Wireless signal noise reduction methods, devices, equipment and storage media
By introducing a machine learning model into the receiving device to reduce noise in the received signal, the impact of channel environment and interference noise on wireless communication is resolved, thereby improving the quality of the received signal and the system performance.
Patent Information
- Application Number
- CN202180084002.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-01-13
- Publication Date
- 2026-11-17
- Estimated Expiration
- 2041-01-13
AI Technical Summary
In existing technologies, the effects of channel environment and interference noise on wireless communication systems cannot be effectively eliminated, resulting in poor received signal quality, low signal-to-noise ratio, and impaired communication performance.
Introducing a machine learning model into the receiving device, the noise reduction model, trained using received and transmitted information samples, processes the received signal to reduce the impact of channel environment and interference noise, and selects an appropriate noise reduction model to match the current channel and noise conditions.
This improves the signal-to-noise ratio of the received signal and the system's receiving gain, thereby enhancing the transmission performance of the wireless communication system.
Smart Images

Figure CN116671024B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of communication technology, and in particular to a wireless signal noise reduction method, apparatus, device, and storage medium. Background Technology
[0002] In wireless communication systems, channel environment and interference noise are among the main problems affecting the wireless transmission performance between communication devices.
[0003] In related technologies, in order to reduce the impact of the channel environment on wireless communication, the receiver usually performs channel estimation by means of pilot signals, and the receiver performs demodulation and other processing on the received signal based on the channel estimation results.
[0004] However, channel estimation in related technologies can only provide certain parameters for receiver demodulation and other operations, and cannot effectively eliminate the impact of channel environment and interference noise on wireless signals during transmission. Summary of the Invention
[0005] This application provides a wireless signal noise reduction method, apparatus, device, and storage medium. The technical solution is as follows:
[0006] On one hand, embodiments of this application provide a wireless signal noise reduction method, which is executed by a receiving device, and the method includes:
[0007] It receives wireless signals sent by the transmitting device and obtains the received information;
[0008] The received information is processed by the first noise reduction model to obtain the noise-reduced received information;
[0009] The first noise reduction model is a machine learning model trained based on received information samples and transmitted information samples; the received information samples are obtained by receiving wireless signals from the transmitted information samples.
[0010] In one possible implementation, the received information is a wireless signal received by the antenna assembly of the receiving device;
[0011] or,
[0012] The received information is information obtained by mathematically transforming the wireless signal received by the antenna assembly of the receiving device.
[0013] In one possible implementation, the method further includes:
[0014] The channel measurement information is acquired, which includes at least one of interference noise information and valid signal reception information; the interference noise information is used to indicate the interference noise situation in the environment where the receiving device is located, and the valid signal reception information is used to indicate the receiving device's reception of the valid signal sent by the transmitting device;
[0015] Based on the channel measurement information, the first noise reduction model is selected from at least two candidate noise reduction models.
[0016] In one possible implementation, when the channel measurement information includes the interference noise information, acquiring the channel measurement information includes:
[0017] Obtain first information about the network configuration, wherein the first information is used to indicate a first resource;
[0018] Measurements are performed on the first resource indicated by the first information to obtain the interference noise information.
[0019] In one possible implementation, the step of measuring on the first resource indicated by the first information to obtain the interference noise information includes:
[0020] Measurements are performed on the first resource indicated by the first information to obtain first measurement information, the first measurement information including at least one of RSRP, RSRQ, RSSI and SINR;
[0021] Based on the first measurement information, the interference noise information is obtained.
[0022] In one possible implementation,
[0023] When the receiving device is a terminal, the first information is information configured by the network-side device through at least one of the following: broadcast message, System Information Block (SIB), Radio Resource Control (RRC) message, RRC reconfiguration signaling, Downlink Control Information (DCI), Media Access Control (MAC) layer control element (CE), and Physical Downlink Control Channel (PDCCH) order.
[0024] When the receiving device is a network-side device, the first information is predefined information; or, the first information is determined by the network-side device based on unused resources.
[0025] In one possible implementation, when the channel measurement information includes the valid signal reception information, acquiring the channel measurement information includes:
[0026] Obtain second information about the network configuration, which is used to indicate a second resource;
[0027] On the second resource indicated by the second information, the specified signal sent by the receiving device is measured to obtain the valid signal reception information.
[0028] In one possible implementation, measuring the specified signal transmitted by the receiving device on the second resource indicated by the second information to obtain the valid signal reception information includes:
[0029] The specified signal transmitted by the receiving device is measured on the second resource indicated by the second information to obtain second measurement information, the second measurement information including at least one of RSRP, RSRQ, RSSI, SINR, packet loss rate, and bit error rate;
[0030] Based on the second measurement information, the valid signal reception information is obtained.
[0031] In one possible implementation,
[0032] When the receiving device is a terminal, the second information is information configured by the network-side device through at least one of broadcast messages, SIB, RRC messages, RRC reconfiguration signaling, DCI, MAC CE, and PDCCH order;
[0033] When the receiving device is a network-side device, the second information is predefined information; or, the second information is determined by the network-side device based on unused resources.
[0034] In one possible implementation,
[0035] When the receiving device is a terminal, the designated signal is a signal predefined by the protocol, or the designated signal is a signal configured by the network-side device through at least one of broadcast messages, SIB, RRC messages, RRC reconfiguration signaling, DCI, MAC CE, and PDCCHorder;
[0036] When the receiving device is a network-side device, the designated signal is a signal predefined by the protocol, or the designated signal is a signal configured by the network-side device to the sending device through at least one of broadcast messages, SIB, RRC messages, RRC reconfiguration signaling, DCI, MACCE, and PDCCH order.
[0037] In one possible implementation, when the receiving device is the network-side device, the method further includes:
[0038] The second information is configured to the sending device via at least one of the following: broadcast message, SIB, RRC message, RRC reconfiguration signaling, DCI, MAC CE, and PDCCH order.
[0039] In one possible implementation, the first noise reduction model is a fully connected neural network model consisting of N fully connected layers, where N ≥ 1 and N is an integer.
[0040] In one possible implementation, the first noise reduction model is a convolutional neural network model consisting of M convolutional layers, where M ≥ 1 and M is an integer.
[0041] In one possible implementation, the first noise reduction model sequentially includes a first fully connected layer, a first dimension adjustment layer, a first convolutional layer, L common layers, a second convolutional layer, a second dimension adjustment layer, and a second fully connected layer;
[0042] The common layer contains a third convolutional layer, a normalization layer, and an activation layer connected in sequence; the L common layers are connected in sequence; L≥1, and A is an integer.
[0043] In one possible implementation, there is a sampling operation between the first dimension adjustment layer and the first convolutional layer, and a stacking operation between the second convolutional layer and the second dimension adjustment layer;
[0044] The sampling operation is used to sample the output of the first dimension adjustment layer and output the sampling result to the first convolutional layer and the superposition operation respectively; the superposition operation is used to superimpose the output of the second convolutional layer and the output of the sampling operation and output it to the second dimension adjustment layer.
[0045] On the other hand, embodiments of this application provide a wireless signal noise reduction device, which is used in a receiving device, and the device includes:
[0046] The receiving module is used to receive wireless signals sent by the transmitting device and obtain the received information;
[0047] The noise reduction processing module is used to process the received information through a first noise reduction model to obtain the noise-reduced received information;
[0048] The first noise reduction model is a machine learning model trained based on received information samples and transmitted information samples; the received information samples are obtained by receiving wireless signals from the transmitted information samples.
[0049] In one possible implementation, the received information is a wireless signal received by the antenna assembly of the receiving device;
[0050] or,
[0051] The received information is information obtained by mathematically transforming the wireless signal received by the antenna assembly of the receiving device.
[0052] In one possible implementation, the device further includes:
[0053] A measurement information acquisition module is used to acquire channel measurement information, which includes at least one of interference noise information and effective signal reception information; the interference noise information is used to indicate the interference noise situation in the environment where the receiving device is located, and the effective signal reception information is used to indicate the reception status of the receiving device for the effective signal sent by the transmitting device.
[0054] The model selection module is used to select the first noise reduction model from at least two candidate noise reduction models based on the channel measurement information.
[0055] In one possible implementation, when the channel measurement information includes the interference noise information, the measurement information acquisition module is used to:
[0056] Obtain first information about the network configuration, wherein the first information is used to indicate a first resource;
[0057] Measurements are performed on the first resource indicated by the first information to obtain the interference noise information.
