Training methods for signal estimation models, signal estimation methods and devices

Abstract

This application proposes a training method, a signal estimation method, and an apparatus for a signal estimation model. The method includes: using the signal sampled by the RF chip in the terminal device to obtain the signal received by the antenna as a signal sample, thereby improving the randomness and authenticity of the sample data; training the signal estimation model with the training sample, enabling the deep learning model to have the ability to estimate signals. Since training the signal estimation model with the training sample is data-driven, it does not depend on the classification and modeling of specific RF impairments and does not change with changes in RF hardware. Therefore, it reduces the limitations of application scenarios and has better versatility.

Description

Technical Field

[0001] This application relates to the field of signal processing technology, and in particular to a training method for a signal estimation model, a signal estimation method, and an apparatus. Background Technology

[0002] In the field of communications, signal distortion occurs during air interface signal transmission. Distortion includes the addition of noise or distortion. In order to ensure the sensitivity and anti-interference capability of baseband demodulation, the digital front-end in the terminal equipment performs signal estimation on the distorted signal, that is, corrects the distorted signal to obtain the estimated signal.

[0003] In related technologies, signal estimation is achieved through the digital front-end in the radio frequency chip. However, the signal estimation algorithm implemented by the digital front-end is driven by mathematical modeling, which has limitations and poor versatility. Summary of the Invention

[0004] This application aims to at least partially address one of the technical problems in the related art.

[0005] To address this, this application proposes a training method, a signal estimation method, and an apparatus for a signal estimation model. The signal estimation is based on a data-driven neural network model, which reduces the limitations of the algorithm and improves its versatility.

[0006] One embodiment of this application proposes a training method for a signal estimation model, including:

[0007] Acquire signal samples; wherein, the signal samples are obtained by the radio frequency chip sampling the signal received by the antenna;

[0008] The signal sample is estimated using the signal estimation model set in the radio frequency chip to obtain the estimated signal.

[0009] The signal estimation model is trained based on the estimated signal and the ground truth signal corresponding to the signal sample to obtain the trained signal estimation model.

[0010] Another embodiment of this application proposes a signal estimation method, including:

[0011] Acquire the signal to be processed;

[0012] The signal is estimated using the trained signal estimation model to obtain the target signal; wherein the signal estimation model is trained using the method described in the preceding aspect.

[0013] Another embodiment of this application proposes a training apparatus for a signal estimation model, comprising:

[0014] An acquisition module is used to acquire signal samples; wherein the signal samples are obtained by the radio frequency chip sampling the signal received by the antenna;

[0015] The processing module is used to perform signal estimation on the signal sample using the signal estimation model set in the radio frequency chip to obtain the estimated signal;

[0016] The training module is used to train the signal estimation model based on the estimated signal and the ground truth signal corresponding to the signal sample, so as to obtain the trained signal estimation model.

[0017] Another embodiment of this application proposes a signal estimation apparatus, comprising:

[0018] The acquisition module is used to acquire the signal to be processed.

[0019] An estimation module is used to estimate the signal using a trained signal estimation model to obtain a target signal; wherein the signal estimation model is trained using a training device for the signal estimation model described in the other aspect above.

[0020] Another embodiment of this application provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the method described in one aspect above and the method described in the other aspect above.

[0021] Another embodiment of this application proposes a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in one aspect above and the method described in the other aspect above.

[0022] Another embodiment of this application proposes a chip including processing circuitry configured to perform the methods described in one aspect and the methods described in the other aspect.

[0023] Another embodiment of this application proposes a computer program product having a computer program stored thereon, which, when executed by a processor, implements the method described in one aspect above and the method described in the other aspect above.

[0024] The signal estimation model training method, signal estimation method, and apparatus proposed in this application use the signal obtained by sampling the signal received by the antenna from the radio frequency chip in the terminal device as the signal sample, which improves the randomness and authenticity of the sample data. The signal estimation model is trained by training the training sample, which enables the deep learning model to have the ability to estimate signals. The training of the signal estimation model by training the training sample is data-driven and does not depend on the classification and modeling of specific radio frequency damage. It is not modified with changes in radio frequency hardware. Therefore, it reduces the limitation of application scenarios and has better versatility.

[0025] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0026] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0027] Figure 1 This is a schematic diagram of the receiver structure in the related technology provided in the embodiments of this application;

[0028] Figure 2 A flowchart illustrating a training method for a signal estimation model provided in an embodiment of this application;

[0029] Figure 3 A flowchart illustrating another training method for a signal estimation model provided in an embodiment of this application;

[0030] Figure 4 A schematic diagram of a digital radio frequency front-end architecture provided for an embodiment of this application;

[0031] Figure 5 A timing diagram of the input and output of a signal estimation model provided in an embodiment of this application;

[0032] Figure 6 A schematic diagram of a multi-head attention layer provided in an embodiment of this application;

[0033] Figure 7 One of the training diagrams of a signal estimation model provided as an example in this application;

[0034] Figure 8 A schematic flowchart illustrating a signal estimation method provided in an embodiment of this application;

[0035] Figure 9 A flowchart illustrating another signal estimation method provided in an embodiment of this application;

[0036] Figure 10A second schematic diagram illustrating the training of a signal estimation model provided in an embodiment of this application;

[0037] Figure 11 A schematic diagram of the structure of a training device for a signal estimation model provided in an embodiment of this application;

[0038] Figure 12 This is a schematic diagram of the structure of a signal estimation device provided in an embodiment of this application;

[0039] Figure 13 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0040] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0041] The following description, with reference to the accompanying drawings, describes a training method, a signal estimation method, and an apparatus for a signal estimation model according to embodiments of this application.

