Noise interference suppression method and system facing wireless channel basic model, terminal and storage medium

Abstract

This invention relates to the field of noise cancellation technology, and discloses a noise interference suppression method, system, terminal, and storage medium for a wireless channel fundamental model. The method includes: constructing the received signal and initial channel estimation results of pilot resource units in the transmission time slot of the transmitted signal; refining the initial channel estimation results to construct a subspace projection matrix and an orthogonal projection matrix, and projecting the transmitted signal into the two matrices to obtain multiple pairs of noise-plus-interference estimation components; using a noise-plus-interference estimation neural network to concatenate and fuse each pair of noise-plus-interference estimation components to obtain a noise-plus-interference estimation result; and using the noise-plus-interference estimation result to perform denoising and completion to obtain a purified received signal. This invention utilizes only the pilot signal contaminated by noise and interference to recover complete channel state information and generate a high-quality channel representation, making the channel representation extracted by the wireless channel fundamental model more accurate and improving the performance of downstream tasks.

Description

Technical Field

[0001] This invention relates to the field of noise cancellation technology, and in particular to a noise interference suppression method, system, terminal, and computer-readable storage medium for a wireless channel fundamental model. Background Technology

[0002] The input to the basic wireless channel model is the wireless channel matrix. After transformation and segmentation, the wireless channel matrix forms a token sequence. The token sequence is then processed through a linear transformation layer, a position coding layer, and multiple transformer layers to extract features, forming a channel feature sequence. The channel feature sequence is a general feature representation of the wireless channel matrix and can be input into downstream models to achieve various communication or sensing tasks.

[0003] In existing communication systems, pilot signals are affected by noise and interference, leading to channel estimation errors, making it impossible to obtain a perfect wireless channel matrix. Furthermore, inputting a noise- and interference-affected channel matrix into the basic wireless channel model results in inaccurate feature extraction, thus impacting the performance of downstream tasks.

[0004] Therefore, existing technologies still need to be improved and developed. Summary of the Invention

[0005] The main objective of this invention is to provide a noise interference suppression method, system, terminal, and computer-readable storage medium for a wireless channel fundamental model. This invention aims to solve the problem in the prior art where inputting a channel matrix affected by noise and interference into a wireless channel fundamental model for feature extraction leads to errors in the extracted feature results.

[0006] To achieve the above objectives, the present invention provides a noise interference suppression method based on a fundamental model of a wireless channel, the method comprising the following steps: Acquire multiple transmitted signals, construct the received signals corresponding to multiple pilot resource units in all transmission time slots of each transmitted signal, and construct the corresponding initial channel estimation results based on each transmitted signal and each received signal; After refining each initial channel estimation result, a corresponding subspace projection matrix and orthogonal projection matrix are constructed, and each transmitted signal is projected onto the subspace projection matrix and orthogonal projection matrix to obtain multiple pairs of noise plus interference estimation components; After splicing each pair of noise-plus-interference estimation components using a noise interference estimation neural network, a fusion network incorporating the system signal-to-interference-plus-noise ratio is used to fuse all splicing results, resulting in multiple noise-plus-interference estimation results. The received signal is denoised and padded using all the noise and interference estimation results to obtain the channel estimation result for each resource unit.

[0007] Optionally, in the noise interference suppression method for the wireless channel fundamental model, the transmission time slot of the transmitted signal includes multiple OFDM symbols; The OFDM symbol represents the basic time unit for transmitting the transmitted signal; Each OFDM symbol comprises multiple subcarriers; The number of pilot resource units in the transmission time slot is the product of the number of subcarriers and the number of OFDM symbols.

[0008] Optionally, the noise interference suppression method for the wireless channel fundamental model, wherein acquiring multiple transmitted signals, constructing received signals corresponding to multiple pilot resource units in all transmission time slots of each transmitted signal, and constructing corresponding initial channel estimation results based on each transmitted signal and each received signal, specifically includes: Acquire multiple transmitted signals, and for each transmitted signal, construct a corresponding wireless channel matrix for all pilot resource elements in all transmission time slots: ; Where H represents the wireless channel matrix, C represents the complex field, M represents the number of transmit antennas, and N represents the number of receive antennas; Based on each transmitted signal and the corresponding wireless channel matrix, a corresponding received signal is constructed: ; ; ; ; Where y represents the received signal, x represents the transmitted signal, n represents the receiver noise, and i represents the received interference signal; For each pilot resource element, an initial channel estimate is constructed for the pilot resource element based on the radio channel matrix: ; in, This represents the initial channel estimate.

