Method and apparatus for optimizing average bit error probability via deep multi-armed bandit in OFDM and index modulation system for low power communication

Patent metadata

Field	Value
Publication number	US-11201694-B2
Application number	US-202016878826-A
Country	US
Kind code	B2
Filing date	May 20, 2020
Priority date	Nov 26, 2019
Publication date	Dec 14, 2021
Grant date	Dec 14, 2021

Abstract

A method and apparatus for optimizing average bit error probability via a deep multi-armed bandit in an orthogonal-frequency division multiplexing and index modulation system for low power communication are proposed. The method proposed in the present invention comprises: detecting BPSK symbols and subcarriers among all subcarriers; defining a combination of selected subcarriers as a subcarrier selection pattern; selecting the subcarrier selection pattern through learning to minimize the average bit error probability for all combinations of selected subcarriers; and updating a learning parameter of the subcarrier selection pattern selected through learning.

First claim

What is claimed is:

1. A method for optimizing average bit error probability via a deep multi-armed bandit in an orthogonal-frequency division multiplexing and index modulation (OFDM-IM) system for low power communication, the method comprising:
detecting binary phase shift keying (BPSK) symbols and a plurality of subcarriers among all subcarriers;
defining a plurality of subcarrier selection patterns, each subcarrier selection pattern (SSP) including a combination of individual subcarriers selected from the plurality of subcarriers;
selecting a set of subcarrier selection patterns from the plurality of subcarrier selection patterns through learning; and
updating a learning parameter of the set of subcarrier selection patterns selected through learning.

2. The method of claim 1 , wherein each subcarrier selection pattern is represented as a vector representing BPSK symbols depending on a use or non-use of each of the individual subcarriers.

3. The method of claim 2 , wherein the method further includes using a maximum likelihood detection to jointly detect a subcarrier selection pattern and its respective BPSK symbols in a received signal, and, in the maximum likelihood detection, determining a transmitted signal based on the Euclidean distance between the received signal and the transmitted signal.

4. The method of claim 1 , wherein the selecting the set of subcarrier selection patterns through learning further includes using a multi-armed bandit (MAB) algorithm to select the set of subcarrier selection patterns, and obtaining a bit error probability and a reward for the selected set of subcarrier selection patterns.

5. The method of claim 4 , wherein the selecting the set of subcarrier selection patterns through learning further includes maximizing the reward for a transmission of each OFDM-IM block of the OFDM-IM system.

6. The method of claim 4 , wherein the selecting the set of subcarrier selection patterns from the plurality of subcarrier selection patterns through learning further includes selecting a set S i of subcarrier selection patterns using an upper confidence bound (UCB), wherein the UCB is defined as:

UCB

S
i

=

R

S
i

T

S
i

+

βlog
⁡

(
n
)

T

S
i

where R S i is the reward for the set S i of subcarrier selection patterns, T S i is a number of times the set S i of subcarrier selection patterns is selected, and β is a value for adjusting a convergence rate of the MAB algorithm and a ratio of exploration and exploitation.

7. The method of claim 6 , wherein the method further includes selecting the set S i having the highest UCB.

8. The method of claim 6 , wherein the updating the learning parameter of the set of subcarrier selection patterns further includes using a deep neural network that learns a channel size of subcarriers in the set of subcarrier selection patterns as an input and the UCB for the set of subcarrier selection patterns as an output.

9. The method of claim 1 , wherein the selecting the set of subcarrier selection patterns through learning further includes setting combinatorial observations and rewards for the set of subcarrier selection patterns to 0.

10. The method of claim 1 , wherein the learning parameter represents a bit error probability of the set of subcarrier selection patterns, and the method further includes performing learning to minimize the average bit error probability for all combinations of individual subcarriers using a deep neural network without storing combinatorial observations of all sets of subcarrier selection patterns.

11. An apparatus for optimizing average bit error probability via a deep multi-armed bandit in an orthogonal-frequency division multiplexing and index modulation (OFDM-IM) system for low power communication, the apparatus comprising:
a detection part that detects binary phase shift keying (BPSK) symbols and a plurality of subcarriers among all subcarriers;
a subcarrier selection pattern determination part that defines a plurality of subcarrier selection patterns, each subcarrier selection pattern including a combination of individual subcarriers selected from the plurality of subcarriers, and selects a set of the subcarrier selection patterns from the plurality of subcarrier selection patterns through learning; and
a learning part that updates a learning parameter of the set of subcarrier selection patterns selected through learning.

12. The apparatus of claim 11 , wherein the subcarrier selection pattern determination part represents each subcarrier selection pattern as a vector representing BPSK symbols depending on a use or non-use of each of the individual subcarriers.

13. The apparatus of claim 11 , wherein the subcarrier selection pattern determination part uses a multi-armed bandit (MAB) algorithm, selects the set of subcarrier selection patterns to minimize bit error probability, and obtains a bit error probability and a reward for the selected set of subcarrier selection patterns.

14. The apparatus of claim 13 , wherein the set of subcarrier selection patterns is determined in such a way as to maximize the reward for a transmission of each OFDM-IM block of the OFDM-IM system.

15. The apparatus of claim 13 , wherein a set S i is selected from the plurality of subcarrier selection patterns using an upper confidence bound (UCB), wherein the UCB is defined as:

UCB

S
i

=

R

S
i

T

S
i

+

βlog
⁡

(
n
)

T

S
i

Where R S i is the reward for the set S i of subcarrier selection patterns, T S i is

Assignees

• Korea Advanced Inst Sci & Tech

Inventors

• Choi Wan
• Lee Jinkyu
• Choi Junil

Classifications

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06N3/047

  Probabilistic or stochastic networks · CPC title

• G06N3/0499

  Feedforward networks · CPC title

• G06N3/092

  Reinforcement learning · CPC title

• H04L5/0058

  Allocation criteria · CPC title