A data compensation method for a radio frequency chip

Abstract

The present invention relates to the field of radio frequency chip data compensation technology, and discloses a data compensation method for radio frequency chips, including: real-time acquisition of I / Q signal status data of the radio frequency chip; calculation based on a fuzzy control algorithm to obtain a first delay compensation parameter; construction of a nonlinear compensation filter to obtain a preliminary delay compensation signal and feedback adjustment of the first delay compensation parameter; construction of a long short-term memory network to predict and obtain a third delay compensation parameter; and inputting the third delay compensation parameter into the nonlinear compensation filter. Compared with the existing technology using fixed interpolation or simple linear feedback methods, it is difficult to achieve high-precision, real-time dynamic delay compensation, especially under conditions of large-scale MIMO and multi-channel ADC clock inconsistency and dynamic changes in signal status; this application introduces a fuzzy control algorithm and combines it with a long short-term network to achieve accurate prediction and real-time update of the delay compensation parameter, thereby improving the data synchronization performance of the radio frequency chip.

Description

Technical Field

[0001] The present invention belongs to the technical field of radio frequency chip data compensation, and in particular relates to a data compensation method for a radio frequency chip. Background Art

[0002] With the rapid development of 5G, 6G, and future communication systems, RF chips are increasingly being used in wireless communications, radar systems, and the Internet of Things. In particular, in Massive MIMO (Massive Multiple-Input Multiple-Output) systems, to meet the demands of high throughput, low latency, and high reliability, RF chips place extremely high demands on signal synchronization and data accuracy during multi-channel parallel processing. However, due to the heterogeneity of RF signal channels and the inconsistency of ADC (analog-to-digital converter) clocks across each channel, subtle but significant fractional delay variations are common in these systems. While minimal, these delay variations can lead to desynchronization between I / Q signals in high-density, multi-channel parallel transmission environments, resulting in signal demodulation errors, increased phase noise, and reduced data processing accuracy. For example, in traditional Massive MIMO systems, methods such as fixed interpolation or piecewise polynomials are often used to achieve delay compensation. However, these methods have drawbacks, including poor real-time performance. Fixed or piecewise methods often rely on pre-set compensation parameters and cannot respond to dynamic changes in channel conditions in real time, resulting in significant synchronization errors. High computational complexity: When faced with a large number of parallel channels, algorithms such as piecewise polynomial interpolation have high computational complexity, making it difficult to achieve fast calculations while maintaining low power consumption and high real-time performance. Traditional methods also lack adaptability: They often use linear models or fixed rules, which make it difficult to fully capture the nonlinear distortion and multidimensional information in the signal. This leads to poor compensation in dynamic and complex channel environments and serious error accumulation. Therefore, there is an urgent need for a method that can achieve real-time and accurate delay compensation in complex and dynamic channel environments to improve the signal synchronization performance and data processing accuracy of RF chips. Summary of the Invention

[0003] In response to the above-mentioned technical deficiencies, the purpose of the present invention is to propose a data compensation method for RF chips, aiming to solve the technical problem that fixed interpolation or simple linear feedback methods are used in the existing technology, especially under the conditions of large-scale MIMO and multi-channel ADC clock inconsistency and dynamic signal state changes, which makes it difficult to achieve high-precision, real-time dynamic delay compensation.

[0004] In order to solve the above technical problems, the present invention adopts the following technical solution: The present invention provides a data compensation method for a radio frequency chip,

[0005] The data compensation method of the radio frequency chip includes:

[0006] Step S10: real-time state data of the I / Q signal of the RF chip is collected in real time, the real-time state data including real-time signal amplitude, real-time signal frequency and real-time noise level, and a first delay compensation parameter is calculated based on the fuzzy control algorithm and the real-time state data;

[0007] Step S20: constructing a nonlinear compensation filter, inputting the first delay compensation parameter into the nonlinear compensation filter, and obtaining a preliminary delay compensation signal;

[0008] Step S30: Calculating a signal synchronization error according to the first delay compensation signal, and performing feedback adjustment on the first delay compensation parameter according to the signal synchronization error to obtain a second delay compensation parameter;

[0009] Step S40: constructing a long short-term memory network and performing pre-training, and dynamically predicting the second delay compensation parameter through the pre-trained long short-term memory network to obtain a third delay compensation parameter;

[0010] Step S50: applying the third delay compensation parameter, inputting the third delay compensation parameter into a nonlinear compensation filter to obtain a final delay compensation signal.

