance on each device” (Wang, 0014). Also, Fig. 4A in Wang shows a packet parameter prediction model being trained, and Fig. 4B in Wang shows real-time prediction of UE packet parameters.
Here, “predict the application performance on each device in the network” maps to “estimating cellular network performance”,
inputting “Live Traffic 412” into the “Trained Model 410” in Fig. 4B maps to “applying the values to a machine learning module”,
inputting “Training Data 402” into the “Supervised Learning” module “405” in Fig. 4A maps to “machine learning module trained using truth data associated with the one or more UE-side features”).
Estimating 5G performance includes estimating a first 5G performance that is specific for a first UE and estimating a second 5G performance that is specific for a second UE
(“The network evaluation service described herein may provide a solution to ‘learn’ about parameters of essentially every packet size of each application on each device in the network.... In one instance, the learning result can be used to predict the application performance on each device in different network conditions” (Wang, 0014).
Here, “predict the application performance” maps to “estimating 5G performance includes estimating a first 5G performance”, and
“application performance on each device” maps to “a first 5G performance that is specific for a first UE and estimating a second 5G performance that is specific for a second UE”).
Estimating 5G performance includes estimating cellular data throughput and one or more of signal strength and level of carrier aggregation based on the one or more UE-side features that are specific to each UE
(“The current condition of the network 502 may include, for each application in each client device ... whether there is a degrading (e.g., data rate, latency, jitter, etc.) ... Network prediction module 510 may learn from the current conditions 502 ... Fig. 6 is an exemplary table which may illustrate prediction output 506.... Entries 630 for field 620 may include one or more predicted performance parameters, such as data speed, latency, jitter, etc” (Wang, 0053-0055).
Here, “predicted performance” which includes “data speed” maps to “estimating 5G performance”,
“includes” maps to “include”, and
“predicted performance parameters, such as data speed” maps to “estimating cellular data throughput”,
“jitter” maps to “signal strength” from the list of “one or more of signal strength and level of carrier aggregation” because jitter is a measure of noise which is inversely proportional to relative signal strength, so measuring jitter is measuring the inverse of signal strength,
“learn from” maps to “based on”, and
“current condition of the network 502” for “each client device” maps to “the one or more UE-side features that are specific to each UE”).
Data traffic
(“The network evaluation service described herein may provide a solution to ‘learn’ about parameters of essentially every packet size of each application on each device in the network.... In one instance, the learning result can be used to predict the application performance on each device” (Wang, 0014).
Here, “packet” maps to “data traffic”).
Wang does not explicitly disclose:
The cellular nodes including two or more 5G panels and at least one 4G tower
Estimating 4G cellular performance for each UE
Estimating 5G cellular performance for each UE
The estimated 4G cellular data performance
The estimated 5G cellular data performance
Determining, for each UE and based on the estimated 4G cellular data performance and the estimated 5G cellular data performance, a combination of 4G ... and 5G ... needed to optimize cellular performance across the plurality of pieces of UE
However, Guha does describe a method for tuning the ratio of network resources allocated to 4G and 5G networks.
Specifically, Guha teaches:
The cellular nodes including two or more 5G panels and at least one 4G tower
(“RAN 110 may include one or more eNBs ... RAN 110 [in Fig. 1A] may receive, from a second UE 105, second traffic associated with a 5G service” (Guha col. 2, lines 27-28, 38-40).
Here, “UE 105” and other parts of the “RAN” map to “the cellular nodes”,
the sending and receiving panels for transmission 125 in Fig. 1A mp to “two or more 5G panels”, and
“eNBs” maps to “at least one 4G tower”).
Estimating 4G cellular performance for each UE
(“The eNB/gNB may calculate a per quality of service class identifier (QCI) or 5G quality of service flow identifier (QFI) split for the first traffic and the second traffic based on the core network data ... The eNB/gNB utilizes a second loop to adjust the spectrum allocation of the RAN resources based on core network data (e.g., network slice data), to adapt to weights assigned by the core network for different technologies, and to tune user experience levels for 4G and 5G traffic” (Guha, col. 1 lines 61-64; col. 2 lines 13-18).
Here, “calculate” maps to “estimating”, and
“a per quality of service class identifier (QCI) ... for 4G ... traffic” maps to “4G cellular performance for each UE”).
Estimating 5G cellular performance for each UE
(“The eNB/gNB may calculate a per quality of service class identifier (QCI) or 5G quality of service flow identifier (QFI) split for the first traffic and the second traffic based on the core network data ... The eNB/gNB utilizes a second loop to adjust the spectrum allocation of the RAN resources based on core network data (e.g., network slice data), to adapt to weights assigned by the core network for different technologies, and to tune user experience levels for 4G and 5G traffic” (Guha, col. 1 lines 61-64; col. 2 lines 13-18).
Here, “calculate” maps to “estimating”, and
“a per quality of service class identifier (QCI) ... for ... 5G traffic” maps to “5G cellular performance for each UE”).
