# US 2020/0403723 A1 · Communication system having a configurable modulation order and an associated method and apparatus

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2020/0403723 A1
- Document ID: US20200403723A1
- Status: Published application
- Kind: A1
- Application number: 16/905,569
- Filed: 2020-06-18
- Published: 2020-12-24
- Earliest claimed: 2019-06-20
- Assignees: Nokia Technologies Oy
- Applicants: Nokia Technologies Oy
- Inventors: Jakob HOYDIS; Faycal AIT AOUDIA; Mathieu GOUTAY
- Web page: https://patentgrep.com/patent/US20200403723A1

## CPC

- H04L 1/0003: {by switching between different modulation schemes}
- G06N 3/02: Neural networks
- G06N 3/0455: Auto-encoder networks; Encoder-decoder networks
- G06N 3/0499: Feedforward networks
- G06N 3/08: Learning methods
- G06N 3/09: Supervised learning
- H04L 1/0009: {by adapting the channel coding (H04L1/1812 takes precedence)}
- H04L 1/0026: {Transmission of channel quality indication}
- H04L 1/0036: {arrangements specific to the receiver}
- H04L 5/0044: {Allocation of payload; Allocation of data channels, e.g. PDSCH or PUSCH}
- H04L 27/0008: {arrangements for allowing a transmitter or receiver to use more than one type of modulation (negotiating modulation type for two-way transmission paths H04L5/1453)}
- H04L 27/0012: {arrangements for identifying the type of modulation}
- G06N 3/044: Recurrent networks, e.g. Hopfield networks
- G06N 3/045: Combinations of networks
- G06N 3/048: Activation functions
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

A method, apparatus, receiver and system provide for configurability of the modulation order of a communications system, such as an autoencoder-based communication system. With respect to a system including a transmitter and a receiver, the transmitter encodes a message into a vector of channel symbols for transmission via a channel. The message is encoded pursuant to a modulation order m that is adjustable up to a maximum modulation order Mmax. The receiver receives a vector of samples generated by the channel and determines a first prediction vector for the message encoded by the transmitter. The receiver includes a slicing layer to eliminate one or more elements of the first prediction vector if the modulation order m is less than the maximum modulation order Mmax so as to generate a second prediction vector tailored to the modulation order from which a prediction of the message encoded by the transmitter is identified.

## Claims

### Claim 1

A system comprising:
a transmitter configured to encode a message into a vector of channel symbols for transmission via a channel, wherein the transmitter is configured to encode the message in accordance with a modulation order m that is adjustable up to a maximum modulation order Mmax; and
a receiver configured to receive a vector of samples generated by the channel in response to the vector of channel symbols transmitted by the transmitter, wherein the receiver is configured to determine a first prediction vector for the message encoded by the transmitter, and wherein the receiver comprises a slicing layer configured to eliminate one or more elements of the first prediction vector in an instance in which the modulation order m is less than the maximum modulation order Mmax so as to generate a second prediction vector tailored to the modulation order from which a prediction of the message encoded by the transmitter is identified.

### Claim 2 (depends on 1)

A system according to claim 1 wherein the receiver is further configured to generate a probability vector associated with the modulation order m, wherein the probability vector assigns probabilities for m-possible messages based upon the second prediction vector, wherein the system further comprises at least one processor configured to determine one or more updated parameters of the receiver and of the transmitter, wherein the updated parameters of the receiver and of the transmitter are based on the probability vector for each of a plurality of modulation orders, and wherein the transmitter and receiver are reconfigured based upon the one or more updated parameters of the transmitter and receiver, respectively.

### Claim 3 (depends on 2)

A system according to claim 2 wherein the receiver further comprises an activation function configured to normalize the second prediction vector into the probability vector representing a probability distribution having m probabilities.

### Claim 4 (depends on 2)

A system according to claim 2 wherein at least one of the transmitter or the receiver comprises a neural network having a plurality of trainable parameters that are updated by determinations of the at least one processor.

