CN118802425A - Channel estimation optimization method, device, equipment and medium for base station MIMO wireless communication system - Google Patents

Abstract

The application relates to a channel estimation optimization method, a device, equipment and a medium of a base station MIMO wireless communication system, which relate to the field of data processing, wherein the method comprises the following steps: acquiring channel lead codes in all communication connection requests; determining pulse release rates of a plurality of neurons corresponding to channel preambles in a preset channel estimation optimization model based on the channel preambles in the communication connection request; in neurons of a channel estimation optimization model, poisson sampling is carried out according to a pulse release rate by using poisson distribution, a time interval of next pulse release is randomly generated, and when the accumulated time interval reaches or exceeds a preset time interval threshold value, channel pulse codes of channel lead codes corresponding to the current time point are recorded; the channel pulse code of the channel preamble is input to a variable self-encoder in a channel estimation optimization model to reconstruct the channel preamble. The application improves the accuracy and efficiency of channel estimation and simultaneously obviously reduces the energy consumption of the MIMO wireless communication system.

Description

Channel estimation optimization method, device, equipment and medium for base station MIMO wireless communication system

Technical Field

The present application relates to the field of data processing, and in particular to a channel estimation optimization method for a base station MIMO wireless communication system, a corresponding device, an electronic device, and a computer-readable storage medium.

Background Art

In the base station MIMO wireless communication system, the signal from the transmitter to the receiver will be reflected and refracted through multiple paths, resulting in the received signal being a superposition of multiple delayed versions, i.e., multipath propagation. As 5G signal systems widely use high-frequency spectrum such as millimeter waves, signals in these frequency bands are more susceptible to multipath propagation and attenuation, and 5G base stations are usually equipped with a large number of antenna units (such as 64 or more), forming a large-scale MIMO wireless communication system. In this configuration, the characteristics of each antenna channel may be different, and accurate channel estimation is essential for achieving beamforming, spatial multiplexing, and improving spectrum efficiency.

In recent years, artificial neural networks (ANN) have developed rapidly and have achieved great success in the fields of computer vision and NLP. However, artificial neural networks usually require a lot of computing resources, which is a challenge when computing resources are limited. Traditional neural network models are data-driven rather than event-driven. Since the base station user connection process takes a short time, and the base station is basically in normal operation before and after the connection, and channel estimation occurs during the user connection to the base station. Therefore, traditional channel prediction networks often consume a lot of power for ineffective calculations during normal operation.

At present, the main problems of traditional artificial neural networks are as follows: traditional artificial neural networks usually process continuous value inputs, and their expression ability in the time dimension is limited, and they cannot naturally capture and utilize the precise time information in time series data; when performing calculations, traditional artificial neural networks will continuously process input data even if there is no significant change in information, which leads to higher energy consumption; traditional artificial neural networks may require additional strategies to handle sparse data to avoid overfitting, etc.

To sum up, in order to adapt to the problems in the prior art that traditional artificial neural networks often consume a lot of power for invalid calculations in normal work, and that traditional artificial neural networks usually deal with continuous value inputs, their expression ability in the time dimension is limited, and they cannot naturally capture and utilize the precise time information in time series data, the applicant has made corresponding explorations to solve these problems.

Summary of the invention

The purpose of the present application is to solve the above-mentioned problem and to provide a channel estimation optimization method for a base station MIMO wireless communication system, a corresponding device, an electronic device and a computer-readable storage medium.

In order to meet the various objectives of this application, this application adopts the following technical solutions:

A channel estimation optimization method for a base station MIMO wireless communication system proposed to meet one of the purposes of the present application includes:

Responding to each communication connection request from a user equipment and acquiring a channel preamble in each communication connection request;

Based on the channel preamble in the communication connection request, determining the pulse firing rates of multiple neurons corresponding to the channel preamble in a preset channel estimation optimization model, wherein each communication connection request includes a channel preamble, and each channel preamble corresponds to multiple neurons;

In the neurons of the channel estimation optimization model, Poisson sampling is performed according to the pulse emission rate using Poisson distribution to randomly generate the time interval for the next pulse emission, and when the accumulated time interval reaches or exceeds a preset time interval threshold, the channel pulse code of the channel preamble code corresponding to the current time point is recorded;

The channel pulse coding of the channel preamble is input into the variational autoencoder in the channel estimation optimization model to reconstruct the channel preamble, so as to complete the channel estimation optimization of the base station MIMO wireless communication system.

Optionally, the step of determining the pulse firing rates of a plurality of neurons corresponding to the channel preamble code in a preset channel estimation optimization model based on the channel preamble code in the communication connection request includes:

Determine a channel preamble in the communication connection request, and extract a characteristic value combination containing key information from the channel preamble, wherein the characteristic value combination includes any multiple items of channel gain, noise level, and transmission delay;

Inputting the extracted eigenvalue combination into an input layer of a preset channel estimation optimization model, and distributing the input eigenvalue combination to a plurality of neurons corresponding to the channel preamble code in the channel estimation optimization model;

The integral pulse emission function is used to multiply each eigenvalue in the input eigenvalue combination with its corresponding weight and sum them up to obtain the membrane potential of each neuron;

The pulse firing rate of each neuron corresponding to the channel preamble is determined according to the membrane potential of each neuron.

Optionally, in the neuron of the channel estimation optimization model, Poisson sampling is performed according to the pulse emission rate using Poisson distribution, and the time interval for the next pulse emission is randomly generated. When the accumulated time interval reaches or exceeds a preset time interval threshold, the step of recording the channel pulse coding of the channel preamble code corresponding to the current time point includes:

Determine the pulse firing rate of each neuron corresponding to the channel preamble in a preset channel estimation optimization model;

Using a preset Poisson distribution probability density function to randomly generate a time interval for the next pulse emission according to the pulse emission rate;

Accumulating the time interval of each pulse emission to determine the accumulated time interval, detecting whether the accumulated time interval reaches or exceeds a preset time interval threshold, and if the accumulated time interval reaches or exceeds the preset time interval threshold, recording the channel pulse code of the channel preamble code corresponding to the current time point;

The accumulated time interval is reset to zero, and the above steps are repeated.

