DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
This office action is a response to an application filed on 10/24/2023 in which claims 1-20 are pending.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 10/24/2023 has been considered by the examiner. The submission is in compliance with the provisions of 37 CFR 1.97.
Claim Rejections - 35 USC § 102
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
Claims 1, 3-4, 11, 13 and 14 are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Choi et al. (US 2020/0084102), hereinafter “Choi”.
As to claim 1, Choi teaches a communication device comprising:
a transceiver (Choi, Fig. 4, [0095], the first wireless communication circuit 412 for transmission and reception of a wireless signals) configured to receive traffic over a wireless network for an application in a target wake time (TWT) operation (Choi, Fig. 10, step 1030, [0148], the first electronic device receive data of a first service type during a service period of the first TWT. [0146], the first service type of a first application (e.g. VI). [0063], video (VI)); and
a processor operably coupled to the transceiver and configured to (Choi, Fig. 4, [0095], “a first processor 413 operatively connected to the first wireless communication circuit 412” ):
determine network statistics from the traffic (Choi, [0068], the latency (delay) is measured to detect deterioration of the quality of service. The frame/packet loss rate is also determined. Additionally, [0071], monitor a data throughput for each service type at a given interval, where the data throughput is the amount of data sent and received during a given time period),
estimate a quality of experience (QoE) value for the application based on the network statistics (Choi, [0068], a call quality score is determined in accordance to the latency (mouth to ear delay) and frame/packet loss rate. The call quality score represents a quality of experience at the user, such as 80 points being an acceptable or good call quality. The latency when the call quality score starts to be lowered to 80 points or less may be considered as a critical latency), and
determine new TWT parameters for the TWT operation based on the estimated QoE value (Choi, [0068], the TWT is configured to be shorter than the given critical latency in consideration of the service quality score (i.e. higher or lower than the 80 points). The critical latency is the latency in which deterioration of the quality of service occurs. Additionally, [0125]-[0126], a second TWT configuration is generated in accordance to the issue of data throughput being equal to or lower than a first critical value).
As to claim 3, Choi teaches wherein the processor is further configured to:
estimate quality of service (QoS) values for the application as a function of the network statistics (Choi, [0068], the latency (delay) is measured to detect deterioration of the quality of service. The frame/packet loss rate is also determined. Additionally, [0071], monitor a data throughput for each service type at a given interval, where the data throughput is the amount of data sent and received during a given time period); and
estimate the QoE value for the application as a function of the estimated QoS values (Choi, [0068], a call quality score is determined in accordance to the latency (mouth to ear delay) and frame/packet loss rate. The call quality score represents a quality of experience at the user, such as 80 points being an acceptable or good call quality. The latency when the call quality score starts to be lowered to 80 points or less may be considered as a critical latency).
As to claim 4, Choi teaches wherein the estimated QoS values of the application include at least one of delay, jitter, or packet loss (Choi, [0068], the latency (delay) is measured to detect deterioration of the quality of service. The frame/packet loss rate is also determined. Additionally, [0071], monitor a data throughput for each service type at a given interval, where the data throughput is the amount of data sent and received during a given time period).
As to claim 11, Choi teaches a method for communication comprising:
receiving traffic over a wireless network for an application in a target wake time (TWT) operation (Choi, Fig. 10, step 1030, [0148], the first electronic device receive data of a first service type during a service period of the first TWT. [0146], the first service type of a first application (e.g. VI). [0063], video (VI));
determining network statistics from the traffic (Choi, [0068], the latency (delay) is measured to detect deterioration of the quality of service. The frame/packet loss rate is also determined. Additionally, [0071], monitor a data throughput for each service type at a given interval, where the data throughput is the amount of data sent and received during a given time period);
estimating a quality of experience (QoE) value for the application based on the network statistics (Choi, [0068], a call quality score is determined in accordance to the latency (mouth to ear delay) and frame/packet loss rate. The call quality score represents a quality of experience at the user, such as 80 points being an acceptable or good call quality. The latency when the call quality score starts to be lowered to 80 points or less may be considered as a critical latency); and
determining new TWT parameters for the TWT operation based on the estimated QoE value (Choi, [0068], the TWT is configured to be shorter than the given critical latency in consideration of the service quality score (i.e. higher or lower than the 80 points). The critical latency is the latency in which deterioration of the quality of service occurs. Additionally, [0125]-[0126], a second TWT configuration is generated in accordance to the issue of data throughput being equal to or lower than a first critical value).