[0011] Preferably, in step S10, real-time status data of the I / Q signal of the RF chip is collected in real time, where the real-time status data includes real-time signal amplitude, real-time signal frequency, and real-time noise level. The step of calculating the first delay compensation parameter based on the fuzzy control algorithm combined with the real-time status data specifically includes:

[0012] Step S101: real-time state data of the I / Q signal of the RF chip at time t is collected from the RF chip in real time, including real-time signal amplitude A(t), real-time signal frequency f(t) and real-time noise level N(t);

[0013] Step S102: defining three levels of membership for the signal amplitude A(t), the real-time signal frequency f(t), and the real-time noise level N(t), respectively, and quantizing the signal amplitude A(t), the real-time signal frequency f(t), and the real-time noise level N(t) using a triangular method to form fuzzy variable inputs;

[0014] Step S103: Derivation is performed using a preset Mamdani model according to a preset fuzzy control rule base to obtain a fuzzy set output;

[0015] Step S104: Defuzzifying the fuzzy set output, using the centroid method or the maximum membership method to calculate and obtain the first delay compensation parameter.

[0016] Preferably, in step S20, the step of constructing a nonlinear compensation filter and inputting the first delay compensation parameter into the nonlinear compensation filter to obtain a preliminary delay compensation signal specifically includes:

[0017] Step S201: designing a nonlinear filter based on the Volterra series expansion method;

[0018] Step S202: inputting the first delay compensation parameter into a nonlinear compensation filter, and performing weighted processing on the linear term and the high-order nonlinear term in the nonlinear compensation filter to obtain a mapped first delay compensation parameter;

[0019] Step S203: Apply time domain compensation to the mapped first delay compensation parameter to obtain a preliminary delay compensation signal.

[0020] Preferably, in step S20, the step of applying time domain compensation to the mapped first delay compensation parameter to obtain a preliminary delay compensation signal adopts the formula:

[0021] ;

[0022] in, is the preliminary delay compensation signal; is the first delay compensation parameter after mapping; is a nonlinear adjustment factor used to further correct the compensation effect according to the signal estimation error; is the input signal; is the estimated value of the input signal; is the integral variable; is the error integral term; is the delay correction term.

[0023] Preferably, in step S30, the step of calculating the signal synchronization error according to the first delay compensation signal, performing feedback adjustment on the first delay compensation parameter according to the signal synchronization error, and obtaining the second delay compensation parameter specifically includes:

[0024] Step S301: obtaining a target synchronization signal from preset prior information, and performing time-baselining processing on the target synchronization signal and the first delay compensation signal;

[0025] Step S302: Calculating a signal synchronization error using a mean square error based on the target synchronization signal after benchmarking and the first delay compensation signal;

[0026] Step S303: Determine the direction and magnitude of the delay compensation parameter that needs to be adjusted based on the signal synchronization error combined with the PID control method, and output a correction value;

[0027] Step S304: Apply the correction amount to perform feedback adjustment on the first delay compensation parameter to obtain a second delay compensation parameter.

[0028] Preferably, in step S40, the steps of constructing a long short-term memory network and pre-training it, and dynamically predicting the second delay compensation parameter using the pre-trained long short-term memory network to obtain the third delay compensation parameter specifically include:

[0029] Step S401: Acquire historical second delay compensation parameter time series data, historical signal amplitude time series data, historical signal frequency time series data, and historical noise level time series data, perform denoising, normalization, and time alignment on the collected time series data, and construct a fusion feature vector;

[0030] Step S402: Construct a long short-term memory network, including an input layer, multiple LSTM layers, a fully connected layer, and an output layer. The input layer is used to receive fused feature vectors, and the multiple LSTM layers are used to capture long-term dependencies and nonlinear relationships in time series data. The fully connected layer is used to map the output of the LSTM layer to predicted values, and the output layer is used to output the predicted values.

[0031] Step S403: obtaining a target synchronization signal from preset prior information, constructing a standard compensation parameter output according to the target synchronization signal, and training a long short-term memory network using the fused feature vector and the standard compensation parameter output;

[0032] Step S404: The second delay compensation parameter is used as the output of the long short-term memory network, and the long short-term memory network outputs a third delay compensation parameter.