Optionally, the step of inputting the channel pulse coding of the channel preamble into a variational autoencoder in the channel estimation optimization model to reconstruct the channel preamble includes:

In the encoder of the variational autoencoder, mapping the channel pulse coding of the channel preamble code to a latent space, and outputting the probability distribution of the latent variables in the latent space;

In the decoder of the variational autoencoder, sampling is performed from the probability distribution of the latent variables output by the encoder to determine a specific latent variable to generate a reconstructed channel preamble;

The difference between the probability distribution of the latent variables in the latent space and the prior distribution is calculated to determine a KL divergence loss value, and when the KL divergence loss value is lower than a preset KL divergence loss value, the reconstruction of the channel preamble code is completed.

Optionally, in the encoder of the variational autoencoder, after the step of mapping the channel pulse coding of the channel preamble code to a latent space and outputting the probability distribution of the latent variables in the latent space, the method further comprises:

In the latent space, sampling is performed from the probability distribution of the latent variables output by the encoder to determine a specific latent variable;

Gaussian noise is added to the specific latent variable, and the specific latent variable after the Gaussian noise is added is input into the decoder of the variational autoencoder to generate a reconstructed channel preamble code.

Optionally, after the step of obtaining the channel preamble code in each communication connection request, the method further includes:

In response to the data preprocessing instruction, convert the channel preamble in the communication connection request into binary time series data;

The binary time series data corresponding to the channel preamble code is input into a preset channel estimation optimization model to determine the pulse firing rates of multiple neurons corresponding to the channel preamble code in the preset channel estimation optimization model.

Optionally, the pulse firing rate represents the number or frequency of pulses generated by neurons in the channel estimation optimization model per unit time, and the basic network architecture of the channel estimation optimization model is a pulse neural network.

A channel estimation optimization device for a base station MIMO wireless communication system provided for another purpose of the present application includes:

A preamble extraction module, configured to respond to each communication connection request from a user equipment and obtain a channel preamble in each communication connection request;

A pulse emission rate determination module is configured to determine the pulse emission rates of multiple neurons corresponding to the channel preamble code in a preset channel estimation optimization model based on the channel preamble code in the communication connection request, wherein each communication connection request includes a channel preamble code, and each channel preamble code corresponds to multiple neurons;

A pulse code determination module is configured to use a Poisson distribution to perform Poisson sampling according to the pulse emission rate in the neurons of the channel estimation optimization model, randomly generate a time interval for the next pulse emission, and when the accumulated time interval reaches or exceeds a preset time interval threshold, record the channel pulse code of the channel preamble code corresponding to the current time point;

The channel estimation optimization module is configured to input the channel pulse coding of the channel preamble code into the variational autoencoder in the channel estimation optimization model to reconstruct the channel preamble code, so as to complete the channel estimation optimization of the base station MIMO wireless communication system.

An electronic device provided to meet another purpose of the present application includes a central processing unit and a memory, wherein the central processing unit is used to call and run a computer program stored in the memory to execute the steps of the channel estimation optimization method of the base station MIMO wireless communication system described in the present application.

A computer-readable storage medium is provided to meet another purpose of the present application, which stores a computer program implemented according to the channel estimation optimization method of the base station MIMO wireless communication system in the form of computer-readable instructions. When the computer program is called and executed by a computer, the steps included in the corresponding method are executed.

It can be seen from the above embodiments that, compared with the prior art, the present application aims at the problems that the conventional artificial neural network in the prior art often consumes a lot of power for ineffective calculations of normal operation, and the conventional artificial neural network usually processes continuous value inputs, and their expression ability in the time dimension is limited, and they cannot naturally capture and utilize the precise time information in the time series data. The present application includes but is not limited to the following beneficial effects:

First, the channel estimation optimization method of the base station MIMO wireless communication system of the present application, the pulse neural network can significantly enhance the expression ability in the time dimension, and can accurately and quickly capture and utilize the precise time information in the time series data;

Secondly, the channel estimation optimization method of the base station MIMO wireless communication system of the present application is that the pulse neural network "releases" pulses only when the neuron reaches the threshold, which is closer to the energy-saving mechanism of the biological nervous system and more suitable for energy-sensitive base station applications, significantly reducing the energy consumption of the MIMO wireless communication system;

Third, the channel estimation optimization method of the base station MIMO wireless communication system of the present application, the base station channel estimation often involves sparse wireless signal measurement, the pulse neural network can encode information through the pulse emission mode, and can effectively utilize the sparsity of data to accurately and efficiently process such sparse data;

Fourthly, in the channel estimation optimization method of the base station MIMO wireless communication system of the present application, the pulse neural network encodes the signal strength by the frequency of pulse emission, and has stronger adaptability to signal changes with a large dynamic range, such as channel fluctuations caused by multipath propagation, than traditional neural networks based on continuous activation values.

Furthermore, the channel estimation optimization method of the base station MIMO wireless communication system of the present application, the variational autoencoder based on the pulse network can often achieve similar or even better performance with a smaller model size, which is conducive to deployment on the base station; the pulse coding process in the MIMO multi-user case, according to the sparsity of user connections, proposes a Poisson coding process based on a pulse neural network, which can greatly improve the robustness of signal transmission to combat multipath interference, noise and other adverse channel conditions; variational autoencoder channel reconstruction based on the pulse neural network, using the efficient computing power of the pulse network and the nonlinear feature extraction capability of the autoencoder;

For channel estimation driven by user connection events, pulse neural networks can optimize device signal connection latency and are easy to implement in hardware, especially using application-specific integrated circuits (ASICs) or field-programmable gate arrays (FPGAs), thereby accelerating calculations and reducing energy consumption.

Furthermore, the channel estimation optimization method of the base station MIMO wireless communication system of the present application ensures that the system can quickly respond to channel changes and maintain connection quality through the real-time channel estimation. Especially in applications that are sensitive to delays, such as vehicle communications and drone control, through effective channel estimation, the 5G system can more accurately match signal transmission and reception, reduce energy waste, and improve the overall energy efficiency of the system. At the same time, optimized resource allocation and signal processing technology improve spectrum utilization and support simultaneous transmission of more users and higher data rates.