As to claim 13, Choi teaches further comprising:
estimating quality of service (QoS) values for the application as a function of the network statistics (Choi, [0068], the latency (delay) is measured to detect deterioration of the quality of service. The frame/packet loss rate is also determined. Additionally, [0071], monitor a data throughput for each service type at a given interval, where the data throughput is the amount of data sent and received during a given time period); and
estimating the QoE value for the application as a function of the estimated QoS values (Choi, [0068], a call quality score is determined in accordance to the latency (mouth to ear delay) and frame/packet loss rate. The call quality score represents a quality of experience at the user, such as 80 points being an acceptable or good call quality. The latency when the call quality score starts to be lowered to 80 points or less may be considered as a critical latency).
As to claim 14, Choi teaches wherein the estimated QoS values of the application include at least one of delay, jitter, or packet loss (Choi, [0068], the latency (delay) is measured to detect deterioration of the quality of service. The frame/packet loss rate is also determined. Additionally, [0071], monitor a data throughput for each service type at a given interval, where the data throughput is the amount of data sent and received during a given time period).
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 2 and 12 are rejected under 35 U.S.C. 103 as being unpatentable over Choi et al. (US 2020/0084102), hereinafter “Choi” in view of Dunne et al. (US 2015/0082130), hereinafter “Dunne”.
Choi teaches the claimed limitations as stated above. Choi does not explicitly teach the following features: regarding claim 2, wherein the processor is further configured to:
determine whether the application is a real-time application based on the network statistics, and
based on a determination that the application is a real-time application, perform the estimation of the QoE value for the application.
As to claim 2, Dunne teaches wherein the processor is further configured to:
determine whether the application is a real-time application based on the network statistics (Dunne, Fig. 4, [0043], during a communication session, a packet loss rate greater than 5% is considered detrimental to voice quality. A burst ratio of 2 is also measured for the communication session. [0023], the communication session affected by this type of qualities include voice and video, which are real-time communications), and
based on a determination that the application is a real-time application, perform the estimation of the QoE value for the application (Dunne, Fig. 4, [0043], based on the packet loss rate greater than 5% being detrimental to voice quality of the communication session, a mean opinion score (“MOS”) of 3.1 is determined, which indicate that user 40 is unsatisfied with the quality of experience of the communication session).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Choi to have the features, as taught by Dunne in order to increase the quality of experience for the participants of the communication session and to maintain the quality of experience within an acceptable level for the participants (Dunne, [0025]).
Choi teaches the claimed limitations as stated above. Choi does not explicitly teach the following features: regarding claim 12, further comprising:
determining whether the application is a real-time application based on the network statistics; and
based on a determination that the application is a real-time application, performing the estimation of the QoE value for the application.
As to claim 12, Dunne teaches further comprising:
determining whether the application is a real-time application based on the network statistics (Dunne, Fig. 4, [0043], during a communication session, a packet loss rate greater than 5% is considered detrimental to voice quality. A burst ratio of 2 is also measured for the communication session. [0023], the communication session affected by this type of qualities include voice and video, which are real-time communications); and
based on a determination that the application is a real-time application, performing the estimation of the QoE value for the application (Dunne, Fig. 4, [0043], based on the packet loss rate greater than 5% being detrimental to voice quality of the communication session, a mean opinion score (“MOS”) of 3.1 is determined, which indicate that user 40 is unsatisfied with the quality of experience of the communication session).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Choi to have the features, as taught by Dunne in order to increase the quality of experience for the participants of the communication session and to maintain the quality of experience within an acceptable level for the participants (Dunne, [0025]).
Claims 5, 7, 15 and 17 are rejected under 35 U.S.C. 103 as being unpatentable over Choi et al. (US 2020/0084102), hereinafter “Choi” in view of Johansson et al. (US 2021/0150393), hereinafter “Johansson”.
Choi teaches the claimed limitations as stated above. Choi does not explicitly teach the following features: regarding claim 5, wherein:
prior to the traffic being received by the transceiver, training traffic for the application is collected over the wireless network in a previous TWT operation,
a set of training QoE values for the application is derived from:
copies of content included in the training traffic that have been impaired by transmission over the wireless network; and
an unimpaired copy of the content,
a set of training network statistics is determined from the training traffic, and
a machine learning (ML) model is trained, using the set of training network statistics as training features and the set of training QoE values as training outcomes, to predict outcome QoE values for the application from network statistic features.