[0033] Preferably, in step S40, during the process of training the long short-term memory network, the synchronization performance index is introduced as a regularization term to design a loss function L, and the loss function L adopts the formula:

[0034] ;

[0035] Among them, i represents the i-th training sample, is the third delay compensation parameter predicted by the long short-term memory network; Output of standard compensation parameters constructed according to target synchronization signal; is the preset regularization coefficient used to balance the two parts of loss; is the measure of the signal synchronization error after compensation based on the prediction parameters; N is the preset number of training samples.

[0036] The present invention also provides a data compensation system for a radio frequency chip, comprising:

[0037] A real-time status data acquisition module is used to collect real-time status data of the I / Q signal of the RF chip in real time. The real-time status data includes real-time signal amplitude, real-time signal frequency and real-time noise level, and calculates the first delay compensation parameter based on the fuzzy control algorithm combined with the real-time status data;

[0038] A fuzzy control delay parameter calculation module is used to construct a nonlinear compensation filter, input the first delay compensation parameter into the nonlinear compensation filter, and obtain a preliminary delay compensation signal;

[0039] a nonlinear compensation filter module, configured to calculate a signal synchronization error based on the first delay compensation signal, and perform feedback adjustment on the first delay compensation parameter based on the signal synchronization error to obtain a second delay compensation parameter;

[0040] A synchronization error feedback and parameter adjustment module is used to construct a long short-term memory network and perform pre-training, and dynamically predict the second delay compensation parameter through the pre-trained long short-term memory network to obtain the third delay compensation parameter;

[0041] The LSTM delay prediction and optimization module is used to apply the third delay compensation parameter, input the third delay compensation parameter into the nonlinear compensation filter, and obtain a final delay compensation signal.

[0042] The present invention also provides a computer program product, including a data compensation program for a radio frequency chip, wherein the data compensation program for the radio frequency chip implements the data compensation method for the radio frequency chip when executed by a processor.

[0043] The beneficial effects of the present invention are as follows: compared with the existing technology that adopts fixed interpolation or simple linear feedback methods, especially under the conditions of inconsistent clocks of large-scale MIMO and multi-channel ADCs and dynamic changes in signal states, it is difficult to achieve high-precision, real-time dynamic delay compensation; this application introduces a fuzzy control algorithm and combines it with long-term and short-term networks to achieve accurate prediction and real-time update of delay compensation parameters, thereby improving the data synchronization performance of the RF chip. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0045] Figure 1 The figure is a flow chart of a first embodiment of a data compensation method for a radio frequency chip according to the present invention.

[0046] Figure 2 The present invention is a schematic diagram of a device for a data compensation method for a radio frequency chip. DETAILED DESCRIPTION

[0047] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0048] Example 1: Figure 1 FIG. 1 is a flow chart of a first embodiment of a data compensation method for a radio frequency chip according to the present invention, and provides a first embodiment of a data compensation method for a radio frequency chip according to the present invention.

[0049] In a first embodiment, the data compensation method of the radio frequency chip includes:

[0050] Step S10: real-time state data of the I / Q signal of the RF chip is collected in real time, the real-time state data including real-time signal amplitude, real-time signal frequency and real-time noise level, and a first delay compensation parameter is calculated based on the fuzzy control algorithm and the real-time state data;

[0051] It should be noted that in step S10, real-time state data of the I / Q signal of the RF chip is collected in real time, and the real-time state data includes real-time signal amplitude, real-time signal frequency, and real-time noise level. The step of calculating the first delay compensation parameter based on the fuzzy control algorithm and the real-time state data specifically includes:

[0052] Step S101: real-time state data of the I / Q signal of the RF chip at time t is collected from the RF chip in real time, including real-time signal amplitude A(t), real-time signal frequency f(t) and real-time noise level N(t);

[0053] Step S102: defining three levels of membership for the signal amplitude A(t), the real-time signal frequency f(t), and the real-time noise level N(t), respectively, and quantizing the signal amplitude A(t), the real-time signal frequency f(t), and the real-time noise level N(t) using a triangular method to form fuzzy variable inputs;

[0054] Step S103: Derivation is performed using a preset Mamdani model according to a preset fuzzy control rule base to obtain a fuzzy set output;

[0055] Step S104: Defuzzifying the fuzzy set output, using the centroid method or the maximum membership method to calculate and obtain the first delay compensation parameter.

