A radio frequency processing circuit control method and system

Abstract

The present application relates to the field of radio frequency communication technology, and discloses a kind of radio frequency processing circuit control method and system, by constructing complex reflection spectrum slice matrix, the local bending degree, phase continuous variation degree etc. Local structure characteristics in frequency domain are analyzed, the wideband component continuously changed along the frequency direction in reflection data is distinguished from the sharp component concentrated on a small number of frequency points and a small number of time slices, global mismatch index and baseline flatness index are generated based on wideband mismatch result, and sharp proportion index is generated based on local sharp result, so that subsequent control decision is closer to real wideband mismatch state, multiple static tentative control states are constructed around current matching control code, same data processing and result evaluation are carried out respectively, to determine target matching control code and target power amplifier bias control code, so that the hypersensitivity of power amplifier bias control to local sharp abnormality is reduced, while ensuring the safety of power amplifier, maintaining better transmission capability and control stability.

Description

Technical Field

[0001] This invention relates to the field of radio frequency communication technology, and in particular to a radio frequency processing circuit control method and system. Background Technology

[0002] Radio frequency (RF) chips are typically used to perform operations such as transmit signal processing, power amplification and driving, impedance matching and adjustment, and transmit status monitoring. In practical applications, the control accuracy of the RF processing circuit directly affects the energy transmission efficiency of the antenna port, transmit power stability, spectral purity, and power amplifier reliability. Therefore, how to accurately control the matching network and power amplifier bias based on the on-chip detected reflection status has always been a key issue in RF control technology.

[0003] In existing technologies, common control methods are typically based on single-point reflection power detection, VSWR detection, return loss detection, or simple lookup table switching. These methods can accomplish basic control tasks in typical operating scenarios. For example, when the antenna's surrounding environment is relatively stable, the boundary conditions near the feed point change little, and there are no obvious local anomalies within the operating frequency band, the reflection detection results can usually directly characterize the overall mismatch. The control circuit can then adjust the matching state or back off the power amplifier bias accordingly, generally achieving relatively stable results.

[0004] However, in high-humidity condensation environments, micro-liquid bridges, localized water film boundaries, or intermittently wetted conductive areas may form near the antenna feed point. These special conditions often do not cause a synchronous, uniform change across the entire operating frequency band, but rather more easily create narrow-band, spike-like, and slightly drifting local anomalies near the center frequency. When the RF chip still uses single-point detection or simple averaging methods for judgment, it is easy to misjudge local spike anomalies as overall broadband mismatch. Furthermore, if the control system directly switches the matching state or prematurely reduces the power amplifier bias based on such misjudgments, it can easily cause problems such as back-and-forth switching of the matching network, fluctuations in transmit power, deterioration of spectral performance, and degradation of link quality. Technically speaking, the key in the above scenario is not whether the reflection value increases, but whether the detected anomaly is caused by a true broadband mismatch or by localized liquid bridge disturbances. The reason why traditional methods can still be used in ordinary scenarios is that most anomalies in ordinary scenarios can be approximately understood as overall mismatch changes, without the need to distinguish between local spike components and broadband continuous components. However, in the high humidity condensate bridge scenario, if local spike anomalies cannot be separated from the true broadband mismatch trend, the subsequent control results will be significantly distorted.

[0005] Therefore, how to conduct a more detailed analysis of the reflection structure near the operating center frequency, distinguish between the true broadband mismatch trend and the local spike disturbance trend, and determine more reasonable matching control codes and power amplifier bias control codes on this basis has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0006] This invention provides a radio frequency processing circuit control method and system to at least solve the problem in the prior art that, in high humidity condensate bridge scenarios, traditional control methods are unable to distinguish between local spike anomalies and the true broadband mismatch trend, which leads to mis-switching of the matching network and mis-back-off of the power amplifier bias.

[0007] To achieve the above objectives, the present invention provides a radio frequency processing circuit control method, the method comprising the following steps: Acquire on-chip complex reflection spectrum sampling data under the current control state, and construct a complex reflection spectrum slice matrix based on the on-chip complex reflection spectrum sampling data; Based on the complex reflection spectrum slice matrix, frequency domain local structural features are extracted, and the complex reflection spectrum slice matrix is ​​separated into broadband mismatched baseline data and local spike perturbation data; Based on the broadband mismatch baseline data, a global mismatch index characterizing the true degree of broadband mismatch and a baseline flatness index characterizing the degree of broadband profile change are determined, and a spike proportion index characterizing the degree of local anomaly proportion is determined based on the local spike perturbation data. Multiple static trial control states are constructed around the current matching control code. Scoring results are determined based on the global mismatch index, the baseline flatness index, and the spike ratio index corresponding to each static trial control state. The target matching control code is then determined based on the scoring results. The target power amplifier bias control code is determined based on the global mismatch index and the spike ratio index corresponding to the target matching control code, and the target control vector for controlling the radio frequency processing circuit is output.

[0008] Optionally, the on-chip complex reflection spectrum sampling data under the current control state is obtained, and a complex reflection spectrum slice matrix is ​​constructed based on the on-chip complex reflection spectrum sampling data, specifically including: Multiple micro-frequency offset detection points are determined on both sides of the current working center frequency point, and on-chip reflection detection data are collected for each micro-frequency offset detection point on multiple time slices within the same control cycle; For each time slice and each micro-offset frequency detection point, a complex reflection spectrum slice matrix consisting of all complex reflection spectrum samples is constructed: The amplitude and phase data at each time slice and each micro-offset frequency detection point are obtained based on the in-phase component sampling value and the quadrature component sampling value, respectively, for subsequent extraction of local structural features in the frequency domain.

[0009] Optionally, based on the complex reflection spectrum slice matrix, frequency domain local structural features are extracted, and the complex reflection spectrum slice matrix is ​​separated into broadband mismatched baseline data and local spike perturbation data, specifically including: Based on the amplitude data corresponding to each position in the complex reflection spectrum slice matrix, the degree of amplitude change and local bending degree between adjacent micro-frequency detection points are determined along the micro-frequency deviation direction. Based on the phase data corresponding to each position in the complex reflection spectrum slice matrix, the degree of phase continuity change between adjacent micro-frequency detection points is determined along the micro-frequency offset direction. Based on the magnitude of the change, the degree of local bending, and the degree of continuous phase change, structural weight data corresponding to each position is generated; The influence of local frequency abrupt changes on the broadband continuous variation results is reduced based on the structural weight data. The part of the complex reflection spectrum slice matrix that changes continuously along the micro-frequency deviation direction is determined as the broadband mismatch baseline data. The abrupt changes that occur in a small number of frequency points and a small number of time slices in the complex reflection spectrum slice matrix are determined as the local spike perturbation data.

[0010] Optionally, structural weight data corresponding to each position is generated based on the magnitude of the change, the degree of local bending, and the degree of continuous phase change, specifically including: The structural weight values ​​corresponding to each position are determined based on the degree of local bending and the degree of continuous phase change at each position. The data participation degree of each position in the complex reflection spectrum slice matrix is ​​constrained according to the structural weight value, so that the proportion of positions with large local bending degree and large phase continuous change degree is reduced in the broadband mismatch baseline data generation process.