[0042] Figure 1 This is a schematic diagram of the receiver structure in the related technology provided in the embodiments of this application, such as... Figure 1 As shown, the air interface signal is coupled to the RF transmission path through the antenna 101. It first passes through the RF front-end module 102, which typically includes components such as a duplexer, antenna switch, and low-noise amplifier (LNA). The RF front-end module 102 can be replaced according to different RF band requirements. After processing the air interface signal, the RF module 102 outputs the signal to the RF processor 103. The analog front-end 104 mixes the signal to baseband and performs filtering and analog-to-digital conversion. Since baseband processing requirements do not change with the RF band, the same analog front-end 104 can be adapted to different RF front-end modules 102, facilitating modular processing of different RF bands. After being converted into a digital signal by the analog front-end 104, the digital front-end (DFE) 105 performs signal estimation on the digital signal to complete tasks such as correcting radio frequency impairments. Finally, the estimated signal is sent to the base-band processor 106 for demodulation.

[0043] The Digital Front End 105 has the following problems in estimating the signal:

[0044] First: The architecture of digital front-end lacks versatility. Digital front-end needs to be built on specific signal modeling and is bound to the design architecture. Digital front-end is usually implemented by application-specific integrated circuits (ASICs). The digital modules inside the chip cannot be changed. When there are changes such as RF front-end or architecture upgrades, it is necessary to redesign and re-batch the chip, which requires a high cost.

[0045] Secondly, the limitations of the algorithm model mean that the design of non-ideal RF algorithms cannot cope with RF damage that is not expected in the modeling.

[0046] To address the aforementioned issues, this application embodiment implements a signal estimation algorithm using a deep learning signal estimation model. It employs a signal sample obtained by sampling the signal received by the antenna using an RF chip in the terminal device, thereby improving the randomness and authenticity of the sample data. The signal estimation model is trained using training samples, enabling the deep learning model to possess signal estimation capabilities. Since training the signal estimation model using training samples is data-driven and does not rely on the classification and modeling of specific RF impairments, nor is it modified with changes in RF hardware, it reduces the limitations of application scenarios and has better versatility.

[0047] Figure 2 This is a flowchart illustrating a training method for a signal estimation model provided in an embodiment of this application.

[0048] As one implementation, the signal estimation model training method of this application embodiment can be configured in a signal estimation model training device. This signal estimation model training device can be applied to any electronic device or chip, enabling the electronic device to perform the signal estimation model training function. The chip, for example, is a microprocessor (General Parallel Unit, GPU) with a highly parallel computing architecture.

[0049] Among them, electronic devices can be any device with computing capabilities, such as computers, and computers can be servers, such as local servers.

[0050] like Figure 2 As shown, the method may include the following steps:

[0051] Step 211: Obtain signal samples.

[0052] In the embodiments of this application, the signal sample includes an air interface signal, which can be a wireless signal in the field of wireless communication, such as a 5G or 4G communication signal, or a radar signal in the field of radar, or a wireless signal in the field of Internet of Things.

[0053] In the embodiments of this application, the signal sample can be obtained by the RF chip in the terminal device sampling the signal received by the antenna. As one implementation, the terminal device receives an air interface signal through the antenna, processes it through the RF front-end, and sends it to the RF chip. The RF chip samples the processed air interface signal to obtain a signal sample. The signal sample is obtained by sampling the air interface signal received in a real-world scenario. The RF impairments in the signal sample are more realistic, random, and complex than artificially added RF impairments, thus improving the offline training effect of the model. By driving the training of the signal estimation model with training data, the signal estimation model can learn to handle any type of air interface signal and any type of RF impairment, unaffected by changes in RF hardware such as the analog front-end and RF module. Even if the analog front-end and RF module change, resampling data and retraining can adapt the model. Simultaneously, the deep learning characteristics of the signal estimation model can learn nonlinear distortions that cannot be represented by mathematical modeling in related technologies, improving the model's signal estimation capability and its versatility.

[0054] In one implementation of this application, the signal sample is a signal sequence, which includes air interface signals sampled at multiple time units, thereby achieving synchronous processing of multiple signals and improving processing efficiency.

[0055] Step 212: The signal sample is estimated using the signal estimation model set in the radio frequency chip to obtain the estimated signal.

[0056] In this embodiment of the application, the signal estimation model integrates DFE (Discrete Element Function) for signal estimation of the received signal. Since the signal received by the antenna has problems such as distortion and noise interference, in order to ensure the sensitivity and anti-interference capability of the subsequent baseband chip demodulation capability, it is necessary to complete the signal estimation process through the signal estimation model. Here, signal estimation refers to estimating various damages of the signal, calculating compensation information based on various losses, and compensating the received signal based on the compensation information to obtain the estimated signal.

[0057] In this embodiment, the signal estimation model is set in the radio frequency (RF) chip of the electronic device. Due to hardware deviations between different batches of RF chips, it is necessary to train the signal estimation model set in the RF chip of the electronic device. By training the signal estimation model, the trained signal estimation model can achieve the mapping between the original signal and the estimated signal, so as to remove distortion information such as RF damage in the signal and obtain the estimated signal. The estimated signal is the signal without distortion after removing RF damage and other distortion information from the signal sample. Since the neural network model does not depend on the specific modeling form of RF damage, nor on the specific RF front-end architecture, the estimation and compensation of RF damage are achieved through data acquisition, thereby improving the versatility of signal processing.

[0058] Step 213: Train the signal estimation model based on the estimated signal and the true signal corresponding to the signal sample to obtain the trained signal estimation model.