[0009] Optionally, the noise interference suppression method for the basic model of wireless channels, wherein refining each initial channel estimation result and constructing the corresponding subspace projection matrix and orthogonal projection matrix specifically includes: The initial channel estimate is input into the channel refiner for processing, and the channel estimate result is output. For each pilot resource element, a subspace projection matrix is ​​constructed based on the channel estimation results: ; in, Represents the subspace projection matrix. This represents the channel estimation result; For each pilot resource element, based on the channel estimation results, an orthogonal projection matrix intersecting with the channel subspace is constructed: ; in, Represents the orthogonal projection matrix. Represents the identity matrix.

[0010] Optionally, the noise interference suppression method for the fundamental model of a wireless channel, wherein projecting each transmitted signal onto the subspace projection matrix and the orthogonal projection matrix to obtain multiple pairs of noise-plus-interference estimation components specifically includes: After projecting each received signal onto the subspace projection matrix, noise extraction is performed to obtain multiple first noise plus interference estimation components: ; in, This represents the first noise plus interference estimation component; Each of the received signals is projected onto the orthogonal projection matrix to obtain multiple second noise plus interference estimation components: ; in, This represents the second noise plus interference estimation component; Each of the first noise plus interference estimation components is combined with the corresponding second noise plus interference estimation component to obtain multiple pairs of noise plus interference estimation components.

[0011] Optionally, the noise interference suppression method for the wireless channel fundamental model includes a noise interference estimation neural network comprising a first sub-network and a second sub-network. After the noise interference estimation neural network concatenates each pair of noise-plus-interference estimation components, a fusion network incorporating the system signal-to-interference-plus-noise ratio is used to fuse all the concatenation results, yielding multiple noise-plus-interference estimation results, specifically including: For each pair of noise-plus-interference estimation components, the first noise-plus-interference estimation component is input into the first sub-network for processing, and the first processing result is output. The second noise-interference estimation component is input into the second sub-network for processing, and the second processing result is output. The system signal-to-interference-plus-noise ratio is added to the fusion network, and the first processing result and the second processing result are concatenated and input into the fusion network; In the fusion network, the residual signal in the splicing result is classified, distinguished, and adjusted using the system signal-to-interference-plus-noise ratio, and the noise-plus-interference estimation result corresponding to each pair of noise-plus-interference estimation components is output.

[0012] Optionally, the noise interference suppression method for the wireless channel fundamental model, wherein the step of using all the noise plus interference estimation results to denoise and pad each received signal to obtain the channel estimation result for each resource unit specifically includes: The noise and interference estimation results are used to denoise each of the received signals to obtain the corresponding cleaned received signals: ; in, This indicates that the received signal has been purified. This represents the estimation result with added noise and interference. The least squares channel is used to complete the channel for each of the cleaned received signals, and the channel estimation result on the current pilot resource unit is obtained.

[0013] Furthermore, to achieve the above objectives, the present invention also provides a noise interference suppression system oriented towards a fundamental model of a wireless channel, wherein the noise interference suppression system oriented towards a fundamental model of a wireless channel includes: The channel estimation module is used to acquire multiple transmitted signals, construct the received signals corresponding to multiple pilot resource units in all transmission time slots of each transmitted signal, and construct the corresponding initial channel estimation results based on each transmitted signal and each received signal. The noise extraction module is used to refine each of the initial channel estimation results, construct the corresponding subspace projection matrix and orthogonal projection matrix, and project each of the transmitted signals into the subspace projection matrix and orthogonal projection matrix to obtain multiple pairs of noise and interference estimation components; The noise estimation module is used to perform splicing processing on each pair of noise plus interference estimation components using a noise interference estimation neural network, and then to perform fusion processing on all splicing results using a fusion network that introduces the system signal-to-interference-plus-noise ratio to obtain multiple noise plus interference estimation results. The interference cancellation and channel completion module is used to perform noise reduction and channel completion on each of the received signals using all the noise plus interference estimation results, so as to obtain the channel estimation result on each of the resource units.

[0014] Furthermore, to achieve the above objectives, the present invention also provides a terminal, wherein the terminal includes: a memory, a processor, and a noise interference suppression program oriented towards a wireless channel fundamental model stored in the memory and executable on the processor, wherein when the noise interference suppression program oriented towards a wireless channel fundamental model is executed by the processor, it implements the steps of the noise interference suppression method oriented towards a wireless channel fundamental model as described above.