[0056] It can be understood that by using triangular quantization, continuously changing signal state data can be mapped into discrete fuzzy sets, providing accurate and stable input data for subsequent fuzzy reasoning. The fuzzy control rule base stores multiple rules, such as "If the signal amplitude is high, the frequency is low, and the noise is low, then the delay compensation parameter should be large." These rules are comprehensively reflected in the output fuzzy set through the Mamdani model, ensuring that the algorithm can flexibly respond to nonlinear signal changes. Defuzzification concretizes the results of fuzzy reasoning, giving the output parameters clear numerical representations, which facilitates parameter input and delay compensation processing in subsequent modules (such as nonlinear compensation filters).

[0057] It should be understood that the present invention dynamically collects signal state data and utilizes fuzzy control technology to comprehensively consider multiple factors, such as signal amplitude, frequency, and noise, to calculate the first delay compensation parameter in real time. Compared with traditional fixed or piecewise interpolation methods, this solution offers greater real-time performance, adaptability, and compensation accuracy. It effectively addresses signal synchronization issues caused by massive MIMO and ADC delay inconsistencies, thereby improving demodulation accuracy and data processing precision throughout the communication process.

[0058] For example, in a certain actual test, the signal data collected by step S101 is: 、 GHz and After quantization in step S102, the corresponding membership degrees are determined to be "medium", "low" and "low" respectively. In step S103, based on the preset fuzzy rules, the fuzzy output obtained is "the delay compensation parameter is large". Finally, the centroid method is used to defuzzify and calculate the specific first delay compensation parameter value. This result shows that the proposed method can accurately estimate the required delay compensation parameters in a dynamic channel environment, thus providing an accurate basis for subsequent compensation.

[0059] Step S20: constructing a nonlinear compensation filter, inputting the first delay compensation parameter into the nonlinear compensation filter, and obtaining a preliminary delay compensation signal;

[0060] It should be noted that, in step S20, the step of constructing a nonlinear compensation filter, inputting the first delay compensation parameter into the nonlinear compensation filter, and obtaining a preliminary delay compensation signal specifically includes:

[0061] Step S201: designing a nonlinear filter based on the Volterra series expansion method;

[0062] Step S202: inputting the first delay compensation parameter into a nonlinear compensation filter, and performing weighted processing on the linear term and the high-order nonlinear term in the nonlinear compensation filter to obtain a mapped first delay compensation parameter;

[0063] Step S203: Apply time domain compensation to the mapped first delay compensation parameter to obtain a preliminary delay compensation signal.

[0064] The step of applying time domain compensation to the mapped first delay compensation parameter to obtain a preliminary delay compensation signal adopts the formula:

[0065] ;

[0066] in, is the preliminary delay compensation signal; is the first delay compensation parameter after mapping; is a nonlinear adjustment factor used to further correct the compensation effect according to the signal estimation error; is the input signal; is the estimated value of the input signal; is the integral variable; is the error integral term; is the delay correction term.

[0067] It can be understood that the above steps, by introducing the Volterra series expansion method into nonlinear filter design, can fully describe and compensate for the slight nonlinear delay differences between ADCs. At the same time, by weighting the linear terms and higher-order nonlinear terms separately, the first input delay compensation parameter is flexibly mapped, providing a more accurate correction basis for time-domain compensation. The time-domain compensation processing formula not only directly corrects the delay of the input signal but also introduces a nonlinear adjustment factor and an error integral term, comprehensively considering historical error accumulation. This allows for fine-tuning the delay and ensures that the resulting preliminary compensated signal is closer to the desired synchronization signal.

[0068] It should be understood that the present invention utilizes a nonlinear filter designed using Volterra series expansions to obtain a more accurate preliminary delay compensation signal after weighted mapping, time-domain compensation, and online tuning of the first delay compensation parameter. Compared to traditional methods, this solution offers higher compensation accuracy, better adaptability, and real-time performance for addressing the nonlinear issues caused by inconsistent ADC delays in massive MIMO systems, significantly improving the signal synchronization performance and data processing quality of the RF chip.