[0059] In this embodiment of the application, the difference between the estimated signal and the true signal is determined based on the estimated signal and the true signal corresponding to the signal sample, and the loss function is determined based on the difference. As one implementation, the loss function can be a loss function determined based on the L2 norm (L2 Loss) or a loss function determined based on the L1 norm (L1 Loss).

[0060] Among them, the loss function expressed using the 1-norm. as follows:

[0061] ;

[0062] Among them, the loss function expressed using the L2 norm. as follows:

[0063] ;

[0064] in, For true signals, To estimate the signal, is the standard deviation of the training sample dataset, which is determined when constructing the training sample dataset.

[0065] The signal estimation model training method of this application embodiment acquires signal samples, estimates the signal from the signal samples using the signal estimation model to obtain an estimated signal, trains the signal estimation model based on the estimated signal and the true signal carried in the signal samples to obtain the trained signal estimation model, realizes the signal estimation function through the deep learning signal estimation model, and uses the signal obtained by sampling the signal received by the antenna by the radio frequency chip in the terminal device as the signal sample to improve the randomness and authenticity of the sample data. The signal estimation model is trained through training samples, so that the deep learning model has the ability to estimate signals. The training of the signal estimation model through training samples is data-driven, does not depend on the classification and modeling of specific radio frequency damage, and does not change with the change of radio frequency hardware. Therefore, it reduces the limitation of the application scenario and has better versatility.

[0066] Based on the above embodiments, Figure 3 A flowchart illustrating another training method for a signal estimation model provided in this application embodiment is shown below. Figure 3 As shown, the method includes the following steps:

[0067] Step 301: Obtain signal samples.

[0068] In this application, a signal sequence is used as an example for illustration. As an example, the model training method of this application embodiment can be applied to the scenario of factory calibration of terminal devices to calibrate the signal estimation model for different terminal devices and improve the accuracy of the signal estimation model configured in each terminal device.

[0069] In this embodiment, the digital radio frequency (RF) front-end architecture includes an RF front-end module 201, an analog front-end 202, and a signal estimation model 203. The signal estimation model 203, trained to its full potential, implements the signal estimation function of the digital front-end (DFE). The signal estimation model 203 is deployed within the RF chip. The over-the-air signal passes through the RF front-end module 201 and the analog front-end 202 before being fed into the signal estimation model 203. The RF front-end module 201 performs signal filtering and amplification (e.g., low-noise amplifier - LNA), while the analog front-end 202 converts the analog signal to a digital signal through downsampling, low-pass filtering (LPF), and analog-to-digital conversion (ADC) to meet the requirements of the baseband signal. As an example, such as... Figure 4 As shown, the implementation methods of the radio frequency front-end module 201 and the analog front-end 202 are not unique, and no limitation is made in this embodiment.

[0070] As an example, after analog-to-digital conversion, the air interface signal becomes in-phase and quadrature signals, i.e., I / Q signals. If the signal sample includes N signals, meaning the antenna receives N signals in different reception time units, these N signals are filled into a buffer of depth N. Thus, the signal estimation model processes N I / Q signals at a time. In other words, the dimension of the signal sample is N×2, i.e., the signal sample is {x}. i, j} N×2 Where i and j are used to indicate the I-dimensional and Q-dimensional data in the signal sample, as an example, such as Figure 4 As shown, the signal estimation model 203 includes an encoding network 2031, a sampling network 2032, an encoder 2033, and an output network 2034.

[0071] Step 302: Use an encoding network to perform position encoding on the signal sequence to obtain the first encoded feature.

[0072] In this application, since the signal sample is a signal sequence, the signal sequence includes signals received at each receiving time unit. In order to identify the receiving order of each signal in the output estimated signal after the signal estimation model is processed, position information is added to each signal in the signal sequence, i.e., position encoding is implemented. As one implementation method, the receiving time unit of each signal included in the signal sequence is obtained, and the position information of each signal in the signal sequence is determined according to the receiving time unit of each signal. The position information is used to indicate the receiving order of each signal in the signal sequence. The corresponding position information is added to each signal in the signal sequence to obtain the first encoding feature.

[0073] As one implementation method, location information can be added using the following formula:

[0074] ;

[0075] in, It is a signal in a signal sequence. This refers to the location information corresponding to a signal in a signal sequence. L is a set constant that is used to add location information to a signal. It can be set according to the needs of the scenario, but is not limited in this embodiment.

[0076] Step 303: The first coding feature is sampled using a sampling network to obtain the second coding feature.

[0077] The second coding feature is a feature that matches the processing dimension of the radio frequency chip.

[0078] In this embodiment of the application, since the dimensions of signal processing and input data processing of the radio frequency chip may be different, it is necessary to perform sampling processing according to the processing dimensions of the radio frequency chip. This sampling can be upsampling or downsampling.

[0079] As an example, sampling is typically downsampling. The sampling network consists of a fully connected network of one or two layers; this embodiment does not impose a limitation. For instance, the signal estimation model operates at a downsampling rate of D1 times, while the output rate of the estimation result operates at a downsampling rate of D1D2 times. The dimensionality change of the signal sequence in the signal estimation model is as follows: Figure 4 As shown in the annotation: The input to the encoding network 2031 is a signal sequence of N signals. After encoding, it is sent to the sampling network 2032 with a dimension of N×2. The output dimension of the sampling network 2032 is N / D1×2, and this dimension is maintained unchanged in the encoder 2033 until it is sent to the output layer 2034. Finally, the dimension of the output layer 2034 after processing is N / (D1D2)×2.

