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 .
Applicant cancelled claims 1-36 prior to the first office action.
Applicant added new claims 37-56 prior to the first office action.
Status of claims:
Claims 37-56 are pending in this office action.
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.
Claim(s) 37, 38, 40, 41, 44-50, 52-56 is/are rejected under 35 U.S.C. 102(a1) as being anticipated by Timo et al. (WO2020/180221) (hereinafter D1)
Per claim 37, D1 discloses a method for a radio access network, RAN, node, (see Section 6.1 "network node") the method comprising :transmitting downlink, DL, reference signals, RS, in accordance with one or more configurations (see Section 6.2.6 "each configured CSI-RS resource"); receiving, from each of one or more user equipment, UEs, feedback representing a DL channel from the RAN node to the UE, wherein the feedback is based on the UE's measurements of the transmitted DL RS (see Fig. 15, (1503) and Section 6.2.6 "downlink channel estimates [ ... ] "derived from each configured CSI-RS resource", wherein the Z compressed versions of the downlink estimates X are "transmitted to the BS", see Section 6.2.6) and is encoded by one or more UE encoders that correspond to respective RAN node decoders (see Fig. 15 and Section 6.2.6 wherein the downlink channel estimates X are encoded into Z using encoders corresponding to respective decoders of the network node in the autoencoder framework)
for each of the one or more UEs, obtaining one or more measurements of uplink, UL, RS transmitted by the UE in accordance with the one or more configurations (see Section 6.3 "Embodiment 13" disclosing measurements at the network node based on uplink SRS, which are necessarily configured); and training the one or more RAN node decoders based on the received feedback and the obtained UL RS measurements (see Section 6.2.10 disclosing the training of the decoder of the network node disclosing the training of the autoencoder, i.e. comprising the network node decoders according to
"Embodiment 1" and Section 6.3 "Embodiment 3" and "Embodiment 13" both dependent on "Embodiment 1" disclosing the training of the decoders based on the CSI-RS measurements transmitted by the UE and based on measured uplink SRS sent by the UE).
Per claim 38, D1 discloses the method of claim 37, further comprising sending the one or more configurations to the one or more UEs, wherein each configuration includes the following: a first configuration for UE measurement of DL RS transmitted by the RAN node, a second configuration for UE transmission of UL RS, and an identifier of one of the UE encoders (, examiner interprets specific encoder/decoder selection and configuration based on the UE type and capabilities, see Sections 6.1.2, 6.2.1 and 6.2. 3, The CSC encoders can be used by the UEs to compress their raw “ explicit ” channel matrices or tx-channel covariance matrices for transmission over the uplink. The CSC decoders can be used by the BSs to decompress the UEs’ channel estimates. The cost functions can be used by the BS and/or UEs to measure the performance of a given compressor-decompressor pair. The CSC parameters (a*, b*) can be tailored by the BS to, for example, match important attributes of the UE and the cell. For example, the parameters (a*, b*) can be tailored by the BS to match important factors in the propagation environment).
Per claim 40, refer to similar language as explained in in claim 37, however using the term estimate one more UL channels, see D1,page 10 lines 17-20, The terminal device is
configured to receive a first set of parameters ,form a compression function based on the first set of parameters, compress downlink channel estimates using the compression function, and transmit the compressed downlink channel estimates, examiner interprets one or more RAN nodes as network nodes as 501 and 501b, page 44 line 25-27, Fig 16).