[0015] Furthermore, to achieve the above objectives, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a noise interference suppression program oriented towards a wireless channel fundamental model, and the noise interference suppression program oriented towards a wireless channel fundamental model, when executed by a processor, implements the steps of the noise interference suppression method oriented towards a wireless channel fundamental model as described above.

[0016] In this invention, multiple transmitted signals are acquired, and received signals corresponding to multiple pilot resource units in all transmission time slots of each transmitted signal are constructed. Initial channel estimation results are constructed based on each transmitted and received signal. After refining each initial channel estimation result, corresponding subspace projection matrices and orthogonal projection matrices are constructed, and each transmitted signal is projected onto these matrices to obtain multiple pairs of noise-plus-interference estimation components. Each pair of noise-plus-interference estimation components is then concatenated using a noise-plus-interference estimation neural network, and all concatenated results are fused using a fusion network incorporating the system signal-to-interference-plus-noise ratio (SINR) to obtain multiple noise-plus-interference estimation results. Finally, each received signal is denoised and padded using all the noise-plus-interference estimation results to obtain the channel estimation result for each resource unit. This invention utilizes only pilot signals contaminated by noise and interference to recover complete channel state information and generate high-quality channel representations, making the channel representation extracted from the wireless channel fundamental model more accurate and improving the performance of downstream tasks. Attached Figure Description

[0017] Figure 1 This is a flowchart of a preferred embodiment of the noise interference suppression method for a fundamental wireless channel model according to the present invention; Figure 2 This is a flowchart of channel characterization extraction after introducing a noise interference suppression module in a preferred embodiment of the noise interference suppression method for a wireless channel fundamental model of the present invention. Figure 3 This is a structural diagram of a preferred embodiment of the noise interference suppression system for a fundamental wireless channel model according to the present invention; Figure 4 This is a structural diagram of a preferred embodiment of the terminal of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer and more explicit, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0019] The noise interference suppression method for a fundamental wireless channel model described in the preferred embodiment of the present invention, such as... Figure 1 As shown, the noise interference suppression method for the basic model of wireless channels includes the following steps: Step S10: Acquire multiple transmitted signals, construct the received signals corresponding to the multiple pilot resource units in all transmission time slots of each transmitted signal, and construct the corresponding initial channel estimation results based on each transmitted signal and each received signal.

[0020] The transmission time slot of the transmitted signal includes multiple OFDM symbols; The OFDM symbol represents the basic time unit for transmitting the transmitted signal; Each OFDM symbol comprises multiple subcarriers; The number of pilot resource units in the transmission time slot is the product of the number of subcarriers and the number of OFDM symbols.

[0021] Specifically, this invention designs a noise and interference suppression method in a Multiple-Input Multiple-Output Orthogonal Frequency Division Multiplexing (MIMO-OFDM) communication system. The method is described in the example of recovering complete channel state information and generating a high-quality channel representation using only pilot signals contaminated by noise and interference. In this invention, N transmit antennas and M receive antennas are disclosed in the MIMO-OFDM system. Transmission is based on transmission time slots as the basic unit. Each transmission time slot contains T OFDM symbols, and each OFDM symbol contains K subcarriers. Therefore, each transmission time slot contains K×T pilot resource elements (REs).

[0022] Specifically, multiple transmitted signals are acquired, and for each transmitted signal, a corresponding wireless channel matrix is ​​constructed for all pilot resource elements in all transmission time slots: ; Where H represents the wireless channel matrix, C represents the complex field, M represents the number of transmit antennas, and N represents the number of receive antennas; Based on each transmitted signal and the corresponding wireless channel matrix, a corresponding received signal is constructed: ; ; ; ; Where y represents the received signal, x represents the transmitted signal, n represents the receiver noise, and i represents the received interference signal; For each pilot resource element, an initial channel estimate is constructed for the pilot resource element based on the radio channel matrix: ; in, This represents the initial channel estimate.

[0023] Within each pilot resource unit, its infinite channel matrix can be represented as H. Within each OFDM time slot, the system inputs the known transmitted signal and the corresponding received signal pair on the pilot resource unit into the basic wireless channel model to generate a general wireless channel representation. Based on this, a corresponding received interference signal can be added to each transmitted signal to construct the corresponding received signal.

[0024] Furthermore, such as Figure 2 As shown, the LS (Least Squares) channel estimator is used to perform initial channel estimation on the transmitted and received signals, thereby constructing the initial channel estimation results.

[0025] Step S20: After refining each initial channel estimation result, construct the corresponding subspace projection matrix and orthogonal projection matrix, and project each transmitted signal into the subspace projection matrix and orthogonal projection matrix to obtain multiple pairs of noise plus interference estimation components.