This application combines the temporal coding capability of pulse neural networks and the generative modeling capability of variational autoencoders (VAEs), which can help understand and optimize channel conditions in complex communication environments, especially in mobile communication systems. While improving the accuracy and efficiency of channel estimation, this application also significantly reduces the energy consumption of MIMO wireless communication systems, which is of great significance for the performance optimization of MIMO wireless communication systems.

BRIEF DESCRIPTION OF THE DRAWINGS

The above and/or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

FIG1 is a flow chart of a channel estimation optimization method for a base station MIMO wireless communication system of the present application;

FIG2 is an exemplary network architecture of low-power channel estimation for pulse networks according to an embodiment of the present application;

FIG3 is a schematic diagram of a process of converting a channel preamble into binary time series data in an embodiment of the present application;

FIG4 is a schematic diagram of a process for determining the pulse firing rate corresponding to each neuron in a channel estimation optimization model in an embodiment of the present application;

FIG5 is a schematic diagram of a process of recording a channel pulse coding of a channel preamble code corresponding to a current time point in an embodiment of the present application;

FIG6 is a schematic diagram of a process for reconstructing a channel preamble based on a variational autoencoder in an embodiment of the present application;

FIG7 is an exemplary network architecture of a variational coding channel reconstruction process of a spiking neural network in an embodiment of the present application;

FIG8 is a schematic diagram of a process of adding Gaussian noise to a latent variable in an embodiment of the present application;

FIG9 is a flowchart of a pulse neural network for channel estimation in an embodiment of the present application;

FIG10 is a schematic block diagram of a channel estimation and optimization device for a base station MIMO wireless communication system according to an embodiment of the present application;

FIG. 11 is a schematic diagram of the structure of a computer device in an embodiment of the present application.

DETAILED DESCRIPTION

The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present application, and cannot be interpreted as limiting the present application.

It will be understood by those skilled in the art that, unless expressly stated, the singular forms "one", "said", and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present application refers to the presence of the features, integers, steps, operations, elements and/or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and/or groups thereof. It should be understood that when we refer to an element as being "connected" or "coupled" to another element, it may be directly connected or coupled to the other element, or there may be an intermediate element. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The term "and/or" used herein includes all or any unit and all combinations of one or more associated listed items.

It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as generally understood by those skilled in the art to which this application belongs. It should also be understood that terms such as those defined in common dictionaries should be understood to have meanings consistent with the meanings in the context of the prior art, and will not be interpreted with idealized or overly formal meanings unless specifically defined as here.

It will be understood by those skilled in the art that the "client", "terminal" and "terminal device" used herein include both devices with wireless signal receivers, which are devices with only wireless signal receivers without transmission capabilities, and devices with receiving and transmitting hardware, which are devices with receiving and transmitting hardware capable of two-way communication on a two-way communication link. Such devices may include: cellular or other communication devices such as personal computers, tablet computers, which have single-line displays or multi-line displays or cellular or other communication devices without multi-line displays; PCS (Personal Communications Service, personal communication system), which can combine voice, data processing, fax and/or data communication capabilities; PDA (Personal Digital Assistant, personal digital assistant), which may include a radio frequency receiver, pager, Internet/intranet access, web browser, notepad, calendar and/or GPS (Global Positioning System, global positioning system) receiver; conventional laptop and/or palmtop computers or other devices, which have and/or include a conventional laptop and/or palmtop computer or other device with and/or including a radio frequency receiver. The "client", "terminal" and "terminal device" used herein may be portable, transportable, installed in a vehicle (air, sea and/or land), or suitable and/or configured to run locally, and/or in a distributed form, at any other location on the earth and/or in space. The "client", "terminal" and "terminal device" used herein may also be a communication terminal, an Internet terminal, a music/video playback terminal, such as a PDA, a MID (Mobile Internet Device) and/or a mobile phone with a music/video playback function, or a smart TV, a set-top box and other devices.

The hardware referred to by the names such as "server", "client", and "service node" in this application is essentially an electronic device with capabilities equivalent to those of a personal computer. It is a hardware device that has the necessary components revealed by the von Neumann principle, such as a central processing unit (including an arithmetic unit and a controller), a memory, an input device, and an output device. The computer program is stored in its memory, and the central processing unit calls the program stored in the external memory into the internal memory for execution, executes the instructions in the program, and interacts with the input and output devices to complete specific functions.

It should be pointed out that the concept of "server" referred to in this application can also be extended to the case of server clusters. According to the network deployment principle understood by those skilled in the art, the servers should be logically divided. In physical space, these servers can be independent of each other but can be called through interfaces, or integrated into a physical computer or a set of computer clusters. Those skilled in the art should understand this flexibility, and should not use it to restrict the implementation of the network deployment method of this application.

Unless expressly specified, one or more technical features of the present application can be deployed on a server for implementation and accessed by a client through a remote call to obtain an online service interface provided by the server, or can be directly deployed and run on a client for access.

The neural network models referenced or may be referenced in this application, unless expressly specified, can be deployed on a remote server and remotely called on the client, or can be deployed and directly called on a client with sufficient device capabilities. In some embodiments, when it runs on the client, its corresponding intelligence can be obtained through transfer learning to reduce the requirements for the client's hardware operating resources and avoid excessive occupation of the client's hardware operating resources.

Unless explicitly specified, the various data involved in this application can be stored remotely on a server or on a local terminal device, as long as it is suitable for being called by the technical solution of this application.

Those skilled in the art should be aware that, although the various methods of the present application are described based on the same concept and thus present commonality to each other, unless otherwise specified, these methods can be independently executed. Similarly, for each embodiment disclosed in the present application, they are all proposed based on the same inventive concept, therefore, concepts with the same expression, and concepts that are appropriately changed for convenience despite different expressions, should be understood as equivalent.

Unless the mutually exclusive relationship between the embodiments to be disclosed in this application is explicitly stated, the relevant technical features involved in each embodiment can be cross-combined to flexibly construct a new embodiment, as long as such combination does not deviate from the creative spirit of this application and can meet the needs of the prior art or solve certain deficiencies in the prior art. Those skilled in the art should be aware of this flexibility.