[0080] As an example, Figure 5 The timing diagram of the input and output of a signal estimation model provided in the embodiments of this application is as follows: Figure 5 As shown, at time t0 of the digital clock, a signal sequence S1 of N signals is processed, and an estimated signal sequence S2 of M signals is output (where M = N / (D1D2)). Subsequently, at time tM' of the digital clock, a signal sequence S3 of N signals is processed, and an estimated signal sequence S4 of M signals is output, where M = N / (D1D2). Under the same digital clock Fs, N points are input and M points are output. The actual sampling rate of the input signal samples is Fs*N, and the sampling rate of the output estimated signal is Fs*M. This enables the processing of signals at different rates.

[0081] Step 304: The encoder is used to estimate the signal based on the second coding feature to obtain the target signal feature.

[0082] In one implementation of this application, the encoder includes a multi-head attention layer and a residual connection layer. The multi-head attention layer in the encoder performs attention encoding on the second encoded feature to obtain a third encoded feature output by the multi-head attention layer. The third encoded feature carries the estimated signal information. Then, the residual connection layer in the encoder fuses and normalizes the third encoded feature and the second encoded feature to obtain the target signal feature. This fusion can be a concatenation. The second and third encoded features are fused because the gradient of the second encoded feature is relatively stable, ensuring backpropagation even when subsequent gradient changes to zero, thus guaranteeing the smooth progress of the training process. In this application embodiment, training samples containing signal sequences with multiple signals can be processed in parallel. Compared to related technologies that process a single data point each time, this improves processing efficiency. Due to the increased sequence length, a multi-head attention layer is used to process longer signal sequences. The multi-head attention layer introduces the dependencies between signal sequences, allowing for flexible processing of longer signal sequences and facilitating parallel training on time signals, thereby improving training effectiveness.

[0083] As an example, Figure 6 This is a schematic diagram of a multi-head attention layer structure provided in an embodiment of this application, as shown below. Figure 6 As shown, the second encoded data X of N / D 1×2 dimensions in 401 input multi-head attention layer, via W Q W K W V Operators (Query402, Key403, and Value404) encode information between different positions in the features of the second encoder, where the dimension of the operators is 2×Da. The output of a single attention head is obtained through the dot product and norm operation in Query405. The operational relationship is as follows:

[0084] ;

[0085] The softmax operation corresponding to 406 can be implemented in several approximate ways, and no limitation is imposed in this embodiment. The classic softmax operation is as follows:

[0086] ;

[0087] This example has H attention heads, where H is a natural number greater than or equal to 2. Concatenating all attention results yields the final output 407, in the form:

[0088] ;

[0089] The dimensions Da and H satisfy the relation H×Da = 2, ensuring that the dimensions of the final concatenated Y are the same as those of the input X. in The dimensions are consistent. After concatenation, the data is transformed by a weight matrix W of dimension N / D1×N / D1 to obtain the final output third encoded feature. The formula is as follows:

[0090] ;

[0091] Signal X after multi-head attention processing out The data is fed into the residual connection layer to complete the residual connection, which fuses the third and second coding features. The formula for the fused result X is as follows:

[0092] ;

[0093] The residual connection layer can also perform a normalization (norm) operation on the fusion result X obtained from the residual connections. This normalization operation can then be used to obtain the target signal features. The classic layer normalization is implemented using the following formula:

[0094] ;

[0095] ;

[0096] Where a and b are the weights and biases of the signal estimation model, and These are the expected value and the variance, respectively. It is a small, fixed constant set to prevent the denominator from being zero.

[0097] Normalization methods include root mean square normalization, NTK parameterization-based normalization, etc. The normalization method is not limited in the embodiments of this application.

[0098] In another implementation of this application, the encoder includes a multi-head attention layer, a residual connection layer, a feedforward layer, and a subsequent residual connection. Similarly, after the multi-head attention layer processing and residual connection layer fusion are completed, the encoder output can be obtained by passing through the feedforward layer and the subsequent residual connection. The structure of the subsequent residual connection layer is completely identical to that of the residual connection layer. The feedforward layer can be composed of a fully connected network or implemented using other types of networks, as long as the output dimension and input dimension remain consistent.

[0099] Among them, such as Figure 4As shown, encoder 2033 can be one or more, such as K, where K is a natural number greater than or equal to 1. As one implementation method, when there are multiple encoders, multiple encoders process in parallel to achieve encoding from different interference or distortion angles to obtain diverse encoding results. The target signal features output by multiple encoders are fused to obtain the target signal features of the final input-output network, thereby improving the accuracy of the encoding results.

[0100] Step 305: The output network is used to obtain the estimated signal based on the characteristics of the target signal.

[0101] In this embodiment, the output network extracts an estimated signal based on the characteristics of the target signal. As one implementation method, in order to adapt to the data processing requirements of the baseband signal, the estimated signal can be sampled and interface adapted. The sampling is downsampling to achieve data matching, so as to facilitate subsequent processing by the baseband chip.

[0102] Step 306: Train the signal estimation model based on the estimated signal and the true signal carried in the signal sample to obtain the trained signal estimation model.

[0103] In one implementation of this application, the training process of the signal estimation model can be regarded as a classification task. Based on the bit width of the estimated signal, a first probability distribution corresponding to the estimated signal and a second probability distribution corresponding to the estimated signal are determined. Based on the difference between the first probability distribution and the second probability distribution, a loss function is determined. The signal estimation model is trained based on the loss function to obtain the trained signal estimation model.