[0069] For example, in a test, assume that the first delay compensation parameter obtained from step S10 is ns, which is mapped after weighted processing in step S202 ns, and at the same time, the measured delay correction term ns, input signal Its estimated value The error integral term is calculated by integration to obtain a certain value, which is adjusted by the nonlinear adjustment factor After correction, the preliminary compensation signal finally obtained in step S203 is The synchronization error was significantly reduced, and its error index was reduced by about 20% compared with the traditional method using fixed compensation parameters, verifying the efficient compensation capability of this scheme for complex nonlinear time delay errors in a real-time dynamic environment.

[0070] Step S30: Calculating a signal synchronization error according to the first delay compensation signal, and performing feedback adjustment on the first delay compensation parameter according to the signal synchronization error to obtain a second delay compensation parameter;

[0071] It should be noted that, in step S30, the steps of calculating the signal synchronization error according to the first delay compensation signal, performing feedback adjustment on the first delay compensation parameter according to the signal synchronization error, and obtaining the second delay compensation parameter specifically include:

[0072] Step S301: obtaining a target synchronization signal from preset prior information, and performing time-baselining processing on the target synchronization signal and the first delay compensation signal;

[0073] Step S302: Calculating a signal synchronization error using a mean square error based on the target synchronization signal after benchmarking and the first delay compensation signal;

[0074] Step S303: Determine the direction and magnitude of the delay compensation parameter that needs to be adjusted based on the signal synchronization error combined with the PID control method, and output a correction value;

[0075] Step S304: Apply the correction amount to perform feedback adjustment on the first delay compensation parameter to obtain a second delay compensation parameter.

[0076] Understandably, in massive MIMO systems, the first delay compensation signal often fails to fully achieve the desired synchronization due to inconsistent ADC clocks and dynamic channel environments. Therefore, it is necessary to calculate the actual signal synchronization error and, based on this error, use feedback control to adjust the first delay compensation parameter. This yields a more accurate second delay compensation parameter, providing accurate input for subsequent compensation.

[0077] It should be understood that the mean square error metric can comprehensively reflect the synchronization deviation of the signal over a period of time, thereby obtaining a more stable and accurate error indicator. The PID control method has the advantages of rapid adjustment and smooth response, and can calculate the appropriate correction amount in real time, so that the compensation parameter quickly approaches the optimal state. The correction amount is applied to the first delay compensation parameter for feedback adjustment to obtain the second delay compensation parameter. The feedback adjustment mechanism enables dynamic correction of the compensation deviation caused by environmental changes or imperfect models, so that the delay compensation parameter continues to approach the optimal value. In summary, the use of real-time feedback and PID control methods to correct the first delay compensation parameter can not only make up for the shortcomings of the traditional fixed compensation method, but also can continuously optimize the compensation effect in a dynamic and complex communication environment, ensuring the high accuracy and stability of RF chip data synchronization.

[0078] For example, in a certain actual test, the target synchronization signal is obtained through step S301. With the first delay compensation signal After time benchmarking, we found that the mean square error between the two was is 0.05. Using pre-set PID parameters 、 and Then, the correction amount is calculated ns. Add this correction to the first delay compensation parameter ns later, the second delay compensation parameter is obtained Experimental results show that after feedback adjustment, the signal synchronization error is reduced by about 15%, significantly improving the signal synchronization effect.

[0079] Step S40: constructing a long short-term memory network and performing pre-training, and dynamically predicting the second delay compensation parameter through the pre-trained long short-term memory network to obtain a third delay compensation parameter;

[0080] It should be noted that in step S40, the steps of constructing a long short-term memory network and pre-training it, and dynamically predicting the second delay compensation parameter using the pre-trained long short-term memory network to obtain the third delay compensation parameter specifically include:

[0081] Step S401: Acquire historical second delay compensation parameter time series data, historical signal amplitude time series data, historical signal frequency time series data, and historical noise level time series data, perform denoising, normalization, and time alignment on the collected time series data, and construct a fusion feature vector;

[0082] Step S402: Construct a long short-term memory network, including an input layer, multiple LSTM layers, a fully connected layer, and an output layer. The input layer is used to receive fused feature vectors, and the multiple LSTM layers are used to capture long-term dependencies and nonlinear relationships in time series data. The fully connected layer is used to map the output of the LSTM layer to predicted values, and the output layer is used to output the predicted values.