Referring to FIG. 1 and FIG. 2 , the channel estimation optimization method of the base station MIMO wireless communication system of the present application, in one embodiment, includes:

Step S10, responding to each communication connection request from the user equipment and acquiring a channel preamble code in each communication connection request;

The receiving end in the base station MIMO wireless communication system can respond to each communication connection request from the user equipment and obtain the channel preamble code in the each communication connection request; the receiving end in the base station MIMO wireless communication system can be a signal processing unit or a baseband processor, the signal processing unit includes a channel estimation module or a digital signal processing module, the channel estimation module is used to estimate the channel state information (CSI) to help correct the signal distortion caused by channel fading and noise; the user equipment can be a terminal device such as a mobile phone, a tablet computer or a laptop, all of which can be used as the user equipment of the present application.

In some embodiments, a receiving end in a base station MIMO wireless communication system receives a communication connection request sent by a user device such as a mobile phone, a tablet computer, or a laptop computer. Each communication connection request may include multiple information, including data, a control signal, a channel preamble code, and other information; the receiving end in the base station MIMO wireless communication system extracts a channel preamble code from each communication connection request. The channel preamble code is a special signal identifier used to identify and estimate the channel state. Using the extracted channel preamble code, the base station can perform channel estimation, thereby understanding the propagation characteristics and signal quality of the signal in the wireless link.

In some embodiments, the base station MIMO wireless communication system is a wireless communication system using multiple input multiple output (MIMO) technology, which uses a base station equipped with multiple transmitting and receiving antennas to achieve higher data transmission rates, enhanced signal coverage and stronger anti-interference capabilities.

In some embodiments, referring to FIG. 3 , after the step of obtaining the channel preamble code in each communication connection request, the method includes:

Step S101, in response to a data preprocessing instruction, converting a channel preamble in the communication connection request into binary time series data;

Step S102: input the binary time series data corresponding to the channel preamble code into a preset channel estimation optimization model to determine the pulse firing rates of multiple neurons corresponding to the channel preamble code in the preset channel estimation optimization model.

Specifically, the basic network architecture of the channel estimation optimization model is a pulse neural network. The input features of the pulse neural network used for base station channel estimation must be binary time series data, and the channel preamble (pilot) in each communication connection request of the user equipment is a complex signal value. Therefore, the channel preamble in the communication connection request is converted into binary time series data; the binary time series data corresponding to the channel preamble is input into the preset channel estimation optimization model to determine the pulse firing rates of multiple neurons corresponding to the channel preamble in the preset channel estimation optimization model.

Step S20, based on the channel preamble in the communication connection request, determining the pulse emission rates of multiple neurons corresponding to the channel preamble in a preset channel estimation optimization model, wherein each communication connection request includes a channel preamble, and each channel preamble corresponds to multiple neurons;

The basic network architecture of the channel estimation optimization model is a pulse neural network. After obtaining the channel preamble code in each communication connection request, the pulse firing rate of multiple neurons corresponding to the channel preamble code in a preset pulse neural network is determined based on the channel preamble code in the communication connection request, wherein each communication connection request includes a channel preamble code, and each channel preamble code corresponds to multiple neurons; the pulse firing rate represents the number or frequency of pulses generated by neurons in the channel estimation optimization model per unit time, that is, the pulse firing rate represents the number or frequency of pulses generated by neurons in the pulse neural network per unit time.

In some embodiments, the concept of spiking neural networks is derived from the observation of neuronal activity in biological brains. In biological neural systems, information is transmitted through electrical signals emitted by neurons - pulses or spikes. This mode of transmission is highly time-dependent and event-driven, which is fundamentally different from the continuous signal transmission in traditional artificial neural networks. As the field of artificial intelligence pays more attention to energy efficiency, spiking neural networks (SNNs) are used for their potential low power consumption. Spiking neural networks (SNNs) are only activated when there is information (pulses), allowing higher computational efficiency and energy efficiency, and are particularly suitable for base station MIMO wireless communication systems.

In some embodiments, in a base station MIMO wireless communication system, a channel preamble generally does not correspond to a channel one-to-one. The channel preamble is usually used for synchronization and channel estimation. The MIMO wireless communication system uses a shared preamble to estimate multiple channels. The channel preamble is reused on different antennas or channels instead of assigning a unique preamble to each channel. The main function of the preamble is to help the receiving end estimate and compensate for channel effects to optimize signal decoding and transmission. Therefore, if multiple channels use the same preamble or preamble for synchronization, then a shared preamble can be used in a neural network to adjust the behavior of multiple neurons in a pulse neural network. Each channel preamble corresponds to multiple neurons, and the preamble of each communication connection request affects the pulse firing rate of multiple neurons. This mapping helps the pulse neural network capture and process dynamic changes in the channel.

Furthermore, each channel preamble code corresponds to multiple neurons. It is not difficult to understand that each channel is associated with multiple neurons, and each channel preamble code will be mapped to multiple neurons. Because each channel corresponds to multiple neurons, this can improve the sensitivity and adaptability of the pulse neural network to channel changes; each communication connection request can include one or more channel preamble codes, and the pulse neural network adjusts the pulse firing rate of neurons according to these channel preamble codes, thereby reflecting different channel conditions.

Referring to FIG. 4 , the step of determining the pulse firing rate corresponding to each neuron in the preset channel estimation optimization model based on the channel preamble in the communication connection request includes:

Step S201, determining a channel preamble in the communication connection request, and extracting a feature value combination containing key information from the channel preamble, wherein the feature value combination includes any multiple items of channel gain, noise level, and transmission delay;

Step S202: input the extracted eigenvalue combination into an input layer of a preset channel estimation optimization model, and distribute the input eigenvalue combination to a plurality of neurons corresponding to the channel preamble in the channel estimation optimization model;

Step S203, using an integral pulse emission function to multiply each eigenvalue in the input eigenvalue combination by its corresponding weight and sum them up to obtain the membrane potential of each neuron;

Step S204: Determine the pulse firing rate of each neuron corresponding to the channel preamble code according to the membrane potential of each neuron.