[0104] As an example, Figure 7 One of the training diagrams of a signal estimation model provided as an example in this application is shown below. Figure 7 As shown, signal source 501 inputs the true value signal s(n) of the signal sample into the RF chip 502. The DFE samples the input data x(n) and groups it with s(n) before sending it to the host computer 502 (usually a PC containing a General Parallel Unit and GPU). The host computer first initializes the parameters Φ510 of the signal estimation model, and then performs batch standardization on {x(n), s(n)} to ensure numerical stability and enhance the convergence speed and generalization ability of the model. Subsequently, forward propagation and backward propagation are implemented in batches within the host computer until the loss function converges. Finally, the trained network parameters are written into the chip register.

[0105] If the bit width of the output estimated signal is specified as 12 bits, then the estimated signal corresponding to the nth signal sample... The value of (n) needs to be One category is selected from the available categories. Therefore, the classic cross-entropy can be used as the loss function, as follows:

[0106] ;

[0107] in, Let cross-entropy be the loss function. The feature function for the fixed-point data category of the data. This is the nth signal sample.

[0108] Furthermore, given a defined loss function, the data output from the intermediate networks during forward propagation are read, including the output layer gradient 507, the encoding layer gradient 508, and the input layer gradient 509. The backpropagation gradient J is then calculated layer by layer. Loss The network weights Φ510 are updated based on the backpropagation results. The training process can be performed multiple times based on different signal samples until the loss function is less than a set threshold, or the number of training iterations reaches a set value, at which point the model training is complete.

[0109] In the signal estimation model training method of this application embodiment, most existing digital front-end architectures rely on mathematical modeling of radio frequency (RF) impairments. Different RF modules and analog front-end designs lead to different RF impairment modeling methods, thus affecting the design of the digital front-end architecture. Therefore, these digital front-end designs lack universality. However, the universal digital front-end architecture based on a multi-head attention mechanism proposed in this application is a signal estimation model trained on data. This model does not depend on specific RF impairment modeling methods or specific RF front-end architectures; it can estimate and compensate for RF impairments solely through data acquisition. Furthermore, it is applicable to variable-rate time-series processing and has better versatility. Moreover, while other neural network models use training data derived from manually added RF impairments, the training data in this application comes from hardware sampling, reducing limitations, meeting the needs of different situations, and improving versatility.

[0110] Based on the above embodiments, in one implementation of this application, multiple encoders are used. Each encoder performs signal estimation based on second coding features to obtain first signal features corresponding to each encoder. Multiple first signal features are then fused to obtain target signal features. By using multiple encoders to perform signal estimation based on second coding features, multiple signal features from different angles or dimensions can be obtained. After fusion, these features can more comprehensively reflect the true state of the signal, thereby improving the accuracy of signal estimation. Especially when processing complex radio frequency signals, a single encoder may not be able to completely capture all the features of the signal, while the combined use of multiple encoders can more effectively address this situation.

[0111] Based on the above embodiments, Figure 8 This is a flowchart illustrating a signal estimation method provided in an embodiment of this application. The execution subject of this embodiment is a terminal device, such as a mobile phone, tablet, or smart wearable device, but this embodiment is not limited to any particular device.

[0112] like Figure 8 As shown, the method includes the following steps:

[0113] Step 801: Obtain the signal to be processed.

[0114] In one scenario of this application embodiment, the signal to be processed is an air interface signal obtained by sampling from the radio frequency chip of the terminal device. The relevant explanations and descriptions in the foregoing embodiments are also applicable to this embodiment, and the principle is the same, so they will not be repeated here.

[0115] Step 802: Use the trained signal estimation model to estimate the signal and obtain the target signal.

[0116] The signal estimation model is trained using the training method described in any of the preceding embodiments. The target signal, relative to the signal to be processed, is a signal with corrected radio frequency impairments. The training process is explained in the preceding embodiments; the principle is the same and will not be repeated here.

[0117] In the signal estimation method of this application embodiment, the signal estimation function of the digital front-end (DFE) is implemented through a deep learning signal estimation model. The signal obtained by sampling the signal received by the antenna from the RF chip in the terminal device is used as a signal sample, improving the randomness and authenticity of the sample data. The signal estimation model is trained using training samples, enabling the deep learning model to possess signal estimation capabilities. Training the signal estimation model using training samples is data-driven, independent of the classification and modeling of specific RF impairments, and does not change with variations in RF hardware. Therefore, it reduces the limitations of application scenarios and has better versatility. Furthermore, using the trained signal estimation model to estimate the signal to be processed allows for rapid and accurate acquisition of signal estimation results, improving the versatility of application scenarios.

[0118] Based on the above embodiments, Figure 9 This is a flowchart illustrating another signal estimation method provided in this application embodiment, specifically demonstrating how the signal estimation model is updated online during actual use, such as... Figure 9 As shown, the method for online updating the signal estimation model includes:

[0119] Step 901: Obtain signal samples, input the signal samples into the baseband processor for signal extraction, and obtain the predicted pilot signal.

[0120] The signal sample is obtained by the terminal device through sampling the signal received by the antenna via the radio frequency chip during use. For details, please refer to the description of the signal sample in the previous embodiments; the principle is similar and will not be repeated here.

[0121] In this embodiment of the application, the signal sample is input into the baseband processor, and the baseband processor extracts the target signal to obtain the predicted pilot signal.

[0122] Step 902: Determine the loss function based on the predicted pilot signal and the reference pilot signal stored in the terminal device.

[0123] The reference pilot signal can be an ideal pilot signal pre-stored in the terminal device.

[0124] Step 903: Update the signal estimation model according to the loss function to obtain the updated signal estimation model.