[0083] Step S403: obtaining a target synchronization signal from preset prior information, constructing a standard compensation parameter output according to the target synchronization signal, and training a long short-term memory network using the fused feature vector and the standard compensation parameter output;

[0084] Step S404: The second delay compensation parameter is used as the output of the long short-term memory network, and the long short-term memory network outputs a third delay compensation parameter.

[0085] Among them, the process of training the long short-term memory network also includes introducing the synchronization performance index as a regular term to design the loss function L. The loss function L adopts the formula:

[0086] ;

[0087] Among them, i represents the i-th training sample, is the third delay compensation parameter predicted by the long short-term memory network; Output of standard compensation parameters constructed according to target synchronization signal; is the preset regularization coefficient used to balance the two parts of loss; is the measure of the signal synchronization error after compensation based on the prediction parameters; N is the preset number of training samples.

[0088] Understandably, traditional methods often only use a single parameter for prediction. However, this step, by integrating multi-dimensional features, can more comprehensively reflect channel dynamics and compensation requirements, providing rich information for subsequent predictions. The use of a deep LSTM network fully learns the temporal dependencies in historical data, far superior to traditional simple linear or polynomial models, and can more accurately capture nonlinear dynamic changes. By introducing a synchronization performance metric as a regularization term, model training not only focuses on the numerical difference between the predicted and true values, but also directly optimizes signal synchronization performance. This can significantly reduce compensation errors in practice, significantly outperforming traditional methods that only consider the Mean Squared Error (MSE). LSTM online inference dynamically predicts delay compensation parameters in real time based on the latest data, offering fast response and high accuracy. It is more adaptable to environmental changes than predictions based on traditional fixed models. An adaptive update mechanism ensures that the LSTM model continuously adapts to changing channel conditions, preventing degradation in model prediction capabilities and ensuring long-term, high-precision delay compensation. This provides greater robustness than traditional one-time trained models.

[0089] It should be understood that this embodiment utilizes an LSTM network to perform deep learning and prediction on historical time series data, dynamically outputting a third delay compensation parameter. This enables precise control of delay compensation in complex and changing communication environments. This solution outperforms traditional methods in terms of accuracy, real-time performance, and adaptability, providing a higher level of technical support for RF chip data compensation.

[0090] Step S50: applying the third delay compensation parameter, inputting the third delay compensation parameter into a nonlinear compensation filter to obtain a final delay compensation signal.

[0091] It should be understood that using the third delay compensation parameter can overcome the cumulative errors caused by noise, channel fluctuations, and ADC differences, achieving dynamic and adaptive delay correction. Compared with traditional methods using fixed compensation parameters, this method has higher real-time performance and adaptability, significantly improving signal synchronization accuracy and overall data processing quality.

[0092] Embodiment 2: In addition, the present invention provides a data compensation system for a radio frequency chip, which adopts the data compensation method for a radio frequency chip in the above embodiment to solve the technical problem of data compensation for a radio frequency chip. Compared with the prior art, the beneficial effects of the data compensation system for a radio frequency chip provided by the present invention are the same as the beneficial effects of the data compensation method for a radio frequency chip provided by the above embodiment. The other technical features of the data compensation system for a radio frequency chip are the same as those disclosed in the above embodiment method, and are not further described here.

[0093] Example 3: The present invention provides a data compensation device for a radio frequency chip, please refer to Figure 2A data compensation device for a radio frequency chip includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the data compensation method for a radio frequency chip described in the first embodiment. The data compensation device for a radio frequency chip in this embodiment of the present invention may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. The data compensation device for a radio frequency chip is merely an example and should not limit the functionality or scope of use of this embodiment of the present invention. The data compensation device for a radio frequency chip may include a processing device 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which may perform various appropriate actions and processes based on a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage device 1003 into a random access memory (RAM) 1004. RAM 1004 also stores various programs and data required for the operation of a data compensation device for a radio frequency chip. The processing device 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems may be connected to the I / O interface 1006: an input device 1007, such as a touch screen, touchpad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; an output device 1008, such as a liquid crystal display (LCD), speaker, vibrator, etc.; a storage device 1003, such as a magnetic tape or hard disk; and a communication device 1009. The communication device 1009 allows the data compensation device for a radio frequency chip to communicate with other devices wirelessly or wired to exchange data. While the figure shows a data compensation device for a radio frequency chip with various systems, it should be understood that implementation or presence of all the illustrated systems is not required. More or fewer systems may alternatively be implemented or present.

