Signal communication method and sending end

Abstract

The invention provides a signal communication method and a sending end, the method is applied to the sending end, and the signal communication method comprises the following steps: obtaining a target signal to be transmitted, and carrying out real-time estimation and tracking on initial channel state information through an estimation tracking model to obtain real-time channel state information; selecting the real-time channel state information and the user demand in the to-be-selected frequency bands through a multi-frequency-band joint optimization model, and determining a target frequency band; performing signal processing on the real-time channel state information to obtain a target transmission strategy; the target to-be-transmitted signal is transmitted in the target frequency band based on the target transmission strategy, so that the receiving end receives the target to-be-transmitted signal, the problem of transmission limitation caused by communication in an improper frequency band is avoided, the adaptability to environment change and user requirements in the communication process is improved, and the communication efficiency is improved. And a better transmission mode is determined based on the target transmission strategy, so that the reliability in the signal transmission process in the communication process is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of signal communication, and in particular, to a signal communication method and a transmitting end. Background Art

[0002] In the current field of wireless communication technology, there are multiple frequency bands for communication, especially the millimeter wave band and the low frequency band. Millimeter wave (mmWave) communication refers to wireless communication within the frequency band of 30 GHz to 300 GHz, with a wavelength between 1 millimeter and 10 millimeters. Low frequency band communication usually refers to the frequency band between 3 MHz and 3 GHz. Currently, the use of a single frequency band has been difficult to meet the growing data transmission requirements and diverse application scenarios. The joint use of multiple frequency bands for the millimeter wave band and the low frequency band has received increasing attention. However, how to reasonably select and appropriately choose communication in the low frequency band and the millimeter wave band is still a problem. And due to the inability to appropriately select the frequency band, the current joint use of multiple frequency bands has caused many problems such as resource waste, limited coverage, and limited data transmission rate. Therefore, when the low frequency band and the millimeter wave band are used for joint communication, how to find an appropriate frequency band for communication is an urgent problem to be solved. Summary of the Invention

[0003] The present invention aims to provide a signal communication method and a transmitting end to solve the above technical problems and improve the reliability of signal communication.

[0004] To solve the above technical problems, the present invention provides a signal communication method, which is applied to a transmitting end. The method includes:

[0005] Obtain a target signal to be transmitted, and perform real-time estimation and tracking on the initial channel state information through an estimation and tracking model to obtain real-time channel state information;

[0006] Select the real-time channel state information and user requirements in the candidate frequency bands through a multi-band joint optimization model to determine the target frequency band;

[0007] Perform signal processing on the real-time channel state information to obtain a target transmission strategy;

[0008] Transmit the target signal to be transmitted in the target frequency band based on the target transmission strategy, so that the receiving end receives the target signal to be transmitted.

[0009] In the above solution, the target signal to be transmitted is obtained, and the initial channel state information is estimated and tracked in real time through an estimation and tracking model to obtain real-time channel state information, ensuring the real-time nature of obtaining real-time channel state information and providing an accurate and timely data basis for obtaining the target transmission strategy subsequently; the real-time channel state information and user requirements are selected from the candidate frequency bands through a multi-band joint optimization model to determine the target frequency band, thereby dynamically determining the frequency band for transmitting the signal, avoiding problems such as transmission limitations caused by communicating in inappropriate frequency bands, improving the reliability of signal communication, and enhancing the adaptability to environmental changes and user requirements during the communication process; then, the real-time channel state information is signal-processed to obtain the target transmission strategy, thereby determining a better transmission method, improving the reliability during the signal transmission process in the communication process, and then transmitting the target signal to be transmitted through the target transmission strategy in the target frequency band so that the receiving end can receive the target signal to be transmitted.

[0010] Further, the multi-band joint optimization model includes a first prediction sub-model and a second prediction sub-model. The process of selecting the real-time channel state information and user requirements from the candidate frequency bands through the multi-band joint optimization model to determine the target frequency band includes:

[0011] The channel quality prediction is performed on the scene characteristics, the real-time location information and moving speed of the user through the first prediction sub-model to obtain the channel quality change information of different frequency bands;

[0012] The user service preference prediction is performed on the real-time location information, moving speed and historical application behavior of the user through the second prediction sub-model to obtain the user service preference information;

[0013] Based on the channel quality change information and the user service preference information of different frequency bands, the target frequency band is determined from the candidate frequency bands.

[0014] In the above solution, based on the first prediction sub-model and the second prediction sub-model, the predicted channel quality change information and user service preference information are obtained, so as to obtain the target frequency band that can dynamically adapt to user requirements and the current environment.

[0015] Further, after the process of selecting the real-time channel state information and user requirements from the candidate frequency bands through the multi-band joint optimization model to determine the target frequency band, it includes:

[0016] Based on the spectrum resources of the candidate frequency bands, an objective function is constructed;

[0017] Based on the constraint conditions, the optimal solution of the objective function is obtained, and a multi-band resource allocation scheme is obtained based on the optimal solution;

[0018] Allocate resources for the target frequency band based on the multi-band resource allocation scheme.

[0019] In the above solution, a target function is constructed through the spectrum resources of the candidate frequency bands, and a multi-band resource allocation scheme is obtained based on the optimal solution, improving the utilization rate of spectrum resources and enhancing the overall spectrum efficiency of the system.

[0020] Furthermore, the signal processing of the real-time channel state information to obtain the target transmission strategy includes:

[0021] Perform beamforming processing on the real-time channel state information according to the beamforming algorithm to obtain beamforming parameters;

[0022] Based on the beamforming parameters and the real-time channel state information, adjust the modulation order and channel coding rate to obtain the target modulation order and target channel coding rate;

[0023] Perform signal conversion on the initial signal to be transmitted based on the target modulation order and target channel coding rate to obtain the target signal to be transmitted, so that the receiving end can obtain the target signal to be transmitted.

[0024] In the above solution, the target modulation order and target channel coding rate are obtained through the beamforming parameters and the real-time channel state information, thereby performing signal conversion on the initial signal to be transmitted, improving the reliability of the signal during transmission.

[0025] Furthermore, the signal processing of the real-time channel state information to obtain the target transmission strategy includes:

[0026] Obtain the current transmission state information;

[0027] Based on the real-time channel state information and the current transmission state information, construct a retransmission risk assessment model;

[0028] Evaluate the target signal to be transmitted based on the retransmission risk assessment model to obtain a retransmission strategy;

[0029] Transmit the target signal to be transmitted based on the retransmission strategy.

[0030] In the above solution, by determining whether the target signal to be transmitted needs to be retransmitted, the probability of transmission failure is reduced and the reliability of communication is improved.

[0031] Furthermore, the signal processing of the real-time channel state information to obtain the target transmission strategy includes:

[0032] Perform path selection on the real-time channel state information through a dynamic routing algorithm to obtain a target path;

[0033] Transmit the target signal to be transmitted based on the target path.

[0034] In the above solution, the real-time channel state information is used for path selection through a dynamic routing algorithm to obtain a target path, thereby improving the efficiency and security of transmitting the target signal to be transmitted.

[0035] Further, the signal processing of the real-time channel state information to obtain a target transmission strategy includes:

[0036] Obtain performance parameters;

[0037] Adjust the preset transmission parameters based on the performance parameters to obtain target transmission parameters;

[0038] Transmit the target signal to be transmitted based on the target transmission parameters.

[0039] In the above solution, the target signal to be transmitted is transmitted based on the target transmission parameters, thereby improving the flexibility of transmission.

[0040] Further, the obtaining of the target signal to be transmitted includes:

[0041] Encode the target signal to be transmitted through a preset space-time coding matrix so that the receiving end can perform signal recovery and interference suppression on the encoded target signal to be transmitted based on a preset decoding algorithm.

[0042] In the above solution, the interference between multiple frequency bands is reduced through the space-time coding matrix.

[0043] Further, the obtaining of the target signal to be transmitted includes:

[0044] Use an iterative optimization algorithm to optimize the transmission beam direction of the transmitting end and the receiving beam direction of the receiving end so that the interference signals in the target frequency band are aligned at the receiving end.

[0045] In the above solution, the interference between multiple frequency bands is reduced through the iterative optimization algorithm.

[0046] The present invention also provides a transmitting end for performing the signal communication method as described above. The transmitting end includes a physical layer, a link layer, a network layer, and an application layer, where: the physical layer is used to obtain a target signal to be transmitted, and perform real-time estimation and tracking on the initial channel state information through an estimation and tracking model to obtain real-time channel state information; the physical layer is used to select from candidate frequency bands the real-time channel state information and user requirements through a multi-band joint optimization model to determine a target frequency band; the link layer, the network layer, and the application layer are used to perform signal processing on the real-time channel state information to obtain a target transmission strategy; the application layer is used to transmit the target signal to be transmitted based on the target frequency band and the target transmission strategy, so that the receiving end receives the target signal to be transmitted.