[0125] In this embodiment, a loss function is determined based on the difference between the predicted pilot signal and the reference pilot signal stored in the RF chip. Data output by each network in the signal estimation model is obtained. The backpropagation gradient is calculated based on the loss function and the data output by each network. The parameters of the signal estimation model are updated based on the backpropagation gradient to adapt to the signal estimation requirements in real-world scenarios, thereby achieving online updates to the signal estimation model and improving its accuracy.

[0126] As an example, Figure 10 This is a second schematic diagram illustrating the training of a signal estimation model provided in an embodiment of this application, as shown below. Figure 10 As shown, after the receiver's RFIC (Radio Frequency Integrated Chip) receives the InputSignal 601, the output network of the signal estimation model outputs the target signal 604. This process is equivalent to forward propagation. The target signal is then processed by the baseband processor 602 (Base-band Integrated Chip, BBIC) to extract the pilot signal. 606, and obtain the expected ideal pilot signal s 605, where the ideal pilot signal matches the communication protocol. The intermediate variable Y output by each network (encoding network, sampling network, encoder) of the signal estimation model is sent to the processing module 603 in the terminal device. For example, the processing module 603 is a neural processing unit (NPU) used to calculate the loss function and perform backpropagation. After completion, the updated parameter Φ is configured into the signal estimation model to realize the online real-time update of the signal estimation model.

[0127] In this embodiment, the online update of the signal estimation model is triggered in three scenarios: First, the terminal device detects that the difference between the current ambient temperature and the temperature during production line calibration is greater than a set temperature threshold, thus indicating a change in radio frequency characteristics and triggering an online update of the signal estimation model. Second, the terminal device triggers an online update of the signal estimation model when the bit error rate of the demodulated baseband signal exceeds a set threshold. Third, after a set usage period, the terminal device triggers an online update of the signal estimation model based on hardware aging, thereby achieving timely updates of the signal estimation model and ensuring the effectiveness of signal estimation.

[0128] To implement the above embodiments, this application also proposes a training device for a signal estimation model.

[0129] Figure 11 This is a schematic diagram of the structure of a training device for a signal estimation model provided in an embodiment of this application.

[0130] like Figure 11 As shown, the device may include:

[0131] The acquisition module 110 is used to acquire signal samples; wherein the signal samples are obtained by the radio frequency chip sampling the signal received by the antenna.

[0132] The processing module 111 is used to perform signal estimation on the signal sample using the signal estimation model set in the radio frequency chip, and obtain the estimated signal.

[0133] The training module 112 is used to train the signal estimation model based on the estimated signal and the ground truth signal corresponding to the signal sample, so as to obtain the trained signal estimation model.

[0134] Furthermore, in one implementation of this application embodiment, the signal sample is a signal sequence, the signal estimation model includes an encoding network, a sampling network, an encoder, and an output network, and the processing module 111 is specifically used for:

[0135] The signal sequence is positionally encoded using the encoding network to obtain a first encoding feature;

[0136] The sampling network is used to sample the first encoded feature to obtain the second encoded feature; wherein the second encoded feature is a feature that matches the processing dimension of the radio frequency chip;

[0137] The encoder is used to estimate the signal based on the second encoded feature to obtain the target signal feature;

[0138] The output network is used to obtain the estimated signal based on the characteristics of the target signal.

[0139] In one implementation of this application embodiment, the processing module 111 is specifically used for:

[0140] The second coded feature is attention-encoded using the multi-head attention layer in the encoder to obtain the third coded feature output by the multi-head attention layer; wherein, the third coded feature carries the estimated signal information;

[0141] The third coding feature and the second coding feature are fused and normalized using the residual connection layer in the encoder to obtain the target signal feature.

[0142] In one implementation of this application embodiment, the processing module 111 is specifically used for:

[0143] Obtain the reception time unit of each signal included in the signal sequence;

[0144] Based on the reception time unit of each signal, the position information of each signal in the signal sequence is determined; wherein, the position information is used to indicate the reception order of each signal in the signal sequence;

[0145] The first coding feature is obtained by adding corresponding position information to each signal in the signal sequence.

[0146] In one implementation of this application, there are multiple encoders, and the processing module 111 is specifically used for:

[0147] Each encoder performs signal estimation based on the second coding feature to obtain the first signal feature corresponding to each encoder;

[0148] The target signal features are obtained by fusing multiple first signal features.

[0149] In one implementation of this application, the training module 112 is specifically used for:

[0150] Based on the bit width of the estimated signal, determine the first probability distribution corresponding to the estimated signal, and determine the second probability distribution corresponding to the estimated signal;

[0151] The loss function is determined based on the difference between the first probability distribution and the second probability distribution;

[0152] The signal estimation model is trained according to the loss function to obtain the trained signal estimation model.

[0153] It should be noted that the foregoing explanation of the method embodiments also applies to the apparatus of this embodiment, and will not be repeated here.

[0154] The signal estimation model training device of this application embodiment acquires signal samples, estimates the signal from the signal samples using the signal estimation model, obtains an estimated signal, and trains the signal estimation model based on the estimated signal and the true signal carried in the signal samples to obtain the trained signal estimation model. By implementing the signal estimation function of the digital front-end (DFE) through a deep learning signal estimation model, and using the signal obtained by sampling the signal received by the antenna from the RF chip in the terminal device as the signal sample, the randomness and authenticity of the sample data are improved. The signal estimation model is trained using training samples, enabling the deep learning model to have the ability to estimate signals. Since training the signal estimation model using training samples is data-driven, it does not depend on the classification and modeling of specific RF impairments and is not modified with changes in RF hardware. Therefore, it reduces the limitations of scenario use and has better versatility.

