CN-122028116-A - Multi-access point distributed cooperative transmission method

CN122028116ACN 122028116 ACN122028116 ACN 122028116ACN-122028116-A

Abstract

The application belongs to the field of wireless network multiple access, and provides a multi-access point distributed cooperative transmission method which comprises the steps of configuring access points into distributed decision entities and initializing a multi-mode model, executing cooperative time division multiple access scheduling in a virtual network environment, taking an access point set which is involved in transmission in each decision period as a label, constructing a training data set training multi-mode model based on local states of the access points corresponding to the label and historical channel observation matrixes of the access points, deploying the trained multi-mode model into the wireless network environment, inputting the current historical channel observation matrixes and the local states into the multi-mode model when each transmission time slot starts, and determining the transmission behavior of the current time slot according to the output of the multi-mode model. Compared with the prior art, the method and the device greatly improve the transmission efficiency of the multi-access point network environment.

Inventors

SUN XINGHUA
HAN MINGQI
ZHAN WEN
QIU JIYUN
WANG XIJUN
CHEN XIANG

Assignees

中山大学·深圳
中山大学

Dates

Publication Date: 20260512
Application Date: 20260114

Claims (10)

1. A multi-access point distributed cooperative transmission method applied to a wireless network environment including a plurality of access points and a plurality of wireless users, the method comprising: configuring the access point as a distributed decision entity and initializing a multi-modal model; executing collaborative time division multiple access scheduling in a virtual network environment by taking a period of a transmission opportunity as a decision period, taking an access point set participating in transmission in each decision period as a tag, and constructing a training data set based on the local state of each access point corresponding to the tag and a history channel observation matrix of each access point; The multi-modal model is used for training the multi-modal model based on the training data set, wherein the multi-modal model receives the history channel observation matrix and the local state, and outputs a predicted access point transmission target set after multi-modal feature fusion; And deploying the trained multi-mode model to the wireless network environment, inputting the current historical channel observation matrix and the local state into the multi-mode model by each access point at the beginning of each transmission time slot, and determining the transmission behavior of the current time slot according to the output of the multi-mode model.
2. The multi-access point distributed cooperative transmission method of claim 1, wherein the multi-modal model is a vision-language model, the history channel observation matrix is used as a vision modal input, and the local state is used as a language modal input.
3. The multi-access point distributed cooperative transmission method according to claim 2, wherein the visual-language model comprises a visual transducer network, a text transducer network, an attention module and a multi-layer perceptron, wherein the visual transducer network receives a history channel observation matrix and outputs visual hidden features, the text transducer network receives a local state and outputs text hidden features, the attention module carries out cross attention fusion on the visual hidden features and the text hidden features to obtain joint hidden features, and the joint hidden features are input into an access point transmission target set predicted by the multi-layer perceptron network.
4. The method of claim 2, wherein the historical channel observation matrix has a column dimension corresponding to an observation time window and a row dimension corresponding to a historical observation time period, and the element values in the matrix are used to characterize the channel state type and signal power value detected by the access point at a specific historical time, and the expression is as follows: Wherein, the size of the observation period H, T is the history observation length, To be specific to the moment Channel observations of (2); In order to detect the power of the transmission signal, For transmitting the indicator.
5. The method of claim 2, wherein the local state comprises observation sub-vectors of the access point to one or more other access points in the network, each of the observation sub-vectors comprising location information of the observed access point, received signal power information from the observed access point, and time interval information from the last time the observed access point acknowledgement frame was received, wherein the expression is as follows: Wherein, the The local state is indicated as such, Representing an access point For an access point Is used for the observation information of the (a), For an access point Is used for the two-dimensional coordinates of (c), For an access point Receiving an access point Is used for the signal power of the (c), For an access point Last time the access point was intercepted Acknowledgement frame to date Time elapsed from the moment.
6. The multi-access point distributed cooperative transmission method according to claim 1, wherein in the training step of the multi-mode model, gradient descent and parameter update are performed by using a multi-label classification loss function, and the expression is as follows: Where M is the number of access points in the network, Representing an access point Estimating access points Whether transmission will occur or not, 0 indicating no transmission, and 1 indicating transmission.
7. The method of claim 1, wherein determining the transmission behavior of the current time slot based on the output of the multi-mode model comprises selecting a maximum MCS rate and a corresponding power that satisfy successful transmission based on the interference signal power of the current access point by the target access point.
8. The multi-access point distributed cooperative transmission method according to any of claims 1 to 7, further comprising the steps of, before the multi-modal model training: the positions of the access points and the users thereof are set randomly, and network communication parameters are configured.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the multi-access point distributed cooperative transmission method of any of claims 1 to 8 when the computer program is executed by the processor.
10. A computer readable storage medium having stored thereon a computer program, wherein the computer program when executed by a processor implements the multi-access point distributed cooperative transmission method according to any of claims 1 to 8.