[0047] In the above solution, obtaining the target signal to be transmitted and performing real-time estimation and tracking on the initial channel state information through an estimation and tracking model to obtain real-time channel state information ensures the real-time nature of obtaining real-time channel state information, providing an accurate and timely data basis for subsequently obtaining a target transmission strategy; selecting from candidate frequency bands the real-time channel state information and user requirements through a multi-band joint optimization model to determine a target frequency band, thereby dynamically determining the frequency band for transmitting the signal, avoiding problems such as transmission limitations caused by communicating in an inappropriate frequency band, improving the reliability of signal communication, and enhancing the adaptability to environmental changes and user requirements during the communication process; then performing signal processing on the real-time channel state information to obtain a target transmission strategy, thereby determining a better transmission method, improving the reliability during the signal transmission process in the communication process, and further transmitting the target signal to be transmitted through the target transmission strategy in the target frequency band so that the receiving end receives the target signal to be transmitted. Brief Description of the Drawings

[0048] Figure 1 It is a schematic flowchart of a signal communication method provided by an embodiment of the present invention;

[0049] Figure 2 It is a schematic diagram of the architecture of a transmitting end provided by an embodiment of the present invention. Detailed Embodiments

[0050] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0051] Please refer to Figure 1, this embodiment provides a signal communication method, which is applied to a sending end. The method includes the following steps:

[0052] Step S1: Obtain a target signal to be transmitted, and perform real-time estimation and tracking on the initial channel state information through an estimation and tracking model to obtain real-time channel state information;

[0053] Step S2: Select the real-time channel state information and user requirements in the candidate frequency bands through a multi-band joint optimization model to determine the target frequency band;

[0054] Step S3: Perform signal processing on the real-time channel state information to obtain a target transmission strategy;

[0055] Step S4: Transmit the target signal to be transmitted in the target frequency band based on the target transmission strategy, so that the receiving end receives the target signal to be transmitted.

[0056] In the above solution, obtaining the target signal to be transmitted and performing real-time estimation and tracking on the initial channel state information through the estimation and tracking model to obtain real-time channel state information ensures the real-time nature of obtaining real-time channel state information and provides an accurate and timely data basis for obtaining the target transmission strategy subsequently; selecting the real-time channel state information and user requirements in the candidate frequency bands through the multi-band joint optimization model to determine the target frequency band, thereby dynamically determining the frequency band for transmitting the signal, avoiding problems such as transmission limitations caused by communicating in inappropriate frequency bands, improving the reliability of signal communication, and enhancing the adaptability to environmental changes and user requirements during the communication process; then performing signal processing on the real-time channel state information to obtain the target transmission strategy, thereby determining a better transmission method, improving the reliability during the signal transmission process in the communication process, and further transmitting the target signal to be transmitted in the target frequency band through the target transmission strategy so that the receiving end receives the target signal to be transmitted.

[0057] In another embodiment, the multi-band joint optimization model includes a first prediction sub-model and a second prediction sub-model. Selecting the real-time channel state information and user requirements in the candidate frequency bands through the multi-band joint optimization model to determine the target frequency band includes:

[0058] Perform channel quality prediction on the scenario characteristics, the real-time location information and moving speed of the user through the first prediction sub-model to obtain channel quality change information of different frequency bands;

[0059] Perform user service preference prediction on the real-time location information, moving speed and historical application behavior of the user through the second prediction sub-model to obtain user service preference information;

[0060] Based on the channel quality change information of different frequency bands and the user service preference information, determine the target frequency band in the candidate frequency bands.

[0061] It should be noted that through pilot signals and channel estimation algorithms, the channel state information (CSI) in the low-frequency and millimeter-wave frequency bands is collected in real time, including parameters such as channel gain, noise level, and interference intensity. Then, statistical analysis and machine learning methods are used to process and analyze the collected CSI data to evaluate the current channel quality change information in each frequency band. Moreover, according to the user's service type (such as video streaming, real-time communication, data download, etc.), the requirements for bandwidth, latency, and reliability of different services are determined; based on the user's location information and moving speed, the channel quality change information and user service preference information of the user in different frequency bands are predicted. In a multi-band wireless communication system, the channel characteristics and service requirements change dynamically with the user's location and movement state. Therefore, by combining geographical location information, moving speed, and trajectory trends, etc., the channel availability and service requirements of the user in different frequency bands can be predicted, thereby supporting resource scheduling, frequency band switching, and QoS optimization. Exemplarily, a change in location will affect occlusion, path loss, and the number of visible base stations, especially in the millimeter-wave frequency band; speed will determine the rapid variability of the channel, Doppler shift, and access / handover frequency; the moving direction determines whether to enter an occluded area, a coverage boundary, or a cross-band area; application habits, such as using video more when stationary and using voice or caching more when moving at high speed.

[0062] Specifically: (1) Collect location information and speed: Obtain the user's current three-dimensional location, such as GPS, Beidou, base station positioning, etc.; obtain the user's speed and direction, such as the terminal IMU, network estimation, so as to output the current geographical location coordinates and motion vectors, such as speed, acceleration, direction angle, etc. (2) Match the geography and scene model: Match the user's location to the environmental map or scene database (such as streets, buildings, indoor / outdoor labels); then judge whether the current is a LOS / NLOS area, and judge whether the millimeter-wave frequency band is available and whether it is easily occluded; (3) Predict the channel quality change information by frequency band: According to the user's current speed / trajectory and combined with the scene characteristics, predict the channel quality change trend in each frequency band in the next period of time. Among them, the key indicators include: the predicted channel gain (Path Loss); the SNR / SINR trend; possible handover requirements (such as about to enter the NLOS area); The channel quality change information is mainly predicted by the following methods: sliding window average and difference, specifically: ΔSNR mmW = SNR t - SNR t-w . Among them, SNR t : The signal-to-noise ratio (Signal-to-Noise Ratio) measured at time t, which reflects the current channel quality; SNR t-w: The signal-to-noise ratio measured at time t - w, where w is the window length (number of time slots or seconds), reflecting the channel quality a short time ago; ΔSNR mmW : The difference between two window points - the increment of the signal-to-noise ratio, indicating whether the channel quality improves or deteriorates during this period. Among them: The moving window average is to perform a mean operation on a time series (such as the SNR measured in each time slot) with a window of length w; the difference (Δ) is to take the difference between the current window mean and the previous window mean: equivalent to the (rate of change) of the channel quality, indicating whether the channel is getting better (Δ > 0) or worse (Δ < 0). Since relying solely on the instantaneous SNR may be misled by occasional noise or measurement errors, using the differential trend can more reliably indicate that the channel is continuously deteriorating, facilitating early band switching or power adjustment to avoid communication interruption; in addition, relying solely on the instantaneous SNR may be misled by occasional noise or measurement errors, while the differential trend can more reliably indicate that the channel is continuously deteriorating, facilitating early band switching or power adjustment to avoid communication interruption. Further, before the network or link layer makes retransmission, modulation switching, or routing decisions, it can first look at ΔSNR mmW :

[0063] If ΔSNR mmW << 0 (substantial decrease), immediately activate the backup band or retransmission mechanism;

[0064] If ΔSNR mmW >> 0 (continuous improvement), consider upgrading the modulation method or reducing redundancy.

[0065] (4) Prediction of user service preference information: Combine the current speed, location type, and historical application behavior, and predict the user's service preference in the next period through the second prediction sub-model; Exemplarily, when the user is driving on the highway, the probability of video call is low, while the probability of voice navigation or cache download is high; When the user enters the subway, it is very likely to switch to offline cache; When the user is stationary in the office, the demand for video conferencing or cloud synchronization increases; Therefore, construct a scenario-service mapping table and perform statistical modeling on application historical behavior; (5) Output comprehensive decision: Output the channel quality change information and user service preference information for each frequency band; for modules such as resource scheduling / frequency band selection / power allocation to call. In summary, early perception of channel availability: Avoid frequent handovers, improve system stability and spectrum utilization; Enhancement of service adaptability: Match the actual usage scenario and dynamically optimize the QoS configuration; More accurate resource scheduling: Provide predictive data for frequency band selection, beamforming, and power control. Among them, for (3), more specifically: To ensure the accuracy of subsequent analysis, first perform data preprocessing, and the operations include: Denoising: Use methods such as moving average filtering and median filtering to eliminate high-frequency noise; Missing value filling: If some CSI samples are missing, they can be filled with interpolation or historical mean; Normalization: Scale CSI parameters (such as amplitude, phase) to a unified interval (such as [0,1]); Frequency domain conversion: Convert time-domain CSI data to the frequency domain for frequency response analysis. After that, to extract statistical features or physical properties that can represent the channel state from the original CSI data, perform feature extraction. Typical features include: Channel Gain: Represents the transmission ability on each subcarrier or frequency point; Instantaneous Signal-to-Noise Ratio (SNR): Estimate the instantaneous signal strength using CSI and noise power; Delay Spread: Reflects the severity of the multipath effect; Frequency Selectivity: Measures the fluctuation of the channel response in the frequency domain; Spatial Correlation (for MIMO): Measures the correlation degree of signals received by different antennas; Path Loss Estimation: Estimate by combining location data and the average gain in CSI; Phase Stability Index: Evaluate the intensity of phase change in high-frequency (such as millimeter wave) channels. After that, to convert the analysis results into a ranking or scoring system that can be used for algorithm decision-making, quantify the channel quality level assessment. The methods include: Set thresholds or scoring models as the first prediction sub-model, and divide levels according to each feature index: Exemplarily, SNR < 10dB is a weak channel; 10dB ≤ SNR ≤ 20dB is a medium channel; SNR > 20dB is an excellent channel; Construct the first prediction sub-model:

