# US 12,015,479 B2 · Method and apparatus to decode packets to compute log likelihood ratio in wireless network > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,015,479 B2 - Document ID: US12015479B2 - Status: Granted - Kind: B2 - Application number: 17/747,485 - Filed: 2022-05-18 - Published: 2022-09-01 - Granted: 2024-06-18 - Earliest claimed: 2019-11-22 - Assignees: Samsung Electronics Co., Ltd. - Applicants: Samsung Electronics Co., Ltd. - Inventors: Satya Kumar Vankayala; Seungil Yoon; Issaac Kommineni; Venkateswarlu Yarramala - Web page: https://patentgrep.com/patent/US12015479B2 ## CPC - H04L 1/0026: {Transmission of channel quality indication} - H04L 25/067: {providing soft decisions, i.e. decisions together with an estimate of reliability (H04L25/068 and H04L25/069 take precedence; sequence estimation techniques H04L25/03178)} - G06F 18/214: Generating training patterns; Bootstrap methods, e.g. bagging or boosting - G06N 3/04: Architecture, e.g. interconnection topology - G06N 3/044: Recurrent networks, e.g. Hopfield networks - G06N 3/045: Combinations of networks - G06N 3/0495: Quantised networks; Sparse networks; Compressed networks - G06N 3/0499: Feedforward networks - G06N 3/08: Learning methods - G06N 3/082: modifying the architecture, e.g. adding, deleting or silencing nodes or connections - G06N 3/09: Supervised learning - G06N 3/098: Distributed learning, e.g. federated learning - H04L 1/0017: {where the mode-switching is based on Quality of Service requirement} - H04L 1/005: {Iterative decoding, including iteration between signal detection and decoding operation} - G06N 3/048: Activation functions - H04B 7/0413: MIMO systems - H04L 1/0045: {Arrangements at the receiver end} - H04L 27/38: Demodulator circuits; Receiver circuits - H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M) - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS - H04B: TRANSMISSION ## Abstract A method for decoding user data by an apparatus in a wireless network is provided. The method includes receiving, by the apparatus, user data associated with a plurality of network parameters. The method includes training, by the apparatus, a neural network using the plurality of received network parameters. Further, the method includes computing, by the apparatus, a Log Likelihood Ratio (LLR) using the trained neural network. Further, the method includes decoding, by the apparatus, the received user data using the computed LLR. ## Claims ### Claim 1 A method for decoding user data by an apparatus in a wireless network, comprising: segregating, by the apparatus, in-phase data and quadrature data (IQ data) streams, based on quadrature amplitude modulation (QAM) scheme of each of the IQ data streams; determining, by the apparatus, a channel condition associated with each of the QAM schemes of the segregated IQ data streams; receiving, by the apparatus, user data associated with a plurality of network parameters, based on the segregated IQ data streams and the determined channel conditions; training, by the apparatus, a neural network using the plurality of network parameters; computing, by the apparatus, a log likelihood ratio (LLR) using the trained neural network; transferring, by the apparatus, the computed LLR to a physical layer; and decoding, by the apparatus, the received user data using the computed LLR. ### Claim 2 (depends on 1) The method of claim 1, wherein the plurality of network parameters comprises at least one of: a Minimum Mean Square Error (MMSE) value, channel gains, a modulation scheme, Channel Quality Indicator (CQI) information, a number of a User Equipment (UE) in range, a location of a user equipment (UE), a distance of the UE from the apparatus, a weather condition, resource block (RB) information, an operating frequency, an RB bandwidth, a Quality of Service (QoS), a QoS Class Identifier (QCI), a bandwidth part (BWP), a Subcarrier Spacing (SCS), a coherence bandwidth (BW), a coherence time, a coherence interference, a noise, an operating frequency, a capability of the UE, a Multiple Input Multiple Output (MIMO) capability, a transmission mode, real time traffic data, remote radio head (RRH) base station (BS) capabilities, in-phase data and quadrature data (I and Q) values of the UE, Quadrature Amplitude Modulation (QAM) modulation details, a size of in-phase data and quadrature data (I and Q) vector, allocated Resource Blocks to the UE, a density of traffic associated with the BS, a distribution of traffic associated with the BS, a category of the BS, climate conditions of present day, information of special occasions on the present day associated with an area of the UE, an event based calendar of the UE, vacation details of a user associated with the UE, a category of the UE, or subscription details of the