In some embodiments, the Integrate-and-Fire (I&F) function is a model in spiking neural networks (SNNs), the core idea of which is to simulate the potential accumulation and firing process of biological neurons. The basic working principle is as follows: the membrane potential of a neuron increases with the input signal (such as synaptic input), and this process is similar to the integration process. When the membrane potential reaches a certain threshold, the neuron fires a pulse (spike) and resets the membrane potential to the initial state.

Specifically, the eigenvalues in the channel preamble are extracted to construct eigenvalue combinations, wherein the eigenvalue combinations include any multiple items of channel gain, noise level, and transmission delay, and these eigenvalue combinations are used as inputs of neurons; the extracted eigenvalue combinations are input into the input layer of a preset pulse neural network, and the input eigenvalue combinations are distributed to multiple neurons corresponding to the channel preamble in the pulse neural network; the integral pulse emission function in the pulse neural network is used to multiply and sum each eigenvalue in the input eigenvalue combination with its corresponding weight to obtain the membrane potential of each neuron; and the pulse emission rate of each neuron corresponding to the channel preamble is determined based on the membrane potential of each neuron.

More specifically, the number or frequency of pulses generated by a neuron per unit time is usually called the spiking rate, which refers to the average number of pulses emitted by a neuron in a given time interval. It can be expressed as:

Pulse firing rate = number of pulses/time interval;

For example, if a neuron fires 50 times in 1 second, its firing rate is 50 Hertz (Hz). This frequency reflects the intensity of the neuron's activity and can be used to encode information and the learning process of neural networks.

From the above steps, we can see that spiking neural networks encode signal strength through the frequency of pulse emission. Compared with traditional artificial neural networks based on continuous activation values, they may have stronger adaptability when dealing with signal changes with a large dynamic range (such as channel fluctuations caused by multipath propagation). The frequency of the pulse can accurately reflect the intensity changes of the input signal, thereby more effectively processing complex signal environments.

Step S30, in the neurons of the channel estimation optimization model, Poisson sampling is performed according to the pulse emission rate using Poisson distribution, and the time interval for the next pulse emission is randomly generated. When the accumulated time interval reaches or exceeds the preset time interval threshold, the channel pulse code of the channel preamble code corresponding to the current time point is recorded;

After determining the pulse firing rates of multiple neurons corresponding to the channel preamble code in a preset pulse neural network, in the neurons of the pulse neural network, Poisson sampling is performed according to the pulse firing rate using Poisson distribution to randomly generate the time interval for the next pulse firing, and when the accumulated time interval reaches or exceeds the preset time interval threshold, the channel pulse code of the channel preamble code corresponding to the current time point is recorded;

In some embodiments, referring to FIG5 , in the neuron of the channel estimation optimization model, Poisson sampling is performed according to the pulse emission rate using Poisson distribution, and the time interval for the next pulse emission is randomly generated. When the accumulated time interval reaches or exceeds the preset time interval threshold, the step of recording the channel pulse coding of the channel preamble code corresponding to the current time point includes:

Step S301, determining the pulse firing rate of each neuron corresponding to the channel preamble code in a preset channel estimation optimization model;

Step S302, using a preset Poisson distribution probability density function to randomly generate a time interval for the next pulse emission according to the pulse emission rate;

Step S303, accumulating the time interval of each pulse emission to determine the accumulated time interval, detecting whether the accumulated time interval reaches or exceeds the preset time interval threshold, and if the accumulated time interval reaches or exceeds the preset time interval threshold, recording the channel pulse code of the channel preamble code corresponding to the current time point;

Step S304: reset the accumulated time interval to zero, and repeat the above steps.

Specifically, based on the above steps, the pulse firing rate of each neuron corresponding to the channel preamble code in the preset pulse neural network is determined; the preset Poisson distribution probability density function is used to randomly generate the time interval for the next pulse firing according to the pulse firing rate; each pulse firing time interval is accumulated to determine the accumulated time interval, and it is detected whether the accumulated time interval reaches or exceeds the preset time interval threshold; if the accumulated time interval reaches or exceeds the preset time interval threshold, the channel pulse code of the channel preamble code corresponding to the current time point is recorded; the accumulated time interval is reset to zero, and the above steps are repeated.

More specifically, in a spiking neural network, it is first necessary to determine the spiking rate of each neuron. The spiking rate can be dynamically determined according to the channel preamble code through the steps in the above embodiment, or it can be pre-set based on the actual service requirements of the base station MIMO wireless communication system. The spiking rate is usually measured in Hertz (Hz), which indicates the average number of spiking times of a neuron per unit time. For example, if the spiking rate of a neuron is 20 Hz, it will emit 20 pulses on average in 1 second.

According to the pulse firing rate of neurons, the Poisson distribution can be used to randomly generate the time interval of pulse firing. Poisson distribution describes the probability distribution of the number of times an event occurs per unit time. The time interval of the next pulse firing is generated according to the following Poisson distribution probability density function, and the expression of the Poisson distribution probability density function is as follows:

,

in, is the pulse firing rate, is the time interval, is the number of pulses; the mean interval time of the Poisson distribution It represents the average time until a neuron fires its next spike at a given spike rate.

After generating the time interval for each pulse emission, these intervals are accumulated to obtain the cumulative time interval. Each time a new time interval is generated, it is added to the cumulative time interval: each time the cumulative time interval is updated, it is necessary to check whether it reaches or exceeds the preset time interval threshold.

The preset time interval threshold can be designed based on the actual business needs of the base station MIMO wireless communication system to determine when to record a pulse code. For example, if the threshold is set to 0.1 seconds, the recording will be triggered when the cumulative time interval reaches or exceeds 0.1 seconds. When the cumulative time interval reaches or exceeds the preset time interval threshold, the channel pulse code of the channel preamble code corresponding to the current time point is recorded, which means that the pulse information at a specific time point will be recorded, which can be used for subsequent processing or analysis;

After recording the pulse code, reset the accumulated time interval to zero so as to restart the calculation of the time interval for the next pulse; repeat the above steps to continuously generate pulse emission time intervals, accumulate time intervals, detect whether the preset time interval threshold is reached or exceeded, and record the pulse code until all required operations are completed or the system stops running.