[0066]

[0067] where Q is the model, SNR is the instantaneous signal-to-noise ratio, Delay Spread is the delay spread, and α iThe Stability Index is a weight that is adjusted according to the scenario. It indicates the degree of fluctuation or stability of the channel quality over time and is used to measure whether the channel is stable and reliable or jitters violently in a short time. The Stability Index is used to calculate the correlation coefficient ρ between two or more adjacent CSI (or SNR) samples. The formula for obtaining it is as follows:

[0068]

[0069] Among them, Cov() is the existing covariance function, t represents the time, SNR is the instantaneous signal-to-noise ratio, is the standard deviation of the signal-to-noise ratio samples at time t, is the standard deviation of the signal-to-noise ratio samples at time t-1. The closer ρ is to 1, the more stable the channel changes in adjacent time slots; the closer it is to 0, the more violent the channel fluctuations. StabilityIndex can reflect the jitter of the channel in a short period of time, which is of great reference value for frequency band switching, beam adjustment and retransmission decisions. The first prediction sub-model is used to quantify the current channel quality change information of each frequency band.

[0070] In another embodiment, the real-time channel state information and user requirements are selected from the candidate frequency bands through a multi-frequency band joint optimization model, and after the target frequency band is determined, the method includes:

[0071] Based on the spectrum resources of the selected frequency band, an objective function is constructed;

[0072] Based on the constraints, the optimal solution of the objective function is obtained, and a multi-band resource allocation scheme is obtained based on the optimal solution;

[0073] Resources are allocated to the target frequency band based on a multi-frequency band resource allocation scheme.

[0074] It should be noted that the priority of low-frequency and millimeter wave bands is set according to the channel quality change information and user service preference information. For example, low-frequency bands are preferred in the case of high user density or low channel quality, while millimeter wave bands are preferred in the case of low user density and high channel quality. Based on the priority setting and real-time channel status information, the frequency band to be used for current communication is dynamically determined, and the frequency band is switched when necessary to continuously maintain good communication performance.

[0075] Frequency band selection can be viewed as a dynamic optimization problem, whose goal is to improve the overall spectrum efficiency of the system and user experience. Assume the following notation: S(t): the set of frequency bands (low frequency or millimeter wave) selected at time t. The optimization objective can be expressed as:

[0076]

[0077] in:

[0078] K is the number of users.

[0079] R L (t) and R M (t) are the transmission rates in the low - frequency and millimeter - wave frequency bands respectively.

[0080] α k and β k are the weight coefficients of user k for the low - frequency and millimeter - wave frequency band services respectively, reflecting its requirements for different service qualities. The constraints include spectrum resource limitations, interference management requirements, and the minimum interval time for frequency band switching, etc.

[0081] Furthermore, in order to optimize the spectrum utilization rate between multiple frequency bands and reduce the interference between frequency bands, a resource allocation strategy is proposed. This strategy is based on optimization theory and machine learning techniques, and dynamically adjusts the resource allocation ratio of different frequency bands to adapt to real - time communication requirements and channel conditions. Specifically:

[0082] (1) Resource allocation model construction:

[0083] Spectrum resource definition: The available spectrum resources are divided into multiple sub - bands, and a certain bandwidth and power are allocated to each sub - band.

[0084] User demand modeling: According to the service type and data requirements of users, a user demand model is established to determine the resource requirements of each user in different frequency bands.

[0085] (2) Optimization algorithm design:

[0086] Optimization objective: Maximize the spectrum efficiency and user satisfaction of the system, and minimize interference and energy consumption.

[0087] Constraints: Include the total spectrum resource limitation, the minimum service quality requirement for each user, the interference tolerance between frequency bands, etc.

[0088] Algorithm selection: Heuristic algorithms and machine learning algorithms (such as deep reinforcement learning) are used for resource allocation optimization.

[0089] (3) Interference management mechanism:

[0090] Frequency band isolation: By frequency planning and power control, reduce the mutual interference between different frequency bands.

[0091] Interference coordination technology: Use technologies such as interference alignment and interference cancellation to further reduce the interference impact between frequency bands.

[0092] (4) Dynamic adjustment and real - time optimization:

[0093] Real - time monitoring: Continuously monitor the channel state and user demand changes in each frequency band.

[0094] Dynamic adjustment: Based on the real-time monitoring results, dynamically adjust the resource allocation ratio to ensure the maximization of spectrum utilization and the minimization of interference.

[0095] Therefore, based on the optimization objectives, the resource allocation strategy for multiple frequency bands is constructed as follows:

[0096]

[0097] Where:

[0098] b L (t) and b M (t) are the bandwidths allocated to the low-frequency and millimeter-wave frequency bands at time t, respectively.

[0099] P L (t) and P M (t) are the powers allocated to the low-frequency and millimeter-wave frequency bands at time t, respectively.

[0100] h L,k (t) and h M,k (t) are the channel gains of user k in the low-frequency and millimeter-wave frequency bands, respectively.

[0101] N0 is the noise power spectral density.

[0102] B total and P total are the total system bandwidth and total power limit, respectively.

[0103] QoS k (t) is the quality of service requirement of user k at time t, and Q k is its minimum quality of service requirement.

[0104] Interference(t) is the total interference level of the system at time t, and I max is the upper limit of the interference tolerance.

[0105] By solving the above optimization problem under the constraint conditions, the bandwidths and powers to be allocated to the low-frequency and millimeter-wave frequency bands at time t can be determined to ensure the maximization of the system's spectrum utilization rate while meeting the quality of service requirements of users and the interference management requirements.

[0106] Furthermore, the resource allocation problem can be formulated as the following objective function:

[0107]

[0108] Where:

[0109] I L (t) and I M (t) are the interference levels of the low-frequency and millimeter-wave frequency bands at time t, respectively.

[0110] I L,max and I M,max are the upper limits of interference tolerance in the low - frequency and millimeter - wave frequency bands, respectively.

[0111] Use a genetic algorithm (GA) or a deep reinforcement learning algorithm (Deep Reinforcement Learning, DRL) to solve the decomposed sub - problems, find an approximate optimal solution, and obtain a multi - band resource allocation scheme based on the optimal solution;

[0112] In summary, the core of resource allocation optimization lies in maximizing the system's spectral efficiency and user satisfaction under the premise of meeting the constraint conditions. By introducing a throughput model in the form of a logarithmic function, the complex multi - band resource allocation problem can be transformed into a convex optimization problem, and thus can be solved using existing optimization algorithms.

[0113] Specific methods for solving include, by way of example, introducing the constraint conditions into the objective function by constructing a Lagrangian function:

[0114]

[0115] By taking the partial derivatives of the Lagrangian function with respect to b L (t), b M (t), P L (t), P M (t) and setting them to zero, the necessary conditions for the optimal solution can be obtained.

[0116] Alternatively, using the dual problem and the Karush - Kuhn - Tucker (KKT) conditions, the optimal strategy for resource allocation can be further solved to ensure that all constraint conditions are met.

[0117] The collaborative communication mechanism and multi-band resource allocation strategy based on low frequency and millimeter wave can achieve the following effects: Spectrum utilization improvement: By means of intelligent frequency band selection and efficient resource allocation, the utilization rate of spectrum resources is maximized, and the overall spectrum efficiency of the system is improved. Interference reduction: Optimize spectrum allocation and interference management, significantly reduce interference between different frequency bands, and improve signal quality and system stability. Enhanced dynamic adaptation ability: Based on real-time CSI and user requirements, dynamically adjust frequency band selection and resource allocation, and enhance the system's adaptability to environmental changes and user needs. User experience optimization: Meet the quality of service requirements of different users and improve the communication experience and satisfaction of users. In practical applications, for example, in a high-density urban environment, the low frequency band can be used for wide coverage and communication that can penetrate obstacles, while the millimeter wave band can be used for high data rate hot spots. Through the collaborative communication mechanism of this case, the system can dynamically allocate resources to different frequency bands according to the real-time channel conditions and user requirements, ensuring the efficient operation and high-quality service of the overall communication system.