Per claim 41, D1 discloses the method of claim 40, wherein :each second configuration identifies a plurality of symbols of a single timeslot for UE transmission of UL RS (page 1, lines 18-26, i.e. Essentially, an antenna port is defined by the reference signal transmitted from the antenna port, where the reference signal can be mapped to the antenna elements (precoded) in an arbitrary fashion. For instance, DMRS antenna ports are precoded in the same way as PDSCH symbols and are
typically mapped to the entire antenna while CSI-RS antenna ports are typically either mapped to individual antenna elements or subarrays or are alternatively also precoded in the similar fashion as PDSCH. It is common practice to partition an antenna array into subarrays of physical antenna elements, where only a single transmit-receive unit (TXRU) is connected to each subarray) ; for each of the one or more UEs, obtaining one or more measurements of UL RS transmitted by the UE (page 37,lines 18-23, i.e. X = represents the (Xt) vector of complex CSI-RS measurements taken by the UE. That is, Xt denotes the UE's estimate 1503 of the lth CSI-RS resource on its antenna port. Alternatively, one can make X a (2Ntx X 1) vector of real-valued CSI-RS measurements by splitting each complex measurement into its real and imaginary parts, also see page 38, lines 10-13, i.e. The (Ncm X 1)-complex output Z (indicated by 1512 in Fig. 15) of this non-linear subfunction 1511 is then quantized 1513 to generate the quantized compressed measurement 1514 that is sent to the BS over the uplink represented by a finite set of bits) comprises receiving the UL RS transmitted by the UE during the plurality of symbols using a respective plurality of different RAN node antennas(page 1 lines 13-16, The input to each antenna port is a sequence of complex-valued modulation symbols e.g., QPSK, 16QAM, 64QAM or 256QAM). Orthogonal frequency division multiplexing (OFDM) is used in LTE/5G networks to encode these symbols onto many orthogonal subcarriers for transmission also see page 1 lines 18-26 for different antenna ports is related to a different RAN node).; obtaining estimates of one or more UL channels from the UE to the RAN node comprises obtaining estimates of a plurality of UL channels corresponding to the respective plurality of different RAN node antennas(page 7 lines 14-17,i.e. instead of feeding-back raw channel estimates over the uplink, the UE can send averages of these channel estimates over several subbands (or resource blocks). For example, in an example scenario with subbands, it is possible to reduce the uplink overhead by a factor of 9 by averaging the channels over all 9 subbands; and training the one or more RAN node decoders is based on the plurality of UL channel estimates (page 31, lines 1-15, i.e. The 3GPP standard can be modified to define a class of channel state compression (CSC) encoders (FA) a matching class of CSC decoders(Gb) and a class cost function d, The CSC encoders can be used by the UEs to compress their raw "explicit" channel matrices or tx-channel covariance matrices for transmission over the uplink. The CSC decoders can be used by the BSs to decompress the UEs' channel estimates. The cost functions can be used by the BS and/or UEs to measure the performance of a given compressor-decompressor pair, also see page 19 lines 9-17, regarding training process of decompression or decoders in the second function).
Per claim 44, D1 discloses the method of claim 37, wherein training the one or more RAN node decoders based on the received feedback and the obtained UL RS measurements (refer to claim 37 rationale) comprises, for each of the one or more UEs: determining an importance weight based on a duration between transmitting the DL RS and obtaining the one or more measurements of UL RS transmitted by the UE( page 33 lines 20-31 and page 40, 6.2.9 initial CSC parameters, The CSC parameters (a*,b*) can be tailored by the BS to, for example, match important attributes of the UE and the cell. For example, the parameters (a*,b*) can be tailored by the BS to match important factors in the propagation environment and UE-specific: For example, (a*,b*) can be chosen specifically to exploit
properties of the UE's antenna array (e.g., antenna coherence, spacing and layout); and the computational abilities of the UE. For example, a specific parameter set can be specified for a given UE model, . The CSC parameters a* and b* are determined by the BS and uniquely specify the initial encoder/decoder weights and biases from the parameter spaces A and B)
;and adjusting, in proportion to the importance weight, effect on the training of the UE's feedback representing the DL channel and of the measurements of UL RS transmitted by the UE(6.2.10 updating CSC parameters, page 40, the encoder/decoder weights and encoder/decoder biases can be updated as described in Section 6.2.5 above. The UEs' channel measurements without normalization can differ by several orders of magnitude (in linear scale). To partially combat this dramatic variation, we have normalized the CSI-RS measurements by the corresponding (wideband) L1-RSRP.The L1-RSRP is measured by the UE for each CSI report and may in some embodiments be reported to the BS. Thus, the normalized values can be used for both training the autoencoder-based CSC and during live operation also see page 36 lines 20-22, UE-initiated CSC parameter updates: The BS can configure the UE to
recommend updates to one-or-more of CSC parameters. The time-domain behavior of these updates can be configured to be periodic or aperiodic).
