US-12627338-B2 - Transmitter

Abstract

An apparatus, method and computer program is described comprising: transmitting signals from a transmitter of a multiple-input single-output transmission system to a receiver of the transmission system, wherein the transmitter communicates with the receiver over a plurality of channels of the transmission system, wherein the transmitter includes a transmitter algorithm having at least some trainable weights, wherein said transmitter algorithm converts a sequence of coded bits into time domain baseband symbols for transmission over said channels; updating weights of said transmitter algorithm based on a loss function, said loss function having a first loss term, a second loss term and a variable defining a weighting of those loss terms, wherein the first parameter relates to an information rate of communications from the transmitter to the receiver; and repeating the transmitting and updating until a first condition is reached.

Inventors

• Faycal AIT AOUDIA
• Jakob HOYDIS
• Stefan Wesemann

Assignees

• NOKIA TECHNOLOGIES OY

Dates

Publication Date

20260512

Application Date

20221012

Claims (20)

1. 1 . An apparatus, comprising: at least one processor; and at least one memory storing instructions that, when executed by the at least one processor, cause the apparatus to perform: transmitting signals from a transmitter of a multiple-input single-output transmission system to a receiver of the transmission system, wherein the transmitter communicates with the receiver over a plurality of channels of the transmission system, wherein the transmitter includes a transmitter algorithm having at least some trainable weights, wherein said transmitter algorithm converts a sequence of coded bits into time domain baseband symbols for transmission over said channels; updating at least some of the trainable weights of said transmitter algorithm based on a loss function, said loss function having a first loss term, a second loss term and a variable defining a weighting of those loss terms, wherein the first loss term relates to an information rate of communications from the transmitter to the receiver and the second loss term of the loss function enforces a constraint of generating a constant envelope of at least one signal comprising at least one of the time domain baseband symbols by performing an s-fold oversampling of the at least one signal; and repeating the transmitting and updating until a first condition is reached.
2. 2 . The apparatus as claimed in claim 1 , wherein the receiver has a fixed receiver algorithm.
3. 3 . The apparatus as claimed in claim 1 , wherein the receiver includes a receiver algorithm having at least some trainable weights.
4. 4 . The apparatus as claimed in claim 3 , wherein the at least one memory and instructions, with the at least one processor, cause the apparatus to update the weights of said receiver algorithm based on said loss function together with the weights of the transmitter algorithm.
5. 5 . The apparatus as claimed in claim 3 , wherein said receiver algorithm is implemented using one or more neural networks.
6. 6 . The apparatus as claimed in claim 1 , wherein the at least one memory and instructions, with the at least one processor, cause the apparatus to perform: pre-processing said sequence of coded bits using a pre-coder prior to application to said transmitter algorithm.
7. 7 . The apparatus as claimed in claim 1 , wherein the at least one memory and instructions, with the at least one processor, cause the apparatus to perform: oversampling said time domain baseband symbols for transmission over said channels, wherein said loss function is computed based on the oversampled time domain baseband symbols.
8. 8 . The apparatus as claimed in claim 1 , wherein the first condition comprises a defined number of iterations.
9. 9 . The apparatus as claimed in claim 1 , wherein the at least one memory and instructions, with the at least one processor, cause the apparatus to initialize said weights.
10. 10 . The apparatus as claimed in claim 1 , wherein the receiver comprises a communication node of a mobile communication system.
11. 11 . The apparatus as claimed in claim 1 , wherein said transmitter algorithm is implemented using one or more neural networks.
12. 12 . A multiple-input single-output transmission system, comprising: a transmitter; a plurality of channels; and a receiver, wherein the transmission system is configured to perform: transmitting signals from the transmitter to the receiver over the plurality of channels, wherein the transmitter includes a transmitter algorithm having at least some trainable weights, wherein said transmitter algorithm converts a sequence of coded bits into baseband symbols for transmission over said channels; receiving the transmitted signals at the receiver; updating weights of said transmitter algorithm based on a loss function, said loss function having a first loss term, a second loss term and a variable defining the weighting of those parameters, wherein the first loss term relates to an information rate of communications from the transmitter to the receiver and the second loss term of the loss function enforces a constraint of generating a constant envelope of at least one signal comprising at least one of the time domain baseband symbols by performing an s-fold oversampling of the at least one signal; and repeating the transmitting, receiving and updating until a first condition is reached.
13. 13 . The multiple-input single-output transmission system as claimed in claim 12 , wherein the receiver has a fixed receiver algorithm.
14. 14 . The multiple-input single-output transmission system as claimed in claim 12 , wherein the receiver includes a receiver algorithm having at least some trainable weights.
15. 15 . The multiple-input single-output transmission system as claimed in claim 14 , further comprising updating the weights of said receiver algorithm based on said loss function together with the weights of the transmitter algorithm.
16. 16 . A method, comprising: transmitting signals from a transmitter of a multiple-input single-output transmission system to a receiver of the transmission system, wherein the transmitter communicates with the receiver over a plurality of channels of the transmission system, wherein the transmitter includes a transmitter algorithm having at least some trainable weights, wherein said transmitter algorithm converts a sequence of coded bits into time domain baseband symbols for transmission over said channels; updating weights of said transmitter algorithm based on a loss function, said loss function having a first loss term, a second loss term and a variable defining a weighting of those loss terms, wherein the first loss term relates to an information rate of communications from the transmitter to the receiver and the second loss term of the loss function enforces a constraint of generating a constant envelope of at least one signal comprising at least one of the time domain baseband symbols by performing an s-fold oversampling of the at least one signal; and repeating the transmitting and updating until a first condition is reached.
17. 17 . The method as claimed in claim 16 , further comprising pre-processing said sequence of coded bits using a pre-coder prior to application to said transmitter algorithm.
18. 18 . The method as claimed in claim 16 , further comprising oversampling said time domain baseband symbols for transmission over said channels, wherein said loss function is computed based on the oversampled time domain baseband symbols.
19. 19 . The method as claimed in claim 16 , wherein the first condition comprises a defined number of iterations.
20. 20 . A non-transitory computer-readable medium comprising computer-readable instructions encoded thereon that, when executed by an apparatus, cause the apparatus to: transmit signals from a transmitter of a multiple-input single-output transmission system to a receiver of the transmission system, wherein the transmitter communicates with the receiver over a plurality of channels of the transmission system, wherein the transmitter includes a transmitter algorithm having at least some trainable weights, wherein said transmitter algorithm converts a sequence of coded bits into time domain baseband symbols for transmission over said channels; update weights of said transmitter algorithm based on a loss function, said loss function having a first loss term, a second loss term and a variable defining a weighting of those loss term, wherein the first loss term relates to an information rate of communications from the transmitter to the receiver and the second loss term of the loss function enforces a constraint of generating a constant envelope of at least one signal comprising at least one of the time domain baseband symbols by performing an s-fold oversampling of the at least one signal; and repeat the transmitting and updating until a first condition is reached.