Step S40: input the channel pulse coding of the channel preamble code into the variational autoencoder in the channel estimation optimization model to reconstruct the channel preamble code, so as to complete the channel estimation optimization of the base station MIMO wireless communication system.

After recording the channel pulse coding of the channel preamble code corresponding to the current time point, the channel pulse coding of the channel preamble code is input into the variational autoencoder in the channel estimation optimization model to reconstruct the channel preamble code to complete the channel estimation optimization of the base station MIMO wireless communication system.

In some embodiments, referring to FIG. 6 and FIG. 7 , the step of inputting the channel pulse coding of the channel preamble into the variational autoencoder in the channel estimation optimization model to reconstruct the channel preamble includes:

Step S401: In the encoder of the variational autoencoder, the channel pulse coding of the channel preamble is mapped to a latent space, and the probability distribution of the latent variables in the latent space is output;

Step S402: In the decoder of the variational autoencoder, sampling is performed from the probability distribution of the latent variables output by the encoder to determine a specific latent variable, so as to generate a reconstructed channel preamble;

Step S403: Calculate and determine the difference between the probability distribution of the latent variables in the latent space and the prior distribution to determine the KL divergence loss value. When the KL divergence loss value is lower than the preset KL divergence loss value, the reconstruction of the channel preamble code is completed.

Specifically, the KL divergence loss value represents the difference between the probability distribution of the latent variable in the latent space and the prior distribution. The variational autoencoder (VAE) is a deep learning model that combines the generative model and the autoencoder idea. It is mainly used to learn the effective representation of data and can generate new samples similar to the training data. The encoder part of VAE is responsible for mapping the input data to the distribution of a latent variable, rather than directly mapping it to a fixed vector like a normal autoencoder. This distribution is usually a Gaussian distribution (mean and variance ), for each input, the encoder outputs the parameters of the distribution (mean vector and logarithmic variance vector). In order to effectively train models with random variables in gradient descent, the variational autoencoder uses a reparameterization technique. It does not directly sample hidden variables , but sampling a standard normal distribution of noise , and then by The hidden variable is calculated by the mean and variance of the encoder output , which ensures that the gradient can flow through the entire network.

,

The error calculation used for the encoder and decoder is:

,

in, The distribution of the channel preamble hidden variable used to measure the encoder output (Posterior distribution) and a predefined prior distribution The difference between Satisfies normal distribution By minimizing this total loss function, the variational autoencoder can optimize the reconstruction quality of the data, while also forcing the latent variable distribution to be close to the preset prior distribution, thereby achieving effective data encoding and decoding.

Further, referring to FIG8 , in the encoder of the variational autoencoder, after the step of mapping the channel pulse coding of the channel preamble to a latent space and outputting the probability distribution of the latent variables in the latent space, the method further comprises:

Step S4001: In the latent space, sampling is performed from the probability distribution of the latent variables output by the encoder to determine a specific latent variable;

Step S4002: Add Gaussian noise to the specific latent variable, and input the specific latent variable after adding Gaussian noise into the decoder of the variational autoencoder to generate a reconstructed channel preamble code.

Specifically, in the process of channel reconstruction, Gaussian noise needs to be added so that the decoder can better learn the data characteristics of pulse coding. The channel preamble is composed of complex numbers, and the Gaussian noise addition step includes the following expressions:

(1) The expression of signal power is:

,

in, represents the signal power,

(2) The expression of noise power is:

,

in, represents the noise power, represents the signal-to-noise ratio;

(3) The expression of Gaussian noise superposition is:

,

in, is the input signal, is the absolute value symbol, is the square root symbol, is a standard normal distribution sample, Characterizes the sum of the original signal and the random Gaussian noise signal.

Based on the above formula, Gaussian noise can be added to the specific latent variables so that the decoder can better learn the data characteristics of the pulse code.

In some embodiments, please refer to FIG. 9 , during the channel preamble reconstruction process, when driven by a user device connection event, process ① will trigger enough pulses so that the pulse network can be in an ignition state to emit pulses. And the node only drives data input when the user connection event is triggered. This process can filter out a large amount of computing power consumption and prevent energy consumption caused by network computing driven by data.

In the process of variational autoencoder reconstruction, process ② is calculated by KL divergence. In theory, if the KL divergence value is 0, it means that the distribution of latent variables learned by the encoder is exactly the same as the prior distribution, which is the most ideal situation because it shows that our model can perfectly match the preset simple prior and the learned representation is very standardized and easy to interpret. However, in practical applications, a KL divergence of 0 is not always an achievable or desired result. On the one hand, too strict pursuit of a KL divergence of 0 may limit the learning ability of the model, causing it to fail to fully capture the complexity of the data. On the other hand, a non-zero KL divergence can actually help the model maintain a certain diversity and promote generation capabilities, and in some cases, a moderate KL divergence value can balance the reconstruction error of the model and the ability to generate new samples. Therefore, in this process, the threshold hyperparameter is set according to the base station load and signal processing quality. , that is, the KL divergence value is lower than It is the value with better channel reconstruction effect.

It can be seen from the above embodiments that, compared with the prior art, the present application aims at the problems that the conventional artificial neural network in the prior art often consumes a lot of power for ineffective calculations of normal operation, and the conventional artificial neural network usually processes continuous value inputs, and their expression ability in the time dimension is limited, and they cannot naturally capture and utilize the precise time information in the time series data. The present application includes but is not limited to the following beneficial effects:

First, the channel estimation optimization method of the base station MIMO wireless communication system of the present application, the pulse neural network can significantly enhance the expression ability in the time dimension, and can accurately and quickly capture and utilize the precise time information in the time series data;

Secondly, the channel estimation optimization method of the base station MIMO wireless communication system of the present application is that the pulse neural network "releases" pulses only when the neuron reaches the threshold, which is closer to the energy-saving mechanism of the biological nervous system and more suitable for energy-sensitive base station applications, significantly reducing the energy consumption of the MIMO wireless communication system;

Third, the channel estimation optimization method of the base station MIMO wireless communication system of the present application, the base station channel estimation often involves sparse wireless signal measurement, the pulse neural network can encode information through the pulse emission mode, and can effectively utilize the sparsity of data to accurately and efficiently process such sparse data;

Fourthly, in the channel estimation optimization method of the base station MIMO wireless communication system of the present application, the pulse neural network encodes the signal strength by the frequency of pulse emission, and has stronger adaptability to signal changes with a large dynamic range, such as channel fluctuations caused by multipath propagation, than traditional neural networks based on continuous activation values.