[0094] Example 4: The present invention also provides a computer program product, including a computer program. When executed by a processor, the computer program implements the steps of the data compensation method for a radio frequency chip as described above. The computer program product provided by the present invention can solve the technical problem of data compensation for a radio frequency chip. Compared with the prior art, the beneficial effects of the computer program product provided by the present invention are the same as those of the data compensation method for a radio frequency chip provided in the above embodiment, and are not further described here.

[0095] In particular, according to the embodiments disclosed in the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments disclosed in the present invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program comprising program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present invention are performed.

[0096] It should be understood that the various parts disclosed in the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any appropriate manner in any one or more embodiments or examples.

[0097] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A data compensation method for a radio frequency chip, characterized in that: Methods include: Step S10: real-time state data of the I / Q signal of the RF chip is collected in real time, the real-time state data including real-time signal amplitude, real-time signal frequency and real-time noise level, and a first delay compensation parameter is calculated based on the fuzzy control algorithm and the real-time state data; Step S20: constructing a nonlinear compensation filter, inputting the first delay compensation parameter into the nonlinear compensation filter, and obtaining a preliminary delay compensation signal; Step S30: Calculating a signal synchronization error according to the first delay compensation signal, and performing feedback adjustment on the first delay compensation parameter according to the signal synchronization error to obtain a second delay compensation parameter; Step S40: constructing a long short-term memory network and performing pre-training, and dynamically predicting the second delay compensation parameter through the pre-trained long short-term memory network to obtain a third delay compensation parameter; Step S50: applying the third delay compensation parameter, inputting the third delay compensation parameter into a nonlinear compensation filter to obtain a final delay compensation signal.

2. The data compensation method of a radio frequency chip according to claim 1, characterized in that: In step S10, real-time state data of the I / Q signal of the RF chip is collected in real time, where the real-time state data includes real-time signal amplitude, real-time signal frequency, and real-time noise level. The step of calculating the first delay compensation parameter based on the fuzzy control algorithm and the real-time state data specifically includes: Step S101: real-time state data of the I / Q signal of the RF chip at time t is collected from the RF chip in real time, including real-time signal amplitude A(t), real-time signal frequency f(t) and real-time noise level N(t); Step S102: defining three levels of membership for the signal amplitude A(t), the real-time signal frequency f(t), and the real-time noise level N(t), respectively, and quantizing the signal amplitude A(t), the real-time signal frequency f(t), and the real-time noise level N(t) using a triangular method to form fuzzy variable inputs; Step S103: Derivation is performed using a preset Mamdani model according to a preset fuzzy control rule base to obtain a fuzzy set output; Step S104: Defuzzifying the fuzzy set output, using the centroid method or the maximum membership method to calculate and obtain the first delay compensation parameter.

3. The data compensation method of a radio frequency chip according to claim 1, characterized in that: In step S20, a nonlinear compensation filter is constructed, and the first delay compensation parameter is input into the nonlinear compensation filter to obtain a preliminary delay compensation signal, which specifically includes: Step S201: designing a nonlinear filter based on the Volterra series expansion method; Step S202: inputting the first delay compensation parameter into a nonlinear compensation filter, and performing weighted processing on the linear term and the high-order nonlinear term in the nonlinear compensation filter to obtain a mapped first delay compensation parameter; Step S203: Apply time domain compensation to the mapped first delay compensation parameter to obtain a preliminary delay compensation signal.

4. The data compensation method of a radio frequency chip according to claim 3, characterized in that: In step S20, the time domain compensation is performed on the mapped first delay compensation parameter to obtain a preliminary delay compensation signal, using the formula: ； in, is the preliminary delay compensation signal; is the first delay compensation parameter after mapping; is a nonlinear adjustment factor used to further correct the compensation effect according to the signal estimation error; is the input signal; is the estimated value of the input signal; is the integral variable; is the error integral term; is the delay correction term.