Description

CROSS REFERENCE TO RELATED APPLICATION This application corresponds to Finnish Application No. 20215104, filed on Feb. 1, 2021 and published on Aug. 2, 2022. Though no priority is claimed, the publication is not available as prior art against this document, under the exceptions of 35 USC § 102(b)(1) and (b)(2). FIELD The present specification relates to transmitters, for example to transmitters in Multiple Input Single Output (MISO) systems. BACKGROUND MISO systems comprising a transmitter and a receiver are known. Although options exist for training modules of such as a system, there remains a desire for further developments in this field. SUMMARY In a first aspect, this specification describes an apparatus comprising means for performing: transmitting signals from a transmitter of a multiple-input single-output transmission system to a receiver of the transmission system, wherein the transmitter communicates with the receiver over a plurality of channels of the transmission system, wherein the transmitter includes a transmitter algorithm having at least some trainable weights, wherein said transmitter algorithm converts a sequence of coded bits into time domain baseband symbols for transmission over said channels; updating at least some of the trainable weights of said transmitter algorithm based on a loss function, said loss function having a first loss term, a second loss term and a variable defining a weighting of those loss terms, wherein the first loss term relates to an information rate of communications from the transmitter to the receiver; and repeating the transmitting and updating until a first condition (e.g. a defined number of iterations or a defined performance level) is reached. The receiver may be a communication node of a mobile communication system. The trainable weights of the transmitter algorithm may enable the transmitter to be trained using machine learning principles (such that the transmitter algorithm is a machine learning algorithm). The second loss term of the loss function may relate to an envelope of at least one of the time domain baseband symbols. In some example embodiments, the receiver has a fixed receiver algorithm. In some other example embodiments, the receiver includes a receiver algorithm having at least some trainable weights. The trainable weights of the receiver algorithm may enable the received to be trained using machine learning principles (such that the receiver algorithm is a machine learning algorithm). In such embodiments, the apparatus may further comprise means configured to perform updating the weights of said receiver algorithm based on said loss function together with the weights of the transmitter algorithm. Some example embodiments further comprise means for performing: pre-processing said sequence of coded bits using a pre-coder prior to application to said transmitter algorithm. Said pre-processing may be implemented using a pre-processor having fixed functionality. Some example embodiments further comprise means for performing: oversampling said time domain baseband symbols for transmission over said channels, wherein said loss function is computed based on the oversampled time domain baseband symbols. Some example embodiments further comprise means for performing: initialising said weights. The transmitter algorithm may be implemented using one or more neural networks. Alternatively, or in addition, the receiver algorithm may be implemented using one or more neural networks. In a second aspect, this specification describes a multiple-input single-output transmission system comprising a transmitter, a plurality of channels and a receiver, the transmission system comprising means for performing: transmitting signals from the transmitter to the receiver over the plurality of channels, wherein the transmitter includes a transmitter algorithm having at least some trainable weights, wherein said transmitter algorithm converts a sequence of coded bits into baseband symbols for transmission over said channels; receiving the transmitted signals at the receiver; updating weights of said transmitter algorithm based on a loss function, said loss function having a first loss term, a second loss term and a variable defining the weighting of those parameters; and repeating the transmitting, receiving and updating until a first condition (e.g. a defined number of iterations or a defined performance level) is reached. The trainable weights of the transmitter algorithm may enable the transmitter to be trained using machine learning principles (such that the transmitter algorithm is a machine learning algorithm). The second loss term of the loss function may relate to an envelope of at least one of the time domain baseband symbols. In some example embodiments, the receiver has a fixed receiver algorithm. In some other example embodiments, the receiver includes a receiver algorithm having at least some trainable weights. The trainable weights of the receiver algori