As to claim 5, Johansson teaches wherein:
prior to the traffic being received by the transceiver (Johansson, Fig. 17, [0151], the receiving device has a pretrained ML model to predict MOS for files to be received from the sending wireless device using RTP/IP protocol over the wireless network), training traffic (Johansson, [0067] ln 1-3, [0068], Fig. 1, reference voice samples) for the application is collected over the wireless network (Johansson, Fig. 2, [0071], “analyzing real life data collected during drive testing”) in a previous TWT operation (Johansson, [0085], the drive test data was measured while using DRX),
a set of training QoE values (Johansson, [0006], the Mean Opinion Score (MOS) is a score indicating the experience of listening voice samples by a number of listeners. Fig. 1, [0067], “MOS/P.863 measures 122”, Fig. 2, [0075], the MOS/P.863 score 232. Fig. 5, MOS values) for the application (Johansson, [0017], QoE (Quality of Experience) predictors for voice/video IP based services on a network) is derived from:
copies of content included in the training traffic that have been impaired by transmission over the wireless network (Johansson, Fig. 2, [0072], “Network errors, in the form of jitter and packet loss patterns 218, are applied to the coded audio to simulate degradations from an IP network”); and
an unimpaired copy of the content (Johansson, Fig. 2, [0075], “the original reference audio 212/230”),
a set of training network statistics is determined from the training traffic (Johansson, Fig. 1, [0067], “The database generator module 118 is configured to receive, as database generator inputs, (a) reference voice sample(s) 112…In response to these database generator inputs, the database generator module 118 is configured to output (a) Jitter files 120 comprising error patterns”), and
a machine learning (ML) model (Johansson, Fig. 1, [0105], the machine learning based module 146) is trained, using the set of training network statistics as training features and the set of training QoE values as training outcomes (Johansson, Fig. 1, [0105], the machine learning receives as learning input the Jitter files 142 and the MOS/P.863. [0067], the MOS/P.863 is the expected MOS output for the corresponding Jitter file), to predict outcome QoE values for the application from network statistic features (Johansson, Fig. 1, [0105], Fig. 7, [0106], to output a machine learning based MOS (MOS estimates/predicted MOS) from the jitter, packet loss, etc. Furthermore, Figs. 15-16 show the comparison of the MOS/P.863 with the predicted MOS, where the predicted MOS outputted by the machine learning module (Fig. 1, [0105])).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Choi to have the features, as taught by Johansson in order to cope with the complexity of the inter-dependencies between network, codec and client parameters as well as their significance in impacting the QoE, which is better described with machine learning algorithms (Johansson, [0164]).
Choi teaches the claimed limitations as stated above. Choi does not explicitly teach the following features: regarding claim 7, wherein the processor is configured to estimate the QoE value for the application by inputting the network statistics to the trained ML model as network statistic features and obtaining predicted outcome QoE values from the trained ML model.
As to claim 7, Johansson teaches wherein the processor is configured to estimate the QoE value for the application by inputting the network statistics to the trained ML model as network statistic features and obtaining predicted outcome QoE values from the trained ML model (Johansson, [0027]-[0028], Table 4, Fig. 17, [0151], the wireless device predicts the MOS for the RTP/IP files based on inputting the features/attributes in Table 4 into the pretrained ML to obtain the predicted MOS. [0110], the features include statistical measures from the RTP stream).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Choi to have the features, as taught by Johansson in order to cope with the complexity of the inter-dependencies between network, codec and client parameters as well as their significance in impacting the QoE, which is better described with machine learning algorithms (Johansson, [0164]).
Choi teaches the claimed limitations as stated above. Choi does not explicitly teach the following features: regarding claim 15, wherein:
prior to the traffic being received, training traffic for the application is collected over the wireless network in a previous TWT operation,
a set of training QoE values for the application is derived from:
copies of content included in the training traffic that have been impaired by transmission over the wireless network; and
an unimpaired copy of the content,
a set of training network statistics is determined from the training traffic, and
a machine learning (ML) model is trained, using the set of training network statistics as training features and the set of training QoE values as training outcomes, to predict outcome QoE values for the application from network statistic features.