[0011] Optionally, based on the broadband mismatch baseline data, a global mismatch index characterizing the true degree of broadband mismatch and a baseline flatness index characterizing the degree of broadband profile change are determined, and based on the local spike perturbation data, a spike proportion index characterizing the degree of local anomaly proportion is determined, specifically including: Based on the local spike perturbation data, the slice weight of each time slice is determined, and the broadband mismatch baseline data is weighted and aggregated to obtain the broadband mismatch baseline vector. The global mismatch index and baseline flatness index are determined based on the broadband mismatch baseline vector, and the spike proportion index is determined based on the local spike perturbation data and the on-chip complex reflection spectrum sampling data.

[0012] Optionally, multiple static trial control states are constructed around the current matching control code, and the scoring result is determined based on the global mismatch index, the baseline flatness index, and the spike ratio index corresponding to each static trial control state, specifically including: Construct a low-side static trial control state, a current static trial control state, and a high-side static trial control state based on the current matching control code; Under each static test control state, the on-chip complex reflection spectrum sampling data is reacquired, and the generation and processing of the broadband mismatch baseline data and local spike perturbation data, as well as the determination and processing of the global mismatch index, the baseline flatness index and the spike proportion index are repeatedly executed. The scoring results for each static test control state are determined based on the global mismatch index, the baseline flatness index, and the spike ratio index corresponding to each static test control state.

[0013] Optionally, a target matching control code is determined based on the scoring results, specifically including: The scoring results corresponding to the low-side static trial control state, the current static trial control state, and the high-side static trial control state are used as local scoring sampling points to construct a local scoring relationship about the matching control code; The local scoring function parameters are determined based on the scoring results corresponding to the low-side static test control state, the current static test control state, and the high-side static test control state. When the local scoring function opens upward, a continuous optimal matching control value is determined according to the local scoring function, and the continuous optimal matching control value is mapped to a preset discrete matching control code set to obtain the target matching control code. When the local scoring function does not open upwards, the matching control code corresponding to the static trial control state with the smallest scoring result is determined as the target matching control code.

[0014] Optionally, the target power amplifier bias control code is determined based on the global mismatch index and the spike ratio index corresponding to the target matching control code, specifically including: Under the target matching control code, the on-chip complex reflectance spectrum sampling data is reacquired, and the global mismatch index and spike ratio index corresponding to the target matching control code are determined based on the reacquired on-chip complex reflectance spectrum sampling data. The effective mismatch index is determined based on the global mismatch index and spike ratio index corresponding to the target matching control code, and the target power amplifier bias control code is determined in combination with the reference bias control data.

[0015] Optionally, the output is a target control vector used to control the RF processing circuitry, specifically including: The target matching control code and the target power amplifier bias control code are combined to form the target control vector; The target control vector is written into the control latch region so that the target control vector remains stable in subsequent launch control cycles; Based on the target control vector, matching network control data and power amplifier bias control data are generated respectively, and output to the corresponding RF processing circuit control interface to complete the control of the RF processing circuit.

[0016] Furthermore, to achieve the above objectives, the present invention also provides a radio frequency processing circuit control system, comprising: The acquisition module is used to acquire on-chip complex reflection spectrum sampling data under the current control state, and construct a complex reflection spectrum slice matrix based on the on-chip complex reflection spectrum sampling data; The separation module is used to extract local structural features in the frequency domain based on the complex reflection spectrum slice matrix, and to separate the complex reflection spectrum slice matrix into broadband mismatched baseline data and local spike perturbation data. The determination module is used to determine a global mismatch index that characterizes the true degree of broadband mismatch and a baseline flatness index that characterizes the degree of broadband profile change based on the broadband mismatch baseline data, and to determine a spike proportion index that characterizes the degree of local anomaly based on the local spike perturbation data. The construction module is used to construct multiple static trial control states around the current matching control code, determine the scoring results based on the global mismatch index, the baseline flatness index and the spike ratio index corresponding to each static trial control state, and then determine the target matching control code based on the scoring results. The output module is used to determine the target power amplifier bias control code based on the global mismatch index and the spike ratio index corresponding to the target matching control code, and output the target control vector for controlling the radio frequency processing circuit.

[0017] The beneficial effects of this invention are as follows: It proposes a radio frequency processing circuit control method and system. By constructing multiple micro-offset detection points around the working center frequency point within the same control cycle and simultaneously acquiring on-chip complex reflection spectrum sampling data on multiple time slices, a complex reflection spectrum slice matrix capable of characterizing local spectral structure changes is formed. This allows for analysis of local frequency domain structural features such as the degree of local bending and the degree of continuous phase change. It can distinguish between broadband components that continuously change along the frequency direction in the reflection data and spike components that appear concentrated on a small number of frequency points and time slices. Based on the broadband mismatch results, a global mismatch index and baseline flatness are generated. The system generates a spike ratio index based on local spike results, making subsequent control decisions closer to the actual broadband mismatch state, rather than being directly driven by local liquid bridge anomalies. Multiple static trial control states are constructed around the current matching control code, and the same data processing and result evaluation are performed on each static trial control state. The target matching control code is then determined based on the scoring results, and the target power amplifier bias control code is determined based on the global mismatch index and spike ratio index corresponding to the target matching control code. This reduces the sensitivity of the power amplifier bias control to local spike anomalies, thereby maintaining better transmission capability and control stability while ensuring the safety of the power amplifier. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the method of an embodiment of the present invention; Figure 2 This is a schematic diagram of the system structure according to an embodiment of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0020] This invention provides a method and system for controlling a radio frequency (RF) processing circuit. The method can be executed by a control processing unit integrated within the RF chip, or by the RF chip in conjunction with an external digital baseband processing unit or a dedicated controller. It is readily understood that this invention does not limit the control logic to be entirely executed by a single processor; any processing steps that can achieve complex reflection spectrum sampling, structural weight generation, broadband mismatch and local spike separation, scoring decision, and control vector output (described later) are considered to fall within the scope of this invention.

[0021] In one executable implementation, the hardware executing the method of the present invention includes at least a transmit local oscillator control unit, a micro-offset frequency detection control unit, a directional coupling detection path, an I / Q demodulation sampling unit, a control processing unit, a matching network driving unit, a power amplifier bias control unit, and a control result latching unit. The transmit local oscillator control unit generates multiple micro-offset frequency points around the current operating frequency; the micro-offset frequency detection control unit triggers short-time detection transmission at each micro-offset frequency point according to a preset timing sequence; the directional coupling detection path extracts reflected signals from the power amplifier output, relevant nodes of the matching network, or other locations that can characterize the reflection state; the I / Q demodulation sampling unit demodulates the reflected signal into in-phase and quadrature components; the control processing unit performs the data calculations, objective function solution, scoring, and control decisions described later; the matching network driving unit loads the target matching code; the power amplifier bias control unit loads the target power amplifier bias code; and the control result latching unit temporarily stores the target control vector and related intermediate indicators generated in the current control cycle for use in the next control cycle. Furthermore, in one executable implementation, the transmit local oscillator control unit is connected to the micro-biased frequency detection control unit, the micro-biased frequency detection control unit is connected to the transmit link, the directional coupling detection path is connected to the I / Q demodulation sampling unit, the I / Q demodulation sampling unit is connected to the control processing unit, and the control processing unit is connected to the matching network drive unit, the power amplifier bias control unit, and the control result latching unit, thereby forming a complete data processing and execution link from the acquisition of reflected signals to the output of control vectors.