[0118] In another embodiment, signal processing is performed on the real-time channel state information to obtain a target transmission strategy, including:

[0119] Perform beamforming processing on the real-time channel state information according to the beamforming algorithm to obtain beamforming parameters;

[0120] Based on the beamforming parameters and the real-time channel state information, adjust the modulation order and channel coding rate to obtain a target modulation order and a target channel coding rate;

[0121] Perform signal conversion on the initial signal to be transmitted based on the target modulation order and the target channel coding rate to obtain a target signal to be transmitted, so that the receiving end can obtain the target signal to be transmitted.

[0122] It should be noted that the signal communication method is applied to the sending end. The sending end includes a physical layer, a link layer, a network layer, and an application layer. The physical layer obtains the target signal to be transmitted, and performs real-time estimation and tracking on the initial channel state information through an estimation and tracking model to obtain real-time channel state information; selects a target frequency band from the candidate frequency bands for the real-time channel state information and user requirements through a multi-band joint optimization model. Joint beamforming and modulation coding aims to utilize the spatial channel information and beam control capabilities provided by the physical layer to optimize the data modulation and coding strategies of the link layer, thereby improving data transmission efficiency and transmission reliability. Traditional link layer designs usually design modulation and coding strategies independently of the beamforming of the physical layer, while this application realizes the joint optimization of the two through cross-layer information sharing. Specifically, the physical layer uses a large-scale antenna array and a beamforming algorithm to determine beamforming parameters according to the real-time channel state information (CSI). The beamforming parameters include beam direction and beam shape. Build a model, let w iDenote the beamforming vector of the \(i\)-th user. The goal is to maximize the signal-to-noise ratio (SNR), and we get:

[0123]

[0124] where \(h\) i is the channel vector of user \(i\), and \(\sigma\) 2 is the noise power. For user \(i\) (or stream \(i\)), \(h\) i is the channel vector of user \(i\), and \(w\) i is the beamforming vector designed by the transmitter for the \(i\)-th user.

[0125] For the index \(j eq i\), it traverses all other users (or streams) in the system except the \(i\)-th user. Each \(w\) j is the beam vector transmitted to user \(j\). The numerator \(|h\) i \(w\) i |\) 2 represents the useful signal power received by user \(i\) under its own beam; the \(\sum\) in the denominator j≠i \(|h\) i \(w\) j |\) 2 is exactly the sum of the multi-user interference powers caused by the beams of "all other" users to user \(i\). Through the term \(\sum\) j≠i when optimizing the \(i\)-th beam \(w\) i , all the interferences generated by all other simultaneously transmitted beams \(\{w\) j \}\) j≠i on the \(i\)-th user are taken into account, so as to achieve interference suppression and SNR maximization in a multi-user environment.

[0126] After that, the physical layer sends the real-time channel state information to the link layer. According to the beamforming parameters output by the physical layer and the real-time channel quality, the modulation order (such as QPSK, 16-QAM, 64-QAM, etc.) and the channel coding rate are dynamically adjusted to adapt to different channel conditions and ensure a balance between high data rate and low bit error rate. It can be understood that modulation is the process of converting digital data or analog data into analog signals. Coding is the process of converting digital data into digital signals. Therefore, the initial signal to be transmitted is converted using the adjusted modulation order and channel coding rate to obtain the target signal to be transmitted. The optimization goal can be measured by system throughput or bit error rate (BER). For example, the system throughput \(T\) can be expressed as:

[0127]

[0128] where \(B\) is the bandwidth, \(I\) i is the interference power, and \(\eta\) is the coding efficiency.

[0129] The physical layer and the link layer achieve joint decision-making by sharing real-time CSI and beamforming parameters. For example, by feeding back the SNR estimation of the physical layer to the link layer, the link layer can select a more suitable modulation and coding scheme in real time. An alternating iterative optimization method is adopted: first, the link layer strategy is fixed to optimize the physical layer beamforming; then, the physical layer beam parameters are fixed, and the link layer modulation and coding are adjusted according to the feedback until the system performance converges to the optimal solution.

[0130] In summary, joint optimization enables the link layer to select a more appropriate modulation and coding for the actual channel conditions, avoiding the performance degradation of traditional fixed schemes in dynamic channels; the precise beamforming of the physical layer can improve signal gain and reduce interference; at the same time, the link layer adaptive modulation and coding ensure that a low bit error rate can still be maintained under low SNR conditions; the cross-layer optimization framework can respond to channel state changes in real time, dynamically adjust the physical and link layer parameters, and improve the overall performance of the system.

[0131] In another embodiment, signal processing is performed on the real-time channel state information to obtain a target transmission strategy, including:

[0132] Obtain the current transmission state information;

[0133] Based on the real-time channel state information and the current transmission state information, construct a retransmission risk assessment model;

[0134] Based on the retransmission risk assessment model, evaluate the target signal to be transmitted to obtain a retransmission strategy;

[0135] Transmit the target signal to be transmitted based on the retransmission strategy.

[0136] It should be noted that the link layer adaptive retransmission mechanism mainly uses the real-time channel state information provided by the physical layer to dynamically adjust the data packet retransmission strategy, thereby reducing the data packet loss rate and improving the system throughput and communication reliability. This mechanism includes the following steps:

[0137] (1) Real-time channel state feedback:

[0138] After the physical layer completes beamforming and channel estimation, it generates accurate CSI and feeds it back to the link layer in real time. This feedback information includes parameters such as the SNR, interference level, and beam direction of the current channel.

[0139] CSI feedback can be transmitted through a dedicated control channel to ensure the real-time and accuracy of the information.

[0140] (2) Adaptive retransmission strategy design:

[0141] Based on the received CSI information, the link layer dynamically determines the reliability of the current data transmission. If a decrease in channel quality or an increase in the bit error rate is detected, the retransmission mechanism is automatically activated.

[0142] Adopt an adaptive ARQ (Automatic Repeat reQuest) mechanism, and dynamically adjust the retransmission interval and the number of retransmissions according to the channel state. Set the relationship between a retransmission factor β and the current channel SNR:

[0143] β = f(SNR). When SNR is low, increase the number of retransmissions or shorten the retransmission interval; conversely, when SNR is high, reduce retransmissions to reduce latency.

[0144] (3) Optimize retransmission parameters:

[0145] The optimization goal of the retransmission strategy is to maximize the probability of successful packet transmission while minimizing the latency caused by retransmissions as much as possible. Dynamic programming or reinforcement learning algorithms can be used to automatically adjust the retransmission parameters based on historical retransmission success rates and channel states.

[0146] For example, by optimizing the following objective function:

[0147]

[0148] where Delay t is the latency at time t, PacketLoss t is the packet loss rate, and α and β are weighting factors.

[0149] Set up a decision-making module at the link layer to comprehensively evaluate the real-time channel state information fed back by the physical layer and the current data transmission situation, and generate a retransmission decision signal. Specifically:

[0150] (1) Obtain the real-time channel state information fed back by the physical layer, including: instantaneous SNR / SINR: reflecting the current channel quality; channel gain attenuation trend (ΔCSI): judging whether the channel is stable or deteriorating; interference intensity estimation: identifying whether there is a risk of adjacent-frequency or co-frequency interference; reception quality indicators (such as EVM, CQI, BLER): assisting in evaluating the transmission success rate.

[0151] (2) Collect the current transmission state information of the link layer, including: ACK / NACK status of the already sent data packets; transmission delay and round-trip time (RTT) of the current packet; packet error rate (PER) and the number of retransmissions;

[0152] current modulation and coding scheme (MCS) level; congestion window and cache queue status (such as TCP window).

[0153] (3) Based on the real-time channel state information and the current transmission state information, construct a retransmission risk assessment model (RiskScore):

[0154]

[0155] R: Retransmission risk assessment model;

[0156] α i : Weight parameter, which can be dynamically adjusted according to network policies;

[0157] If R≥θ, it is considered that there is a high risk of error in the transmission, and pre-triggered retransmission is recommended.

[0158] Multiple retransmission policies can be set according to different evaluation results: Exemplarily, for low risk, continue to wait for ACK; for medium risk, reduce the MCS level; for high risk, immediately initiate pre-retransmission or request new resources.