As to claim 15, Johansson teaches wherein:
prior to the traffic being received (Johansson, Fig. 17, [0151], the receiving device has a pretrained ML model to predict MOS for files to be received from the sending wireless device using RTP/IP protocol over the wireless network), training traffic (Johansson, [0067] ln 1-3, [0068], Fig. 1, reference voice samples) for the application is collected over the wireless network (Johansson, Fig. 2, [0071], “analyzing real life data collected during drive testing”) in a previous TWT operation (Johansson, [0085], the drive test data was measured while using DRX),
a set of training QoE values (Johansson, [0006], the Mean Opinion Score (MOS) is a score indicating the experience of listening voice samples by a number of listeners. Fig. 1, [0067], “MOS/P.863 measures 122”, Fig. 2, [0075], the MOS/P.863 score 232. Fig. 5, MOS values) for the application (Johansson, [0017], QoE (Quality of Experience) predictors for voice/video IP based services on a network) is derived from:
copies of content included in the training traffic that have been impaired by transmission over the wireless network (Johansson, Fig. 2, [0072], “Network errors, in the form of jitter and packet loss patterns 218, are applied to the coded audio to simulate degradations from an IP network”); and
an unimpaired copy of the content (Johansson, Fig. 2, [0075], “the original reference audio 212/230”),
a set of training network statistics is determined from the training traffic (Johansson, Fig. 1, [0067], “The database generator module 118 is configured to receive, as database generator inputs, (a) reference voice sample(s) 112…In response to these database generator inputs, the database generator module 118 is configured to output (a) Jitter files 120 comprising error patterns”), and
a machine learning (ML) model (Johansson, Fig. 1, [0105], the machine learning based module 146) is trained, using the set of training network statistics as training features and the set of training QoE values as training outcomes (Johansson, Fig. 1, [0105], the machine learning receives as learning input the Jitter files 142 and the MOS/P.863. [0067], the MOS/P.863 is the expected MOS output for the corresponding Jitter file), to predict outcome QoE values for the application from network statistic features (Johansson, Fig. 1, [0105], Fig. 7, [0106], to output a machine learning based MOS (MOS estimates/predicted MOS) from the jitter, packet loss, etc. Furthermore, Figs. 15-16 show the comparison of the MOS/P.863 with the predicted MOS, where the predicted MOS outputted by the machine learning module (Fig. 1, [0105])).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Choi to have the features, as taught by Johansson in order to cope with the complexity of the inter-dependencies between network, codec and client parameters as well as their significance in impacting the QoE, which is better described with machine learning algorithms (Johansson, [0164]).
Choi teaches the claimed limitations as stated above. Choi does not explicitly teach the following features: regarding claim 17, further comprising estimating the QoE value for the application by inputting the network statistics to the trained ML model as network statistic features and obtaining predicted outcome QoE values from the trained ML model.
As to claim 17, Johansson teaches further comprising estimating the QoE value for the application by inputting the network statistics to the trained ML model as network statistic features and obtaining predicted outcome QoE values from the trained ML model (Johansson, [0027]-[0028], Table 4, Fig. 17, [0151], the wireless device predicts the MOS for the RTP/IP files based on inputting the features/attributes in Table 4 into the pretrained ML to obtain the predicted MOS. [0110], the features include statistical measures from the RTP stream).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Choi to have the features, as taught by Johansson in order to cope with the complexity of the inter-dependencies between network, codec and client parameters as well as their significance in impacting the QoE, which is better described with machine learning algorithms (Johansson, [0164]).
Claims 6 and 16 are rejected under 35 U.S.C. 103 as being unpatentable over Choi et al. (US 2020/0084102), hereinafter “Choi” in view of Johansson et al. (US 2021/0150393), hereinafter “Johansson” and further in view of Wang et al. (US 2021/0233259), hereinafter “Wang”.
Choi teaches the claimed limitations as stated above. Choi does not explicitly teach the following features: regarding claim 6, wherein:
the content is an image, and
the training QoE values and the outcome QoE values are multi-scale structural similarity index measure (MS-SSIM) values.
As to claim 6, Johansson teaches wherein:
the content is an image (Johansson, [0066], video services), and
the training QoE values and the outcome QoE values are measure values (Johansson, [0006], the Mean Opinion Score (MOS) is a score indicating the experience of listening voice samples by a number of listeners. Fig. 1, [0066]-[0067], the MOS/P.863 (learning data) 144 and ML MOS 150 are values determined for voice and video services).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Choi to have the features, as taught by Johansson in order to cope with the complexity of the inter-dependencies between network, codec and client parameters as well as their significance in impacting the QoE, which is better described with machine learning algorithms (Johansson, [0164]).
Choi and Johansson teach the claimed limitations as stated above. Choi and Johansson do not explicitly teach the following underlined features: regarding claim 6, the QoE values are multi-scale structural similarity index measure (MS-SSIM) values.