Furthermore, the channel estimation optimization method of the base station MIMO wireless communication system of the present application, the variational autoencoder based on the pulse network can often achieve similar or even better performance with a smaller model size, which is conducive to deployment on the base station; the pulse coding process in the MIMO multi-user case, according to the sparsity of user connections, proposes a Poisson coding process based on a pulse neural network, which can greatly improve the robustness of signal transmission to combat multipath interference, noise and other adverse channel conditions; variational autoencoder channel reconstruction based on the pulse neural network, using the efficient computing power of the pulse network and the nonlinear feature extraction capability of the autoencoder;

For channel estimation driven by user connection events, pulse neural networks can optimize device signal connection latency and are easy to implement in hardware, especially using application-specific integrated circuits (ASICs) or field-programmable gate arrays (FPGAs), thereby accelerating calculations and reducing energy consumption.

Furthermore, the channel estimation optimization method of the base station MIMO wireless communication system of the present application ensures that the system can quickly respond to channel changes and maintain connection quality through the real-time channel estimation. Especially in applications that are sensitive to delays, such as vehicle communications and drone control, through effective channel estimation, the 5G system can more accurately match signal transmission and reception, reduce energy waste, and improve the overall energy efficiency of the system. At the same time, optimized resource allocation and signal processing technology improve spectrum utilization and support simultaneous transmission of more users and higher data rates.

This application combines the temporal coding capability of pulse neural networks and the generative modeling capability of variational autoencoders (VAEs), which can help understand and optimize channel conditions in complex communication environments, especially in mobile communication systems. While improving the accuracy and efficiency of channel estimation, this application also significantly reduces the energy consumption of MIMO wireless communication systems, which is of great significance for the performance optimization of MIMO wireless communication systems.

Please refer to Figure 10, which shows a channel estimation optimization device for a base station MIMO wireless communication system provided to meet one of the purposes of the present application, including a preamble code extraction module 1100, a pulse emission rate determination module 1200, a pulse coding determination module 1300 and a channel estimation optimization module 1400. Among them, the preamble code extraction module 1100 is configured to respond to each communication connection request from the user equipment and obtain the channel preamble code in each communication connection request; the pulse emission rate determination module 1200 is configured to determine the pulse emission rates of multiple neurons corresponding to the channel preamble code in the preset channel estimation optimization model based on the channel preamble code in the communication connection request, wherein each communication connection request includes a channel preamble code, and each channel preamble code corresponds to multiple neurons; the pulse coding determination module 1300 is configured to use Poisson distribution to perform Poisson sampling according to the pulse emission rate in the neurons of the channel estimation optimization model, randomly generate the time interval for the next pulse emission, and when the accumulated time interval reaches or exceeds the preset time interval threshold, record the channel pulse coding of the channel preamble code corresponding to the current time point; the channel estimation optimization module 1400 is configured to input the channel pulse coding of the channel preamble code into the variational autoencoder in the channel estimation optimization model to reconstruct the channel preamble code, so as to complete the channel estimation optimization of the base station MIMO wireless communication system.

On the basis of any embodiment of the present application, please refer to FIG. 11. Another embodiment of the present application further provides an electronic device, which can be implemented by a computer device, as shown in FIG. 11, which is a schematic diagram of the internal structure of the computer device. The computer device includes a processor, a computer-readable storage medium, a memory, and a network interface connected via a system bus. Among them, the computer-readable storage medium of the computer device stores an operating system, a database, and a computer-readable instruction, and the database may store a control information sequence. When the computer-readable instruction is executed by the processor, the processor can implement a channel estimation optimization method for a base station MIMO wireless communication system. The processor of the computer device is used to provide computing and control capabilities to support the operation of the entire computer device. The memory of the computer device may store computer-readable instructions, and when the computer-readable instructions are executed by the processor, the processor may execute the channel estimation optimization method for the base station MIMO wireless communication system of the present application. The network interface of the computer device is used to connect and communicate with the terminal. It can be understood by those skilled in the art that the structure shown in FIG. 11 is only a block diagram of a partial structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

In this embodiment, the processor is used to execute the specific functions of each module and its submodule in Figure 10, and the memory stores the program code and various data required to execute the above modules or submodules. The network interface is used to transmit data between user terminals or servers. The memory in this embodiment stores the program code and data required to execute all modules/submodules in the channel estimation optimization device of the base station MIMO wireless communication system of the present application, and the server can call the program code and data of the server to execute the functions of all submodules.

The present application also provides a storage medium storing computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the channel estimation optimization method of the base station MIMO wireless communication system described in any embodiment of the present application.

The present application also provides a computer program product, including a computer program/instruction, which, when executed by one or more processors, implements the steps of the channel estimation optimization method for the base station MIMO wireless communication system described in any embodiment of the present application.

A person skilled in the art can understand that all or part of the processes in the above-mentioned embodiments of the present application can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the aforementioned storage medium can be a computer-readable storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

The above description is only a partial implementation method of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

In summary, this application combines the temporal coding capability of pulse neural networks and the generative modeling capability of variational autoencoders (VAEs), which can help understand and optimize channel conditions in complex communication environments, especially in mobile communication systems. This method can improve the accuracy and efficiency of channel estimation, which is of great significance for the performance optimization of MIMO wireless communication systems.