[0155] To implement the above embodiments, this application also proposes a signal estimation device.

[0156] Figure 12 This is a schematic diagram of a signal estimation device provided in an embodiment of this application.

[0157] like Figure 12 As shown, the device may include:

[0158] Acquisition module 120 is used to acquire the signal to be processed.

[0159] The estimation module 121 is used to estimate the signal using a trained signal estimation model to obtain the target signal; wherein the signal estimation model is trained using the training device for the signal estimation model in the aforementioned embodiments. The training method can be referred to the relevant explanations in the aforementioned embodiments, as the principle is the same, and will not be repeated here.

[0160] In one implementation of this application, the apparatus further includes an update module, which is configured to:

[0161] Acquire signal samples, input the signal samples into the baseband processor for signal extraction, and obtain the predicted pilot signal;

[0162] The loss function is determined based on the predicted pilot signal and the reference pilot signal stored in the terminal device;

[0163] The signal estimation model is updated based on the loss function to obtain the updated signal estimation model.

[0164] It should be noted that the foregoing explanation of the method embodiments also applies to the apparatus of this embodiment, and will not be repeated here.

[0165] In the signal estimation device of this application embodiment, the signal estimation function of the digital front-end (DFE) is implemented through a deep learning signal estimation model. The signal obtained by sampling the signal received by the antenna from the RF chip in the terminal device is used as a signal sample, improving the randomness and authenticity of the sample data. The signal estimation model is trained using training samples, enabling the deep learning model to possess signal estimation capabilities. Training the signal estimation model using training samples is data-driven, independent of the classification and modeling of specific RF impairments, and does not change with variations in RF hardware. Therefore, it reduces the limitations of application scenarios and has better versatility. Furthermore, using the trained signal estimation model to estimate the signal to be processed allows for rapid and accurate acquisition of signal estimation results, improving the versatility of application scenarios.

[0166] To implement the above embodiments, this application also proposes a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the method described in the foregoing method embodiments.

[0167] To implement the above embodiments, this application also proposes a computer program product having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method described in the foregoing method embodiments.

[0168] To implement the above embodiments, this application also proposes an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, it implements the method described in the foregoing method embodiments.

[0169] Figure 13 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. For example, the electronic device 800 may be a mobile phone, computer, digital broadcasting terminal, messaging device, game console, tablet device, medical device, fitness equipment, personal digital assistant, etc.

[0170] Reference Figure 13The electronic device 800 may include one or more of the following components: processing component 802, memory 804, power component 806, multimedia component 808, audio component 810, input / output (I / O) interface 812, sensor component 814, and communication component 816.

[0171] Processing component 802 typically controls the overall operation of electronic device 800, such as operations associated with display, telephone calls, data communication, camera operation, and recording operations. Processing component 802 may include one or more processors 820 to execute instructions to complete all or part of the steps of the methods described above. Furthermore, processing component 802 may include one or more modules to facilitate interaction between processing component 802 and other components. For example, processing component 802 may include a multimedia module to facilitate interaction between multimedia component 808 and processing component 802.

[0172] Memory 804 is configured to store various types of data to support the operation of electronic device 800. Examples of such data include instructions for any application or method operating on electronic device 800, contact data, phonebook data, messages, pictures, videos, etc. Memory 804 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0173] Power component 806 provides power to various components of electronic device 800. Power component 806 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to electronic device 800.

[0174] Multimedia component 808 includes a screen that provides an output interface between the electronic device 800 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touchscreen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensors may sense not only the boundaries of the touch or swipe action but also the duration and pressure associated with the touch or swipe operation. In some embodiments, multimedia component 808 includes a front-facing camera and / or a rear-facing camera. When the electronic device 800 is in an operating mode, such as a shooting mode or a video mode, the front-facing camera and / or the rear-facing camera may receive external multimedia data. Each front-facing camera and rear-facing camera may be a fixed optical lens system or have focal length and optical zoom capabilities.

[0175] Audio component 810 is configured to output and / or input audio signals. For example, audio component 810 includes a microphone (MIC) configured to receive external audio signals when electronic device 800 is in an operating mode, such as call mode, recording mode, and voice recognition mode. The received audio signals may be further stored in memory 804 or transmitted via communication component 816. In some embodiments, audio component 810 also includes a speaker for outputting audio signals.

[0176] I / O interface 812 provides an interface between processing component 802 and peripheral interface modules, such as keyboards, click wheels, buttons, etc. These buttons may include, but are not limited to, home buttons, volume buttons, power buttons, and lock buttons.

[0177] Sensor assembly 814 includes one or more sensors for providing state assessments of various aspects of electronic device 800. For example, sensor assembly 814 can detect the on / off state of electronic device 800, the relative positioning of components such as the display and keypad of electronic device 800, changes in position of electronic device 800 or a component of electronic device 800, the presence or absence of user contact with electronic device 800, orientation or acceleration / deceleration of electronic device 800, and temperature changes of electronic device 800. Sensor assembly 814 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. Sensor assembly 814 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, sensor assembly 814 may also include an accelerometer, gyroscope, magnetometer, pressure sensor, or temperature sensor.

[0178] Communication component 816 is configured to facilitate wired or wireless communication between electronic device 800 and other devices. Electronic device 800 can access wireless networks based on communication standards, such as WiFi, 4G, or 5G, or combinations thereof. In one exemplary embodiment, communication component 816 receives broadcast signals or broadcast-related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, communication component 816 also includes a near-field communication (NFC) module to facilitate short-range communication. For example, the NFC module may be implemented based on radio frequency identification (RFID) technology, Infrared Data Association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.