[0058] In one possible implementation, the measurement information acquisition module is used for:
[0059] Measurements are performed on the first resource indicated by the first information to obtain first measurement information, the first measurement information including at least one of RSRP, RSRQ, RSSI and SINR;
[0060] Based on the first measurement information, the interference noise information is obtained.
[0061] In one possible implementation, when the receiving device is a terminal, the first information is information configured by the network-side device through at least one of broadcast messages, SIB, RRC messages, RRC reconfiguration signaling, DCI, MAC CE, and PDCCH order.
[0062] When the receiving device is a network-side device, the first information is predefined information; or, the first information is determined by the network-side device based on unused resources.
[0063] In one possible implementation, when the channel measurement information includes the valid signal reception information, the measurement information acquisition module is used to:
[0064] Obtain second information about the network configuration, which is used to indicate a second resource;
[0065] On the second resource indicated by the second information, the specified signal sent by the receiving device is measured to obtain the valid signal reception information.
[0066] In one possible implementation, the measurement information acquisition module is used for:
[0067] The specified signal transmitted by the receiving device is measured on the second resource indicated by the second information to obtain second measurement information, the second measurement information including at least one of RSRP, RSRQ, RSSI, SINR, packet loss rate, and bit error rate;
[0068] Based on the second measurement information, the valid signal reception information is obtained.
[0069] In one possible implementation,
[0070] When the receiving device is a terminal, the second information is information configured by the network-side device through at least one of broadcast messages, SIB, RRC messages, RRC reconfiguration signaling, DCI, MAC CE, and PDCCH order;
[0071] When the receiving device is a network-side device, the second information is predefined information; or, the second information is determined by the network-side device based on unused resources.
[0072] In one possible implementation,
[0073] When the receiving device is a terminal, the designated signal is a signal predefined by the protocol, or the designated signal is a signal configured by the network-side device through at least one of broadcast messages, SIB, RRC messages, RRC reconfiguration signaling, DCI, MAC CE, and PDCCHorder;
[0074] When the receiving device is a network-side device, the designated signal is a signal predefined by the protocol, or the designated signal is a signal configured by the network-side device to the sending device through at least one of broadcast messages, SIB, RRC messages, RRC reconfiguration signaling, DCI, MACCE, and PDCCH order.
[0075] In one possible implementation, when the receiving device is the network-side device, the apparatus further includes:
[0076] The configuration module is used to configure the second information to the sending device via at least one of broadcast messages, SIB, RRC messages, RRC reconfiguration signaling, DCI, MAC CE, and PDCCHorder.
[0077] In one possible implementation, the first noise reduction model is a fully connected neural network model consisting of N fully connected layers, where N ≥ 1 and N is an integer.
[0078] In one possible implementation, the first noise reduction model is a convolutional neural network model consisting of M convolutional layers, where M ≥ 1 and M is an integer.
[0079] In one possible implementation, the first noise reduction model sequentially includes a first fully connected layer, a first dimension adjustment layer, a first convolutional layer, L common layers, a second convolutional layer, a second dimension adjustment layer, and a second fully connected layer;
[0080] The common layer contains a third convolutional layer, a normalization layer, and an activation layer connected in sequence; the L common layers are connected in sequence; L≥1, and L is an integer.
[0081] In one possible implementation, there is a sampling operation between the first dimension adjustment layer and the first convolutional layer, and a stacking operation between the second convolutional layer and the second dimension adjustment layer;
[0082] The sampling operation is used to sample the output of the first dimension adjustment layer and output the sampling result to the first convolutional layer and the superposition operation respectively; the superposition operation is used to superimpose the output of the second convolutional layer and the output of the sampling operation and output it to the second dimension adjustment layer.
[0083] In another aspect, embodiments of this application provide a receiving device, which includes a processor, a memory, and a transceiver. The memory stores a computer program, which is executed by the processor / transceiver to implement the above-described wireless signal noise reduction method.
[0084] In another aspect, embodiments of this application also provide a computer-readable storage medium storing a computer program, which is loaded and executed by a processor / transceiver to implement the above-described wireless signal noise reduction method.
[0085] In another aspect, this application provides a computer program product or computer program that includes computer instructions stored in a computer-readable storage medium. The processor / transceiver of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the aforementioned wireless signal noise reduction method.
[0086] The technical solution provided in this application can bring the following beneficial effects:
[0087] By adding a machine learning model trained on received and transmitted information samples to the receiving device, noise reduction is first performed on the received information before source information recovery. This reduces the impact of channel environment and interference noise on the wireless signal, thereby improving the signal-to-noise ratio of the received signal and the system's receiving gain, and ultimately enhancing the transmission performance of the wireless communication system. Attached Figure Description
[0088] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0089] Figure 1 This is a schematic diagram of a network architecture provided in one embodiment of this application;
[0090] Figure 2 This is a schematic diagram of a communication system provided in one embodiment of this application;
[0091] Figure 3 This is a schematic diagram of a neural network provided in one embodiment of this application;
[0092] Figure 4 This is a schematic diagram of a neural network provided in another embodiment of this application;
[0093] Figure 5 This is a flowchart of a wireless signal noise reduction method provided in one embodiment of this application;
[0094] Figure 6 yes Figure 5 The illustrated embodiment is a schematic diagram of a communication system.
[0095] Figure 7 This is a flowchart of a wireless signal noise reduction method provided in one embodiment of this application;
[0096] Figure 8 yes Figure 7 The illustrated embodiment is a schematic diagram of the noise reduction process on the terminal side.
[0097] Figure 9 yes Figure 7 The illustrated embodiment is a schematic diagram of the noise reduction process on the terminal side.
[0098] Figure 10 yes Figure 7 The illustrated embodiment is a schematic diagram of the noise reduction process on the network side.
[0099] Figure 11 yes Figure 7 The illustrated embodiment is a structural diagram of a noise reduction model.
[0100] Figure 12 yes Figure 7 A schematic diagram of another noise reduction model involved in the embodiment shown;
[0101] Figure 13 yes Figure 7 A schematic diagram of another noise reduction model involved in the embodiment shown;
[0102] Figure 14 This is a block diagram of a wireless signal noise reduction device provided in one embodiment of this application;
[0103] Figure 15 This is a schematic diagram of the structure of a receiving device provided in one embodiment of this application. Detailed Implementation
[0104] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.
[0105] The network architecture and business scenarios described in the embodiments of this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided in the embodiments of this application. As those skilled in the art will know, with the evolution of network architecture and the emergence of new business scenarios, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.
[0106] Please refer to Figure 1 This diagram illustrates a network architecture of a communication system provided in one embodiment of this application. The network architecture may include: a terminal 10 and a base station 20.
[0107] The number of terminals 10 is typically multiple, with one or more terminals 10 distributed within the cell managed by each base station 20. Terminals 10 may include various handheld devices, vehicle-mounted devices, wearable devices, computing devices, or other processing devices connected to a wireless modem, as well as various forms of user equipment (UE), mobile station (MS), terminal device, etc. For ease of description, in this embodiment, the devices mentioned above are collectively referred to as terminals.
[0108] Base station 20 is a device deployed in an access network to provide wireless communication functionality to terminal 10. Base station 20 can include various forms of macro base stations, micro base stations, relay stations, access points, etc. In systems employing different wireless access technologies, the name of the device with base station functionality may differ; for example, in a 5th-Generation (5G) NR system, it is called gNodeB or gNB. As communication technologies evolve, the name "base station" may change. For ease of description, in this embodiment, the device providing wireless communication functionality to terminal 10 is collectively referred to as a base station.
[0109] Optional, Figure 1 As not shown, the above network architecture also includes other network devices, such as: Central Network Control (CNC), Access and Mobility Management Function (AMF) devices, Session Management Function (SMF) or User Plane Function (UPF) devices, etc.
[0110] The "5G NR system" in this disclosure can also be referred to as a 5G system or an NR system, but those skilled in the art will understand its meaning. The technical solutions described in this disclosure are applicable to 4G systems, 5G NR systems, and also to subsequent evolution systems of 5G NR systems.
[0111] To facilitate understanding, the following is an introduction to some relevant terms or background concepts involved in this application:
[0112] I. Wireless Communication.