Description

Multi-access point distributed cooperative transmission method Technical Field The application belongs to the technical field of wireless network multiple access, and provides a multi-access point distributed cooperative transmission method. Background With the rapid growth in the number of wireless devices, wi-Fi technology has become one of the primary means of providing efficient communication services. In order to improve network transmission efficiency, the IEEE 802.11be standard introduces a multi-Access Point (AP) cooperative technology, and the core mechanism thereof is cooperative spatial multiplexing, so that interference among access points is reduced while parallel transmission of multiple access points is supported by methods of cooperative time division multiple access, joint transmission, cooperative beam forming and the like, thereby improving overall network performance. The scheme based on collaborative time division multiple access spatial multiplexing (c-TDMA/SR) is of great interest because of relatively simple deployment, and the scheme generally uses shared access points to collect information such as received signal strength indication of each node in a centralized way, and optimizes system throughput by coordinating transmission timing and power allocation between each access point and a terminal. However, the above-mentioned centralized collaboration still faces some challenges in practical applications. On one hand, the scheme needs to acquire global channel state information, and significant signaling overhead is possibly introduced in the scenes of short data packet transmission and the like to influence the transmission efficiency, and on the other hand, in a dense deployment environment, when the positions of terminals associated with different access points are close, even though the complete channel information is relied on, the anti-interference performance of the system is still possibly reduced, so that the further improvement of the throughput is limited. Disclosure of Invention The invention provides a multi-access point distributed cooperative transmission method for overcoming the defect that the transmission efficiency is reduced due to high communication overhead in the prior art. In order to achieve the technical effects, the technical scheme of the invention is as follows: a multi-access point distributed cooperative transmission method applied to a wireless network environment including a plurality of access points and a plurality of wireless users, the method comprising: configuring the access point as a distributed decision entity and initializing a multi-modal model; executing collaborative time division multiple access scheduling in a virtual network environment by taking a period of a transmission opportunity as a decision period, taking an access point set participating in transmission in each decision period as a tag, and constructing a training data set based on the local state of each access point corresponding to the tag and a history channel observation matrix of each access point; The multi-modal model is used for training the multi-modal model based on the training data set, wherein the multi-modal model receives the history channel observation matrix and the local state, and outputs a predicted access point transmission target set after multi-modal feature fusion; And deploying the trained multi-mode model to the wireless network environment, inputting the current historical channel observation matrix and the local state into the multi-mode model by each access point at the beginning of each transmission time slot, and determining the transmission behavior of the current time slot according to the output of the multi-mode model. Preferably, the multi-modal model is a visual-language model, the history channel observation matrix is used as visual modal input, and the local state is used as language modal input. The visual-language model comprises a visual transducer network, a text transducer network, an attention module and a multi-layer perceptron, wherein the visual transducer network receives a historical channel observation matrix and outputs visual hidden features, the text transducer network receives a local state and outputs text hidden features, the attention module carries out cross attention fusion on the visual hidden features and the text hidden features to obtain joint hidden features, and the joint hidden features are input into the multi-layer perceptron network to output a predicted access point transmission target set. Preferably, the column dimension of the history channel observation matrix corresponds to an observation time window, the row dimension corresponds to a history observation time length, and element values in the matrix are used for representing the channel state type and signal power value detected by the access point at a specific history time, and the expression is as follows: Wherein, the size of the observation pe