[0159] After determining retransmission, it includes the following forms: immediately retransmit the current packet; suspend and continue to listen for feedback; retransmit after reducing the modulation level; request frequency band switching for retransmission; transfer the signal to the link layer ARQ module or scheduler.

[0160] (4) Linkage feedback mechanism, sending back to the physical layer: request to switch beam / power; sending feedback to the network layer: determine whether to change the route or shunt; record statistics: used for dynamic learning and threshold update (such as using RL self-learning to adjust the scoring model).

[0161] A dynamic adjustment mechanism can also be constructed: use a machine learning model to train historical data and predict retransmission parameters suitable for the current channel conditions under different channel conditions. The system periodically evaluates and updates the retransmission mechanism to adapt to the changing channel conditions. Use the reinforcement learning algorithm to construct a state-action-reward model and continuously optimize the retransmission strategy so that the system can automatically select appropriate retransmission parameters in different environments.

[0162] In summary, the retransmission strategy can be adjusted in real time according to the channel state, ensuring that data can be efficiently retransmitted under poor channel conditions, reducing the packet loss rate; when the channel conditions are good, reducing unnecessary retransmissions, reducing latency, and improving the overall system throughput; cross-layer information sharing enables the link layer to respond in a timely manner to physical layer channel changes, enhancing the adaptability of the entire communication system to the dynamic environment.

[0163] In another embodiment, signal processing is performed on the real-time channel state information to obtain a target transmission strategy, including:

[0164] Perform path selection on the real-time channel state information through a dynamic routing algorithm to obtain a target path;

[0165] Transmit the target signal to be transmitted based on the target path.

[0166] It should be noted that in cross-layer optimization design, in order to further improve the performance of the overall communication system, it is necessary not only to optimize the physical layer and the link layer, but also to conduct adaptive design for the network layer and the application layer to achieve collaborative optimization at a high level for the entire system. Specifically, it includes: network layer routing optimization based on real-time channel state information, using the real-time channel state information (CSI) provided by the physical layer to dynamically adjust the data routing strategy of the network layer. This strategy monitors the channel quality of each link in real time, selects a better transmission path, thereby reducing the packet delay and packet loss rate, and improving the transmission efficiency and stability of the entire network. Specifically:

[0167] (1) Acquisition and processing of real-time channel state information:

[0168] CSI acquisition: The physical layer regularly performs channel estimation and feeds back the CSI data between each node to the network layer.

[0169] Data preprocessing: Filter, smooth, and normalize the real-time channel state information to eliminate noise and outliers, ensuring the accuracy and stability of the data.

[0170] (2) Optimization of routing strategy:

[0171] Path selection algorithm: Based on real-time channel state information, adopt dynamic routing algorithms (such as the shortest path algorithm, dynamic weighted routing algorithm) for path selection.

[0172] Exemplarily: Suppose there are multiple paths in the network, and the delay D p and packet loss rate L p of each path p can be estimated through the CSI of each link:

[0173] D p =∑ (i,j)∈p d ij ,L p =1 - ∏ (i,j)∈p (1 - l ij );

[0174] where d ij is the delay of link (i, j), and l ij is the packet loss rate of link (i, j).

[0175] Objective function: The goal of routing optimization is to select a target path that minimizes the delay and packet loss rate:

[0176]

[0177] where ω1 and ω2 are weighting factors, reflecting the importance of delay and packet loss rate to the system performance.

[0178] (3) Dynamic routing update:

[0179] Real-time feedback mechanism: The network layer receives the physical layer CSI updates in a periodic or event-driven manner and adjusts the routing table in real time to ensure that data packets are always transmitted along the current optimal path.

[0180] Load balancing: In addition to single-path routing, multi-path transmission and load balancing technologies can also be adopted to disperse the data stream onto multiple preferred paths, further reducing the congestion risk of a single path.

[0181] In summary, the CSI-based routing optimization can effectively reduce the transmission delay and packet loss rate of data packets, improve the robustness and stability of network transmission, and adapt to the rapidly changing channel conditions in the wireless environment.

[0182] In another embodiment, signal processing is performed on the real-time channel state information to obtain a target transmission strategy, including:

[0183] Obtain performance parameters;

[0184] Adjust the preset transmission parameters based on the performance parameters to obtain target transmission parameters;

[0185] Transmit the target signal to be transmitted based on the target transmission parameters.

[0186] It should be noted that according to the network layer performance feedback (such as end-to-end delay, packet loss rate, throughput, etc.), the application layer parameters are dynamically adjusted to obtain target transmission parameters to improve the user experience and application performance. This mechanism realizes the collaborative interaction between the application layer and the network layer, enabling the application layer to adapt to the current network conditions. Specifically:

[0187] (1) Performance monitoring and feedback:

[0188] Network performance metric collection: Continuously monitor performance parameters such as end-to-end data transmission delay, packet loss rate, and bandwidth utilization in the network layer.

[0189] Feedback channel: Feed the real-time network performance data back to the application layer through a standardized interface or control signaling.

[0190] (2) Dynamically adjust the parameters to obtain target transmission parameters:

[0191] Adaptive data transmission rate: Dynamically adjust the data transmission rate of the application layer according to the current network bandwidth and delay. For example, reduce the rate when the network is congested to reduce packet loss; increase the rate when the network condition is good to optimize the user experience.

[0192] QoS Parameter Adjustment: Adjust the QoS parameters (such as latency requirements, jitter tolerance, data priority, etc.) at the application layer to match the current transmission capacity and quality of service of the network layer. For example, in a real-time video conference, ensure that the latency and packet loss rate are within an acceptable range.

[0193] Algorithm Implementation: The application layer can utilize the feedback data to adjust the transmission rate and QoS parameters through control theory or machine learning methods (e.g., an adaptive algorithm based on PID control or deep reinforcement learning).

[0194] Exemplarily: Let the transmission rate of the application layer be R(t), and the goal is to make the actual throughput T(t) close to the maximum available throughput T fed back by the network layer max , and the following objective function can be constructed: And update R(t) through real-time feedback to achieve the goal. It can be understood that the target transmission parameters include data transmission rate, QoS parameters, etc.

[0195] (3) Closed-loop Control System:

[0196] Real-time Adjustment and Monitoring: Construct a closed-loop control system to achieve dynamic adjustment at the application layer while continuously monitoring the network performance, forming a feedback loop to ensure that the system can quickly respond to changes in the network state.

[0197] Multi-dimensional Feedback Mechanism: Comprehensively consider multiple metrics such as latency, packet loss rate, and throughput for multi-objective optimization to ensure that the adjusted parameters can overall improve the application performance.

[0198] In summary, by dynamically adjusting the application layer parameters, the system can automatically optimize the data transmission strategy during network congestion and network quality fluctuations, providing a smoother user experience and more efficient application performance, especially for real-time applications (such as video conferencing, online games) with significant advantages.

[0199] In another embodiment, obtaining the target signal to be transmitted includes:

[0200] Encoding the target signal to be transmitted through a preset space-time coding matrix so that the receiving end can perform signal recovery and interference suppression on the encoded target signal to be transmitted based on a preset decoding algorithm.

[0201] In another embodiment, obtaining the target signal to be transmitted includes:

[0202] Using an iterative optimization algorithm to optimize the transmit beam direction at the transmitting end and the receive beam direction at the receiving end so that the interference signals in the target frequency band are aligned at the receiving end.

[0203] It should be noted that during the process of obtaining the target signal to be transmitted, frequency band interference suppression needs to be carried out. The interference suppression technology aims to effectively reduce the interference impact between different frequency bands through advanced algorithms and signal processing methods, and improve the signal quality and system capacity. This application adopts advanced interference suppression technologies such as Interference Alignment (IA) and Space-Time Coding (STC), combined with intelligent beamforming and multi-antenna technologies, to achieve efficient interference management and suppression. Specifically:

[0204] (1) Interference Alignment (IA):

[0205] Interference Alignment makes all interference signals occupy the same spatial dimension at the receiving end by designing the spatial dimension of the transmitted and received signals, so as to achieve interference-free signal reception in other dimensions. The implementation method is as follows: Design the channel matrix: Use multi-antenna technology and beamforming to design the interference channel matrix so that the interference signals are aligned at the receiving end; Adopt an iterative optimization algorithm (such as the Min-Sum Interference Alignment algorithm) to optimize the transmitted and received beam directions to achieve interference alignment. More specifically: Suppose multiple users share the same frequency band, and the goal of interference alignment is to make the interference signals occupy the same subspace at the receiving end. Mathematically, let the received matrix H of the interference signal I satisfy: rank(H I ) ≤ d; where d is the dimension of the interference signal at the receiving end, and the alignment of the interference signals is achieved by optimizing the transmitted and received matrices.