[0113] Please refer to Figure 2 This illustrates a schematic diagram of a communication system provided in one embodiment of this application. Figure 2 As shown, in a wireless communication system, the basic workflow is as follows: the transmitter encodes, modulates, and encrypts the source information at the transmitting end to form the information to be transmitted. The transmitted information is transmitted to the receiving end via wireless space, where the receiving end decodes, decrypts, and demodulates the received information to ultimately recover the source information.
[0114] In the aforementioned process, the encoding, modulation, encryption, decoding, demodulation, and decryption operations at the transmitting and receiving ends are controllable. However, the channel conditions and noise in the space environment are uncontrollable, complex, and variable. Currently, to better recover source information, it is necessary to estimate the channel environment in space to match the appropriate algorithm for better information reception at the receiving end. For interference noise in space, there is a lack of necessary processing solutions, resulting in significant differences in source recovery performance under different signal-to-noise ratios.
[0115] II. Artificial Intelligence.
[0116] In recent years, artificial intelligence research, represented by neural networks, has achieved remarkable results in many fields, and it will play an important role in people's production and life for a long time to come.
[0117] Please refer to Figure 3 This illustrates a schematic diagram of a neural network provided in one embodiment of this application. Figure 3 As shown, the basic structure of a simple neural network includes: an input layer, hidden layers, and an output layer, as follows: Figure 2 As shown, the input layer receives data, the hidden layers process the data, and the final result is generated in the output layer. Each node represents a processing unit, which can be considered as simulating a neuron. Multiple neurons form a layer of the neural network, and the information transmission and processing across multiple layers construct a complete neural network.
[0118] With the continuous development of neural network research, deep learning algorithms for neural networks have been proposed in recent years. More hidden layers have been introduced, and feature learning is carried out by training the neural network layer by layer through multiple hidden layers. This has greatly improved the learning and processing capabilities of the neural network and has been widely used in pattern recognition, signal processing, optimization and combination, anomaly detection and other fields.
[0119] Similarly, with the development of deep learning, convolutional neural networks have been further studied. Please refer to... Figure 4This illustrates a neural network diagram provided in another embodiment of this application. Figure 4 As shown, its basic structure includes: an input layer, multiple convolutional layers, multiple pooling layers, a fully connected layer, and an output layer. The introduction of convolutional and pooling layers effectively controls the dramatic increase in network parameters, limits the number of parameters, and taps into the characteristics of local structures, thereby improving the robustness of the algorithm.
[0120] Based on the above introduction to the basic principles of wireless communication transceiver systems, it is clear that channel environment and noise conditions are currently the biggest problems affecting transmission schemes. Regarding channel environment, channel estimation schemes considering methods such as pilot signals can mitigate the impact of channel uncertainty on communication performance to some extent. However, spatial noise is much more difficult to eliminate. Traditional communication algorithms struggle to process and eliminate noise in the received signal. While some methods are beneficial for noise cancellation, their algorithmic complexity is extremely high, and their effectiveness is relatively limited.
[0121] To address the aforementioned issues, this solution considers incorporating a noise reduction network at the receiving end. This network performs noise reduction on the received signal to achieve a better signal-to-noise ratio and improved receiving gain.
[0122] Please refer to Figure 5 The diagram illustrates a flowchart of a wireless signal noise reduction method according to an embodiment of this application. This method can be executed by a receiving device, wherein the receiving device can be implemented as... Figure 1 For terminals or base stations in the communication system corresponding to the network architecture shown, the method may include the following steps:
[0123] Step 501: Receive the wireless signal sent by the transmitting device and obtain the received information.
[0124] In this embodiment, the receiving device can receive the wireless signal transmitted by the transmitting device through an antenna assembly to obtain the original received information. This received information typically carries some noise due to channel conditions and interference noise.
[0125] Step 502: Process the received information using the first noise reduction model to obtain the noise-reduced received information.
[0126] The first noise reduction model is a machine learning model trained based on received information samples and transmitted information samples; the received information samples are obtained by receiving wireless signals from the transmitted information samples.
[0127] Please refer to Figure 6, which shows a schematic diagram of a communication system according to an embodiment of this application. Figure 6As shown in this embodiment, a noise reduction processing unit is added to the receiving end of the communication system. The noise reduction processing unit is used to perform noise reduction processing on the received information, and it may include one or more noise reduction models. Optionally, the received information can be improved after noise reduction processing, for example, from 10dB to 16dB.
[0128] The output of the above noise reduction model can be used by the receiving device to perform corresponding processing to recover the source information, such as performing corresponding decoding, demodulation, and decryption operations.
[0129] In summary, the solution shown in this application adds a machine learning model trained on received and transmitted information samples to the receiving device. Before recovering the source information of the received information, the machine learning model is used to perform noise reduction on the received information, thereby reducing the impact of channel environment and interference noise on the wireless signal, improving the signal-to-noise ratio of the received signal, as well as the receiving gain of the system, and improving the transmission performance of the wireless communication system.
[0130] In practical applications, the channel environment and interference noise conditions vary depending on factors such as different geographical locations and time periods. A single noise reduction model is often insufficient to effectively handle all channel environments and interference noise conditions. Therefore, in the solution presented in this application, the system can set up multiple noise reduction models (for example, multiple model structures, each model structure corresponding to multiple model parameters, and each model result combined with a set of model parameters constitutes a noise reduction model). The receiving device can select a noise reduction model that matches the current channel environment and / or interference noise conditions for noise reduction processing based on the channel measurement results.
[0131] Please refer to Figure 7 The diagram illustrates a flowchart of a wireless signal noise reduction method according to an embodiment of this application. This method can be executed by a receiving device, wherein the receiving device can be implemented as... Figure 1 For terminals or base stations in the communication system corresponding to the network architecture shown, the method may include the following steps:
[0132] Step 701: Receive the wireless signal sent by the transmitting device and obtain the received information.
[0133] The received information is the wireless signal received by the antenna assembly of the receiving device;
[0134] or,
[0135] The received information is obtained by mathematically transforming the wireless signal received by the antenna assembly of the receiving device.
[0136] In this embodiment, the input to the noise reduction model may include: the original wireless signal received by the receiving device through the antenna assembly, or the information after the receiving device has processed the original wireless signal through mathematical transformation (such as at least one of Fourier transform and singular decomposition). The output of the noise reduction network includes: the noise-reduced information of the receiving end, and the information after the receiving end's received information has undergone mathematical transformation and then further processed by the noise reduction network.
[0137] Step 702: Obtain channel measurement information, which includes at least one of interference noise information and effective signal reception information.
[0138] The interference noise information is used to indicate the interference noise situation in the environment where the receiving device is located, and the valid signal reception information is used to indicate the reception status of the receiving device of the valid signal sent by the transmitting device.
[0139] In the embodiments of this application, the receiving device can measure the interference noise situation or the reception situation of the valid signal by measuring the channel, so as to accurately select the noise reduction model used for noise reduction processing of the received signal.
[0140] In one possible implementation, when the channel measurement information includes the interference noise information, acquiring the channel measurement information includes:
[0141] Obtain first information about the network configuration, which is used to indicate the first resource;
[0142] The interference noise information is obtained by measuring the first resource indicated by the first information.
[0143] In this embodiment of the application, the communication resources (i.e. the first resource mentioned above) used by the receiving device to measure the interference noise of the channel can be determined or indicated by the network.
[0144] For example, when the receiving device is a UE, the network side configures the first information to the UE.
[0145] In one possible implementation, the first information is information configured by the network-side device through at least one of broadcast messages, SIBs, RRC messages, RRC reconfiguration signaling, DCI, MAC CE, and PDCCH orders.
[0146] In other words, during the UE's network access process, or after accessing the network, the network-side equipment (such as the base station) configures the aforementioned first information to the UE through at least one of the following: broadcast message, SIB, RRC message, RRC reconfiguration signaling, DCI, MAC CE, and PDCCH order.
[0147] For example, please refer to Figure 8 This illustrates a schematic diagram of the noise reduction process on the terminal side according to an embodiment of this application. For example... Figure 8 As shown, the noise reduction process on the terminal side is as follows:
[0148] S81, the network side indicates the first information to the UE;
[0149] S82, the UE performs measurements on the resources indicated by the first information to estimate the current environmental noise situation, that is, to obtain interference noise information;
[0150] S83, the UE determines the noise reduction processing unit (corresponding noise reduction model) based on the current environmental noise conditions and performs noise reduction processing on the received information.
[0151] For example, when the receiving device is a network-side device (such as a base station), the first information is predefined information; or, the first information is determined by the network-side device based on unused resources.