However, Wang teaches the QoE values are multi-scale structural similarity index measure (MS-SSIM) values (Wang, [0030], “the fully-connected neural network 930 may be trained jointly by back-propagation of a loss function applied at the network output”, where the loss function may be defined based on mean squared error (MSE), SSIM, etc., and the SSIM including MS-SSIM (e.g., as discussed in “Multi-scale structural similarity for image quality assessment”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Choi and Johansson to have the features, as taught by Wang in order to combine the advantages of both data-driven (using DNN and machine learning) and knowledge-driven (based on HVS models, viewer device, viewing condition, video content, and video distortion analysis) approaches (Wang, [0020]).
Choi teaches the claimed limitations as stated above. Choi does not explicitly teach the following features: regarding claim 16, wherein:
the content is an image, and
the training QoE values and the outcome QoE values are multi-scale structural similarity index measure (MS-SSIM) values.
As to claim 16, Johansson teaches wherein:
the content is an image (Johansson, [0066], video services), and
the training QoE values and the outcome QoE values are measure values (Johansson, [0006], the Mean Opinion Score (MOS) is a score indicating the experience of listening voice samples by a number of listeners. Fig. 1, [0066]-[0067], the MOS/P.863 (learning data) 144 and ML MOS 150 are values determined for voice and video services).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Choi to have the features, as taught by Johansson in order to cope with the complexity of the inter-dependencies between network, codec and client parameters as well as their significance in impacting the QoE, which is better described with machine learning algorithms (Johansson, [0164]).
Choi and Johansson teach the claimed limitations as stated above. Choi and Johansson do not explicitly teach the following underlined features: regarding claim 16, the QoE values are multi-scale structural similarity index measure (MS-SSIM) values.
However, Wang teaches the QoE values are multi-scale structural similarity index measure (MS-SSIM) values (Wang, [0030], “the fully-connected neural network 930 may be trained jointly by back-propagation of a loss function applied at the network output”, where the loss function may be defined based on mean squared error (MSE), SSIM, etc., and the SSIM including MS-SSIM (e.g., as discussed in “Multi-scale structural similarity for image quality assessment”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Choi and Johansson to have the features, as taught by Wang in order to combine the advantages of both data-driven (using DNN and machine learning) and knowledge-driven (based on HVS models, viewer device, viewing condition, video content, and video distortion analysis) approaches (Wang, [0020]).
Claims 8 and 18 are rejected under 35 U.S.C. 103 as being unpatentable over Choi et al. (US 2020/0084102), hereinafter “Choi” in view of Johansson et al. (US 2021/0150393), hereinafter “Johansson” and further in view of Vannithamby et al. (US 2023/0199669), hereinafter “Vannithamby”.
Choi and Johansson teach the claimed limitations as stated above. Choi and Johansson do not explicitly teach the following features: regarding claim 8, wherein:
the network statistic features are a set of network statistic feature vectors, and
the processor is further configured to:
generate a network statistic feature vector from the network statistics that are determined from the traffic within a time period, for each time period of a consecutive set of time periods; and
combine the network statistic feature vectors from the consecutive set of time periods to form the set of network statistic feature vectors.
As to claim 8, Vannithamby teaches wherein:
the network statistic features are a set of network statistic feature vectors (Vannithamby, Fig. 9, [0104], [0176], the feature vectors obtained from the attributes. [0087]-[0088], [0104], [0144], the attributes include information indicating network characteristics, such as QoS requirements, QoE requirements, latency, RSRPs, etc.), and
the processor is further configured to:
generate a network statistic feature vector from the network statistics that are determined from the traffic (Vannithamby, Fig. 9, [0104], [0176], the feature vectors obtained from the attributes. [0087]-[0088], [0104], [0144], the attributes include information indicating network characteristics, such as QoS requirements, QoE requirements, latency, RSRPs, etc. for the running applications and signals received) within a time period (Vannithamby, Fig. 9, [0103]-[0104], the feature vectors are determined at an instance of time (903)), for each time period of a consecutive set of time periods (Vannithamby, Fig. 9, [0103]-[0104], the feature vectors are determined at an instance of time (903) during consecutive instances of time defined by a window size of 10 ms); and
combine the network statistic feature vectors from the consecutive set of time periods to form the set of network statistic feature vectors (Vannithamby, Fig. 9, [0104], [0176], the feature vectors obtained from the attributes. As shown in Fig. 9, the features vector (F#1-F#N) of each instance of time instance (t-10 to t-1) are combined as an input to the predictor 904. [0087]-[0088], [0104], [0144], the attributes include information indicating network characteristics, such as QoS requirements, QoE requirements, latency, RSRPs, etc.).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Choi and Johansson to have the features, as taught by Vannithamby in order to predict appropriate TX power levels for terminal devices in order to reduce frequent PHR feedback overhead over corresponding radio communication channels (Vannithamby, [0141]).