[0206] (2) Space-Time Coding (STC): Space-Time Coding improves the reliability and anti-interference ability of the signal by introducing coding in the spatial and time dimensions at the transmitting and receiving ends. The implementation method is as follows: Design the coding matrix: Design an appropriate space-time coding matrix (such as Alamouti coding, spatial multiplexing coding), and encode the signal at the transmitting end; At the receiving end, adopt the corresponding decoding algorithm (such as maximum likelihood decoding, linear decoding) to recover the original signal and suppress interference. More specifically: Let the transmitted signal matrix be X, the received signal matrix be Y, the interference signal matrix be I, and the noise matrix be N, then the received signal can be expressed as: Y = HX + I + N. Through space-time coding and the corresponding decoding algorithm, the useful signal can be effectively separated and recovered, and at the same time, the interference signal can be suppressed.

[0207] (3) It can also combine intelligent beamforming with multi-antenna technology: Beamforming optimization: Combining intelligent beamforming algorithms, dynamically adjust the beam direction to make the spatial distribution of useful signals and interference signals more conducive to interference suppression. Multi-antenna cooperation: Utilize multi-antenna cooperation technology, and through the coordinated operation of multiple antennas at the transmitting and receiving ends, achieve spatial interference suppression and signal enhancement. More specifically: The core of interference suppression lies in reducing the impact of interference signals on useful signals through signal processing in space and time. Taking interference alignment as an example, assume there are k users, and each user is equipped with M transmitting antennas and N receiving antennas. The goal of interference alignment is to align all interference signals to the same spatial dimension at the receiving end, so as to achieve interference-free signal reception in other dimensions. The specific mathematical description is as follows:

[0208]

[0209] where: H ij is the channel matrix from user j to user i. v j is the transmit beam vector of user j. By solving the above equation, determine the transmit beam vector v j to achieve spatial alignment of interference signals.

[0210] More specifically:

[0211] (1) The process of the interference alignment algorithm is as follows: Initialization: Randomly initialize the transmit and receive beam vectors. Iterative optimization: Through iterative algorithms (such as the Min-Sum IA algorithm), optimize the transmit and receive beam vectors to align interference signals at the receiving end. Convergence determination: Set the convergence condition, and when the interference alignment error is less than the preset threshold, stop the iteration.

[0212] (2) The process of the space-time coding and decoding algorithm is as follows: Encoding step: At the transmitting end, according to the designed space-time coding matrix, encode the signal to generate the encoded signal matrix. Transmission and reception: Transmit the signal through multiple antennas, and the receiving end receives the encoded signal matrix. Decoding step: At the receiving end, use the corresponding decoding algorithm to recover the original signal and suppress interference signals.

[0213] The implementation steps are as follows:

[0214] (1) Channel modeling and analysis:

[0215] Utilize the channel modeling method in multi-band joint optimization to obtain the channel state information of the low-frequency and millimeter-wave bands.

[0216] Analyze the interference characteristics in a multi-band environment and identify the main interference sources and interference paths. The steps to identify the main interference sources and interference paths are as follows: A. Collect channel state information (CSI): Obtain the received signal strength (RSS), channel gain matrix H, signal-to-noise ratio (SNR), signal-to-interference-plus-noise ratio (SINR), channel delay, and angle information (such as AoA / AoD) for the low-frequency and millimeter-wave bands. Among them, the millimeter-wave band usually has a higher spatial resolution, which is suitable for accurately judging the interference direction.

[0217] B. Identify abnormal signal characteristics:

[0218] Analyze the following indicator changes to judge the presence and intensity of interference: If the SINR decreases but the signal power is stable, it is considered a suspected external interference; if the RSS fluctuates abnormally large, it is considered that there may be multipath interference or adjacent strong interference; if the inter-subcarrier frequency-domain correlation decreases, it is considered that frequency-selective interference occurs; if the CSI changes violently over time, it is considered a fast-moving interference source or directional interference.

[0219] C. Frequency-domain interference path analysis:

[0220] In a multi-band system, perform adjacent-frequency interference analysis: Identify whether there is strong signal leakage in adjacent bands (such as harmonic interference caused by millimeter-wave signals to low frequencies); Spectrum leakage detection: Based on the frequency-domain power map, judge whether there is abnormal energy distribution.

[0221] D. Interference source classification and tagging:

[0222] Classify the interference sources into the following categories to facilitate subsequent suppression by the system: The type is co-frequency interference, the characteristics are high-power signals in the same frequency band, the SINR decreases, and the common sources are adjacent users and reused cells; The type is adjacent-frequency interference, the characteristics are strong signals appearing in nearby bands, and the common sources are non-coordinated base stations and radar signals; The type is multipath interference, the characteristics are large delay spread and strong frequency selectivity, and the common sources are wall reflections and vehicles; The type is spatial overlap interference, the characteristics are overlapping beam directions, and the common sources are non-orthogonal multi-user beams.

[0223] E. Visualize the interference path map:

[0224] Using CSI + AoA / AoD + path loss modeling, an interference topology map can be generated: user → interference direction → interference node → receiving point, and each path is marked with parameters such as power, delay, and frequency band.

[0225] (2) Interference alignment and space-time coding design:

[0226] Design an interference alignment algorithm and a space-time coding scheme suitable for a multi-band environment to ensure spatial alignment and suppression of interference signals at the receiving end.

[0227] Using an optimization algorithm, determine appropriate transmit and receive beam vectors to achieve interference signal alignment.

[0228] (3) Intelligent beamforming and multi-antenna cooperation implementation:

[0229] Deploy a multi-antenna array and combine it with an intelligent beamforming algorithm to dynamically adjust the beam direction.

[0230] Achieve multi-antenna cooperation. Through the collaborative work of the transmitter and receiver, enhance the useful signal and suppress the interference signal.

[0231] (4) System integration and testing:

[0232] Integrate the interference suppression technology into the multi-band communication system and conduct system-level testing and verification.

[0233] According to the test results, optimize the interference suppression algorithm and the signal processing flow to improve the system performance.

[0234] Consider a dual-band system, the low-frequency band (F1) and the millimeter-wave band (F2). Set the optimization goal of interference suppression as maximizing the quality of the useful signal and minimizing the impact of the interference signal. The optimization problem can be expressed as:

[0235]

[0236] Where:

[0237] V L 、V M are the sets of transmit beam vectors for the low-frequency and millimeter-wave bands respectively. H ij is the channel matrix from user j to user i. P j is the transmit power limit of user j. QoS i is the quality of service requirement of user i. Q i is the minimum quality of service requirement of user i. By solving the above optimization problem, the target beam vectors for the low-frequency and millimeter-wave bands can be obtained to achieve interference signal alignment and suppression.

[0238] In summary, through interference alignment and space-time coding, effectively reduce the impact of interference signals, improve the quality and reliability of useful signals; optimize beamforming and interference management, reduce interference between frequency bands, and improve the overall capacity and spectral efficiency of the system; combine intelligent beamforming and multi-antenna cooperation, the system can dynamically adapt to different frequency bands and interference environments, improve the flexibility and adaptability of communication; through efficient beamforming and interference suppression, reduce power waste, optimize the energy efficiency of the system, and support the development of green communication.

[0239] Exemplarily, in a multi-user dense urban environment, the low-frequency band is used for wide coverage and communication through obstacles, while the millimeter-wave band is used for high-data-rate hotspots. Through the spectrum sharing and interference management technology of this case, the system can intelligently allocate resources to the low-frequency and millimeter-wave bands, and use interference alignment and space-time coding technologies to reduce interference between different bands and ensure high-quality communication services. The specific effects include a significant improvement in the user experience, an effective increase in system capacity, and an efficient utilization of spectrum resources.

[0240] Furthermore, the target signal to be transmitted is obtained, and the initial channel state information is estimated and tracked in real time through the estimation and tracking model to obtain the real-time channel state information.

[0241] It should be noted that as the wireless communication system develops towards the goals of high speed, high capacity, and high reliability, the dynamic changes in the channel environment pose higher requirements for the real-time performance and accuracy of channel estimation. Especially in 5G and future 6G networks, the large mobility, large number of users, and frequent channel fluctuations make traditional channel estimation methods face many challenges. To address this issue, this case proposes a real-time channel estimation algorithm based on the combination of Compressed Sensing (CS) and Machine Learning (ML), thereby obtaining the real-time channel state information. It can significantly reduce the pilot overhead while ensuring the accuracy of channel estimation, and improve the real-time performance and robustness of channel estimation. The estimation and tracking model includes a compressed sensing model, a machine learning-assisted estimation model, and a channel tracking model. Specifically:

[0242] (1) Compressed sensing model

[0243] Compressed Sensing (CS) is an emerging signal processing theory. Based on the sparsity of signals, it can reconstruct high-quality signals with far fewer sampling points than required by the traditional Nyquist sampling theorem. In channel estimation, using compressed sensing can significantly reduce the demand for pilot signals, reduce transmission overhead, and still ensure the accuracy of channel estimation.