[0152] Among them, the network-side devices can determine the first information in a predefined way, including protocol predefinition or operator predefinition. In this way, one or more network devices can use a set of resources predefined by the protocol to estimate the corresponding environmental noise.
[0153] Alternatively, the network-side device can determine the first information by analyzing the resource allocation of the network-side device. For example, the network device can select resources that are not used for data transmission (e.g., resources not allocated for uplink reception and / or resources not allocated for downlink transmission) to estimate the environmental noise situation.
[0154] In one exemplary scheme, the resource information indicated by the first information may include time-domain resource information and frequency-domain resource information. For example, the first information may indicate the start position, duration, end position, period information, start position, range, end position, mode information, etc. in the time domain.
[0155] In one exemplary scheme, the resource information indicated by the first information mentioned above can be single resource information, multiple resource information, or periodic resource information.
[0156] In one possible implementation, the measurement performed on the first resource indicated by the first information to obtain the interference noise information includes:
[0157] Measurements are performed on the first resource indicated by the first information to obtain first measurement information, which includes at least one of Reference Signal Receiving Power (RSRP), Reference Signal Receiving Quality (RSRQ), Received Signal Strength Indication (RSSI), and Signal to Interference plus Noise Ratio (SINR).
[0158] Based on the first measurement information, the interference noise information is obtained.
[0159] In this embodiment of the application, when the receiving device measures interference noise, it can measure parameters such as RSRP, RSRQ, RSSI and SINR on the first resource indicated by the first information, and then determine the interference noise information for model selection based on the measured parameters.
[0160] In one exemplary scheme, the receiving device can directly use at least one of the measured parameters RSRP, RSRQ, RSSI, and SINR as the aforementioned interference noise information.
[0161] In another exemplary scheme, the receiving device can perform a preset fusion calculation or mapping based on the measured RSRP, RSRQ, RSSI, and SINR information to obtain the interference noise information corresponding to the measured information. For example, based on the measured RSRP, RSRQ, RSSI, and SINR information, an interference noise level can be obtained.
[0162] For example, the receiving device can perform a weighted summation on the measured RSRP, RSRQ, RSSI, and SINR information, and use the result of the weighted summation as the aforementioned interference noise information, or use the interference noise level corresponding to the result of the weighted summation as the aforementioned interference noise information.
[0163] In one possible implementation, when the channel measurement information includes the valid signal reception information, the acquisition of the channel measurement information includes:
[0164] Obtain second information about the network configuration, which is used to indicate a second resource;
[0165] On the second resource indicated by the second information, the specified signal sent by the receiving device is measured to obtain the valid signal reception information.
[0166] In this embodiment of the application, the communication resources (i.e. the second resource mentioned above) used by the receiving device to measure the effective signal reception status of the channel can be determined or indicated by the network.
[0167] For example, when the receiving device is a UE, the network side configures second information for the UE.
[0168] In one possible implementation, the second information is information configured by the network-side device through at least one of broadcast messages, SIBs, RRC messages, RRC reconfiguration signaling, DCI, MAC CE, and PDCCH orders.
[0169] In other words, during the UE's network access process, or after accessing the network, the network-side equipment (such as the base station) configures the aforementioned second information to the UE through at least one of the following: broadcast message, SIB, RRC message, RRC reconfiguration signaling, DCI, MAC CE, and PDCCH order.
[0170] In one possible implementation, when the receiving device is a terminal, the designated signal is a signal predefined by the protocol, or the designated signal is a signal configured by the network-side device through at least one of broadcast messages, SIB, RRC messages, RRC reconfiguration signaling, DCI, MAC CE, and PDCCH order.
[0171] In this process, the receiving device and the transmitting device need to predetermine the specified signal and transmit it on the second resource indicated by the second information, so that the receiving device can measure the effective signal reception of the specified signal.
[0172] In one exemplary scheme, when the receiving device is a terminal, the designated signal is predefined by the protocol, so that the terminal and the network-side device can determine the same designated signal through the predefined information of the protocol.
[0173] In another exemplary scheme, when the receiving device is a terminal, the specified signal is configured to be sent to the terminal by the network-side device when the terminal accesses the network, or after the terminal accesses the network, by the network-side device through at least one of the following: broadcast message, SIB, RRC message, RRC reconfiguration signaling, DCI, MAC CE, and PDCCH order.
[0174] For example, please refer to Figure 9 This illustrates a schematic diagram of the noise reduction process on the terminal side according to an embodiment of this application. For example... Figure 9 As shown, the noise reduction process on the terminal side is as follows:
[0175] S91, the network side instructs the UE to provide the second information;
[0176] S92, the UE performs measurements on the resources indicated by the second information to estimate the current effective signal reception status, that is, obtains the above-mentioned effective signal reception information;
[0177] S93, the UE determines the noise reduction processing unit (corresponding noise reduction model) based on the current valid signal reception, and performs noise reduction processing on the received information.
[0178] For example, when the receiving device is a network-side device (such as a base station), the second information is predefined information; or, the second information is determined by the network-side device based on unused resources.
[0179] The network-side device can determine the second information in a predefined way, including protocol predefinition or operator predefinition. In this way, one or more network devices can use a set of resources predefined by the protocol to estimate the corresponding effective signal reception.
[0180] Alternatively, the network-side device can determine the second information by analyzing its resource allocation. For example, the network device can select resources that are not used for data transmission (e.g., resources not allocated for uplink reception and / or resources not allocated for downlink transmission) to estimate the effective signal reception status.
[0181] In this embodiment, when the receiving device is a network-side device (such as a base station), the network-side device also needs to instruct the sending device (such as a UE) to send the specified signal on the resource corresponding to the second information, so that the network-side device can measure the effective signal reception of the specified signal. Therefore, in this embodiment, when the receiving device is a network-side device, the network-side device also configures the second information to the sending device (such as a terminal) through at least one of the following: broadcast message, SIB, RRC message, RRC reconfiguration signaling, DCI, MAC CE, and PDCCH order.
[0182] Correspondingly, when the receiving device is a network-side device, the designated signal is a signal predefined by the protocol, or the designated signal is a signal configured by the network-side device to the sending device through at least one of broadcast messages, SIB, RRC messages, RRC reconfiguration signaling, DCI, MACCE, and PDCCH order.
[0183] For example, please refer to Figure 10 This illustrates a schematic diagram of the network-side noise reduction process involved in an embodiment of this application. For example... Figure 10 As shown, the noise reduction process on the terminal side is as follows:
[0184] S1001, The network side instructs the UE to provide the second information;
[0185] S1002, the UE sends a specified signal on the resource indicated by the second information;
[0186] S1003, the network side performs measurements on the resources indicated by the second information to estimate the current effective signal reception status, that is, to obtain the above-mentioned effective signal reception information.
[0187] S1004, the network side determines the noise reduction processing unit (corresponding noise reduction model) based on the current effective signal reception, and performs noise reduction processing on the received information.
[0188] In one possible implementation, measuring the specified signal transmitted by the receiving device on the second resource indicated by the second information to obtain the valid signal reception information includes:
[0189] The second measurement information is obtained by measuring the specified signal transmitted by the receiving device on the second resource indicated by the second information. The second measurement information includes at least one of RSRP, RSRQ, RSSI, SINR, packet loss rate, and bit error rate.
[0190] Based on the second measurement information, the valid signal reception information is obtained.
[0191] In this embodiment of the application, when the receiving device measures the effective signal condition, it can measure and obtain parameters such as RSRP, RSRQ, RSSI, SINR, packet loss rate, and bit error rate on the first resource indicated by the first information, and then determine the effective signal reception information for model selection based on the measured parameters.
[0192] In one exemplary scheme, the receiving device can directly use at least one of the measured parameters RSRP, RSRQ, RSSI, SINR, packet loss rate, and bit error rate as the aforementioned valid signal reception information.
[0193] In another exemplary scheme, the receiving device can perform preset fusion calculations or mappings based on the measured RSRP, RSRQ, RSSI, SINR, packet loss rate, and bit error rate to obtain the effective signal reception information corresponding to the measured information. For example, based on the measured RSRP, RSRQ, RSSI, SINR, packet loss rate, and bit error rate, an effective signal reception level can be obtained.