Choi and Johansson teach the claimed limitations as stated above. Choi and Johansson do not explicitly teach the following features: regarding claim 18, wherein:
the network statistic features are a set of network statistic feature vectors, and
the method further comprises:
generating a network statistic feature vector from the network statistics that are determined from the traffic within a time period, for each time period of a consecutive set of time periods; and
combining the network statistic feature vectors from the consecutive set of time periods to form the set of network statistic feature vectors.
As to claim 18, Vannithamby teaches wherein:
the network statistic features are a set of network statistic feature vectors (Vannithamby, Fig. 9, [0104], [0176], the feature vectors obtained from the attributes. [0087]-[0088], [0104], [0144], the attributes include information indicating network characteristics, such as QoS requirements, QoE requirements, latency, RSRPs, etc.), and
the method further comprises:
generating a network statistic feature vector from the network statistics that are determined from the traffic within a time period, for each time period of a consecutive set of time periods (Vannithamby, Fig. 9, [0104], [0176], the feature vectors obtained from the attributes. [0087]-[0088], [0104], [0144], the attributes include information indicating network characteristics, such as QoS requirements, QoE requirements, latency, RSRPs, etc. for the running applications and signals received) within a time period (Vannithamby, Fig. 9, [0103]-[0104], the feature vectors are determined at an instance of time (903)), for each time period of a consecutive set of time periods (Vannithamby, Fig. 9, [0103]-[0104], the feature vectors are determined at an instance of time (903) during consecutive instances of time defined by a window size of 10 ms); and
combining the network statistic feature vectors from the consecutive set of time periods to form the set of network statistic feature vectors (Vannithamby, Fig. 9, [0104], [0176], the feature vectors obtained from the attributes. As shown in Fig. 9, the features vector (F#1-F#N) of each instance of time instance (t-10 to t-1) are combined as an input to the predictor 904. [0087]-[0088], [0104], [0144], the attributes include information indicating network characteristics, such as QoS requirements, QoE requirements, latency, RSRPs, etc.).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Choi and Johansson to have the features, as taught by Vannithamby in order to predict appropriate TX power levels for terminal devices in order to reduce frequent PHR feedback overhead over corresponding radio communication channels (Vannithamby, [0141]).
Claims 9 and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Choi et al. (US 2020/0084102), hereinafter “Choi” in view of Johansson et al. (US 2021/0150393), hereinafter “Johansson” and further in view of Wang et al. (US 2021/0233259), hereinafter “Wang” and further in view of Lee et al. (US 2022/0366212), hereinafter “Lee”.
Choi teaches the claimed limitations as stated above. Choi does not explicitly teach the following features: regarding claim 9, wherein:
the trained ML model is a regression model,
the predicted outcome QoE values are in a range between 0 and 1, and
the processor is further configured to determine the new TWT parameters for the TWT operation based on an average of a number of the most recently obtained predicted outcome QoE values.
As to claim 9, Johansson teaches wherein:
the trained ML model is a regression model (Johansson, Fig. 1, Fig. 7, [0108], “other algorithms like Support Vector Regressor and neural networks will work as well, as the machine learning model 714”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Choi to have the features, as taught by Johansson in order to cope with the complexity of the inter-dependencies between network, codec and client parameters as well as their significance in impacting the QoE, which is better described with machine learning algorithms (Johansson, [0164]).
Choi and Johansson teach the claimed limitations as stated above. Choi and Johansson do not explicitly teach the following features: regarding claim 9, the predicted outcome QoE values are in a range between 0 and 1, and
the processor is further configured to determine the new TWT parameters for the TWT operation based on an average of a number of the most recently obtained predicted outcome QoE values.
However, Wang teaches the processor is further configured to determine the new TWT parameters for the TWT operation based on an average of a number of the most recently obtained predicted outcome QoE values (Wang, Figs. 1-2 and 9, [0030], “After training, the DNN may be applied to many 2D or 3D patches extracted from an image or video input using a pixel-by-pixel, frame-by-frame sliding window approach or with jumping steps, and the scores may be aggregated by averaging or weighted averaging to summarize the evaluation of the visual media input”. The DNN output scores of multiple patches are aggregated by averaging or weighted averaging. The teachings of Wang regarding averaging the output scores of the DNN in combination with Choi regarding the configuration of TWT based on determined score discloses the claimed feature).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Choi and Johansson to have the features, as taught by Wang in order to combine the advantages of both data-driven (using DNN and machine learning) and knowledge-driven (based on HVS models, viewer device, viewing condition, video content, and video distortion analysis) approaches (Wang, [0020]).