[0022] like Figure 1 As shown in the embodiment of the present invention, a radio frequency processing circuit control method includes the following steps: S1: Obtain the on-chip complex reflection spectrum sampling data under the current control state, and construct a complex reflection spectrum slice matrix based on the on-chip complex reflection spectrum sampling data.

[0023] Specifically, multiple micro-frequency offset detection points are determined on both sides of the current working center frequency, and on-chip reflection detection data are collected for each micro-frequency offset detection point on multiple time slices within the same control cycle. For the on-chip complex reflection spectrum sampling values ​​corresponding to each time slice and each micro-frequency offset detection point, a complex reflection spectrum slice matrix composed of all complex reflection spectrum sampling values ​​is constructed. The amplitude data and phase data at each time slice and each micro-frequency offset detection point are obtained based on the in-phase component sampling values ​​and the quadrature component sampling values, respectively, for subsequent extraction of local structural features in the frequency domain.

[0024] It should be noted that existing technologies typically acquire a single reflection value at the operating frequency and use this single reflection value directly for control determination. However, this invention does not adopt this approach. Instead, under the current control state, it performs local slice sampling of the reflection spectrum in both the frequency and time dimensions to form a data basis for subsequent differentiation between broadband mismatch and local spike anomalies.

[0025] In practical applications, the control processing unit first reads the current control vector latched in the current control cycle or the previous control cycle. The expression can be: ; in, Indicates the current control vector. Indicates the current matching code. This represents the current amplifier bias code, where T represents the transpose. The current matching code is easy to understand. and the current power amplifier bias code It is not an isolated control parameter, but rather the starting point for subsequent micro-frequency detection and scoring comparison.

[0026] After obtaining the current control vector, the transmit local oscillator control unit rotates around the operating frequency. To generate multiple micro-offset frequency points, the expression can be: ; in, This represents the nth micro-offset frequency point. This represents the nth micro-offset frequency, where N represents the number of micro-offset frequency points. In practical applications, The micro-offset frequency points can be set symmetrically around the operating frequency point. For example, several equally spaced or unequally spaced micro-offset frequency points can be set on both sides of the operating frequency point to allow for more complete observation of the reflection structure within the local frequency band. It should be further noted that the number and spacing of the micro-offset frequency points are not limited to a fixed value. They can be set according to the operating frequency band width, chip computing resources, control cycle length, and expected anomaly resolution, as long as a set of discrete sampling points on the local frequency band can be formed.

[0027] Subsequently, in the current control vector Under the corresponding control state, complex reflection spectrum sampling values ​​are obtained at multiple time slices and multiple micro-offset frequency points, and the expression can be: ; in, This represents the complex reflection spectrum sample value corresponding to the m-th time slice and the n-th micro-offset frequency point. Indicates in-phase components, Let represent the orthogonal components, j represent the imaginary unit, and m represent the time slice number. In practical applications, the specific calculation and acquisition method here can be as follows: Within each time slice, the directional coupling detection path extracts the reflected signal from the transmission link, and the I / Q demodulation sampling unit obtains the in-phase component and the orthogonal component of the reflected signal, respectively. Then, the two are combined into a complex reflection spectrum sample value in complex form. That is, each complex reflection spectrum sample value corresponds to a local reflection characterization result at a time position and a frequency position.

[0028] After sampling all time slices and all micro-offset frequency points, a complex reflection spectrum slice matrix is ​​constructed based on all complex reflection spectrum sample values. The expression can be: ; Here, X represents the complex reflection spectrum slice matrix. Specifically, each row of matrix X corresponds to multiple sampled values ​​of a micro-offset frequency point on a time slice, and each column corresponds to the sampled values ​​of a micro-offset frequency point on multiple time slices. Thus, matrix X simultaneously retains frequency variation information within a local frequency band and time variation information within a short time window.

[0029] It should be noted that when the RF chip is in a high-humidity condensation or local liquid bridge formation environment, the original complex reflection spectrum slice matrix usually shows a sudden increase in amplitude at a few slightly off-frequency points. Furthermore, these sudden increases may drift slightly between adjacent time slices. In existing technologies, observing only the reflection results at a single point on the operating frequency often leads to misinterpretation of this phenomenon as overall mismatch deterioration. This invention, by establishing a complex reflection spectrum slice matrix, provides the necessary data for subsequently distinguishing between broadband mismatch changes and local spike anomalies.

[0030] S2: Extract local structural features in the frequency domain based on the complex reflection spectrum slice matrix, and separate the complex reflection spectrum slice matrix into broadband mismatched baseline data and local spike perturbation data.

[0031] Specifically, based on the amplitude data corresponding to each position in the complex reflection spectrum slice matrix, the degree of amplitude variation and local bending between adjacent micro-frequency offset detection points are determined along the micro-frequency offset direction; based on the phase data corresponding to each position in the complex reflection spectrum slice matrix, the degree of continuous phase variation between adjacent micro-frequency offset detection points is determined along the micro-frequency offset direction; structural weight data corresponding to each position is generated based on the degree of amplitude variation, the degree of local bending, and the degree of continuous phase variation; the influence of local frequency point abrupt changes on the broadband continuous variation result is reduced based on the structural weight data, and the part of the complex reflection spectrum slice matrix that continuously varies along the micro-frequency offset direction is determined as the broadband mismatch baseline data, and the abrupt changes that occur concentrated in a small number of frequency points and a small number of time slices in the complex reflection spectrum slice matrix are determined as the local spike perturbation data.

[0032] Furthermore, structural weight data corresponding to each position is generated based on the magnitude of the amplitude change, the degree of local bending, and the degree of phase continuity change. Specifically, this includes: determining the structural weight value corresponding to each position based on the degree of local bending and the degree of phase continuity change; and constraining the data participation degree of each position in the complex reflection spectrum slice matrix based on the structural weight value, so that the proportion of positions with a large degree of local bending and a large degree of phase continuity change is reduced in the process of generating broadband mismatched baseline data.

[0033] In this embodiment of the invention, step S2 is used to extract structured information from the original complex reflection spectrum slice matrix obtained in step S1, and to separate the broadband continuous variation part and the local spike anomaly part in the original reflection data. It is easy to understand that this invention does not directly perform mean processing or threshold judgment on the original complex reflection spectrum slice matrix, but rather first distinguishes the variation components of different physical properties, and then generates control criteria accordingly. This is fundamentally different from existing technologies.

[0034] Specifically, the control processing unit first calculates the sampled values ​​of each complex reflection spectrum. Calculate the amplitude and phase at the corresponding position. The calculation methods can be as follows: ; ; in, This represents the reflection amplitude corresponding to the m-th time slice and the n-th micro-offset frequency point. Indicates the corresponding reflection phase. This represents the arctangent function in the four quadrants. The calculation process here can be understood as follows: using the in-phase and quadrature components to determine the magnitude and angle of the complex reflection vector, thereby obtaining the amplitude distribution and phase distribution. It should be noted that the amplitude distribution mainly reflects the local reflection intensity, while the phase distribution mainly reflects the directional changes of the local reflection vector. For narrowband anomalies caused by local liquid bridges, water film boundaries, or local electromagnetic disturbances, observing only the amplitude magnitude is usually insufficient to reliably distinguish it from the true broadband mismatch. However, the continuous change pattern of the phase at adjacent micro-offset frequency points can provide additional evidence for this distinction. Therefore, this step retains and calculates both amplitude and phase data simultaneously.