[0194] For example, the receiving device can perform a weighted summation on the measured RSRP, RSRQ, RSSI, SINR, packet loss rate, and bit error rate, and use the result of the weighted summation as the aforementioned valid signal reception information, or use the valid signal reception level corresponding to the result of the weighted summation as the aforementioned valid signal reception information.
[0195] Step 703: Based on the channel measurement information, select the first noise reduction model from at least two candidate noise reduction models.
[0196] In this embodiment of the application, the receiving device can pre-set multiple noise reduction processing units, each noise reduction processing unit can correspond to a candidate noise reduction model; and the receiving device can also pre-set the correspondence between each noise reduction processing unit and the channel measurement information; after the receiving device obtains the channel measurement information, it can determine the noise reduction processing unit corresponding to the first noise reduction model currently in use according to the pre-set correspondence.
[0197] Alternatively, in another exemplary scheme, each noise reduction processing unit can correspond to a set of noise reduction model structures and multiple sets of model parameters. Each set of model parameters can constitute a noise reduction model. Furthermore, the receiving device can pre-set the model structure corresponding to each noise reduction processing unit, as well as the correspondence between each set of model parameters corresponding to the model structure and the channel measurement information. After obtaining the aforementioned channel measurement information, the receiving device can determine the corresponding noise reduction processing unit and the model parameters used by the model in the noise reduction processing unit according to the pre-set correspondence. The determined model structure in the noise reduction unit combined with the determined model parameters constitutes the aforementioned first noise reduction model.
[0198] In this embodiment of the application, the first information and the second information can be configured separately or uniformly.
[0199] The noise estimation based on the first information and the effective signal transmission estimation based on the second information can be used individually or in combination to determine the environmental noise and effective signal reception.
[0200] In other words, when the receiving device determines the first noise reduction model, it can use the interference noise information alone, or it can use the effective signal reception information alone, or it can combine the interference noise information and the effective signal reception information to determine the first noise reduction model.
[0201] For example, when determining the first noise reduction model by combining interference noise information and effective signal reception information, the receiving device can fuse the interference noise information and effective signal reception information (e.g., by weighted summation) to obtain the fused information, and determine the first noise reduction model based on the correspondence between the fused information and the noise reduction model.
[0202] In one possible implementation, the first noise reduction model is a fully connected neural network model consisting of N fully connected layers, where N ≥ 1 and N is an integer.
[0203] Please refer to Figure 11 The diagram illustrates a structural schematic of a noise reduction model according to an embodiment of this application. Figure 11 As shown, in an exemplary scheme, the denoising model can employ a fully connected network, meaning that the denoising processing unit is composed of fully connected networks. This fully connected network consists of N fully connected layers, with each layer containing Cn neurons. Figure 11 In the process, the noise-reduced information 1101 is output after passing through N fully connected layers 1102.
[0204] In one possible implementation, the first noise reduction model is a convolutional neural network model consisting of M convolutional layers, where M ≥ 1 and M is an integer.
[0205] Please refer to Figure 12 This illustrates a structural schematic diagram of another noise reduction model involved in an embodiment of this application. For example... Figure 12 As shown, in an exemplary scheme, the noise reduction processing unit is composed of a convolutional neural network. This convolutional neural network consists of M convolutional layers, with the m-th layer having Km convolutional kernels. The kernel dimension Pm and kernel size of the m-th layer are Dm = [D_1(m), D_2(m), ..., D_Pm(m)]. Figure 12 In the process, the noise-reduced information 1201 is passed through the M-layer convolutional layer 1202 and then outputs the noise-reduced information 1203.
[0206] In one possible implementation, the first noise reduction model sequentially includes a first fully connected layer, a first dimension adjustment layer, a first convolutional layer, L common layers, a second convolutional layer, a second dimension adjustment layer, and a second fully connected layer.
[0207] The common layer contains a third convolutional layer, a normalization layer, and an activation layer connected in sequence; the L common layers are connected in sequence; L≥1, and A is an integer.
[0208] The first fully connected layer may contain a single fully connected layer or may be composed of at least two fully connected layers connected in sequence. In other words, the first fully connected layer may have one, two, or more layers.
[0209] Similarly, the first convolutional layer described above contains a single convolutional layer, or is composed of at least two convolutional layers connected in sequence; the second convolutional layer contains a single convolutional layer, or is composed of at least two convolutional layers connected in sequence; the third convolutional layer contains a single convolutional layer, or is composed of at least two convolutional layers connected in sequence; and the second fully connected layer contains a single fully connected layer, or is composed of at least one fully connected layer connected in sequence.
[0210] In one possible implementation, a sampling operation is performed between the first dimension adjustment layer and the first convolutional layer, and a superposition operation is performed between the second convolutional layer and the second dimension adjustment layer. The sampling operation is used to sample the output of the first dimension adjustment layer and output the sampled result to the first convolutional layer and the superposition operation respectively. The superposition operation is used to superimpose the output of the second convolutional layer with the output of the sampling operation and output the result to the second dimension adjustment layer, thereby forming a deep residual network (ResNet).
[0211] Please refer to Figure 13 This illustrates a structural schematic diagram of another noise reduction model involved in an embodiment of this application. For example... Figure 13 As shown, in an exemplary scheme, the information to be denoised 1301 first passes through at least one fully connected layer ( Figure 13 (As shown in the diagram, 1), then after dimensionality adjustment, sampling is performed, and it enters at least one convolutional layer ( Figure 13 (As shown in the diagram, 1), then enters L serially connected common layer modules, and then passes through at least one convolutional layer ( Figure 13 The sample is shown as 1 and superimposed with the original sample, and after dimensionality adjustment, it is passed through at least one fully connected layer. Figure 13 The image shows one (as shown) undergoing post-processing, finally yielding the denoised information 1302, which is then output. The common layer module here consists of at least one convolutional layer (…). Figure 13 It consists of one layer, a normalization layer, and an activation function.
[0212] The dimension adjustment operation of the first dimension adjustment layer is the opposite of the dimension adjustment operation of the second dimension adjustment layer. For example, the dimension adjustment operation of the first dimension adjustment layer adjusts the dimension of the input signal from 1×1024 to 32×32; the dimension adjustment operation of the second dimension adjustment layer adjusts the dimension of the input signal from 32×32 to 1×1024.
[0213] For example, the noise reduction information 1301 input in the first noise reduction model mentioned above can be modulation symbols arranged in chronological order and with a fixed length, such as 1024 modulation symbols input each time. These input modulation symbols can be modulation symbols received by the antenna assembly, or they can be the result of mathematical transformation of the modulation symbols received by the antenna assembly. The noise reduction information 1301 first passes through one or more fully connected layers, and then undergoes dimension adjustment to obtain data with a dimension of 32×32. After subsequent sampling processing, one or more convolutional layers, and common layer processing, it is superimposed with the original sampled signal obtained from the sampling processing. Then, after dimension adjustment and one or more fully connected layers, 1024 noise-reduced modulation symbols are obtained.
[0214] Step 704: Process the received information using the first noise reduction model to obtain the noise-reduced received information.
[0215] Wherein, when the input of the above-mentioned noise reduction network includes: the wireless signal received by the receiving device through the antenna component, the output of the first noise reduction model includes: the information after noise reduction processing of the wireless signal.
[0216] When the input to the aforementioned noise reduction network includes: information of the wireless signal received by the receiving device through the antenna assembly after mathematical transformation, the output of the first noise reduction model includes: information after noise reduction processing of the information after mathematical transformation.
[0217] The noise reduction model involved in this application embodiment is trained using received information samples and transmitted information samples. The transmitted information samples can be information transmitted by the transmitting device via an antenna assembly, while the received information samples can be information received by the receiving device via an antenna assembly. The transmitted information samples can be modulation symbols obtained by the transmitting device's transmitter through encryption, encoding, and modulation processing of the information source. The received information samples can be quasi-modulation symbols received by the receiving device via an antenna.
[0218] In this embodiment of the application, since the receiving device is equipped with multiple candidate noise reduction models, correspondingly, during the model training phase, multiple sample sets can be set for different effective signal reception conditions and / or interference noise conditions, and candidate noise reduction models corresponding to different effective signal reception conditions and / or interference noise conditions can be trained according to the sample sets corresponding to the different effective signal reception conditions and / or interference noise conditions.
[0219] In one possible training process for the noise reduction model, the received information samples can be input into a pre-set model architecture (e.g., Figure 13 In the machine learning model (shown in the model architecture), the machine learning model processes and outputs the predicted denoised information. The loss function value is calculated based on the denoised information and the corresponding sent information sample. The model parameters of the machine learning model are then updated using the loss function value. The above process is repeated until the model converges, thus obtaining the candidate denoising model.