[0026] Specifically, the initial channel estimate is input into the channel refiner for processing, and the channel estimate result is output. For each pilot resource element, a subspace projection matrix is ​​constructed based on the channel estimation results: ; in, Represents the subspace projection matrix. This represents the channel estimation result; For each pilot resource element, based on the channel estimation results, an orthogonal projection matrix intersecting with the channel subspace is constructed: ; in, Represents the orthogonal projection matrix. Represents the identity matrix.

[0027] In this invention, based on the constructed initial channel estimation results, a channel refiner is further used to refine the channel, and a channel subspace projection matrix and an orthogonal projection matrix orthogonal to the channel subspace are constructed for each pilot resource unit.

[0028] In this invention, the entire space is divided into a signal subspace containing useful channel information and an orthogonal subspace containing only noise and interference through subspace decomposition. After projecting the observed signal, the noise and interference components can be directly extracted and then filtered out in a targeted manner, thereby fundamentally suppressing the noise amplification problem caused by basic estimation methods such as LS.

[0029] Furthermore, after projecting each of the received signals onto the subspace projection matrix, noise extraction is performed to obtain multiple first noise plus interference estimation components: ; in, This represents the first noise plus interference estimation component; Each of the received signals is projected onto the orthogonal projection matrix to obtain multiple second noise plus interference estimation components: ; in, This represents the second noise plus interference estimation component; Each of the first noise plus interference estimation components is combined with the corresponding second noise plus interference estimation component to obtain multiple pairs of noise plus interference estimation components.

[0030] Furthermore, this invention performs preliminary extraction of noise interference components and projects the received signal onto two projection matrices respectively. This yields the noise-plus-interference estimated components (i.e., the first noise-plus-interference estimated component) in the channel subspace projection matrix and the noise-plus-interference estimated components (i.e., the second noise-plus-interference estimated component) corresponding to the orthogonal projection matrix. The projection matrix only needs to be constructed based on the initial channel estimation results, without requiring prior information such as the channel's statistical correlation matrix and signal-to-noise ratio. Compared to algorithms like MMSE (Minimum Mean Square Error), which require prior information, this invention has a wider range of applications and lower implementation costs. In multi-user MIMO scenarios, inter-user interference tends to concentrate in the orthogonal subspace. After projecting and extracting the interference components, multi-user interference within or between cells can be effectively eliminated. Especially in low signal-to-noise ratio and deep channel fading scenarios, the improvement in estimation performance is more significant than traditional methods.

[0031] Step S30: After splicing each pair of noise and interference estimation components using a noise interference estimation neural network, a fusion network that incorporates the system signal-to-interference-plus-noise ratio is used to fuse all splicing results to obtain multiple noise and interference estimation results.

[0032] Since the noise-plus-interference estimation still contains residual signal components, this invention further introduces a noise interference estimation neural network to remove these residual components, achieving accurate estimation of the noise interference components. The noise interference estimation neural network includes two sub-networks, the first sub-network (i.e....) Figure 2 The noise interference estimation neural network 1) and the second sub-network (i.e. Figure 2 The noise interference estimation neural network 2) processes the two noise and interference estimation components mentioned above to obtain the corresponding results.

[0033] Specifically, two independent sub-networks are used to process the inputs of two different statistical features: the noise-plus-interference estimation component in the channel subspace projection matrix and the noise-plus-interference estimation component corresponding to the orthogonal projection matrix. Then, the features are integrated through a fusion network. This ensures targeted processing of features of different dimensions and avoids the inadequacy of a single network to fit different distribution features. Compared with a fully connected integrated network, the modular design has lower training difficulty, faster inference speed, and is more suitable for engineering implementation.

[0034] Specifically, for each pair of noise-plus-interference estimation components, the first noise-plus-interference estimation component is input into the first sub-network for processing, and the first processing result is output. The second noise-interference estimation component is input into the second sub-network for processing, and the second processing result is output. The system signal-to-interference-plus-noise ratio is added to the fusion network, and the first processing result and the second processing result are concatenated and input into the fusion network; In the fusion network, the residual signal in the splicing result is classified, distinguished, and adjusted using the system signal-to-interference-plus-noise ratio, and the noise-plus-interference estimation result corresponding to each pair of noise-plus-interference estimation components is output.