Claims (10)

1. A channel estimation optimization method for a base station MIMO wireless communication system, comprising:

Acquiring channel preambles in respective communication connection requests in response to the respective communication connection requests from the user equipment;

determining pulse release rates of a plurality of neurons corresponding to channel preambles in a preset channel estimation optimization model based on the channel preambles in the communication connection requests, wherein each communication connection request comprises a channel preamble, and each channel preamble corresponds to a plurality of neurons;

In the neuron of the channel estimation optimization model, poisson sampling is carried out according to the pulse release rate by using poisson distribution, a next pulse release time interval is randomly generated, and when the accumulated time interval reaches or exceeds a preset time interval threshold value, the channel pulse code of the channel preamble corresponding to the current time point is recorded;

and inputting the channel pulse codes of the channel preamble codes to a variable self-encoder in the channel estimation optimization model to reconstruct the channel preamble codes so as to complete the channel estimation optimization of the base station MIMO wireless communication system.

2. The channel estimation optimization method of a base station MIMO wireless communication system according to claim 1, wherein the step of determining the burst firing rate of a plurality of neurons corresponding to the channel preamble in a preset channel estimation optimization model based on the channel preamble in the communication connection request comprises:

determining a channel lead code in the communication connection request, and extracting a characteristic value combination containing key information from the channel lead code, wherein the characteristic value combination comprises any multiple items of channel gain, noise level and transmission delay;

Inputting the extracted characteristic value combination into an input layer of a preset channel estimation optimization model, and distributing the input characteristic value combination to a plurality of neurons corresponding to the channel lead codes in the channel estimation optimization model;

Multiplying and summing each characteristic value in the input characteristic value combination with the corresponding weight by adopting an integral pulse issuing function to obtain the membrane potential of each neuron;

And determining the pulse emission rate of each neuron corresponding to the channel preamble according to the membrane potential of each neuron.

3. The channel estimation optimization method of a base station MIMO wireless communication system according to claim 1, wherein in neurons of the channel estimation optimization model, poisson sampling is performed according to the burst release rate using poisson distribution, a next burst release time interval is randomly generated, and when the accumulated time interval reaches or exceeds a preset time interval threshold, a step of recording channel burst coding of the channel preamble corresponding to the current time point includes:

Determining the pulse release rate of each neuron corresponding to the channel preamble in a preset channel estimation optimization model;

randomly generating a next pulse issuing time interval according to the pulse issuing rate by adopting a preset Poisson distribution probability density function;

Accumulating the time intervals of each pulse emission to determine an accumulated time interval, detecting whether the accumulated time interval reaches or exceeds a preset time interval threshold, and recording the channel pulse codes of the channel preamble codes corresponding to the current time point if the accumulated time interval reaches or exceeds the preset time interval threshold;

resetting the accumulated time interval to zero and repeating the above steps.

4. The channel estimation optimization method of a base station MIMO wireless communication system according to claim 1, wherein the step of inputting channel pulse codes of the channel preambles to a variant self-encoder in the channel estimation optimization model to reconstruct the channel preambles comprises:

Mapping, in an encoder of the variational self-encoder, channel pulse codes of the channel preambles to a potential space and outputting probability distributions of potential variables in the potential space;

sampling in a decoder of the variational self-encoder from the probability distribution of the latent variable output by the encoder to determine a specific latent variable so as to generate a reconstructed channel preamble;

Calculating and determining a difference value between probability distribution and prior distribution of potential variables in the potential space to determine a KL divergence loss value, and completing reconstruction of the channel preamble when the KL divergence loss value is lower than a preset KL divergence loss value.

5. The channel estimation optimization method of a base station MIMO wireless communication system according to claim 4, wherein after the step of mapping the channel pulse code of the channel preamble to one potential space in the encoder of the variation self-encoder and outputting the probability distribution of the potential variable in the potential space, comprising:

Sampling from the probability distribution of the latent variable output by the encoder in the latent space to determine a specific latent variable;

and adding Gaussian noise to the specific latent variable, and inputting the specific latent variable added with the Gaussian noise into a decoder of the variable self-encoder to generate a reconstructed channel preamble.

6. The channel estimation optimization method of a base station MIMO wireless communication system according to claim 1, wherein after the step of acquiring the channel preamble in the respective communication connection requests, comprising:

responding to a data preprocessing instruction, and converting a channel lead code in the communication connection request into binary time sequence data;

And inputting the binary time series data corresponding to the channel preamble into a preset channel estimation optimization model to determine the pulse emission rates of a plurality of neurons corresponding to the channel preamble in the preset channel estimation optimization model.

7. The channel estimation optimization method of a base station MIMO wireless communication system according to any one of claims 1 to 6, wherein the burst release rate characterizes the number of times or frequency of generation of bursts per unit time of neurons in a channel estimation optimization model, and the basic network architecture of the channel estimation optimization model is a burst neural network.

8. A channel estimation optimizing apparatus of a base station MIMO wireless communication system, comprising:

A preamble extraction module configured to obtain channel preambles in respective communication connection requests from a user equipment in response to the respective communication connection requests;

The pulse issuing rate determining module is configured to determine pulse issuing rates of a plurality of neurons corresponding to channel preambles in a preset channel estimation optimization model based on the channel preambles in the communication connection requests, wherein each communication connection request comprises one channel preamble, and each channel preamble corresponds to a plurality of neurons;

The pulse code determining module is arranged for carrying out poisson sampling according to the pulse release rate by adopting poisson distribution in neurons of the channel estimation optimization model, randomly generating a time interval of next pulse release, and recording channel pulse codes of the channel preamble corresponding to the current time point when the accumulated time interval reaches or exceeds a preset time interval threshold;

and the channel estimation optimization module is used for inputting the channel pulse codes of the channel preambles to a variable self-encoder in the channel estimation optimization model so as to reconstruct the channel preambles, so as to complete the channel estimation optimization of the base station MIMO wireless communication system.

9. An electronic device comprising a central processor and a memory, characterized in that the central processor is arranged to invoke a computer program stored in the memory for performing the steps of the method according to any of claims 1 to 7.

10. A computer-readable storage medium, characterized in that it stores in the form of computer-readable instructions a computer program implemented according to the method of any one of claims 1 to 7, which, when invoked by a computer, performs the steps comprised by the corresponding method.