### Claim 5 (depends on 1)

A system according to claim 1 wherein the transmitter is configured to encode a message and the receiver is configured to predict the message encoded by the transmitter in accordance with a plurality of modulation orders, and wherein a respective modulation order defines a number of different messages capable of being encoded by the transmitter and predicted by the receiver.

### Claim 6 (depends on 5)

A system according to claim 5 wherein the transmitter is configured to encode the message as an Mmax-dimensional vector having an element that corresponds to the message being encoded set to a first value and all other elements set to a second value, different than the first value.

### Claim 7 (depends on 6)

A system according to claim 6 wherein the transmitter is configured to encode a message and the receiver is configured to predict the message encoded by the transmitter in accordance with a respective modulation order m, wherein only elements 1, 2, . . . m of the Mmax-dimensional vector are able to be set to the first value, and wherein elements m+1, m+2, . . . Mmax of the Mmax-dimensional vector are set to the second value.

### Claim 8 (depends on 1)

A system according to claim 1 wherein the slicing layer of the receiver is configured to eliminate elements of the first prediction vector such that the second prediction vector has a number of elements that equals the modulation order.

### Claim 9 (depends on 1)

A system according to claim 1 wherein the slicing layer of the receiver has a variable width that is dependent upon the modulation order, wherein the width of the slicing layer defines a number of elements of the second prediction vector.

### Claim 10 (depends on 1)

A system according to claim 1 wherein the transmitter is further configured to change the modulation order based upon feedback provided by the receiver.

### Claim 11 (depends on 1)

A system according to claim 1 wherein the transmitter is further configured to provide a message to the receiver to inform the receiver of the modulation order.

### Claim 12

A receiver comprising:
a communication interface configured to receive a vector of samples generated by a channel in response to a vector of channel symbols transmitted by a transmitter; and
at least one processor configured to determine a first prediction vector for a message encoded by the transmitter into the vector of channel symbols in accordance with a modulation order m that is adjustable up to a maximum modulation order Mmax, wherein the at least one processor comprises a slicing layer configured to eliminate one or more elements of the first prediction vector in an instance in which the modulation order m is less than the maximum modulation order Mmax so as to generate a second prediction vector tailored to the modulation order from which a prediction of the message encoded by the transmitter is identified.

### Claim 13 (depends on 12)

A receiver according to claim 12 wherein the at least one processor is further configured to generate a probability vector associated with the modulation order m, wherein the probability vector assigns probabilities for m-possible messages based upon the second prediction vector, wherein the at least one processor is further configured to determine one or more updated parameters of the receiver, wherein the updated parameters of the receiver are based on the probability vector for each of a plurality of modulation orders, and wherein the receiver is reconfigured based upon the one or more updated parameters.

### Claim 14 (depends on 13)

A receiver according to claim 13 wherein the at least one processor comprises an activation function configured to normalize the second prediction vector into the probability vector representing a probability distribution having m probabilities.

### Claim 15 (depends on 13)

A receiver according to claim 13 wherein the at least one processor comprises a neural network having a plurality of trainable parameters that are updated by determinations of the at least one processor.

### Claim 16 (depends on 12)

A receiver according to claim 12 wherein the receiver is configured to predict the message encoded by the transmitter in accordance with a plurality of modulation orders, and wherein a respective modulation order defines a number of different messages capable of being encoded by the transmitter and predicted by the receiver.

### Claim 17 (depends on 12)

A receiver according to claim 12 wherein the slicing layer is configured to eliminate elements of the first prediction vector such that the second prediction vector has a number of elements that equals the modulation order.

### Claim 18 (depends on 12)

A receiver according to claim 12 wherein the slicing layer has a variable width that is dependent upon the modulation order, wherein the width of the slicing layer defines a number of elements of the second prediction vector.