[0035] In this process, after processing the two noise-plus-interference estimation components separately, the two results are first concatenated and then input into the fusion network. The system signal-to-interference-plus-noise ratio (SINR) is introduced into the fusion network as an auxiliary input to distinguish residual signal components and adaptively adjust the output amplitude. Finally, the complete noise-plus-interference estimation result is input. The SINR is input into the fusion network to obtain an output gating coefficient. This gating coefficient is multiplied by the feature vector in the fusion network to adjust the output amplitude of the fusion network. The output amplitude is determined by the SINR and can be automatically adjusted under different channel conditions.

[0036] This invention introduces a dedicated neural network to fit the distribution characteristics of residual components, which can accurately distinguish between noise interference and residual useful signals, solving the problem of missed estimation in traditional methods and ultimately further reducing the error of subsequent channel estimation. It also introduces the actual signal-to-interference-plus-noise ratio (SIR) of the system as additional prior information, allowing the network to adaptively adjust its output according to the interference level of the current communication environment. In high SIR scenarios (less interference), it compresses the estimation amplitude and avoids excessive filtering of useful signals; in low SIR scenarios (strong interference), it enhances the suppression of residual interference. Compared with networks without auxiliary input, the estimation performance is more stable in dynamic and complex channels.

[0037] Step S40: Use all the noise and interference estimation results to denoise and complete each of the received signals to obtain the channel estimation results for each resource unit.

[0038] Specifically, the noise and interference estimation results are used to denoise each of the received signals to obtain the corresponding cleaned received signals: ; in, This indicates that the received signal has been purified. This represents the estimation result with added noise and interference. The least squares channel is used to complete the channel for each of the cleaned received signals, and the channel estimation result on the current pilot resource unit is obtained.

[0039] In a complete embodiment of this invention, after acquiring the transmitted and received signals, a noise interference suppression model is directly inserted to accurately estimate the noise. This eliminates the complex steps of multiple rounds of iterative estimation and repeated filtering in traditional schemes, resulting in a simpler process, lower computational load and processing latency, and better meeting the engineering requirements of real-time communication. Then, the noise and interference estimation results of the accurately estimated signal are removed using the wireless channel matrix to obtain a purified received signal. Subsequently, the purified received signal is channel-completed using the basic wireless channel model. The channel completion directly outputs a general channel representation that can simultaneously support both communication and sensing tasks, eliminating the need to design different channel processing procedures for communication and sensing, thus improving the versatility of the solution and reducing the cost of multi-task deployment.

[0040] This invention utilizes only pilot signals contaminated by noise and interference to recover complete channel state information and generate a high-quality channel representation, making the channel representation extracted by the wireless channel fundamental model more accurate and improving the performance of downstream tasks.

[0041] Furthermore, such as Figure 3 As shown, based on the above-described noise interference suppression method for a fundamental wireless channel model, this invention also provides a noise interference suppression system for a fundamental wireless channel model, wherein the noise interference suppression system for a fundamental wireless channel model includes: The channel estimation module 51 is used to acquire multiple transmitted signals, construct the received signals corresponding to multiple pilot resource units in all transmission time slots of each transmitted signal, and construct the corresponding initial channel estimation results based on each transmitted signal and each received signal. The noise extraction module 52 is used to refine each of the initial channel estimation results, construct the corresponding subspace projection matrix and orthogonal projection matrix, and project each of the transmitted signals into the subspace projection matrix and orthogonal projection matrix to obtain multiple pairs of noise and interference estimation components; The noise estimation module 53 is used to perform splicing processing on each pair of noise plus interference estimation components using a noise interference estimation neural network, and then to perform fusion processing on all splicing results using a fusion network that introduces the system signal-to-interference-plus-noise ratio to obtain multiple noise plus interference estimation results. The interference cancellation and channel completion module 54 is used to perform noise reduction and channel completion on each of the received signals using all the noise plus interference estimation results, so as to obtain the channel estimation result on each of the resource units.

[0042] Furthermore, such as Figure 4 As shown, based on the above-mentioned noise interference suppression method and system for the basic model of wireless channels, the present invention also provides a terminal, which includes a processor 10, a memory 20 and a display 30. Figure 4 Only some of the terminal components are shown; however, it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0043] In some embodiments, the memory 20 may be an internal storage unit of the terminal, such as a hard disk or memory. In other embodiments, the memory 20 may be an external storage device of the terminal, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc. Further, the memory 20 may include both internal and external storage devices. The memory 20 is used to store application software and various types of data installed on the terminal, such as the program code installed on the terminal. The memory 20 can also be used to temporarily store data that has been output or will be output. In one embodiment, the memory 20 stores a noise interference suppression program 40 oriented towards a wireless channel fundamental model, which can be executed by the processor 10 to implement the noise interference suppression method oriented towards a wireless channel fundamental model in this application.