[0220] For example, in one possible implementation, the training process of the noise reduction model is as follows:
[0221] 1) Initialize the noise reduction model;
[0222] The training device corresponding to the denoising model initializes the weight parameters of the denoising model according to the pre-set model structure to obtain an untrained initial denoising model. This initialization process can involve randomly assigning values to the weight parameters of the denoising model, or inputting pre-set initial values into the denoising model.
[0223] 2) Input the received information samples into the noise reduction model;
[0224] Obtain the training sample set corresponding to the noise reduction model. The training sample set contains the received information sample corresponding to the noise reduction model. Input the received information sample into the noise reduction model to obtain the predicted sample value output by the noise reduction model corresponding to the received information sample. Then input the predicted sample value and the transmitted information sample corresponding to the received information sample into the loss function to obtain the loss function value corresponding to the received information sample.
[0225] 3) Update the weight parameters of the noise reduction model;
[0226] After obtaining the loss function value corresponding to the received information sample, the denoising model can be updated using the backpropagation algorithm based on this loss function value. This can be done using a single loss function value or multiple loss function values (e.g., the sum or average of multiple loss function values) simultaneously via backpropagation. The loss function can be chosen appropriately based on the signal type and model structure, such as the cross-entropy loss function; no restrictions are imposed here.
[0227] 4) Obtain the trained noise reduction model;
[0228] Repeat the above process until the training meets the specified conditions. The trained model is then selected as a candidate denoising model to achieve denoising of the received information. These specified conditions can be either reaching a training threshold or the denoising model's accuracy exceeding a validation threshold when validated on a validation set.
[0229] The model training process described above can be applied to models with different structures. For example, the fully connected neural network model, convolutional neural network model, and deep residual network can all be trained using the above model training process to train the network model weights.
[0230] This solution introduces Artificial Intelligence (AI) technology into the wireless communication system to construct a noise reduction processing unit, which can perform corresponding noise reduction processing on the received signal and achieve better data recovery results. Experimental tests have confirmed that this solution, by introducing the aforementioned... Figure 13 After the basic structure of the model shown is completed, the noise reduction processing of the received signal can improve the signal-to-noise ratio by more than 6dB, and at the same time, better data recovery results can be obtained.
[0231] The aforementioned gain is due to the fact that the denoising network can reduce the impact of noise in the received signal, and this reduction is directly related to the subsequent data recovery effect. Traditional methods struggle to eliminate this type of near-white noise, while AI-based solutions, with specific model structures trained on a large training set, can achieve a certain degree of noise reduction.
[0232] In summary, the solution shown in this application adds a machine learning model trained on received and transmitted information samples to the receiving device. Before recovering the source information of the received information, the machine learning model is used to perform noise reduction on the received information, thereby reducing the impact of channel environment and interference noise on the wireless signal, improving the signal-to-noise ratio of the received signal, as well as the receiving gain of the system, and improving the transmission performance of the wireless communication system.
[0233] The following are embodiments of the apparatus described in this application, which can be used to execute the embodiments of the method described in this application. For details not disclosed in the apparatus embodiments of this application, please refer to the embodiments of the method described in this application.
[0234] Please refer to Figure 14 This diagram illustrates a block diagram of a wireless signal noise reduction apparatus according to an embodiment of this application. This apparatus can be one of the receiving devices described above. Figure 14 As shown, the device may include:
[0235] The receiving module 1401 is used to receive the wireless signals sent by the transmitting device and obtain the received information;
[0236] The noise reduction processing module 1402 is used to process the received information through a first noise reduction model to obtain the noise-reduced received information;
[0237] The first noise reduction model is a machine learning model trained based on received information samples and transmitted information samples; the received information samples are obtained by receiving wireless signals from the transmitted information samples.
[0238] In one possible implementation, the received information is a wireless signal received by the antenna assembly of the receiving device;
[0239] or,
[0240] The received information is information obtained by mathematically transforming the wireless signal received by the antenna assembly of the receiving device.
[0241] In one possible implementation, the device further includes:
[0242] A measurement information acquisition module is used to acquire channel measurement information, which includes at least one of interference noise information and effective signal reception information; the interference noise information is used to indicate the interference noise situation in the environment where the receiving device is located, and the effective signal reception information is used to indicate the reception status of the receiving device for the effective signal sent by the transmitting device.
[0243] The model selection module is used to select the first noise reduction model from at least two candidate noise reduction models based on the channel measurement information.
[0244] In one possible implementation, when the channel measurement information includes the interference noise information, the measurement information acquisition module is used to:
[0245] Obtain first information about the network configuration, wherein the first information is used to indicate a first resource;
[0246] Measurements are performed on the first resource indicated by the first information to obtain the interference noise information.
[0247] In one possible implementation, the measurement information acquisition module is used for:
[0248] Measurements are performed on the first resource indicated by the first information to obtain first measurement information, the first measurement information including at least one of RSRP, RSRQ, RSSI and SINR;
[0249] Based on the first measurement information, the interference noise information is obtained.
[0250] In one possible implementation, when the receiving device is a terminal, the first information is information configured by the network-side device through at least one of broadcast messages, SIB, RRC messages, RRC reconfiguration signaling, DCI, MAC CE, and PDCCH order.
[0251] When the receiving device is a network-side device, the first information is predefined information; or, the first information is determined by the network-side device based on unused resources.
[0252] In one possible implementation, when the channel measurement information includes the valid signal reception information, the measurement information acquisition module is used to:
[0253] Obtain second information about the network configuration, which is used to indicate a second resource;
[0254] On the second resource indicated by the second information, the specified signal sent by the receiving device is measured to obtain the valid signal reception information.
[0255] In one possible implementation, the measurement information acquisition module is used for:
[0256] The specified signal transmitted by the receiving device is measured on the second resource indicated by the second information to obtain second measurement information, the second measurement information including at least one of RSRP, RSRQ, RSSI, SINR, packet loss rate, and bit error rate;
[0257] Based on the second measurement information, the valid signal reception information is obtained.
[0258] In one possible implementation, when the receiving device is a terminal, the second information is information configured by the network-side device through at least one of broadcast messages, SIB, RRC messages, RRC reconfiguration signaling, DCI, MAC CE, and PDCCH order.
[0259] When the receiving device is a network-side device, the second information is predefined information; or, the second information is determined by the network-side device based on unused resources.
[0260] In one possible implementation, when the receiving device is a terminal, the designated signal is a signal predefined by the protocol, or the designated signal is a signal configured by the network-side device through at least one of broadcast messages, SIB, RRC messages, RRC reconfiguration signaling, DCI, MAC CE, and PDCCH order.
[0261] When the receiving device is a network-side device, the designated signal is a signal predefined by the protocol, or the designated signal is a signal configured by the network-side device to the sending device through at least one of broadcast messages, SIB, RRC messages, RRC reconfiguration signaling, DCI, MACCE, and PDCCH order.
[0262] In one possible implementation, when the receiving device is the network-side device, the apparatus further includes:
[0263] The configuration module is used to configure the second information to the sending device via at least one of broadcast messages, SIB, RRC messages, RRC reconfiguration signaling, DCI, MAC CE, and PDCCHorder.
[0264] In one possible implementation, the first noise reduction model is a fully connected neural network model consisting of N fully connected layers, where N ≥ 1 and N is an integer.
[0265] In one possible implementation, the first noise reduction model is a convolutional neural network model consisting of M convolutional layers, where M ≥ 1 and M is an integer.
[0266] In one possible implementation, the first noise reduction model sequentially includes a first fully connected layer, a first dimension adjustment layer, a first convolutional layer, L common layers, a second convolutional layer, a second dimension adjustment layer, and a second fully connected layer;
[0267] The common layer contains a third convolutional layer, a normalization layer, and an activation layer connected in sequence; the L common layers are connected in sequence; L≥1, and L is an integer.
[0268] In one possible implementation, there is a sampling operation between the first dimension adjustment layer and the first convolutional layer, and a stacking operation between the second convolutional layer and the second dimension adjustment layer;
[0269] The sampling operation is used to sample the output of the first dimension adjustment layer and output the sampling result to the first convolutional layer and the superposition operation respectively; the superposition operation is used to superimpose the output of the second convolutional layer and the output of the sampling operation and output it to the second dimension adjustment layer.