5. The data compensation method of a radio frequency chip according to claim 1, characterized in that: In step S30, the steps of calculating the signal synchronization error according to the first delay compensation signal, performing feedback adjustment on the first delay compensation parameter according to the signal synchronization error, and obtaining the second delay compensation parameter specifically include: Step S301: obtaining a target synchronization signal from preset prior information, and performing time-baselining processing on the target synchronization signal and the first delay compensation signal; Step S302: Calculating a signal synchronization error using a mean square error based on the target synchronization signal after benchmarking and the first delay compensation signal; Step S303: Determine the direction and magnitude of the delay compensation parameter that needs to be adjusted based on the signal synchronization error combined with the PID control method, and output a correction value; Step S304: Apply the correction amount to perform feedback adjustment on the first delay compensation parameter to obtain a second delay compensation parameter.

6. The data compensation method of a radio frequency chip according to claim 1, characterized in that: In step S40, a long short-term memory network is constructed and pre-trained, and the second delay compensation parameter is dynamically predicted by the pre-trained long short-term memory network to obtain the third delay compensation parameter, specifically comprising: Step S401: Acquire historical second delay compensation parameter time series data, historical signal amplitude time series data, historical signal frequency time series data, and historical noise level time series data, perform denoising, normalization, and time alignment on the collected time series data, and construct a fusion feature vector; Step S402: Construct a long short-term memory network, including an input layer, multiple LSTM layers, a fully connected layer, and an output layer. The input layer is used to receive fused feature vectors, and the multiple LSTM layers are used to capture long-term dependencies and nonlinear relationships in time series data. The fully connected layer is used to map the output of the LSTM layer to predicted values, and the output layer is used to output the predicted values. Step S403: obtaining a target synchronization signal from preset prior information, constructing a standard compensation parameter output according to the target synchronization signal, and training a long short-term memory network using the fused feature vector and the standard compensation parameter output; Step S404: The second delay compensation parameter is used as the output of the long short-term memory network, and the long short-term memory network outputs a third delay compensation parameter.

7. The data compensation method of a radio frequency chip according to claim 6, characterized in that: In step S40, the process of training the long short-term memory network also includes introducing the synchronization performance index as a regularization term to design the loss function L. The loss function L adopts the formula: ； Among them, i represents the i-th training sample, is the third delay compensation parameter predicted by the long short-term memory network; Output of standard compensation parameters constructed according to target synchronization signal; is the preset regularization coefficient used to balance the two parts of loss; is the measure of the signal synchronization error after compensation based on the prediction parameters; N is the preset number of training samples.

8. A data compensation system for a radio frequency chip, applied to a data compensation method for a radio frequency chip according to any one of claims 1 to 7, characterized in that: The data compensation system of the radio frequency chip includes: A real-time status data acquisition module is used to collect real-time status data of the I / Q signal of the RF chip in real time. The real-time status data includes real-time signal amplitude, real-time signal frequency and real-time noise level, and calculates the first delay compensation parameter based on the fuzzy control algorithm combined with the real-time status data; A fuzzy control delay parameter calculation module is used to construct a nonlinear compensation filter, input the first delay compensation parameter into the nonlinear compensation filter, and obtain a preliminary delay compensation signal; a nonlinear compensation filter module, configured to calculate a signal synchronization error based on the first delay compensation signal, and perform feedback adjustment on the first delay compensation parameter based on the signal synchronization error to obtain a second delay compensation parameter; A synchronization error feedback and parameter adjustment module is used to construct a long short-term memory network and perform pre-training, and dynamically predict the second delay compensation parameter through the pre-trained long short-term memory network to obtain the third delay compensation parameter; The LSTM delay prediction and optimization module is used to apply the third delay compensation parameter, input the third delay compensation parameter into the nonlinear compensation filter, and obtain a final delay compensation signal.

9. A data compensation device for a radio frequency chip, characterized in that: The data compensation device of the radio frequency chip includes: a memory, a processor, and a data compensation program for the radio frequency chip stored in the memory and runnable on the processor. When the data compensation program for the radio frequency chip is executed by the processor, a data compensation method for the radio frequency chip according to any one of claims 1 to 7 is implemented.

10. A computer program product, characterized in that The computer program product includes a data compensation program for a radio frequency chip, and when the data compensation program for a radio frequency chip is executed by a processor, a data compensation method for a radio frequency chip according to any one of claims 1 to 7 is implemented.

Images