The estimated 4G cellular data performance
(“The eNB/gNB may calculate a per quality of service class identifier (QCI) or 5G quality of service flow identifier (QFI) split for the first traffic and the second traffic based on the core network data ... The eNB/gNB utilizes a second loop to adjust the spectrum allocation of the RAN resources based on core network data (e.g., network slice data), to adapt to weights assigned by the core network for different technologies, and to tune user experience levels for 4G and 5G traffic” (Guha, col. 1 lines 61-64; col. 2 lines 13-18).
Here, “calculate” maps to “estimated”, and
“a per quality of service class identifier (QCI) ... for 4G ... traffic” maps to “4G cellular data performance”).
The estimated 5G cellular data performance
(“The eNB/gNB may calculate a per quality of service class identifier (QCI) or 5G quality of service flow identifier (QFI) split for the first traffic and the second traffic based on the core network data ... The eNB/gNB utilizes a second loop to adjust the spectrum allocation of the RAN resources based on core network data (e.g., network slice data), to adapt to weights assigned by the core network for different technologies, and to tune user experience levels for 4G and 5G traffic” (Guha, col. 1 lines 61-64; col. 2 lines 13-18).
Here, “calculate” maps to “estimated”, and
“a per quality of service class identifier (QCI) ... for ... 5G traffic” maps to “5G cellular data performance”).
Determining, for each UE and based on the estimated 4G cellular data performance and the estimated 5G cellular data performance, a combination of 4G ... and 5G ... needed to optimize cellular performance across the plurality of pieces of UE
(“If the incoming traffic includes 20% 4G traffic and 80% 5G traffic, the scheduler will target RAN resource allocation in the same ratio (e.g., 20% of the RAN resources will be allocated to the 4G traffic and 80% of the RAN resources will be allocated to the 5G traffic ... RAN 110 may adapt to weights assigned by core network 115 for different technologies, and may tune user experience levels of UEs 105 for 4G and 5G traffic” (Guha col. 1, lines 40-45; col. 6, lines 5-9).
Here, “target” maps to “determining”,
“of UEs” maps to “for each piece of UE”,
“adapt to ... 4G traffic” maps to “based on the estimated 4G cellular data performance”,
“adapt to ... 5G traffic” maps to “based on ... the estimated 5G cellular data performance”,
“20% of the RAN resources will be allocated to the 4G traffic and 80% of the RAN resources will be allocated to the 5G traffic” maps to “a combination of 4G ... and 5G”
“tune” maps to “needed to optimize”, and
“user experience levels of UEs” maps to “cellular performance across the plurality of pieces of UE”).
Thus, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to extend the machine learning model for predicting network traffic taught in Wang to both 4G and 5G traffic like the resource allocation method taught in Guha. Many modern networks contain a mix of 4G and 5G traffic, so it would be obvious to one of ordinary skill in the art already using machine learning to use it to account for the mixing of 4G and 5G traffic and predict the optimal split ratio.
The combination of Wang and Guha also does not explicitly disclose:
The UE-side features include mobility features, the mobility features including direction and speed of movement relative to the 3D space that are specific for each UE
However, Ottersten does teach:
The UE-side features include mobility features, the mobility features including direction and speed of movement relative to the 3D space that are specific for each UE
(“The input data may comprise one or more input parameters such as received signal strength, angle of arrival, measured or estimated UE speed, target block or bit error rates, just to give some examples” (Ottersten, 0103).
Here, “input data” maps to “the UE-side features”,
“angle of arrival” and “UE speed” map to “mobility features ... that are specific for each UE”,
“angle of arrival” maps to “the mobility features including direction”, and
“UE speed” maps to “speed of movement relative to the 3D space”).
Thus, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to use the mobility features described in Ottersten as the inputs for the machine learning model in Wang. UE mobility features can help to predict future throughput and noise, just as past throughput and noise measurements can.
The combination of Wang, Guha, and Ottersten does not explicitly disclose:
Estimating cellular data throughput ... independent of the one or more UE-side features of other UEs
However, Fu does teach:
Estimating cellular data throughput ... independent of the one or more UE-side features of other UEs
(“[T]he traffic throughput prediction module 200b [in Fig. 6] receives ... previous traffic throughput for the wireless device as inputs ... [M]essages from the traffic throughput prediction module 200b can include predicted traffic throughput for the wireless device 140” (Fu, 0078, 0080).
Here, “throughput prediction” maps to “estimating cellular data throughput”, and
“previous traffic for the wireless device as inputs” maps to “independent of the one or more UE-side features of other UEs”).
Thus, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to incorporate Fu’s practice of using device-specific input data into Wang’s method for predicting wireless device throughput. The past throughput values for a device would be an obvious candidate input to predict that device’s throughput in the future.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Benjamin Peter Welte whose telephone number is (703)756-5965. The examiner can normally be reached Monday - Friday, EST.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Chirag Shah, can be reached at (571)272-3144. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/B.P.W./Examiner, Art Unit 2477
/CHIRAG G SHAH/Supervisory Patent Examiner, Art Unit 2477