[0035] After obtaining the amplitude and phase distributions, the local spectral curvature and phase spread are further calculated along the direction of the micro-offset frequency points. For a position located inside the matrix, the local spectral curvature can be determined by the amplitude relationship between adjacent micro-offset frequency points. For example, for the nth micro-offset frequency point in the mth time slice, its local spectral curvature... The phase spread can be calculated by comparing the amplitude at this point with the amplitudes of adjacent micro-frequency offset points. Specifically, it compares whether the current point shows a significant bulge, a significant dip, or a smooth transition with its surroundings relative to the points on either side. If the amplitude at the current position shows a significant bulge or dip relative to the points on either side, it indicates that this position is more likely to be a local spike anomaly, with a correspondingly large local spectral curvature. This is used to characterize whether the phase change between adjacent micro-offset frequency points is smooth. The calculation idea is to expand and reverse the phase difference between adjacent micro-offset frequency points to keep the phase difference within a reasonable angle range, and then use the absolute value of the phase difference to characterize the phase continuity between the current point and the adjacent points. If a certain position shows a large phase reversal compared with the adjacent positions, it indicates that there is a local phase discontinuity feature at that position, which is more likely to correspond to local spike anomalies rather than broadband continuous variation components.

[0036] Based on this, the weight values ​​in the structure weight matrix are generated according to the local spectral curvature and phase spread, and the expression can be: ; in, This represents the structural weight corresponding to the m-th time slice and the n-th micro-offset frequency point. This represents the local spectral curvature at that location. This indicates the phase spread at that position. and This represents the adjustment coefficient. The specific meaning of this formula is: when the local spectral curvature and phase spread at a certain location are both small, it indicates that the frequency change at that location is relatively gentle and the phase change is relatively continuous, making it more likely to belong to a broadband continuous variation component, thus its structural weight is larger; when the local spectral curvature or phase spread at a certain location increases significantly, it indicates that the location is more likely to belong to a local spike anomaly, thus its structural weight decreases. In other words, the structural weight does not simply represent the magnitude of the original reflection value, but rather the degree of reliability of that location in the broadband mismatch extraction process.

[0037] Once the structural weight values ​​for all positions have been calculated, a structural weight matrix can be constructed based on these values. The expression can be: ; Here, W represents the structure weight matrix. It should be noted that the structure weight matrix will participate in the subsequent complex reflection spectrum decomposition process as a position-related constraint. Its role is not to directly replace the original data, but rather to control the degree of influence of data at different locations on the broadband mismatch baseline extraction results.

[0038] After generating the structural weight matrix, the complex reflection spectrum slice matrix is ​​further decomposed into a broadband mismatched baseline matrix and a spike perturbation matrix based on the structural weight matrix. Specifically, let the original complex reflection spectrum slice matrix be X, the broadband mismatched baseline matrix be L, and the spike perturbation matrix be E, and establish the decomposition objective function, the expression of which can be: ; in, The objective function for decomposition is denoted as: The objective function for decomposition can be: ; Here, ⊙ represents element-wise multiplication. Describing the Frobenius norm, This represents the second-order difference result of the broadband mismatch baseline matrix along the frequency direction. This represents the first-order difference result of the broadband mismatch baseline matrix along the time direction. This represents the phase continuity constraint term. This represents the sum of the absolute values ​​of the elements of the spike perturbation matrix. to This represents the adjustment coefficient.

[0039] The specific execution and calculation logic of the decomposition objective function are as follows: The first term constrains the superposition result of the broadband mismatch baseline matrix and the spike perturbation matrix to reconstruct the original complex reflection spectrum slice matrix as much as possible, while ensuring higher reconstruction accuracy at high-confidence positions and less traction on reconstruction at low-confidence positions. The second term constrains the broadband mismatch baseline matrix to remain smooth in the frequency direction, avoiding excessive bending of the broadband baseline between adjacent micro-offset frequency points. The second-order difference here can be understood as: calculating the difference between the baseline values ​​of three adjacent micro-offset frequency points in the same time slice to reflect the magnitude of the baseline curvature at that position. The larger the second-order difference, the more obvious the bending at that position, and therefore it will be suppressed in the optimization. The third term constrains the broadband mismatch baseline matrix to remain continuous in the time direction, avoiding unreasonable jumps in the broadband baseline between adjacent time slices. Its calculation method can be understood as accumulating the difference between the baseline values ​​of the same micro-offset frequency point in adjacent time slices. The fourth term constrains the phase change of the broadband mismatch baseline matrix to be continuous, preventing mathematically feasible but physically unreasonable abrupt changes in the phase dimension of the solution result. The fifth term is used to make the spike perturbation matrix as sparse as possible, thereby causing local anomalies to be concentrated in a few locations rather than spreading to a wide range of locations.

[0040] After solving the objective function described above, the broadband mismatch baseline matrix L and the spike perturbation matrix E are obtained. The broadband mismatch baseline matrix L mainly represents the relatively continuous and smooth broadband mismatch components in both the frequency and time directions of the complex reflection spectrum; the spike perturbation matrix E mainly represents the local burst anomaly components that appear only at a few micro-offset frequency points and a few time slices. Thus, step S2 completes the transformation from the original complex reflection spectrum slice matrix to the broadband mismatch and local spike separation results, and provides input data for subsequent index extraction.

[0041] S3: Based on the broadband mismatch baseline data, determine the global mismatch index that characterizes the true degree of broadband mismatch and the baseline flatness index that characterizes the degree of broadband profile change, and based on the local spike perturbation data, determine the spike proportion index that characterizes the degree of local anomaly proportion.

[0042] Specifically, the slice weights of each time slice are determined based on the local spike perturbation data, and the broadband mismatch baseline data are weighted and aggregated to obtain a broadband mismatch baseline vector; the global mismatch index and the baseline flatness index are determined based on the broadband mismatch baseline vector, and the spike proportion index is determined based on the local spike perturbation data and the on-chip complex reflectance spectrum sampling data.

[0043] In this embodiment of the invention, step S3 is used to further transform the matrix result obtained in step S2 into vectors and scalar indices that can be directly involved in control decisions. It should be noted that although step S2 has completed the separation of the broadband mismatch baseline and the local spike anomaly, the broadband mismatch baseline matrix and the spike disturbance matrix are still two-dimensional data. If they are directly used for matching code selection and power amplifier bias control, the computational and execution complexity is high. Therefore, this step further transforms them into core indices that can stably characterize the current control state.

[0044] Specifically, the slice weights for each time slice are first determined based on the spike perturbation matrix. This can be understood as follows: for each time slice, the spike perturbation amplitudes corresponding to all micro-offset frequency points within that time slice are accumulated or weighted, and the overall strength of local spike anomalies in that time slice is determined accordingly. If the overall spike perturbation in a certain time slice is strong, it indicates that the time slice is significantly affected by local anomalies, and its influence should be reduced when generating the broadband mismatch baseline vector. Conversely, if the overall spike perturbation in a certain time slice is weak, it indicates that the time slice better represents a relatively stable broadband baseline state, and its influence should be increased. In one feasible implementation, the spike intensity corresponding to each time slice can be calculated first, and then the spike intensity of each time slice can be processed through reciprocal transformation and normalization to obtain the slice weights. The essence of this calculation method is to improve the stability of broadband baseline aggregation results by assigning greater weight to time slices with weaker spike perturbations.