[0044] In some embodiments, the processor 10 may be a central processing unit (CPU), a microprocessor, or other data processing chip, used to run program code stored in the memory 20 or process data, such as executing the noise interference suppression method for the wireless channel basic model.

[0045] In some embodiments, the display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. The display 30 is used to display information on the terminal and to display a visual user interface. The components of the terminal communicate with each other via a system bus.

[0046] In one embodiment, when processor 10 executes noise interference suppression program 40 for a wireless channel underlying model in memory 20, the following steps are performed: Acquire multiple transmitted signals, construct the received signals corresponding to multiple pilot resource units in all transmission time slots of each transmitted signal, and construct the corresponding initial channel estimation results based on each transmitted signal and each received signal; After refining each initial channel estimation result, a corresponding subspace projection matrix and orthogonal projection matrix are constructed, and each transmitted signal is projected onto the subspace projection matrix and orthogonal projection matrix to obtain multiple pairs of noise plus interference estimation components; After splicing each pair of noise-plus-interference estimation components using a noise interference estimation neural network, a fusion network incorporating the system signal-to-interference-plus-noise ratio is used to fuse all splicing results, resulting in multiple noise-plus-interference estimation results. The received signal is denoised and padded using all the noise and interference estimation results to obtain the channel estimation result for each resource unit.

[0047] The transmission time slot of the transmitted signal includes multiple OFDM symbols; The OFDM symbol represents the basic time unit for transmitting the transmitted signal; Each OFDM symbol comprises multiple subcarriers; The number of pilot resource units in the transmission time slot is the product of the number of subcarriers and the number of OFDM symbols.

[0048] Specifically, the steps of acquiring multiple transmitted signals, constructing received signals corresponding to multiple pilot resource units in all transmission time slots of each transmitted signal, and constructing corresponding initial channel estimation results based on each transmitted signal and each received signal include: Acquire multiple transmitted signals, and for each transmitted signal, construct a corresponding wireless channel matrix for all pilot resource elements in all transmission time slots: ; Where H represents the wireless channel matrix, C represents the complex field, M represents the number of transmit antennas, and N represents the number of receive antennas; Based on each transmitted signal and the corresponding wireless channel matrix, a corresponding received signal is constructed: ; ; ; ; Where y represents the received signal, x represents the transmitted signal, n represents the receiver noise, and i represents the received interference signal; For each pilot resource element, an initial channel estimate is constructed for the pilot resource element based on the radio channel matrix: ; in, This represents the initial channel estimate.

[0049] Specifically, the refinement of each initial channel estimation result, followed by the construction of a corresponding subspace projection matrix and orthogonal projection matrix, includes: The initial channel estimate is input into the channel refiner for processing, and the channel estimate result is output. For each pilot resource element, a subspace projection matrix is ​​constructed based on the channel estimation results: ; in, Represents the subspace projection matrix. This represents the channel estimation result; For each pilot resource element, based on the channel estimation results, an orthogonal projection matrix intersecting with the channel subspace is constructed: ; in, Represents the orthogonal projection matrix. Represents the identity matrix.

[0050] Specifically, projecting each transmitted signal onto the subspace projection matrix and the orthogonal projection matrix to obtain multiple pairs of noise-interference estimation components includes: After projecting each received signal onto the subspace projection matrix, noise extraction is performed to obtain multiple first noise plus interference estimation components: ; in, This represents the first noise plus interference estimation component; Each of the received signals is projected onto the orthogonal projection matrix to obtain multiple second noise plus interference estimation components: ; in, This represents the second noise plus interference estimation component; Each of the first noise plus interference estimation components is combined with the corresponding second noise plus interference estimation component to obtain multiple pairs of noise plus interference estimation components.

[0051] The noise interference estimation neural network includes: a first sub-network and a second sub-network; After the noise interference estimation neural network concatenates each pair of noise-plus-interference estimation components, a fusion network incorporating the system signal-to-interference-plus-noise ratio is used to fuse all the concatenation results, yielding multiple noise-plus-interference estimation results, specifically including: For each pair of noise-plus-interference estimation components, the first noise-plus-interference estimation component is input into the first sub-network for processing, and the first processing result is output. The second noise-interference estimation component is input into the second sub-network for processing, and the second processing result is output. The system signal-to-interference-plus-noise ratio is added to the fusion network, and the first processing result and the second processing result are concatenated and input into the fusion network; In the fusion network, the residual signal in the splicing result is classified, distinguished, and adjusted using the system signal-to-interference-plus-noise ratio, and the noise-plus-interference estimation result corresponding to each pair of noise-plus-interference estimation components is output.