[0270] In summary, the solution shown in this application adds a machine learning model trained on received and transmitted information samples to the receiving device. Before recovering the source information of the received information, the machine learning model is used to perform noise reduction on the received information, thereby reducing the impact of channel environment and interference noise on the wireless signal, improving the signal-to-noise ratio of the received signal, as well as the receiving gain of the system, and improving the transmission performance of the wireless communication system.
[0271] It should be noted that the device provided in the above embodiments is only illustrated by the division of the above functional modules when implementing its functions. In actual applications, the above functions can be assigned to different functional modules according to actual needs, that is, the content structure of the device can be divided into different functional modules to complete all or part of the functions described above.
[0272] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.
[0273] Please refer to Figure 15The diagram illustrates the structure of a receiving device 150 according to an embodiment of this application. The receiving device 150 may include: a processor 151, a receiver 152, a transmitter 153, a memory 154, and a bus 155.
[0274] The processor 151 includes one or more processing cores, and the processor 151 executes various functional applications and information processing by running software programs and modules.
[0275] The receiver 152 and the transmitter 153 can be implemented as a communication component, which can be a communication chip. This communication chip can also be called a transceiver.
[0276] The memory 154 is connected to the processor 151 via the bus 155.
[0277] The memory 154 can be used to store a computer program, and the processor 151 is used to execute the computer program to implement the various steps performed by the receiving device in the above method embodiments.
[0278] Furthermore, the memory 154 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, including but not limited to: magnetic disks or optical disks, electrically erasable programmable read-only memory, erasable programmable read-only memory, static on-demand memory, read-only memory, magnetic memory, flash memory, and programmable read-only memory.
[0279] In an exemplary embodiment, the receiving device includes a processor, a memory, and a transceiver (the transceiver may include a receiver and a transmitter, the receiver being used to receive information and the transmitter being used to send information);
[0280] The transceiver is used to receive wireless signals sent by the transmitting device and obtain received information.
[0281] The processor is used to process the received information through a first noise reduction model to obtain the noise-reduced received information;
[0282] The first noise reduction model is a machine learning model trained based on received information samples and transmitted information samples; the received information samples are obtained by receiving wireless signals from the transmitted information samples.
[0283] The various method steps performed by the receiving device in this application embodiment can be referred to the above. Figure 5 or Figure 7 All or part of the steps performed by the receiving device in the illustrated embodiment will not be described in detail here.
[0284] This application also provides a computer-readable storage medium storing a computer program, which is loaded and executed by a processor / transceiver to implement the above-described embodiments. Figure 5 or Figure 7 Each step in any of the wireless signal noise reduction methods shown.
[0285] This application also provides a computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the above-described actions. Figure 5 or Figure 7 Each step in any of the wireless signal noise reduction methods shown.
[0286] Those skilled in the art will recognize that the functions described in the embodiments of this application in one or more of the above examples can be implemented using hardware, software, firmware, or any combination thereof. When implemented using software, these functions can be stored in a computer-readable medium or transmitted as one or more instructions or code on a computer-readable medium. Computer-readable media include computer storage media and communication media, wherein communication media include any medium that facilitates the transfer of a computer program from one place to another. Storage media can be any available medium that can be accessed by a general-purpose or special-purpose computer.
[0287] The above description is merely an exemplary embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.
Claims
1. A method of wireless signal noise reduction, the method comprising: The method is executed by the receiving device, and the method includes: It receives wireless signals sent by the transmitting device and obtains the received information; The received information is processed by the first noise reduction model to obtain the noise-reduced received information; The first noise reduction model is a machine learning model trained based on received information samples and transmitted information samples; the received information samples are obtained by receiving wireless signals from the transmitted information samples. The method further includes: Acquire channel measurement information, which includes interference and noise information, and the interference and noise information is used to indicate the interference and noise situation in the environment where the receiving device is located; Based on the channel measurement information, the first noise reduction model is selected from at least two candidate noise reduction models. Each candidate noise reduction model includes a corresponding model structure and a set of model parameters. The receiving device includes multiple noise reduction processing units, each corresponding to a model structure and multiple sets of model parameters. The acquisition of channel measurement information includes: Obtain first information about the network configuration, wherein the first information is used to indicate a first resource; Measurements are performed on the first resource indicated by the first information to obtain first measurement information, which includes reference signal received power (RSRP), reference signal received quality (RSRQ), received signal strength indication (RSSI), and signal-to-interference-plus-noise ratio (SINR). The measured reference signal received power (RSRP), reference signal received quality (RSRQ), received signal strength indication (RSSI), and signal-to-interference-plus-noise ratio (SINR) are weighted and summed, and the result of the weighted summation is used as the interference-noise information. Selecting the first noise reduction model from the at least two candidate noise reduction models includes: The noise reduction processing unit and the corresponding model structure corresponding to the first noise reduction model are determined based on a first preset correspondence between each noise reduction processing unit and the result of the weighted summation, and Then, a set of model parameters corresponding to the first noise reduction model is selected based on the second preset correspondence between each set of model parameters and the weighted summation result.
2. The method according to claim 1, characterized in that, The received information is the wireless signal received by the antenna assembly of the receiving device; or, The received information is information obtained by mathematically transforming the wireless signal received by the antenna assembly of the receiving device.
3. The method according to claim 1, characterized in that, The channel measurement information also includes valid signal reception information, which is used to indicate the reception status of the receiving device for the valid signals sent by the transmitting device.
4. The method according to claim 1, characterized in that, When the receiving device is a terminal, the first information is information configured by the network-side device through at least one of the following: broadcast message, system information block (SIB), radio resource control (RRC) message, RRC reconfiguration signaling, downlink control information (DCI), media access control layer control unit (MAC CE), and physical downlink control channel command (PDCCH order). When the receiving device is a network-side device, the first information is predefined information; or, the first information is determined by the network-side device based on unused resources.
5. The method according to claim 3, characterized in that, The acquisition of channel measurement information also includes: Obtain second information about the network configuration, which is used to indicate a second resource; On the second resource indicated by the second information, the specified signal sent by the receiving device is measured to obtain the valid signal reception information.
6. The method according to claim 5, characterized in that, The step of measuring the specified signal transmitted by the receiving device on the second resource indicated by the second information to obtain the valid signal reception information includes: The specified signal transmitted by the receiving device is measured on the second resource indicated by the second information to obtain second measurement information, the second measurement information including at least one of RSRP, RSRQ, RSSI, SINR, packet loss rate, and bit error rate; Based on the second measurement information, the valid signal reception information is obtained.
7. The method according to claim 5 or 6, characterized in that, When the receiving device is a terminal, the second information is information configured by the network-side device through at least one of broadcast messages, SIB, RRC messages, RRC reconfiguration signaling, DCI, MAC CE, and PDCCH order; When the receiving device is a network-side device, the second information is predefined information; or, the second information is determined by the network-side device based on unused resources.
8. The method according to claim 5 or 6, characterized in that, When the receiving device is a terminal, the designated signal is a signal predefined by the protocol, or the designated signal is a signal configured by the network-side device through at least one of broadcast messages, SIB, RRC messages, RRC reconfiguration signaling, DCI, MAC CE, and PDCCH order; When the receiving device is a network-side device, the designated signal is a signal predefined by the protocol, or the designated signal is a signal configured by the network-side device to the sending device through at least one of broadcast messages, SIB, RRC messages, RRC reconfiguration signaling, DCI, MAC CE, and PDCCH order.
9. The method according to claim 5 or 6, characterized in that, When the receiving device is the network-side device, the method further includes: The second information is configured to the sending device via at least one of the following: broadcast message, SIB, RRC message, RRC reconfiguration signaling, DCI, MAC CE, and PDCCH order.
10. The method according to any one of claims 1 to 9, characterized in that, The first noise reduction model is a fully connected neural network model consisting of N fully connected layers, where N≥1 and N is an integer.
11. The method according to any one of claims 1 to 9, characterized in that, The first noise reduction model is a convolutional neural network model consisting of M convolutional layers, where M ≥ 1 and M is an integer.
12. The method according to any one of claims 1 to 9, characterized in that, The first noise reduction model sequentially includes a first fully connected layer, a first dimension adjustment layer, a first convolutional layer, L common layers, a second convolutional layer, a second dimension adjustment layer, and a second fully connected layer; The common layer contains a third convolutional layer, a normalization layer, and an activation layer connected in sequence; the L common layers are connected in sequence; L≥1, and A is an integer.