[0045] After obtaining the slice weights, the broadband mismatch baseline matrix is ​​aggregated according to the slice weights to obtain the broadband mismatch baseline vector, which can be expressed as: ; in, This represents the element value of the broadband mismatch baseline vector at the nth micro-offset frequency point. This represents the slice weight corresponding to the m-th time slice. This represents the corresponding element in the broadband mismatch baseline matrix. For the same micro-offset frequency point, the broadband baseline values ​​on each time slice are weighted and averaged according to the slice weights, thereby compressing the two-dimensional broadband mismatch baseline matrix into a one-dimensional broadband mismatch baseline vector along the time direction. It should be noted that this broadband mismatch baseline vector retains the broadband mismatch distribution results corresponding to each micro-offset frequency point in the local frequency band, and at the same time, the influence of time slices with large spike interference is suppressed by the slice weights. Therefore, it can represent the true broadband matching state better than directly averaging the original reflection data.

[0046] Meanwhile, the spike proportion index is determined based on the spike perturbation matrix and the complex reflection spectrum slice matrix, and the expression can be: ; Where C represents the percentage of spikes. This represents the corresponding element in the spike perturbation matrix. This represents the corresponding element in the complex reflection spectrum slice matrix. This indicates a positive number to prevent the denominator from being zero. The calculation method for this index is to first count the total amount of local spike anomalies in all complex reflectance spectral data, and then calculate the ratio with the total amount of the original complex reflectance spectrum to obtain the proportion of local spike components in the current overall reflectance structure. If C is large, it indicates that a large proportion of the current reflectance anomalies originates from local spike components; if C is small, it indicates that the current reflectance anomalies originate more from broadband mismatch components.

[0047] After obtaining the broadband mismatch baseline vector, a global mismatch index and a baseline flatness index are further generated based on this vector. The global mismatch index is calculated by statistically summing the amplitudes of the elements corresponding to each micro-offset frequency point in the broadband mismatch baseline vector. This can be done using absolute value summation, weighted absolute value summation, or the mean of frequency point weights, resulting in an index characterizing the overall strength of the current broadband mismatch. The baseline flatness index is calculated as follows: the differences between adjacent micro-offset frequency points in the broadband mismatch baseline vector are statistically summarized. If there are large fluctuations between adjacent micro-offset frequency points, it indicates that the broadband baseline still has significant unevenness within the local frequency band, and the baseline flatness index increases accordingly. If the transition between adjacent micro-offset frequency points is smooth, it indicates that the broadband baseline within the local frequency band is relatively flat, and the baseline flatness index is small. It should be noted that the global mismatch index reflects the overall strength, while the baseline flatness index reflects whether the frequency band shape is flat. The two are not redundant but together form the basis for evaluating the matching status.

[0048] After generating the above indicators, the control processing unit focuses on the current matching code. Multiple trial matching codes are generated. Specifically, the current matching code from the previous control cycle or the current control cycle can be used as the center, and one or more adjacent valid matching codes can be taken forward and backward to form multiple discrete trial points. Then, each trial matching code is combined with the current power amplifier bias code to form a trial control state. For each trial control state, steps S1 and S2, as well as the aforementioned index generation process, are repeated to obtain the corresponding global mismatch index. Baseline straightness index and the proportion of spikes .

[0049] After obtaining the indices corresponding to each trial control state, the score value corresponding to each trial control state is determined according to the following formula, which can be expressed as: ; in, This represents the score value corresponding to the i-th trial control state. This represents the global mismatch index corresponding to the i-th trial control state. This represents the baseline flatness index corresponding to the i-th trial control state. This represents the spike percentage index corresponding to the i-th trial control state. This represents the scoring weighting coefficient. The specific meaning of this scoring formula is that the quality of the current matching code is not determined by a single reflection intensity, but rather by the overall intensity of broadband mismatch, the flatness of the broadband baseline, and the degree of local spike anomalies. By weighting and summing these three indicators into a single score, the overall performance of different trial matching codes can be compared in a unified manner. It should be noted that the smaller the score value, the more effective the corresponding trial matching code is in reducing broadband mismatch, maintaining baseline flatness, and suppressing the impact of local spikes in the current scenario.

[0050] S4: Construct multiple static trial control states around the current matching control code, determine the scoring results based on the global mismatch index, the baseline flatness index, and the spike ratio index corresponding to each static trial control state, and then determine the target matching control code based on the scoring results.

[0051] Specifically, a low-side static test control state, a current static test control state, and a high-side static test control state are constructed based on the current matching control code. On-chip complex reflectance spectrum sampling data are reacquired under each static test control state, and the generation and processing of the broadband mismatch baseline data and local spike perturbation data, as well as the determination and processing of the global mismatch index, the baseline flatness index, and the spike proportion index, are repeatedly executed. The scoring results of each static test control state are determined based on the global mismatch index, the baseline flatness index, and the spike proportion index corresponding to each static test control state.

[0052] Subsequently, the scoring results corresponding to the low-side static trial control state, the current static trial control state, and the high-side static trial control state are used as local scoring sampling points to construct a local scoring relationship regarding the matching control code. The local scoring function parameters are determined based on the scoring results corresponding to the low-side static trial control state, the current static trial control state, and the high-side static trial control state. When the local scoring function opens upwards, continuous optimal matching control values ​​are determined based on the local scoring function, and these continuous optimal matching control values ​​are mapped to a preset discrete matching control code set to obtain the target matching control code. When the local scoring function does not open upwards, the matching control code corresponding to the static trial control state with the smallest scoring result is determined as the target matching control code.

[0053] In this embodiment of the invention, step S4 is used to determine the target matching code for the current control cycle based on the scoring results of multiple trial matching codes, and to generate the effective mismatch index required for subsequent power amplifier bias control. It is easy to understand that this step does not simply select the code with the lowest score from multiple trial matching codes, but rather prioritizes analyzing the local trend of score changes within the neighborhood of the current matching code, so as to make the adjustment of the target matching code smoother and more stable.

[0054] Specifically, a local scoring function is constructed based on multiple trial matching codes and their corresponding score values: ; in, This represents a local scoring function. This represents the fitting parameters. The specific calculation method for this function is as follows: Using several discrete trial matching codes near the current matching code as the x-axis and the corresponding score values ​​as the y-axis, a quadratic function fitting method is used to obtain the coefficients that characterize the local score trend. In one feasible implementation, the parameters of a quadratic function can be solved using a least-squares fitting method based on several discrete trial points. The essence of this fitting process is to transform the discrete scoring results into a local continuous trend, thereby enabling not only the comparison of existing trial points but also the estimation of potentially better matching positions between trial points.

[0055] Once the local scoring function is fitted, the continuous target matching value is determined based on the local scoring function. The expression can be: ; in, This represents the matching value of continuous targets. This calculation method originates from determining the vertex position of a quadratic function. When the quadratic function opens upwards, it indicates the existence of a local low score point near the current matching code. Therefore, the target matching value in a continuous sense can be obtained by calculating the vertex position. Then, the continuous target matching value is mapped to the set of valid matching codes to obtain the target matching code. Its specific execution logic can be understood as: if... If it is not a valid matching code that can be directly loaded, then select with The discrete legal matching code that is closest to and satisfies the matching network coding rules is used as the target matching code. If the local scoring function does not meet the preset conditions, for example... Alternatively, if the fitting residual is too large, it indicates that the scoring trend in the current neighborhood is not stable enough. In this case, instead of using the continuous vertex method, the trial matching code with the smallest score value is directly selected as the target matching code. In this way, the target matching code can utilize the discrete trial scoring results and achieve a smoother target selection when the local trend is clear.