Per claim 45, D1 discloses the method of claim 37, wherein one or more of the following applies :the one or more UE encoders are not trained during the training of the corresponding one or more RAN node decoders; the one or more UE encoders are fixed or configured as non-trainable; and the one or more UE encoders and the corresponding one or more RAN node decoders are part of respective one or more neural network autoencoders (NNAEs), see page 36 lines 30 and 31and page 37, lines 1, 2 and 7-10, i.e. , Fig. 15 shows a compression function 1501 and a decompression function 1502 provided in the form of a neural network 1500, according an embodiment. In the present embodiment, compression and decompression is based on an autoencoder framework. The class of encoders (which act as compression functions) and the decoders {Fa} (which act as decompression functions) can be defined by the autoencoder structure in Fig. 15) .
Per claim 46, D1 discloses the method of claim 37, further comprising, using a trained RAN node decoder, performing one or more of the following based on further feedback, from one of the UEs, that represents a DL channel from the RAN node to the UE and that is encoded by a UE encoder corresponding to the trained RAN node decoder(refer to claim 37 rationale) :determining one or more beam directions for the DL channel, and estimating an UL channel from the UE to the RAN node (page 4, lines 23-26, i.e. Time division duplex (TDD) network deployments in which uplink-downlink channel reciprocity holds are an ideal candidate for 5G systems and MIMO precoding/beamforming. Uplink-downlink channel reciprocity implies that the BS can directly estimate the downlink channels from uplink reference signals (e.g., SRS)) .
Per claim 47, D1 discloses the method of claim 37, wherein: the one or more UE encoders include a plurality of UE encoders corresponding to a plurality of RAN node decoders (page 30, line 4-8, i.e. The CSC encoders can be used by the UEs to compress their raw "explicit" channel matrices or tx-channel covariance matrices for transmission over the uplink. The CSC decoders can be used by the BSs to decompress the UEs' channel estimates) ; the received feedback encoded by each UE encoder includes a different number of bits than the received feedback encoded by other UE encoders(page 31, i.e. The integer M defines how many bits (i.e., [log2 (M)]) are fed back over the uplink. The subscripts a E A and b E B represent parameters that define the encoder fa and decoder gb respectively [and, therefore, the integers N and M], a is a first set of parameters, and b is a second set of parameters) ; and one of the following applies: the transmitting, receiving, obtaining, and training operations are performed for each of the RAN node decoders in sequence; or transmitting the DL RS and obtaining the measurements of the UL RS transmitted by the UE are performed once for the plurality of RAN node decoders, wherein the received feedback is encoded by the plurality of UE encoders based on the UE's measurements of the DL RS.(page 35, i.e., which the UE reports the compressed channel estimates fa(cn,cn,,,,,Cn) using PUCCH and/or PUSCH. If configured for Type-I or Type-II CSI feedback, the ReportConfig contains a CodebookConfig that specifies configuration parameters for Type-1 and Type-II CSI feedback. To enable the proposed CSC framework, the CodebookConfig can be modified to specify the encoding and decoding parameters [a*, b*]), examiner interprets CSI-RS includes measurements or L1-RSRP which is measured by the UE).
Per claim 48, refer to same rationale as explained in claim 37, however from the UE perspective.
Per claim 49, refer to same rationale as explained in claim 37 and 38, transmitting further feedback includes sending one or more configurations to the UE and responding by transmitting the configuration to the network node.
Per claim 50, refer to the same rationale as explained in claim 38.
Per claim 52, refer to the same rationale as explained in claim 41.
Per claim 53, refer to same rationale as explained in claim 45.
Per claim 54, refer to same rationale as explained in claim 47.
Per claim 55, refer to same rationale as explained in claim 37.
Per claim 56, refer to same rationale as explained in claim 48.
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.
Claim(s) 39 , 42, 43 and 51 is/are rejected under 35 U.S.C. 103 as being unpatentable over D1 in view of Machine Learning Enhanced CSI Acquisition and Training Strategy for FDD Massive MIMO(Nokia Bell Labs China) (hereinafter D2).
Per claim 39, D1 discloses the method of claim 38, further comprising but fails to explicitly disclose receiving, from each of the one or more UEs, an indication of UE DL channel feedback encoding capabilities, wherein one of the following applies: the indication comprises at least one identifier of an encoder supported by the UE, including the one or more identifiers included in the respective configurations; or encoders supported by the UE are indicated by one or more of the following: UE vendor, UE model, UE chipset vendor, and UE chipset model.