13. The method according to claim 12, characterized in that, There is a sampling operation between the first dimension adjustment layer and the first convolutional layer, and there is a stacking operation between the second convolutional layer and the second dimension adjustment layer; The sampling operation is used to sample the output of the first dimension adjustment layer and output the sampling result to the first convolutional layer and the superposition operation respectively; the superposition operation is used to superimpose the output of the second convolutional layer and the output of the sampling operation and output it to the second dimension adjustment layer.
14. A wireless signal noise reduction device, characterized in that, The device is used in a receiving end device, and the device includes: The receiving module is used to receive wireless signals sent by the transmitting device and obtain the received information; The noise reduction processing module is used to process the received information through a first noise reduction model to obtain the noise-reduced received information; The first noise reduction model is a machine learning model trained based on received information samples and transmitted information samples; the received information samples are obtained by receiving wireless signals from the transmitted information samples. The device further includes: The measurement information acquisition module is used to acquire channel measurement information, which includes interference noise information, and the interference noise information is used to indicate the interference noise situation in the environment where the receiving device is located. The model selection module is used to select the first noise reduction model from at least two candidate noise reduction models based on the channel measurement information. Each candidate noise reduction model includes a corresponding model structure and a set of model parameters. The receiving device includes multiple noise reduction processing units, each corresponding to a model structure and multiple sets of model parameters. The measurement information acquisition module is further configured to: Obtain first information about the network configuration, wherein the first information is used to indicate a first resource; Measurements are performed on the first resource indicated by the first information to obtain first measurement information, which includes reference signal received power (RSRP), reference signal received quality (RSRQ), received signal strength indication (RSSI), and signal-to-interference-plus-noise ratio (SINR). The measured reference signal received power (RSRP), reference signal received quality (RSRQ), received signal strength indication (RSSI), and signal-to-interference-plus-noise ratio (SINR) are weighted and summed, and the result of the weighted summation is used as the interference-noise information. Selecting the first noise reduction model from the at least two candidate noise reduction models includes: The noise reduction processing unit and the corresponding model structure corresponding to the first noise reduction model are determined based on a first preset correspondence between each noise reduction processing unit and the result of the weighted summation, and Then, a set of model parameters corresponding to the first noise reduction model is selected based on the second preset correspondence between each set of model parameters and the weighted summation result.
15. The apparatus according to claim 14, characterized in that, The received information is the wireless signal received by the antenna assembly of the receiving device; or, The received information is information obtained by mathematically transforming the wireless signal received by the antenna assembly of the receiving device.
16. The apparatus according to claim 14, characterized in that, The channel measurement information also includes valid signal reception information, which is used to indicate the reception status of the receiving device for the valid signals sent by the transmitting device.
17. The apparatus according to claim 14, characterized in that, When the receiving device is a terminal, the first information is information configured by the network-side device through at least one of broadcast messages, SIB, RRC messages, RRC reconfiguration signaling, DCI, MAC CE, and PDCCH order; When the receiving device is a network-side device, the first information is predefined information; or, the first information is determined by the network-side device based on unused resources.
18. The apparatus according to claim 16, characterized in that, The measurement information acquisition module is also used for, Obtain second information about the network configuration, which is used to indicate a second resource; On the second resource indicated by the second information, the specified signal sent by the receiving device is measured to obtain the valid signal reception information.
19. The apparatus according to claim 18, characterized in that, The measurement information acquisition module is used for, The specified signal transmitted by the receiving device is measured on the second resource indicated by the second information to obtain second measurement information, the second measurement information including at least one of RSRP, RSRQ, RSSI, SINR, packet loss rate, and bit error rate; Based on the second measurement information, the valid signal reception information is obtained.
20. The apparatus according to claim 18 or 19, characterized in that, When the receiving device is a terminal, the second information is information configured by the network-side device through at least one of broadcast messages, SIB, RRC messages, RRC reconfiguration signaling, DCI, MAC CE, and PDCCH order; When the receiving device is a network-side device, the second information is predefined information; or, the second information is determined by the network-side device based on unused resources.
21. The apparatus according to claim 18 or 19, characterized in that, When the receiving device is a terminal, the designated signal is a signal predefined by the protocol, or the designated signal is a signal configured by the network-side device through at least one of broadcast messages, SIB, RRC messages, RRC reconfiguration signaling, DCI, MAC CE, and PDCCH order; When the receiving device is a network-side device, the designated signal is a signal predefined by the protocol, or the designated signal is a signal configured by the network-side device to the sending device through at least one of broadcast messages, SIB, RRC messages, RRC reconfiguration signaling, DCI, MAC CE, and PDCCH order.
22. The apparatus according to claim 18 or 19, characterized in that, When the receiving device is the network-side device, the apparatus further includes: The configuration module is used to configure the second information to the sending device via at least one of broadcast messages, SIB, RRC messages, RRC reconfiguration signaling, DCI, MAC CE, and PDCCHorder.
23. The apparatus according to any one of claims 14 to 22, characterized in that, The first noise reduction model is a fully connected neural network model consisting of N fully connected layers, where N≥1 and N is an integer.
24. The apparatus according to any one of claims 14 to 22, characterized in that, The first noise reduction model is a convolutional neural network model consisting of M convolutional layers, where M ≥ 1 and M is an integer.
25. The apparatus according to any one of claims 14 to 22, characterized in that, The first noise reduction model sequentially includes a first fully connected layer, a first dimension adjustment layer, a first convolutional layer, L common layers, a second convolutional layer, a second dimension adjustment layer, and a second fully connected layer; The common layer contains a third convolutional layer, a normalization layer, and an activation layer connected in sequence; the L common layers are connected in sequence; L≥1, and L is an integer.
26. The apparatus according to claim 25, characterized in that, There is a sampling operation between the first dimension adjustment layer and the first convolutional layer, and there is a stacking operation between the second convolutional layer and the second dimension adjustment layer; The sampling operation is used to sample the output of the first dimension adjustment layer and output the sampling result to the first convolutional layer and the superposition operation respectively; the superposition operation is used to superimpose the output of the second convolutional layer and the output of the sampling operation and output it to the second dimension adjustment layer.
27. A receiving device, characterized in that, The receiving device includes a processor, a memory, and a transceiver; The transceiver is used to receive wireless signals sent by the transmitting device and obtain received information. The processor is used to process the received information through a first noise reduction model to obtain the noise-reduced received information; The first noise reduction model is a machine learning model trained based on received information samples and transmitted information samples; the received information samples are obtained by receiving wireless signals from the transmitted information samples. The receiving device is further configured to acquire channel measurement information, which includes interference noise information, and the interference noise information is used to indicate the interference noise situation in the environment where the receiving device is located. Based on the channel measurement information, the first noise reduction model is selected from at least two candidate noise reduction models. Each candidate noise reduction model includes a corresponding model structure and a set of model parameters. The receiving device includes multiple noise reduction processing units, each corresponding to a model structure and multiple sets of model parameters. The acquisition of channel measurement information includes: Obtain first information about the network configuration, wherein the first information is used to indicate a first resource; Measurements are performed on the first resource indicated by the first information to obtain first measurement information, which includes reference signal received power (RSRP), reference signal received quality (RSRQ), received signal strength indication (RSSI), and signal-to-interference-plus-noise ratio (SINR). The measured reference signal received power (RSRP), reference signal received quality (RSRQ), received signal strength indication (RSSI), and signal-to-interference-plus-noise ratio (SINR) are weighted and summed, and the result of the weighted summation is used as the interference-noise information. Selecting the first noise reduction model from the at least two candidate noise reduction models includes: The noise reduction processing unit and the corresponding model structure corresponding to the first noise reduction model are determined based on a first preset correspondence between each noise reduction processing unit and the result of the weighted summation, and Then, a set of model parameters corresponding to the first noise reduction model is selected based on the second preset correspondence between each set of model parameters and the weighted summation result.
28. A computer-readable storage medium, characterized in that, The storage medium stores a computer program that is executed by a processor / transceiver to implement the wireless signal noise reduction method as described in any one of claims 1 to 13.
Citation Information
Patent Citations
Visible light communication MIMO anti-interference and noise-reduction method based on BP neutral network
CN108494710A
Audio signal processing method and device, computer equipment and medium
CN111833895A