[0052] Specifically, the step of using all the noise and interference estimation results to denoise and pad each received signal to obtain the channel estimation result for each resource unit includes: The noise and interference estimation results are used to denoise each of the received signals to obtain the corresponding cleaned received signals: ; in, This indicates that the received signal has been purified. This represents the estimation result with added noise and interference. The least squares channel is used to complete the channel for each of the cleaned received signals, and the channel estimation result on the current pilot resource unit is obtained.

[0053] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a noise interference suppression program oriented towards a wireless channel fundamental model, and the noise interference suppression program oriented towards a wireless channel fundamental model, when executed by a processor, implements the steps of the noise interference suppression method oriented towards a wireless channel fundamental model as described above.

[0054] In summary, this invention provides a noise and interference suppression method and related equipment for a wireless channel fundamental model. The method includes: acquiring multiple transmitted signals; constructing received signals corresponding to multiple pilot resource units in all transmission time slots of each transmitted signal; constructing corresponding initial channel estimation results based on each transmitted signal and each received signal; refining each initial channel estimation result; constructing corresponding subspace projection matrices and orthogonal projection matrices; projecting each transmitted signal onto the subspace projection matrix and the orthogonal projection matrix to obtain multiple pairs of noise and interference estimation components; using a noise and interference estimation neural network to concatenate each pair of noise and interference estimation components; and using a fusion network that incorporates the system signal-to-interference-plus-noise ratio (SINR) to fuse all concatenated results to obtain multiple noise and interference estimation results; and using all the noise and interference estimation results to denoise and complete each received signal to obtain a channel estimation result for each resource unit. This invention utilizes only pilot signals contaminated by noise and interference to recover complete channel state information and generate high-quality channel representations, making the channel representation extracted from the wireless channel fundamental model more accurate and improving the performance of downstream tasks.

[0055] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal that includes that element.

[0056] Of course, those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.). The program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The computer-readable storage medium can be a memory, magnetic disk, optical disk, etc.

[0057] It should be understood that the application of the present invention is not limited to the examples above. Those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. A noise interference suppression method for a fundamental wireless channel model, characterized in that, The noise interference suppression method for the wireless channel fundamental model includes: Acquire multiple transmitted signals, construct the received signals corresponding to multiple pilot resource units in all transmission time slots of each transmitted signal, and construct the corresponding initial channel estimation results based on each transmitted signal and each received signal; After refining each initial channel estimation result, a corresponding subspace projection matrix and orthogonal projection matrix are constructed, and each transmitted signal is projected onto the subspace projection matrix and orthogonal projection matrix to obtain multiple pairs of noise plus interference estimation components; After splicing each pair of noise-plus-interference estimation components using a noise interference estimation neural network, a fusion network incorporating the system signal-to-interference-plus-noise ratio is used to fuse all splicing results, resulting in multiple noise-plus-interference estimation results. The received signal is denoised and padded using all the noise and interference estimation results to obtain the channel estimation result for each resource unit.

2. The noise interference suppression method for a fundamental wireless channel model according to claim 1, characterized in that, The transmission time slot of the transmitted signal includes multiple OFDM symbols; The OFDM symbol represents the basic time unit for transmitting the transmitted signal; Each OFDM symbol comprises multiple subcarriers; The number of pilot resource units in the transmission time slot is the product of the number of subcarriers and the number of OFDM symbols.

3. The noise interference suppression method for a fundamental wireless channel model according to claim 2, characterized in that, The process of acquiring multiple transmitted signals, constructing received signals corresponding to multiple pilot resource units in all transmission time slots of each transmitted signal, and constructing corresponding initial channel estimation results based on each transmitted signal and each received signal specifically includes: Acquire multiple transmitted signals, and for each transmitted signal, construct a corresponding wireless channel matrix for all pilot resource elements in all transmission time slots: ； Where H represents the wireless channel matrix, C represents the complex field, M represents the number of transmit antennas, and N represents the number of receive antennas; Based on each transmitted signal and the corresponding wireless channel matrix, a corresponding received signal is constructed: ； ； ； ； Where y represents the received signal, x represents the transmitted signal, n represents the receiver noise, and i represents the received interference signal; For each pilot resource element, an initial channel estimate is constructed for the pilot resource element based on the radio channel matrix: ； in, This represents the initial channel estimate.