UE. ### Claim 3 (depends on 1) The method of claim 1, wherein decoding, by the apparatus, of the received user data using the computed LLR comprises performing one of: receiving, by the apparatus, a size of a user data vector, Quadrature Amplitude Modulation (QAM) information, Multiple Input Multiple Output (MIMO) rank details, and a number of resource blocks (RBs) assigned to the user data sequentially, and serially decoding, by the apparatus, the received user data on a per packet basis using the computed LLR based on the size of the user data vector, the QAM information, the MIMO rank details, and the number of RBs assigned to the user data sequentially, receiving, by the apparatus, the QAM information, a size of in-phase data and quadrature data (IQ data) of each user data, and allocated RBs to each user data, and parallel decoding, by the apparatus, the received user data per symbol basic using the computed LLR based on the QAM information, the size of IQ data of each user data, and allocated RBs to each user data, or providing, by the apparatus, the user data until all IQ data is decoded sequentially or in parallel, assigning, by the apparatus, a size of an input vector as a function of the number of RBs to the user data, loading, by the apparatus, a weight function using a lookup table in the neural network, and sequentially decoding, by the apparatus, the user data using the computed LLR, the size of the input vector and the weight function, and wherein the LLR is computed for at least one of all packets, each bit serially, all user data, one bit at a time, or one bit at a time and for all user data at once. ### Claim 4 (depends on 1) The method of claim 1, wherein the neural network is trained by: creating multiple batches of training samples, each training sample using at least one modulation scheme, training Quadrature Amplitude Modulation (QAM) schemes, wherein each QAM scheme comprises a unique code embedded in the user data, generating shuffled training samples for each batch from multiple batches, computing an LLR corresponding to each shuffled training sample, and shuffling the computed LLR across the batches while training. ### Claim 5 (depends on 1) The method of claim 1, wherein the apparatus comprises at least one of a base station (BS), a User Equipment (UE), an Open Radio Access Network (ORAN), a centralized Radio Access Network (CRAN), or a virtual Radio Access Network (VRAN). ### Claim 6 (depends on 1) The method of claim 1, wherein the computing, by the apparatus, of the LLR comprises: receiving at least one of a number of resource blocks (RBs), an operating frequency, a modulation and coding scheme (MCS), a number of independent streams, Multiple Input Multiple Output (MIMO) details, layer information associated with the wireless network, layer information associated with a user equipment (UE), a sub-set of network parameters, or a signal to interference and noise ratio (SINR), and sequentially or parallel computing the LLR based on the at least one of the number of RBs, the operating frequency, the MCS, the number of independent streams, the MIMO details, the layer information associated with the wireless network, the layer information associated with the UE, the sub-set of network parameters, or the SINR. ### Claim 7 (depends on 1) The method of claim 1, wherein the neural network comprises an input layer, one or more hidden layers each comprising a plurality of nodes, and an output layer. ### Claim 8 (depends on 7) The method of claim 7, wherein a size of the output layer is log 2 (M) for an M Quadrature Amplitude Modulation (QAM) scheme. ### Claim 9 (depends on 7) The method of claim 7, further comprising using a piecewise approximated Swish activation function, f(x), after the one or more hidden layers, where f ⁡ ( x ) = { 0 for ⁢ x ≤ - 2 1 4 ⁢ ( x + 2 ) ⁢ x for ⁢ ❘ "\[LeftBracketingBar]" x ❘ "\[RightBracketingBar]" < 2 x for ⁢ x ≥ - 2 . ### Claim 10 An apparatus for decoding user data in a wireless network, the apparatus comprising: memory; and at least one processor coupled with the memory, wherein the memory store one or more computer programs including computer-executable instructions that, when executed by the at least one processor, cause the apparatus to: segregate in-phase data and quadrature data (IQ data) streams based on quadrature amplitude modulation (QAM) scheme of each of the IQ data streams, determine a channel condition associated with each of the QAM schemes of the segregated IQ data streams, receive the user data associated with a plurality of network parameters, based on the segregated IQ data streams and the determined channel conditions, train a neural network using the plurality of network parameters, compute a log likelihood ratio (LLR) using the trained neural network, transfer the