Choi, Johansson and Wang teach the claimed limitations as stated above. Choi, Johansson and Wang do not explicitly teach the following features: regarding claim 9, the predicted outcome QoE values are in a range between 0 and 1.
However, Lee teaches the predicted outcome QoE values are in a range between 0 and 1 (Lee, [0026], the last layer of the deep neural network model utilizes the sigmoid function which has an output value range from zero to one. Lee combined with the prediction of MOS (QoE) by Johansson discloses the claimed feature).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Choi, Johansson and Wang to have the features, as taught by Lee in order to generate a confidence level of the output generated by the deep neural network model based on the input of the deep neural network model (Lee, [0026]).
Choi teaches the claimed limitations as stated above. Choi does not explicitly teach the following features: regarding claim 19, wherein:
the trained ML model is a regression model,
the predicted outcome QoE values are in a range between 0 and 1, and
the method further comprises determining the new TWT parameters for the TWT operation based on an average of a number of the most recently obtained predicted outcome QoE values.
As to claim 19, Johansson teaches wherein:
the trained ML model is a regression model (Johansson, Fig. 1, Fig. 7, [0108], “other algorithms like Support Vector Regressor and neural networks will work as well, as the machine learning model 714”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Choi to have the features, as taught by Johansson in order to cope with the complexity of the inter-dependencies between network, codec and client parameters as well as their significance in impacting the QoE, which is better described with machine learning algorithms (Johansson, [0164]).
Choi and Johansson teach the claimed limitations as stated above. Choi and Johansson do not explicitly teach the following features: regarding claim 19, the predicted outcome QoE values are in a range between 0 and 1, and
the method further comprises determining the new TWT parameters for the TWT operation based on an average of a number of the most recently obtained predicted outcome QoE values.
However, Wang teaches the method further comprises determining the new TWT parameters for the TWT operation based on an average of a number of the most recently obtained predicted outcome QoE values (Wang, Figs. 1-2 and 9, [0030], “After training, the DNN may be applied to many 2D or 3D patches extracted from an image or video input using a pixel-by-pixel, frame-by-frame sliding window approach or with jumping steps, and the scores may be aggregated by averaging or weighted averaging to summarize the evaluation of the visual media input”. The DNN output scores of multiple patches are aggregated by averaging or weighted averaging. The teachings of Wang regarding averaging the output scores of the DNN in combination with Choi regarding the configuration of TWT based on determined score discloses the claimed feature).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Choi and Johansson to have the features, as taught by Wang in order to combine the advantages of both data-driven (using DNN and machine learning) and knowledge-driven (based on HVS models, viewer device, viewing condition, video content, and video distortion analysis) approaches (Wang, [0020]).
Choi, Johansson and Wang teach the claimed limitations as stated above. Choi, Johansson and Wang do not explicitly teach the following features: regarding claim 19, the predicted outcome QoE values are in a range between 0 and 1.
However, Lee teaches the predicted outcome QoE values are in a range between 0 and 1 (Lee, [0026], the last layer of the deep neural network model utilizes the sigmoid function which has an output value range from zero to one. Lee combined with the prediction of MOS (QoE) by Johansson discloses the claimed feature).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Choi, Johansson and Wang to have the features, as taught by Lee in order to generate a confidence level of the output generated by the deep neural network model based on the input of the deep neural network model (Lee, [0026]).
Claims 10 and 20 are rejected under 35 U.S.C. 103 as being unpatentable over Choi et al. (US 2020/0084102), hereinafter “Choi” in view of Johansson et al. (US 2021/0150393), hereinafter “Johansson” and further in view of Vannithamby et al. (US 2023/0199669), hereinafter “Vannithamby” and further in view of Gubanov (US 2022/0129486).
Choi and Johansson teach the claimed limitations as stated above. Choi and Johansson do not explicitly teach the following features: regarding claim 10 wherein:
the trained ML model is a binary classification model,
the predicted outcome QoE values are a binary indication of anomalous or not anomalous, and
the processor is further configured to determine the new TWT parameters for the TWT operation by applying a voting scheme to a number of the most recently obtained predicted outcome QoE values.