[0244] Method overview:

[0245] Channel sparsity assumption: In wireless communication systems, especially in millimeter-wave and large-scale MIMO systems, the channel usually has sparse characteristics, that is, the gains of most paths are close to zero, and only a few paths have a significant impact on the received signal. Compressed sensing technology makes full use of this characteristic to estimate the channel through a small number of sampling points, avoiding the need for a large number of pilot signals in traditional methods.

[0246] Pilot Design and Sampling: According to the theory of compressive sensing, initially, a small number of samples are taken on the channel through pilot signals. The design requirements for pilot signals are that they can cover the sparse subspace of the channel and ensure the full acquisition of effective information during the channel estimation process.

[0247] Channel Reconstruction: By applying compressive sensing algorithms (such as Matching Pursuit (MP) or algorithms based on minimizing the L1 norm, such as Basis Pursuit), the state of the entire channel is reconstructed from a small number of pilot signals. Specifically, using the solution strategy of compressive sensing, through optimizing the model, the sparse channel matrix is restored to obtain an accurate channel estimate.

[0248] Technical Means:

[0249] L1 Norm Minimization: Using the method of minimizing the L1 norm, compressive sensing algorithms can effectively approximate the sparse solution of the channel matrix. The objective function is shown as follows:

[0250]

[0251] Where: y is the observed value of the pilot signal, A is the pilot matrix (the matrix composed of pilot signals), h is the channel estimation result (the channel matrix to be solved), and λ is the regularization parameter that controls the sparsity constraint.

[0252] Matching Pursuit Algorithm (MP): Through an iterative approach, the most matching dictionary elements (the elements of the channel matrix) are gradually selected to quickly approximate the sparse solution, which is suitable for the rapid estimation of sparse channels.

[0253] (2) Machine Learning-Assisted Estimation Model

[0254] Although compressive sensing technology can effectively reduce pilot overhead and improve channel estimation efficiency, due to the high complexity and non-linear characteristics of channel state changes, traditional compressive sensing algorithms may still have inaccurate estimations or be affected by noise. To further improve the accuracy of channel estimation, this case combines deep learning technology and post-processes and optimizes the compressive sensing estimation results by training a deep neural network (DNN).

[0255] Method Overview: A. Deep Neural Network (DNN) Model: Design and train a deep neural network to extract features from and optimize the rough channel estimates obtained from compressive sensing algorithms. The DNN can automatically learn complex non-linear mapping relationships from a large amount of channel data through a multi-layer neural network structure, thereby optimizing the channel estimation results and improving the estimation accuracy. B. Training and Optimization: Based on the historical channel estimation dataset, train the DNN using supervised learning. Optimize the network parameters by minimizing the estimation error (such as mean square error, MSE) to enhance the network's robustness to channel estimation. C. Post-processing and Optimization: In real-time communication, post-process the channel matrix obtained from compressive sensing estimation through the DNN to eliminate noise and errors and enhance the accuracy of the estimation results. Specifically, use the DNN to correct the input rough estimated channel to obtain an estimation result close to the true channel.

[0256] Technical Means:

[0257] DNN Training Process: Train the deep neural network by using a large amount of simulated channel data and actual measurement data. The goal is to minimize the channel estimation error and train the weights and biases in the network. The loss function can be selected as mean square error (MSE) or other adaptive loss functions:

[0258]

[0259] Where: is the channel estimation result predicted by the DNN, h i is the true channel, and θ is the network parameter.

[0260] (3) Channel Tracking Model

[0261] In a wireless communication environment, the channel state often fluctuates continuously with time and space, especially in high-speed mobile scenarios where the channel changes particularly violently. To track these rapidly changing channel states in real time, this case introduces a Recurrent Neural Network (RNN) for channel tracking to ensure the consistency and accuracy of the channel estimation results, that is, the real-time channel state information at different time periods.

[0262] Method Overview:

[0263] RNN Model Design: Design an RNN-based channel tracking mechanism for modeling and predicting rapidly changing channel states. The RNN can effectively capture the temporal changes in the channel state and is particularly suitable for modeling long time series data.

[0264] State Prediction and Update: By recursively updating the channel state, the RNN can predict the channel state at each moment and correct the current channel estimation result. This mechanism can dynamically update the channel at the current moment based on the estimation result of the previous moment, ensuring the timeliness and continuity of the estimated real-time channel state information.

[0265] Technical Means:

[0266] RNN Training and Update: Based on the known channel estimation sequence, the RNN network is trained and the weights in the network are continuously updated. The training process optimizes the network parameters by minimizing the prediction error (such as the mean square error).

[0267] Temporal Feature Extraction: The RNN network is used to extract the temporal features in the channel state to ensure that the channel estimation can timely reflect the changing trend of the channel.

[0268] Principle:

[0269] RNN Model: Set the channel state sequence as h t , where t is the time index. The RNN calculates the output h t-1 at the current moment through the state h t at the previous moment and the input signal x t :

[0270] h t = f(W h h t-1 + W x x t + b);

[0271] Where: f is the activation function; W x , W h are weight matrices; b is the bias term.

[0272] By combining compressive sensing, machine learning, and recurrent neural networks, the estimation and tracking model proposed in this case can improve the accuracy and robustness of channel estimation while ensuring low pilot overhead, adapt to the rapidly changing wireless channel environment, and is particularly suitable for large-scale MIMO and high-speed mobile scenarios.

[0273] Further, the sender is set in the full-link optimization framework, which includes the physical layer, link layer, network layer, and application layer. Among them, the physical layer serves as the underlying support, providing real-time CSI, available frequency bands, interference status, etc.; the link layer has a dynamic retransmission mechanism and modulation and coding adjustment, and makes decisions based on the physical layer information; the network layer dynamically switches paths based on the routing policy according to the physical layer channel conditions and link status; the application layer dynamically adjusts QoS parameters and transmission rates, etc. Among them, in the step of performing signal processing on the real-time channel state information to obtain the target transmission strategy, the target transmission strategy includes signal conversion of the target signal to be transmitted based on the target modulation order and target channel coding rate; whether to retransmit the target signal to be transmitted; transmitting the target signal to be transmitted based on the target path; and the target signal to be transmitted follows the target transmission parameters during the transmission process.

[0274] To achieve full-link optimization from the physical layer to the application layer, this case constructs a comprehensive information sharing and collaborative optimization framework, which can establish a unified optimization goal and an efficient information transfer mechanism among layers, so as to coordinate the resources and parameters of each layer within the global scope and improve the overall system performance.

[0275] (1) Unified optimization goal

[0276] In the full-link optimization framework, first set a unified system performance optimization goal, such as minimizing end-to-end delay, maximizing system throughput, or minimizing energy consumption. This goal guides the entire system to adjust parameters and allocate resources among different layers, making the optimization strategies of each layer coordinated and ultimately achieving global optimality.

[0277] The specific steps are as follows: A. Target function setting: Define the comprehensive performance index J, for example: Among them, α, β, and γ are weight coefficients, respectively reflecting the impacts of delay, throughput, and energy consumption on the overall performance. B. Full-link collaborative index: The performance indicators of each layer (such as SNR of the physical layer, bit error rate of the link layer, routing delay of the network layer, QoS of the application layer) are all incorporated into the unified target function, and through weight adjustment, the optimization measures of each layer can contribute to the improvement of the global performance. C. Dynamic adjustment of the optimization goal: According to real-time application requirements and environmental changes, dynamically adjust the weight coefficients, so that the full-link optimization goal can flexibly adapt to different scenarios.

[0278] (2) Information sharing mechanism

[0279] Establish an efficient information sharing mechanism to achieve real-time data exchange and collaborative optimization among different layers. Through standardized interfaces and data protocols, key information (such as CSI, user requirements, network status, etc.) in the physical layer, link layer, network layer, and application layer is uniformly managed and shared to form a closed-loop feedback system.

[0280] The specific steps are as follows: A. Standardized interface: Design a unified interface protocol to enable different layers to exchange information in a standard format. For example, utilize the control interface of software-defined network (SDN) to integrate real-time CSI, routing status, transmission rate, and QoS parameters. B. Real-time data transmission mechanism: Build a high-bandwidth and low-latency data transmission channel to ensure that information sharing among different layers can meet real-time requirements. Adopt technologies such as message queues and subscription / publishing mechanisms to achieve fast data transfer and processing. C. Data fusion and processing: Use data fusion algorithms to integrate and process information from different layers to form a unified view of the system state, providing an accurate basis for optimization decisions at each layer.