[0056] After the target matching code is determined, it is combined with the current power amplifier bias code to form the target matching control state, and the corresponding global mismatch index is obtained based on this control state. and the proportion of spikes Subsequently, the effective mismatch index is determined according to the following formula: ; Where R represents the effective mismatch index, This represents the spike proportion suppression coefficient. In this formula, if the global mismatch index is high under the target matching state, it indicates that there is still a significant broadband mismatch trend, and R should be increased accordingly; however, if the spike proportion index is also high under the target matching state, it indicates that there is still a large proportion of local spike components in the current anomaly. In this case, if we directly rely on... Adjusting the power amplifier bias can easily lead to an overreaction to local anomalies, therefore, through the denominator... Suppression is performed to make the effective mismatch index more biased towards reflecting the true degree of broadband mismatch rather than the degree of local spikes. In other words, the effective mismatch index R is a control quantity obtained by introducing local spike correction on the basis of broadband mismatch intensity, and it will be directly used as the basis for determining the target power amplifier bias code in step S5.

[0057] S5: Determine the target power amplifier bias control code based on the global mismatch index and the spike ratio index corresponding to the target matching control code, and output the target control vector for controlling the radio frequency processing circuit.

[0058] Specifically, under the target matching control code, on-chip complex reflection spectrum sampling data is reacquired, and the global mismatch index and spike ratio index corresponding to the target matching control code are determined based on the reacquired on-chip complex reflection spectrum sampling data; the effective mismatch index is determined according to the global mismatch index and spike ratio index corresponding to the target matching control code, and the target power amplifier bias control code is determined in combination with the reference bias control data.

[0059] Subsequently, the target matching control code and the target power amplifier bias control code are combined to form a target control vector; the target control vector is written into the control latch area so that the target control vector remains stable in subsequent transmit control cycles; matching network control data and power amplifier bias control data are generated according to the target control vector and output to the corresponding RF processing circuit control interface to complete the control of the RF processing circuit.

[0060] In this embodiment of the invention, step S5 is used to determine the target power amplifier bias code and output the control vector based on the obtained target matching code and effective mismatch index. It should be noted that in the prior art, the power amplifier bias is often directly backed up based on the single-point reflection value, while in this invention, the power amplifier bias control is not based on single-point anomalies, but on the effective mismatch index after the influence of local spikes has been removed, thus enabling a more realistic response to broadband mismatch changes.

[0061] Specifically, the target power amplifier bias code is determined based on the effective mismatch index, and the expression can be: ; Where b represents the target power amplifier bias code. This represents a mapping function to the set of valid power amplifier bias codes. This indicates the reference power amplifier bias code. This represents the bias reduction factor. This represents the reference mismatch threshold. The formula first compares the effective mismatch index R with the reference mismatch threshold. The size; if R is not higher than This indicates that the current broadband mismatch is still within an acceptable range. Setting R to 0 keeps the target power amplifier bias code near the reference power amplifier bias code; if R is higher than 0, the target power amplifier bias code remains near the reference power amplifier bias code. Then, depending on the degree of excess With bias reduction factor The product of these factors yields the offset level to be reduced, which is then rounded down to obtain the discrete offset adjustment steps. These steps are then reduced accordingly based on the reference power amplifier offset code. Finally, a mapping function is used... The calculation results are mapped to a set of actually available valid power amplifier bias codes. Thus, the target power amplifier bias code not only has an explicit implementation of mapping continuous control quantities to discrete codes, but also reduces excessive bias backoff caused by local spike anomalies because the input quantities use effective mismatch indices instead of original outliers.

[0062] After obtaining the target power amplifier bias code, the target control vector is constructed based on the target matching code and the target power amplifier bias code. The expression can be: ; Where u represents the target control vector, t represents the target matching code, and b represents the target power amplifier bias code. Subsequently, the target control vector is associated and stored with the broadband mismatch baseline vector, spike ratio index, global mismatch index, baseline flatness index, and effective mismatch index to generate the control result data for the current control cycle. This associated storage not only saves the final output matching code and bias code but also the main intermediate indices corresponding to these control results, so that the results of the previous cycle can be directly called as the new current control state and scoring reference benchmark in the next control cycle. Furthermore, in one executable implementation, the control result latching unit can use the target control vector u output in the current control cycle as the current control vector for the next control cycle. .

[0063] It should be further noted that this invention does not limit the target control vector to be loaded immediately after calculation. In practical applications, the target matching code can be written to the matching network driver unit and the target power amplifier bias code can be written to the power amplifier bias control unit at a single control clock edge. Alternatively, they can be temporarily stored in a register and then synchronously activated at an appropriate time according to the system's transmission timing. As long as the target matching code and the target power amplifier bias code can be guaranteed to correspond to the corresponding effective mismatch index, global mismatch index, and spike ratio index within the same control cycle, they should be considered as implementations conforming to this invention.

[0064] To further illustrate the implementation process of the technical solution of the present invention, a specific embodiment is described below. It should be noted that the following embodiment is only used to illustrate the technical concept of the present invention and does not constitute a limitation on the scope of protection of the present invention.

[0065] In one executable implementation, the RF chip operates at a target transmission frequency point and already possesses the current matching code output and latched from the previous control cycle at the start of the current control cycle. and the current power amplifier bias code When the control processing unit detects the need to reassess the current transmission status, it first controls the transmission local oscillator control unit to generate multiple micro-offset frequency points around the operating frequency, and then collects the reflection signals corresponding to these micro-offset frequency points in multiple consecutive time slices. At this time, the directional coupling detection path extracts the reflection signal from the transmission link, and the I / Q demodulation sampling unit outputs the corresponding in-phase and quadrature components. The control processing unit combines the in-phase and quadrature components corresponding to each time slice and each micro-offset frequency point into a complex reflection spectrum sampling value, thereby forming a complex reflection spectrum slice matrix. Thus, the single-point reflection detection in the prior art is extended to two-dimensional joint sampling of local frequency band and local time.

[0066] After obtaining the complex reflection spectrum slice matrix, the control processing unit calculates the amplitude and phase for each complex reflection spectrum sample value. Specifically, for each sampling position, the squares of the in-phase component and the quadrature component are summed, and the square root is taken to obtain the reflection amplitude; simultaneously, the reflection phase is obtained using the four-quadrant angular relationship between the in-phase and quadrature components. Subsequently, the control processing unit compares the amplitude and phase changes of adjacent positions along the micro-offset frequency direction to obtain the local spectral curvature and phase spread. If a position exhibits a significant amplitude spike relative to the two micro-offset frequency points on both sides, and a large phase shift relative to adjacent micro-offset frequency points, the control processing unit determines that this position is structurally closer to the local spike anomaly position and assigns a smaller weight in the subsequent structural weight calculation; if the amplitude change at a position is relatively gentle and the phase change is relatively continuous, a larger weight is assigned. In this way, the control processing unit completes the conversion from the original complex reflection spectrum data to the structural weight matrix.