computed LLR to a physical layer, and decode the received user data using the computed LLR. ### Claim 11 (depends on 10) The apparatus of claim 10, wherein the plurality of network parameters comprises at least one of: a Minimum Mean Square Error (MMSE) value, channel gains, a modulation scheme, Channel Quality Indicator (CQI) information, a number of a User Equipment (UE) in range, a location of a UE, a distance of the UE from the apparatus, a weather condition, resource block (RB) information, an operating frequency, an RB bandwidth, a Quality of Service (QoS), a QoS Class Identifier (QCI), a bandwidth part (BWP), a Subcarrier Spacing (SCS), a coherence bandwidth (BW), a coherence time, a coherence interference, noise, an operating frequency, a capability of the UE, a Multiple Input Multiple Output (MIMO) capability, a transmission mode, a real time traffic data, remote radio head (RRH) base station (BS) capabilities, in-phase data and quadrature data (I and Q) values of the UE, Quadrature Amplitude Modulation (QAM) modulation details, a size of in-phase data and quadrature data (I and Q) vector, allocated Resource Blocks (RBs) to a user, a density of traffic associated with the BS, a distribution of traffic associated with the BS, a category of the BS, climate conditions of present day, information of special occasions on the present day associated with an area of the UE, an event based calendar of the UE, vacation details of a user associated with the UE, a category of the UE, or subscription details of the UE. ### Claim 12 (depends on 10) The apparatus of claim 10, wherein the one or more computer programs further comprise computer-executable instructions that, when executed by the at least one processor, cause the apparatus to perform at least one of: receive a size of a user data vector, Quadrature Amplitude Modulation (QAM) information, Multiple Input Multiple Output (MIMO) rank details, or a number of resource blocks (RBs) assigned to the user data sequentially; and serially decoding the received user data on a per packet basis using the computed LLR based on the size of the user data vector, the QAM information, and the number of RBs assigned to the user data sequentially, receive QAM information, a size of in-phase data and quadrature data (IQ data) of each user data, and allocated RBs to each user data; and parallel decoding the received user data on a per symbol basis using the computed LLR based on the QAM information, the size of the IQ data of each user data, and the allocated RBs to each user data, or provide the user data until all IQ data is decoded sequentially, assigning a size of an input vector as a function of the number of RBs assigned to the user data, loading a weight function using a lookup table in the neural network, and sequentially decoding the user data using the computed LLR, the size of the input vector, and the weight function, and wherein the LLR is computed for at least one of all packets, each bit serially, all user data, one bit at a time, or one bit at a time and for all user data at once. ### Claim 13 (depends on 10) The apparatus of claim 10, wherein the neural network is trained by: creating multiple batches of training samples, each training sample using at least one modulation scheme, training Quadrature Amplitude Modulation (QAM) schemes, each QAM scheme comprising a unique code embedded in the user data, generating shuffled training samples for each batch from multiple batches, calculating an LLR corresponding to each shuffled training sample, and shuffling the calculated LLR across the batches while training. ### Claim 14 (depends on 10) The apparatus of claim 10, wherein the apparatus comprises at least one of a base station (BS), a User Equipment (UE), an Open Radio Access Network (ORAN), a Centralized Radio Access Network (CRAN), or a Virtual Radio Access Network (VRAN). ### Claim 15 (depends on 10) The apparatus of claim 10, wherein the neural network is implemented in a radio access network (RAN) intelligent controller (RIC) module of at least one of an Open Radio Access Network (O-RAN) or a Virtual Radio Access Network (VRAN). ### Claim 16 (depends on 10) The apparatus of claim 10, computing the LLR comprises: receiving at least one of a number of resource blocks (RBs), an operating frequency, a modulation and coding scheme (MCS), a number of independent streams, or a signal to interference and noise ratio (SINR), and sequentially or parallel computing the LLR based on at least one of the number of RBs, the operating frequency, the MCS, the number of independent streams, Multiple Input Multiple Output (MIMO) details, a layer information associated with the wireless network, a layer information associated with a user equipment (UE), a sub-set of network parameters, or the SINR.