As to claim 10, Vasudevan teaches wherein:
the trained ML model is a binary classification model (Vasudevan, [0088], the traffic discriminator 320 is a binary classifier neural network that look at estimated application layer metrics, and make a prediction about the connection (e.g., whether traffic is anomalous or non-anomalous). The trained discriminator 320 is then used in standalone mode during production as the ML-based anomaly detector 174),
the predicted outcome QoE values are a binary indication of anomalous or not anomalous (Vasudevan, [0088], the traffic discriminator 320 is the binary classifier neural network and looks at estimated application layer metrics, and make a prediction about the connection (e.g., whether traffic is anomalous or non-anomalous). [0089], the trained discriminator 320 is deployed and classifies the connections as either non-anomalous or anomalous).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Choi and Johansson to have the features, as taught by Vasudevan in order to ensure robustness of the system and to improve accuracy of classification (Vasudevan, [0009]).
Choi, Johansson and Vasudevan teach the claimed limitations as stated above. Choi, Johansson and Vasudevan do not explicitly teach the following features: regarding claim 10, the processor is further configured to determine the new TWT parameters for the TWT operation by applying a voting scheme to a number of the most recently obtained predicted outcome QoE values.
However, Gubanov teaches the processor is further configured to determine the new TWT parameters for the TWT operation by applying a voting scheme to a number of the most recently obtained predicted outcome QoE values (Gubanov, [0032], [0037], the majority voting ensemble generates a final prediction based on the most frequently prediction. The teachings of Gubanov regarding the voting and most frequently prediction in combination with Choi regarding the configuration of TWT based on determined score discloses the claimed feature).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Choi and Johansson to have the features, as taught by Gubanov in order to provide a final prediction based on the most frequently prediction, which is a well know technique in the art (Gubanov, [0032]).
Choi and Johansson teach the claimed limitations as stated above. Choi and Johansson do not explicitly teach the following features: regarding claim 20, wherein:
the trained ML model is a binary classification model,
the predicted outcome QoE values are a binary indication of anomalous or not anomalous, and
the method further comprises determining the new TWT parameters for the TWT operation by applying a voting scheme to a number of the most recently obtained predicted outcome QoE values.
As to claim 20, Vasudevan teaches wherein:
the trained ML model is a binary classification model (Vasudevan, [0088], the traffic discriminator 320 is a binary classifier neural network that look at estimated application layer metrics, and make a prediction about the connection (e.g., whether traffic is anomalous or non-anomalous). The trained discriminator 320 is then used in standalone mode during production as the ML-based anomaly detector 174),
the predicted outcome QoE values are a binary indication of anomalous or not anomalous (Vasudevan, [0088], the traffic discriminator 320 is the binary classifier neural network and looks at estimated application layer metrics, and make a prediction about the connection (e.g., whether traffic is anomalous or non-anomalous). [0089], the trained discriminator 320 is deployed and classifies the connections as either non-anomalous or anomalous).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Choi and Johansson to have the features, as taught by Vasudevan in order to ensure robustness of the system and to improve accuracy of classification (Vasudevan, [0009]).
Choi, Johansson and Vasudevan teach the claimed limitations as stated above. Choi, Johansson and Vasudevan do not explicitly teach the following features: regarding claim 20, the method further comprises determining the new TWT parameters for the TWT operation by applying a voting scheme to a number of the most recently obtained predicted outcome QoE values.
However, Gubanov teaches the method further comprises determining the new TWT parameters for the TWT operation by applying a voting scheme to a number of the most recently obtained predicted outcome QoE values (Gubanov, [0032], [0037], the majority voting ensemble generates a final prediction based on the most frequently prediction. The teachings of Gubanov regarding the voting and most frequently prediction in combination with Choi regarding the configuration of TWT based on determined score discloses the claimed feature).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the invention of Choi and Johansson to have the features, as taught by Gubanov in order to provide a final prediction based on the most frequently prediction, which is a well know technique in the art (Gubanov, [0032]).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure:
Grois et al. U.S. Patent Application Publication no. 2023/0088688 – Methods, systems, and apparatuses for adaptive bitrate ladder construction based on dynamically adjustable neural networks.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to RICARDO H CASTANEYRA whose telephone number is (571)272-2486. The examiner can normally be reached M-F 9:00am - 5:30pm.
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, Kwang bin Yao can be reached at 571-272-3182. 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.
/RICARDO H CASTANEYRA/Primary Examiner, Art Unit 2473