[0067] After the structural weight matrix is ​​generated, the control processing unit further performs decomposition operations on the broadband mismatch baseline and spike perturbation. Specifically, the original complex reflection spectrum slice matrix is ​​represented as the sum of the broadband mismatch baseline matrix and the spike perturbation matrix, and solved using an objective function that includes position-related weight constraints, frequency-direction smoothing constraints, time-direction continuity constraints, phase continuity constraints, and spike sparsity constraints. During this solution process, if a position has a higher weight, its fit to the broadband baseline is stronger; if a position has a lower weight, its anomalous changes are more easily assigned to the spike perturbation matrix. Simultaneously, the second-order difference constraint in the frequency direction keeps the broadband baseline smooth between micro-offset frequencies, the first-order difference constraint in the time direction keeps the broadband baseline continuous between adjacent time slices, the phase continuity constraint ensures the broadband baseline conforms to physically continuous changes in the complex phase dimension, and the spike sparsity constraint concentrates local anomalies in a few locations as much as possible. Through the combined effect of these constraints, the control processing unit obtains the broadband mismatch baseline matrix and the spike perturbation matrix.

[0068] After obtaining the decomposition results, the control processing unit further determines the slice weights based on the overall perturbation intensity of each time slice in the spike perturbation matrix. For time slices with a large total spike perturbation, the slice weights are reduced; for time slices with a small total spike perturbation, the slice weights are increased. Then, the control processing unit performs weighted aggregation of the broadband mismatch baseline matrix along the time direction according to the slice weights, thereby obtaining the broadband mismatch baseline vector. Simultaneously, the control processing unit obtains the spike proportion index based on the ratio between the total spike perturbation and the total original complex reflection spectrum, obtains the global mismatch index based on the overall magnitude relationship of each micro-offset frequency element in the broadband mismatch baseline vector, and obtains the baseline flatness index based on the difference relationship between adjacent micro-offset frequency points. Thus, the two-dimensional matrix result is further compressed into several key indicators that can directly participate in control decisions.

[0069] After generating the aforementioned key indicators, the control processing unit generates multiple trial matching codes around the current matching code, and combines these trial matching codes with the current power amplifier bias code to form multiple trial control states. For each trial control state, the control processing unit re-executes the complex reflection spectrum sampling, structural weight generation, baseline decomposition, and indicator generation processes to obtain the corresponding global mismatch indicator, baseline flatness indicator, and spike ratio indicator. Then, the control processing unit performs a comprehensive score on each trial control state using a scoring formula. Since the scoring formula considers broadband mismatch intensity, broadband baseline flatness, and local spike severity simultaneously, the resulting score is not a simple ranking result of a single abnormal amplitude, but rather an evaluation result of the comprehensive adaptability of different control states in the neighborhood of the current matching code.

[0070] After multiple trial control state scores are completed, the control processing unit constructs a local scoring function using these discrete score values. If the local scoring function indicates a clear local low score point near the current matching code, a continuous target matching value is calculated and mapped to a valid target matching code. If the local scoring function does not meet preset conditions, the code with the smallest score value from the multiple trial matching codes is directly selected as the target matching code. In this way, the present invention avoids both the instability caused by directly switching the matching code based on the result of only a single trial point and the problem of overly coarse matching adjustment granularity that may result from simply selecting the smallest discrete point.

[0071] After the target matching code is determined, the control processing unit continues to acquire the corresponding global mismatch index and spike percentage index under the target matching state, and performs spike suppression correction on the global mismatch index using the effective mismatch index formula. If the global mismatch index increases slightly under the target matching state, but the spike percentage index is large, the effective mismatch index will be reduced due to the spike suppression term in the denominator, thus avoiding excessive back-off of the power amplifier bias. If the global mismatch index increases significantly under the target matching state and the spike percentage index is small, it indicates that the current anomaly mainly originates from the actual bandwidth mismatch. In this case, the effective mismatch index will remain at a high level, and the subsequent power amplifier bias back-off will be more significant.

[0072] Finally, the control processing unit determines the target power amplifier bias code based on the relationship between the effective mismatch index and the reference mismatch threshold, and combines the target matching code and the target power amplifier bias code to form the target control vector. This target control vector is sent to the matching network driver unit and the power amplifier bias control unit for execution, and is also written into the control result latch unit along with the broadband mismatch baseline vector, spike ratio index, global mismatch index, baseline flatness index, and effective mismatch index generated in the current control cycle, serving as the current control basis for the next control cycle.

[0073] In this specific implementation, the complex reflection spectrum sampling values ​​acquired in the current control cycle first form a complex reflection spectrum slice matrix. The complex reflection spectrum slice matrix is ​​further used to calculate the amplitude distribution and phase distribution. The amplitude distribution and phase distribution are further used to generate local spectral curvature and phase spread. The local spectral curvature and phase spread are further used to generate a structure weight matrix. The structure weight matrix is ​​further used to decompose and obtain a broadband mismatch baseline matrix and a spike disturbance matrix. The broadband mismatch baseline matrix and spike disturbance matrix are further used to generate a broadband mismatch baseline vector, a spike proportion index, a global mismatch index, and a baseline flatness index. Multiple indices are further used to form multiple test control states corresponding to multiple score values. Multiple score values ​​are further used to determine the target matching code. The target matching code is further used with the corresponding global mismatch index and spike proportion index to generate an effective mismatch index. The effective mismatch index is further used to determine the target power amplifier bias code, and finally, the target control vector is output.

[0074] Reference Figure 2 , Figure 2 This is a schematic diagram of the radio frequency processing circuit control system according to an embodiment of the present invention.

[0075] like Figure 2 As shown, the radio frequency processing circuit control system proposed in this embodiment of the invention includes: The acquisition module 10 is used to acquire on-chip complex reflection spectrum sampling data under the current control state, and construct a complex reflection spectrum slice matrix based on the on-chip complex reflection spectrum sampling data; Separation module 20 is used to extract frequency domain local structural features based on the complex reflection spectrum slice matrix and separate the complex reflection spectrum slice matrix into broadband mismatched baseline data and local spike perturbation data; The determination module 30 is used to determine a global mismatch index that characterizes the true degree of broadband mismatch and a baseline flatness index that characterizes the degree of broadband profile change based on the broadband mismatch baseline data, and to determine a spike proportion index that characterizes the degree of local anomaly based on the local spike perturbation data. The construction module 40 is used to construct multiple static trial control states around the current matching control code, determine the scoring results according to the global mismatch index, the baseline flatness index and the spike ratio index corresponding to each static trial control state, and then determine the target matching control code based on the scoring results. The output module 50 is used to determine the target power amplifier bias control code based on the global mismatch index and the spike ratio index corresponding to the target matching control code, and output the target control vector for controlling the radio frequency processing circuit.

[0076] Other embodiments or specific implementations of the radio frequency processing circuit control system of the present invention can be referred to the above-described method embodiments, and will not be repeated here.

[0077] It is understood that in the description of this specification, references to terms such as "one embodiment," "another embodiment," "other embodiments," or "first embodiment to Nth embodiment," 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 the present invention. In this specification, 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.

[0078] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0079] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. A method for controlling a radio frequency processing circuit, characterized in that, The method includes the following steps: Acquire on-chip complex reflection spectrum sampling data under the current control state, and construct a complex reflection spectrum slice matrix based on the on-chip complex reflection spectrum sampling data; Based on the complex reflection spectrum slice matrix, frequency domain local structural features are extracted, and the complex reflection spectrum slice matrix is ​​separated into broadband mismatched baseline data and local spike perturbation data; Based on the broadband mismatch baseline data, a global mismatch index characterizing the true degree of broadband mismatch and a baseline flatness index characterizing the degree of broadband profile change are determined, and a spike proportion index characterizing the degree of local anomaly proportion is determined based on the local spike perturbation data. Multiple static trial control states are constructed around the current matching control code. Scoring results are determined based on the global mismatch index, the baseline flatness index, and the spike ratio index corresponding to each static trial control state. The target matching control code is then determined based on the scoring results. The target power amplifier bias control code is determined based on the global mismatch index and the spike ratio index corresponding to the target matching control code, and the target control vector for controlling the radio frequency processing circuit is output.