In an analogous field of endeavor, D2 discloses receiving, from each of the one or more UEs, an indication of UE DL channel feedback encoding capabilities, wherein one of the following applies: the indication comprises at least one identifier of an encoder supported by the UE, including the one or more identifiers included in the respective configurations; or encoders supported by the UE are indicated by one or more of the following: UE vendor, UE model, UE chipset vendor, and UE chipset model. (Section 111.B, disclosing training based on an dataset with suitable characteristics, [The NN is used to recover the downlink CSI from the processed downlink CSI feedback. To train the NN, the BS estimates the accurate uplink CSI from the uplink RS and generates the processed uplink CSI. This requires analogous functions at both the BS and the UE to process CSI counterparts. Both the accurate uplink CSI and the processed uplink CSI are considered as the whole training input to the NN] ).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing
date of the claimed invention to have incorporated the teachings of D2 into the invention D1, where D1 provides A network node (501) determines parameters (503) indicating a compression function for compressing downlink channel estimates, and a decompression function. The network node transmits the parameters, receives compressed downlink channel estimates and decompresses the compressed downlink channel estimates using the decompression function and D2 provides Abstract-Massive Multiple Input Multiple Output (MIMO) is able to boost the system throughput. A key challenge is a large overhead of Channel State Information (CSI) feedback with the increased number of antenna ports in Frequency Division
Duplexing (FDD) massive MIMO systems in order to provide a better quality of service where Neutral networks can be trained using the data with same statistics and characteristics where simulation results provide good performance when FDD reciprocity doesn’t hold.
Per claim 42, D1 discloses The method of claim 37, further comprising: receiving, from a second RAN node (network Node 501b or 501) and plurality of decoders of UE DL feedback (refer to claim 37 rationale), but fails to explicitly disclose a training dataset; and selectively training based the received training dataset.
In an analogous field of endeavor, D2 discloses a training dataset; and selectively training based the received training dataset.. (Section II1.B, disclosing training based on an dataset with suitable characteristics, [The NN is used to recover the downlink CSI from the processed downlink CSI feedback. To train the NN, the BS estimates the accurate uplink CSI from the uplink RS and generates the processed uplink CSI. This requires analogous functions at both the BS and the UE to process CSI counterparts. Both the accurate uplink CSI and the processed uplink CSI are considered as the whole training input to the NN], examiner interprets that the neutral node (NN) can be second RAN or NN mode 1, also page 1 in the introduction, states “ Existing work have shown the potential of using ML to reduce the CSI feedback overhead, where is purely fully-connected layer based and requires MI. functions at UE due to the auto-encoder like Neural Network (NN) architecture[Fig 3]), refer to claim 39 motivation..
Per claim 43, refer to the same rationale as explained in claim 42, see claim limitation in claim 43 “training the RAN node decoders only based on portions of the received training dataset that are associated with channel characteristics similar to channel characteristics measured by the RAN node” see D2, Section II1.B, disclosing training based on an dataset with suitable characteristics, [The NN is used to recover the downlink CSI from the processed downlink CSI feedback. To train the NN, the BS estimates the accurate uplink CSI from the uplink RS and generates the processed uplink CSI. This requires analogous functions at both the BS and the UE to process CSI counterparts. Both the accurate uplink CSI and the processed uplink CSI are considered as the whole training input to the NN].
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing
date of the claimed invention to have incorporated the teachings of D2 into the invention D1 in order to Neutral networks can be trained using the data with same statistics and characteristics where simulation results provide good performance which requires analogous functions at both the BS and the UE to process CSI counterparts. Both the accurate uplink CSI and the processed uplink CSI are considered as the whole training input to the NN], when FDD reciprocity doesn’t hold.
Per claim 51 refer to same rationale as explained in claim 39.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JOSEPH E DEAN, JR whose telephone number is (571)270-7116. The examiner can normally be reached Mon-Fri 7:30-3:30.
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, Alison Slater can be reached at 571-270-0375. 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.
/JOSEPH E DEAN, JR/Primary Examiner, Art Unit 2647
Prosecution Insights