### Claim 19

An apparatus comprising:
at least one processor; and
at least one memory including computer program code,
the at least one memory and the computer program code configured to, with the at least one processor, cause the apparatus at least to perform:
for a plurality of messages transmitted by a communications system configured to have a plurality of modulation orders, receive a plurality of probability vectors with each probability vector associated with a respective one of the plurality of modulation orders, wherein the probability vector assigns probabilities to a vector of samples that identifies the probability that a respective message corresponds to the vector of samples;
determine one or more updated parameters of a receiver of the communications system that receives the vector of samples generated by a channel in response to a vector of channel symbols transmitted by a transmitter of the communications system, wherein the updated parameters of the receiver are based on the plurality of probability vectors; and
determine one or more updated parameters of the transmitter that provides the vector of channel symbols for transmission over the channel, wherein the updated parameters of the transmitter are based on the plurality of probability vectors.

### Claim 20 (depends on 19)

An apparatus according to claim 19 wherein the at least one memory and the computer program code are further configured to, with the at least one processor, cause the apparatus to repeatedly receive an additional plurality of probability vectors for a different plurality of messages transmitted by the communications system and to repeatedly determine the one or more updated parameters of the receiver and the transmitter based on the additional plurality of probability vectors.

### Claim 21 (depends on 20)

An apparatus according to claim 20 wherein the at least one memory and the computer program code are further configured to, with the at least one processor, cause the apparatus to evaluate a stopping criterion and to continue to repeatedly receive the additional plurality of probability vectors for the different plurality of messages transmitted by the communications system and to repeatedly determine the one or more updated parameters of the receiver and the transmitter based on the additional plurality of probability vectors until the stopping criterion is satisfied.

### Claim 22 (depends on 19)

An apparatus according to claim 19 wherein the at least one memory and the computer program code are further configured to, with the at least one processor, cause the apparatus to determine a loss function based on the plurality of probability vectors and to perform a gradient descent on the loss function, wherein the at least one memory and the computer program code are also configured to, with the at least one processor, cause the apparatus to determine one or more updated parameters of the receiver and the transmitter based on the gradient descent.

### Claim 23 (depends on 19)

An apparatus according to claim 19 wherein the modulation order defines a number of different messages capable of being encoded by the transmitter and predicted by the receiver, and wherein the plurality of modulation orders define different numbers of messages capable of being encoded by the transmitter and predicted by the receiver.

### Claim 24

A method comprising:
for a plurality of messages transmitted by a communications system configured to have a plurality of modulation orders, receiving a plurality of probability vectors with each probability vector associated with a respective one of the plurality of modulation orders, wherein the probability vector assigns probabilities to a vector of samples that identifies the probability that a respective message corresponds to the vector of samples;
determining one or more updated parameters of a receiver of the communications system that receives the vector of samples generated by a channel in response to a vector of channel symbols transmitted by a transmitter of the communications system, wherein the updated parameters of the receiver are based on the plurality of probability vectors; and
determining one or more updated parameters of the transmitter that provides the vector of channel symbols for transmission over the channel, wherein the updated parameters of the transmitter are based on the plurality of probability vectors.

### Claim 25 (depends on 24)

A method according to claim 24 further comprising repeatedly receiving an additional plurality of probability vectors for a different plurality of messages transmitted by the communications system and repeatedly determining the one or more updated parameters of the receiver and the transmitter based on the additional plurality of probability vectors.

### Claim 26 (depends on 25)

A method according to claim 25 further comprising evaluating a stopping criterion and continuing to repeatedly receive the additional plurality of probability vectors for the different plurality of messages transmitted by the communications system and to repeatedly determine the one or more updated parameters of the receiver and the transmitter based on the additional plurality of probability vectors until the stopping criterion is satisfied.

### Claim 27 (depends on 24)

A method according to claim 24 further comprising determining a loss function based on the plurality of probability vectors and performing a gradient descent on the loss function, wherein determining one or more updated parameters of the receiver and the transmitter comprises determining one or more updated parameters of the receiver and the transmitter based on the gradient descent.

### Claim 28 (depends on 24)

A method according to claim 24 wherein the modulation order defines a number of different messages capable of being encoded by the transmitter and predicted by the receiver, and wherein the plurality of modulation orders define different numbers of messages capable of being encoded by the transmitter and predicted by the receiver.