4. The noise interference suppression method for a fundamental wireless channel model according to claim 3, characterized in that, After refining each initial channel estimation result, the corresponding subspace projection matrix and orthogonal projection matrix are constructed, specifically including: The initial channel estimate is input into the channel refiner for processing, and the channel estimate result is output. For each pilot resource element, a subspace projection matrix is ​​constructed based on the channel estimation results: ； in, Represents the subspace projection matrix. This represents the channel estimation result; For each pilot resource element, based on the channel estimation results, an orthogonal projection matrix intersecting with the channel subspace is constructed: ； in, Represents the orthogonal projection matrix. Represents the identity matrix.

5. The noise interference suppression method for a fundamental wireless channel model according to claim 4, characterized in that, The step of projecting each transmitted signal onto the subspace projection matrix and the orthogonal projection matrix to obtain multiple pairs of noise-plus-interference estimation components specifically includes: After projecting each received signal onto the subspace projection matrix, noise extraction is performed to obtain multiple first noise plus interference estimation components: ； in, This represents the first noise plus interference estimation component; Each of the received signals is projected onto the orthogonal projection matrix to obtain multiple second noise plus interference estimation components: ； in, This represents the second noise plus interference estimation component; Each of the first noise plus interference estimation components is combined with the corresponding second noise plus interference estimation component to obtain multiple pairs of noise plus interference estimation components.

6. The noise interference suppression method for a fundamental wireless channel model according to claim 5, characterized in that, The noise interference estimation neural network includes: a first sub-network and a second sub-network; After the noise interference estimation neural network concatenates each pair of noise-plus-interference estimation components, a fusion network incorporating the system signal-to-interference-plus-noise ratio is used to fuse all the concatenation results, yielding multiple noise-plus-interference estimation results, specifically including: For each pair of noise-plus-interference estimation components, the first noise-plus-interference estimation component is input into the first sub-network for processing, and the first processing result is output. The second noise-interference estimation component is input into the second sub-network for processing, and the second processing result is output. The system signal-to-interference-plus-noise ratio is added to the fusion network, and the first processing result and the second processing result are concatenated and input into the fusion network; In the fusion network, the residual signal in the splicing result is classified, distinguished, and adjusted using the system signal-to-interference-plus-noise ratio, and the noise-plus-interference estimation result corresponding to each pair of noise-plus-interference estimation components is output.

7. The noise interference suppression method for a fundamental wireless channel model according to claim 1, characterized in that, The step of using all the noise and interference estimation results to denoise and complete each received signal to obtain the channel estimation result for each resource unit specifically includes: The noise and interference estimation results are used to denoise each of the received signals to obtain the corresponding cleaned received signals: ； in, This indicates that the received signal has been purified. This represents the estimation result with added noise and interference. The least squares channel is used to complete the channel for each of the cleaned received signals, and the channel estimation result on the current pilot resource unit is obtained.

8. A noise interference suppression system for a fundamental wireless channel model, characterized in that, The noise interference suppression system for the wireless channel fundamental model is used to implement the noise interference suppression method for the wireless channel fundamental model as described in any one of claims 1-7, wherein the noise interference suppression system for the wireless channel fundamental model includes: The channel estimation module is used to acquire multiple transmitted signals, construct the received signals corresponding to multiple pilot resource units in all transmission time slots of each transmitted signal, and construct the corresponding initial channel estimation results based on each transmitted signal and each received signal. The noise extraction module is used to refine each of the initial channel estimation results, construct the corresponding subspace projection matrix and orthogonal projection matrix, and project each of the transmitted signals into the subspace projection matrix and orthogonal projection matrix to obtain multiple pairs of noise and interference estimation components; The noise estimation module is used to perform splicing processing on each pair of noise plus interference estimation components using a noise interference estimation neural network, and then to perform fusion processing on all splicing results using a fusion network that introduces the system signal-to-interference-plus-noise ratio to obtain multiple noise plus interference estimation results. The interference cancellation and channel completion module is used to perform noise reduction and channel completion on each of the received signals using all the noise plus interference estimation results, so as to obtain the channel estimation result on each of the resource units.

9. A terminal, characterized in that, The terminal includes: a memory, a processor, and a noise interference suppression program for a wireless channel fundamental model stored in the memory and executable on the processor. When the noise interference suppression program for a wireless channel fundamental model is executed by the processor, it implements the steps of the noise interference suppression method for a wireless channel fundamental model as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a noise interference suppression program for a wireless channel fundamental model, which, when executed by a processor, implements the steps of the noise interference suppression method for a wireless channel fundamental model as described in any one of claims 1-7.

Images