[0281] (3) Distributed optimization algorithms

[0282] To achieve collaborative optimization of the entire link, this case adopts distributed optimization algorithms, enabling each layer to perform independent optimization under local conditions while coordinating the improvement of global performance through the information sharing mechanism.

[0283] A. Distributed algorithm design:

[0284] Decompose the entire link optimization problem into several sub-problems, each corresponding to a certain layer or functional module, and achieve the global goal through local optimization.

[0285] Apply distributed optimization methods such as Distributed Gradient Descent and Alternating Direction Method of Multipliers (ADMM), share gradient information among nodes, and achieve global optimality.

[0286] B. Coordination mechanism:

[0287] Through the information sharing mechanism, each layer exchanges local optimization results and gradient information in real time to form a globally coordinated optimization process. Use dual decomposition technology to transform the global goal into local goals of each layer, ensuring that local optimization can jointly promote the improvement of global performance.

[0288] C. Adaptive adjustment:

[0289] According to real-time feedback, dynamically adjust the step size and weight of the optimization algorithm for each layer to ensure that the distributed optimization algorithm can quickly converge to the global optimal solution in a dynamic environment.

[0290] Model principle: The distributed optimization problem can be expressed in the following form:

[0291]

[0292] where f i (x i ) represents the local objective function of the i-th layer, and A i is the corresponding constraint matrix. Through dual decomposition and the ADMM method, collaborative optimization can be achieved between layers.

[0293] In summary, the full-link optimization framework of this case realizes collaborative optimization from the physical layer to the application layer through the following key components: Unified optimization goal: Set and dynamically adjust the global performance indicators to ensure that the optimization goals of each layer are consistent and promote the improvement of the overall system performance; Information sharing mechanism: Establish standardized interfaces and real-time data transmission channels to achieve effective information exchange and data fusion between layers, providing reliable data support for full-link optimization; Distributed optimization algorithm: Adopt the distributed optimization method to decompose the global optimization problem into local sub-problems. Each layer optimizes independently and coordinates simultaneously to finally achieve the global optimum. Through this cross-layer optimization design, not only can CSI-based routing optimization be achieved at the network layer and dynamic adjustment of application layer parameters, but also through full-link information sharing and distributed optimization algorithms, comprehensive improvement of key indicators such as latency, throughput, and energy consumption of the entire communication system can be realized, meeting the requirements of 5G / 6G and future wireless communication systems for high efficiency, high reliability, and low energy consumption.

[0294] Please refer to Figure 2 , this embodiment also provides a transmitter for executing the signal communication method as described above. The transmitter includes a physical layer, a link layer, a network layer, and an application layer, where: The physical layer is used to obtain the target signal to be transmitted and perform real-time estimation and tracking of the initial channel state information through an estimation and tracking model to obtain real-time channel state information; The physical layer is used to select the real-time channel state information and user requirements in the candidate frequency bands through a multi-band joint optimization model to determine the target frequency band; The link layer, network layer, and application layer are used to perform signal processing on the real-time channel state information to obtain the target transmission strategy; The application layer is used to transmit the target signal to be transmitted based on the target frequency band and the target transmission strategy so that the receiver can receive the target signal to be transmitted.

[0295] The transmitter provided in this embodiment can well implement the above signal communication method. By acquiring the target signal to be transmitted and estimating and tracking the initial channel state information through an estimation and tracking model, the real-time channel state information is obtained, ensuring the real-time nature of acquiring the real-time channel state information and providing an accurate and timely data basis for obtaining the target transmission strategy subsequently. By selecting the real-time channel state information and user requirements in the candidate frequency bands through a multi-band joint optimization model, the target frequency band is determined, thereby dynamically determining the frequency band for transmitting the signal, avoiding problems such as transmission limitation caused by communicating in inappropriate frequency bands, improving the reliability of signal communication, and enhancing the adaptability to environmental changes and user requirements during the communication process. Subsequently, signal processing is performed on the real-time channel state information to obtain the target transmission strategy, thereby determining an appropriate transmission method, improving the reliability during the signal transmission process in the communication process, and then transmitting the target signal to be transmitted through the target transmission strategy in the target frequency band so that the receiver can receive the target signal to be transmitted.

[0296] The above is the preferred embodiment of the present invention. It should be noted that for those of ordinary skill in the art of this technology, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements are also regarded as the protection scope of the present invention.

Claims

1. A signal communication method, characterized in that, Applied to the sending end, the method includes: Obtain the target signal to be transmitted, and perform real-time estimation and tracking on the initial channel state information through an estimation and tracking model to obtain the real-time channel state information; Select in the candidate frequency bands the real-time channel state information and user requirements through a multi-band joint optimization model to determine the target frequency band; Perform signal processing on the real-time channel state information to obtain the target transmission strategy; Transmit the target signal to be transmitted in the target frequency band based on the target transmission strategy, so that the receiving end receives the target signal to be transmitted.

2. The signal communication method according to claim 1, characterized in that The multi-band joint optimization model includes a first prediction sub-model and a second prediction sub-model. The step of selecting in the candidate frequency bands the real-time channel state information and user requirements through the multi-band joint optimization model to determine the target frequency band includes: Perform channel quality prediction on the scenario characteristics, the real-time location information and moving speed of the user through the first prediction sub-model to obtain the channel quality change information of different frequency bands; Perform user service preference prediction on the real-time location information, moving speed and historical application behavior of the user through the second prediction sub-model to obtain the user service preference information; Based on the channel quality change information and the user service preference information of different frequency bands, determine the target frequency band in the candidate frequency bands.

3. The signal communication method according to claim 2, characterized in that, After determining the target frequency band by selecting in the candidate frequency bands the real-time channel state information and user requirements through the multi-band joint optimization model, it includes: Construct an objective function based on the spectrum resources of the candidate frequency bands; Based on the constraint conditions, obtain the optimal solution of the objective function, and obtain a multi-band resource allocation scheme based on the optimal solution; Allocate resources to the target frequency band based on the multi-band resource allocation scheme.

4. The signal communication method according to claim 1, wherein The step of performing signal processing on the real-time channel state information to obtain the target transmission strategy includes: Perform beamforming processing on the real-time channel state information according to the beamforming algorithm to obtain beamforming parameters; Based on the beamforming parameters and the real-time channel state information, adjust the modulation order and channel coding rate to obtain the target modulation order and target channel coding rate; Perform signal conversion on the initial signal to be transmitted based on the target modulation order and target channel coding rate to obtain the target signal to be transmitted, so that the receiving end obtains the target signal to be transmitted.

5. The signal communication method according to claim 1, wherein The step of performing signal processing on the real-time channel state information to obtain the target transmission strategy includes: Obtain the current transmission state information; Construct a retransmission risk assessment model based on the real-time channel state information and the current transmission state information; Evaluate the target signal to be transmitted based on the retransmission risk assessment model to obtain a retransmission strategy; Transmit the target signal to be transmitted based on the retransmission strategy.

6. The signal communication method according to claim 1, wherein The step of performing signal processing on the real-time channel state information to obtain the target transmission strategy includes: Perform path selection on the real-time channel state information through a dynamic routing algorithm to obtain the target path; Transmit the target signal to be transmitted based on the target path.

7. The signal communication method according to claim 1, wherein Performing signal processing on the real-time channel state information to obtain a target transmission strategy includes: Obtaining performance parameters; Adjusting preset transmission parameters based on the performance parameters to obtain target transmission parameters; Transmitting the target signal to be transmitted based on the target transmission parameters.

8. The signal communication method according to claim 1, wherein The obtaining of the target signal to be transmitted includes: Encoding the target signal to be transmitted through a preset space-time coding matrix so that the receiving end can perform signal recovery and interference suppression on the encoded target signal to be transmitted based on a preset decoding algorithm.

9. The signal communication method according to claim 1, wherein The obtaining of the target signal to be transmitted includes: Using an iterative optimization algorithm to optimize the transmission beam direction of the transmitting end and the receiving beam direction of the receiving end so that the interference signals in the target frequency band are aligned at the receiving end.

10. A transmitting end, characterized in that, For implementing the signal communication method according to any one of claims 1-9, the transmitting end includes a physical layer, a link layer, a network layer, and an application layer, wherein: The physical layer is used to obtain the target signal to be transmitted and perform real-time estimation and tracking on the initial channel state information through an estimation and tracking model to obtain real-time channel state information; The physical layer is used to select the real-time channel state information and user requirements in candidate frequency bands through a multi-band joint optimization model to determine the target frequency band; The link layer, the network layer, and the application layer are used to perform signal processing on the real-time channel state information to obtain a target transmission strategy; The application layer is used to transmit the target signal to be transmitted in the target frequency band based on the target transmission strategy so that the receiving end can receive the target signal to be transmitted.

Images