[0179] In an exemplary embodiment, the electronic device 800 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the methods described above.

[0180] In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory 804 including instructions, which can be executed by a processor 820 of an electronic device 800 to perform the above-described method. For example, the non-transitory computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device, etc.

[0181] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0182] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0183] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0184] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0185] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0186] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0187] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0188] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of this application.

Claims

1. A training method for a signal estimation model, characterized in that, The method comprises: acquiring a signal sample; wherein the signal sample is obtained by sampling a signal received by an antenna by a radio frequency chip; performing signal estimation on the signal sample by using a signal estimation model arranged in the radio frequency chip to obtain an estimated signal; training the signal estimation model according to a true value signal corresponding to the estimated signal and the signal sample, to obtain a trained signal estimation model, wherein the signal estimation model realizes mapping between an original signal and the estimated signal to remove distortion information in the signal; the signal sample is a signal sequence, the signal estimation model comprises an encoding network, a sampling network, an encoder and an output network, and the performing signal estimation on the signal sequence by using the signal estimation model arranged in the radio frequency chip to obtain an estimated signal comprises: performing position encoding on the signal sequence by using the encoding network to obtain first encoding features; sampling the first encoding features by using the sampling network to obtain second encoding features; wherein the second encoding features are features matching a processing dimension of the radio frequency chip; performing signal estimation according to the second encoding features by using the encoder to obtain target signal features; obtaining the estimated signal according to the target signal features by using the output network.

2. The method of claim 1, wherein, the performing signal estimation according to the second encoding features by using the encoder to obtain target signal features comprises: performing attention encoding on the second encoding features by using a multi-head attention layer in the encoder to obtain third encoding features output by the multi-head attention layer; wherein the third encoding features carry estimated signal information; fusing and normalizing the third encoding features and the second encoding features by using a residual connection layer in the encoder to obtain the target signal features.

3. The method of claim 1, wherein, the performing position encoding on the signal sequence by using the encoding network to obtain first encoding features comprises: acquiring a receiving time unit of each signal included in the signal sequence; determining position information of each signal in the signal sequence according to the receiving time unit of each signal; wherein the position information is used to indicate a receiving order of each signal in the signal sequence; adding corresponding position information to each signal in the signal sequence to obtain the first encoding features.

4. The method of claim 1, wherein, there are multiple encoders, and the performing signal estimation according to the second encoding features by using the encoder to obtain target signal features comprises: performing signal estimation according to the second encoding features by using each of the encoders to obtain first signal features corresponding to each of the encoders; fusing multiple first signal features to obtain the target signal features.

5. The method of claim 1, wherein, the training the signal estimation model according to a true value signal corresponding to the estimated signal and the signal sample to obtain a trained signal estimation model comprises: determining a first probability distribution corresponding to the estimated signal and a second probability distribution corresponding to the true value signal according to a bit width of the estimated signal; determining a loss function according to a difference between the first probability distribution and the second probability distribution; According to the loss function, the signal estimation model is trained to obtain a trained signal estimation model.

6. A signal estimation method characterized by, The method comprises: Obtaining a signal to be processed; Using the trained signal estimation model to perform signal estimation on the signal to obtain a target signal; wherein the signal estimation model is trained by the method of any one of claims 1-5.

7. The method of claim 6, wherein, The method further comprises: Obtaining a signal sample; Inputting the signal sample into a baseband processor to perform signal extraction to obtain a predicted pilot signal; According to the predicted pilot signal and the reference pilot signal stored in the terminal device, a loss function is determined; According to the loss function, the signal estimation model is updated to obtain an updated signal estimation model.

8. An apparatus for training a signal estimation model, the apparatus comprising: a processor configured to: receive a plurality of training signals; and train the signal estimation model using the plurality of training signals. The method comprises: An obtaining module is configured to obtain a signal sample; wherein the signal sample is obtained by sampling a signal received by an antenna by a radio frequency chip; A processing module is configured to perform signal estimation on the signal sample by a signal estimation model arranged in the radio frequency chip to obtain an estimated signal; A training module is configured to train a signal estimation model according to the estimated signal and a true value signal corresponding to the signal sample to obtain a trained signal estimation model, wherein the signal estimation model realizes mapping between an original signal and an estimated signal to remove distortion information in the signal; The signal sample is a signal sequence, and the signal estimation model comprises an encoding network, a sampling network, an encoder and an output network, and the signal estimation model is used to perform signal estimation on the signal sequence to obtain an estimated signal, which comprises: The encoding network is used to perform position encoding on the signal sequence to obtain first encoding features; The sampling network is used to sample the first encoding features to obtain second encoding features; wherein the second encoding features are features matching a processing dimension of the radio frequency chip; The encoder is used to perform signal estimation according to the second encoding features to obtain target signal features; The output network is used to obtain the estimated signal according to the target signal features.

9. A signal estimation device, characterized by, The method comprises: An obtaining module is configured to obtain a signal to be processed; An estimation module is configured to use a trained signal estimation model to perform signal estimation on the signal to obtain a target signal; wherein the signal estimation model is trained by the device of claim 8.

10. An electronic device, comprising: The computer program is executed by the processor to implement the method of any one of claims 1-5, or the method of any one of claims 6-7.

11. A non-transitory computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the method of any one of claims 1-5, or the method of any one of claims 6-7.

12. A chip, characterized by The chip comprises a processing circuit configured to execute the method of any one of claims 1-5, or the method of any one of claims 6-7.

13. A computer program product, characterised in that, A computer program comprising computer program elements which, when executed by a processor, implement the method of any one of claims 1-5, or implement the method of any one of claims 6-7.

Images