2. The radio frequency processing circuit control method as described in claim 1, characterized in that, Acquire on-chip complex reflection spectrum sampling data under the current control state, and construct a complex reflection spectrum slice matrix based on the on-chip complex reflection spectrum sampling data, specifically including: Multiple micro-frequency offset detection points are determined on both sides of the current working center frequency point, and on-chip reflection detection data are collected for each micro-frequency offset detection point on multiple time slices within the same control cycle; For each time slice and each micro-offset frequency detection point, a complex reflection spectrum slice matrix consisting of all complex reflection spectrum samples is constructed: The amplitude and phase data at each time slice and each micro-offset frequency detection point are obtained based on the in-phase component sampling value and the quadrature component sampling value, respectively, for subsequent extraction of local structural features in the frequency domain.

3. The radio frequency processing circuit control method as described in claim 1, characterized in that, Based on the complex reflection spectrum slice matrix, local structural features in the frequency domain are extracted, and the complex reflection spectrum slice matrix is ​​separated into broadband mismatched baseline data and local spike perturbation data, specifically including: Based on the amplitude data corresponding to each position in the complex reflection spectrum slice matrix, the degree of amplitude change and local bending degree between adjacent micro-frequency detection points are determined along the micro-frequency deviation direction. Based on the phase data corresponding to each position in the complex reflection spectrum slice matrix, the degree of phase continuity change between adjacent micro-frequency detection points is determined along the micro-frequency offset direction. Based on the magnitude of the change, the degree of local bending, and the degree of continuous phase change, structural weight data corresponding to each position is generated; The influence of local frequency abrupt changes on the broadband continuous variation results is reduced based on the structural weight data. The part of the complex reflection spectrum slice matrix that changes continuously along the micro-frequency deviation direction is determined as the broadband mismatch baseline data. The abrupt changes that occur in a small number of frequency points and a small number of time slices in the complex reflection spectrum slice matrix are determined as the local spike perturbation data.

4. The radio frequency processing circuit control method as described in claim 3, characterized in that, Based on the magnitude of the change, the degree of local bending, and the degree of continuous phase change, structural weight data corresponding to each position is generated, specifically including: The structural weight values ​​corresponding to each position are determined based on the degree of local bending and the degree of continuous phase change at each position. The data participation degree of each position in the complex reflection spectrum slice matrix is ​​constrained according to the structural weight value, so that the proportion of positions with large local bending degree and large phase continuous change degree is reduced in the broadband mismatch baseline data generation process.

5. The radio frequency processing circuit control method as described in claim 1, characterized in that, Based on the aforementioned broadband mismatch baseline data, a global mismatch index characterizing the true degree of broadband mismatch and a baseline flatness index characterizing the degree of broadband profile change are determined. Furthermore, based on the aforementioned local spike perturbation data, a spike proportion index characterizing the proportion of local anomalies is determined, specifically including: Based on the local spike perturbation data, the slice weight of each time slice is determined, and the broadband mismatch baseline data is weighted and aggregated to obtain the broadband mismatch baseline vector. The global mismatch index and baseline flatness index are determined based on the broadband mismatch baseline vector, and the spike proportion index is determined based on the local spike perturbation data and the on-chip complex reflection spectrum sampling data.

6. The radio frequency processing circuit control method as described in claim 1, characterized in that, Multiple static trial control states are constructed around the current matching control code. Scoring results are determined based on the global mismatch index, baseline flatness index, and spike proportion index corresponding to each static trial control state. Specifically, this includes: Construct a low-side static trial control state, a current static trial control state, and a high-side static trial control state based on the current matching control code; Under each static test control state, the on-chip complex reflection spectrum sampling data is reacquired, and the generation and processing of the broadband mismatch baseline data and local spike perturbation data, as well as the determination and processing of the global mismatch index, the baseline flatness index and the spike proportion index are repeatedly executed. The scoring results for each static test control state are determined based on the global mismatch index, the baseline flatness index, and the spike ratio index corresponding to each static test control state.

7. The radio frequency processing circuit control method as described in claim 6, characterized in that, The target matching control code is determined based on the scoring results, specifically including: The scoring results corresponding to the low-side static trial control state, the current static trial control state, and the high-side static trial control state are used as local scoring sampling points to construct a local scoring relationship about the matching control code; The local scoring function parameters are determined based on the scoring results corresponding to the low-side static test control state, the current static test control state, and the high-side static test control state. When the local scoring function opens upward, a continuous optimal matching control value is determined according to the local scoring function, and the continuous optimal matching control value is mapped to a preset discrete matching control code set to obtain the target matching control code. When the local scoring function does not open upwards, the matching control code corresponding to the static trial control state with the smallest scoring result is determined as the target matching control code.

8. The radio frequency processing circuit control method as described in claim 1, characterized in that, The target power amplifier bias control code is determined based on the global mismatch index and the spike ratio index corresponding to the target matching control code, specifically including: Under the target matching control code, the on-chip complex reflectance spectrum sampling data is reacquired, and the global mismatch index and spike ratio index corresponding to the target matching control code are determined based on the reacquired on-chip complex reflectance spectrum sampling data. The effective mismatch index is determined based on the global mismatch index and spike ratio index corresponding to the target matching control code, and the target power amplifier bias control code is determined in combination with the reference bias control data.

9. The radio frequency processing circuit control method as described in claim 8, characterized in that, The output is a target control vector used to control the RF processing circuit, specifically including: The target matching control code and the target power amplifier bias control code are combined to form the target control vector; The target control vector is written into the control latch region so that the target control vector remains stable in subsequent launch control cycles; Based on the target control vector, matching network control data and power amplifier bias control data are generated respectively, and output to the corresponding RF processing circuit control interface to complete the control of the RF processing circuit.

10. A radio frequency processing circuit control system, characterized in that, The system includes: The acquisition module is used to acquire on-chip complex reflection spectrum sampling data under the current control state, and construct a complex reflection spectrum slice matrix based on the on-chip complex reflection spectrum sampling data; The separation module is used to extract local structural features in the frequency domain based on the complex reflection spectrum slice matrix, and to separate the complex reflection spectrum slice matrix into broadband mismatched baseline data and local spike perturbation data. The determination module is used to determine a global mismatch index that characterizes the true degree of broadband mismatch and a baseline flatness index that characterizes the degree of broadband profile change based on the broadband mismatch baseline data, and to determine a spike proportion index that characterizes the degree of local anomaly based on the local spike perturbation data. The construction module is used to construct multiple static trial control states around the current matching control code, determine the scoring results based on the global mismatch index, the baseline flatness index and the spike ratio index corresponding to each static trial control state, and then determine the target matching control code based on the scoring results. The output module is used to determine the target power amplifier bias control code based on the global mismatch index and the spike ratio index corresponding to the target matching control code, and output the target control vector for controlling the radio frequency processing circuit.

Images