Prediction compensation method for single-stage common-source structure and inductive load output stage circuit
Through the deep learning model, the performance parameters of single-stage common source amplifier and inductive load output stage circuit are solved in real time, and the problem that the existing technology cannot adapt to dynamic changes is achieved, and the precise control and stability of circuit performance are achieved.
Patent Information
- Application Number
- CN202411039145.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-31
- Publication Date
- 2024-12-03
AI Technical Summary
The existing technology cannot effectively monitor and adaptively adjust the performance parameters of single-stage common source amplifiers and inductive load output stage circuits in real time, making it difficult to maintain high performance and reliability under dynamic changing conditions, especially in ambient temperature changes, power supply voltage fluctuations and In the case of device aging, etc.
By monitoring performance parameters in real time, building feature vectors and using deep learning models for prediction and dynamic adjustment, it realizes precise control of gain, phase margin and load resistance, and uses Huber loss function and Adam optimization algorithm for model training and parameter updates.
A comprehensive optimization of circuit performance is achieved, stability and reliability are improved, stable predictive performance can be maintained at different error levels, and the performance of circuits in high-frequency or broadband applications.
Smart Images

Figure CN119067017A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the field of single-stage common-source amplifier signal processing, in particular to a prediction compensation method for a single-stage common-source structure and an inductive load output stage circuit. Background Art
[0002] In modern electronic circuit design, the single-stage common-source amplifier is widely used in the field of analog signal processing as a basic amplification structure. This amplifier is valued for its simple circuit structure and high performance indicators. The common-source amplifier uses the amplification effect of the transistor to input the signal through the gate and obtain the amplified signal from the drain. The source is usually grounded. In the output stage circuit, the use of inductive loads can provide the necessary current and impedance characteristics to meet specific signal processing requirements. However, in order to ensure that the circuit can maintain the expected performance throughout the entire operating range, precise control and dynamic adjustment of circuit parameters become particularly important.
[0003] In existing technologies, the control of performance parameters such as gain, phase margin and load resistance of a single-stage common-source amplifier usually relies on manual adjustment and offline optimization. For example, the gain adjustment may be achieved by changing the operating point of the transistor or the load resistance; the phase margin control may be achieved by adding compensation capacitors or adjusting the circuit topology; and the load resistance adjustment may require changing the physical connection of the circuit or using a variable resistor. Although these methods are effective to a certain extent, they lack flexibility and real-time performance and cannot adapt to the dynamic changes of the circuit under actual working conditions.
[0004] As the complexity of electronic systems increases, higher requirements are placed on real-time monitoring and adaptive adjustment of circuit performance. The limitations of existing technologies are that they are static and passive, and cannot effectively cope with the various dynamic changes encountered by circuits in actual operation. For example, changes in ambient temperature, fluctuations in power supply voltage, aging of device parameters, etc., may all affect the working state of the circuit. In addition, manual adjustment methods are inefficient and difficult to achieve precise control. These limitations are particularly prominent in high-frequency or broadband applications, because the performance parameters of the circuit need to remain stable over a wide frequency range. Therefore, existing technologies face challenges in achieving high-performance and high-reliability electronic circuit design. Summary of the invention
[0005] In view of the defects of the prior art, the present invention proposes a predictive compensation method for a single-stage common-source structure and an inductive load output stage circuit. By real-time monitoring the performance parameters of the single-stage common-source amplifier and the inductive load output stage circuit, a feature vector is constructed, and a deep learning model is used for prediction and dynamic adjustment, so as to achieve precise control of circuit parameters and improve the stability and reliability of the circuit.
[0006] Among them, a predictive compensation method for a single-stage common source structure and an inductive load output stage circuit comprises the following steps:
[0007] S1. Monitoring real-time performance parameters of a single-stage common-source amplifier and an inductive load output stage circuit, wherein the performance parameters include gain parameters, phase margin parameters, and load resistance parameters;
[0008] S2. construct a feature vector according to the real-time performance parameters, and construct a multi-task learning model according to the constructed feature vector;
[0009] S3. Calculate the Hunber loss of each prediction task, obtain the loss values of the gain parameter, the phase margin parameter and the load resistance parameter, calculate the joint loss function, train the multi-task learning model by minimizing the joint loss function, and obtain the trained multi-task learning model;
[0010] S4. Input new data and dynamically adjust circuit parameters through the circuit control interface according to the results predicted by the multi-task learning model;
[0011] The Hunber loss specifically includes a gain adjustment Huber loss for gain prediction, a phase compensation Huber loss for phase margin prediction, and a load modulation Huber loss for load resistance prediction.
[0012] Furthermore, in step S4, dynamically adjusting the circuit parameters through the circuit control interface includes:
[0013] adjusting a control voltage of a variable gain amplifier according to the predicted gain result;
[0014] adjusting a control voltage of a voltage control capacitor according to a predicted phase margin result;
[0015] Based on the predicted load resistance result, the variable resistor connected in parallel with the inductor is adjusted.
[0016] Furthermore, the gain-adjusted Huber loss for gain prediction is specifically expressed as:
[0017]
[0018] Among them, the represents the gain-adjusted Huber loss, the G i,pred represents the prediction gain value of the i-th sample, the G i,true represents the true gain value of the i-th sample, N represents the number of samples, and δ G Represents the threshold parameter in the Huber loss function for gain prediction, which is used to balance the sensitivity of the loss function at different error levels.
[0019] Furthermore, the phase compensation Huber loss for phase margin prediction is specifically expressed as:
[0020]
[0021] Among them, the represents the phase compensation Huber loss, the φ i,pred represents the predicted phase margin value of the i-th sample, the φ i,true represents the true phase margin value of the i-th sample, N represents the number of samples, and δ G Represents the threshold parameter in the Huber loss function for phase margin prediction, which is used to balance the sensitivity of the loss function at different error levels.
[0022] Furthermore, the load modulation Huber loss predicted for the load resistance is specifically expressed as:
[0023]
[0024] Among them, the represents the load modulation Huber loss, the R i,pred represents the predicted load resistance value of the ith sample, the R i,true represents the true load resistance value of the i-th sample, N represents the number of samples, and δ G Represents the threshold parameter in the Huber loss function for load modulation prediction, which is used to balance the sensitivity of the loss function at different error levels.
[0025] Furthermore, the joint loss function is calculated based on weighting, which is specifically expressed as:
[0026]
[0027] Among them, the represents the joint loss function, represents the gain-adjusted Huber loss, represents the phase compensation Huber loss, represents the load modulation Huber loss, the ω G ,ω φ and ω R Represent the weight coefficients for gain, phase margin, and load resistance tasks respectively.
[0028] Furthermore, the multi-task learning model includes three task-specific layers, including a gain adjustment task layer, a phase compensation layer and a load modulation layer.
[0029] Furthermore, in step S1, the gain parameter is calculated by comparing the amplitude of the input signal and the output signal, the phase margin parameter is calculated by measuring the phase delay of the output signal relative to the input signal using a network analyzer, and the load resistance parameter is obtained by directly measuring the load resistance value in parallel with the inductor using a resistance measuring instrument.
[0030] Furthermore, the multi-task learning model is constructed through a deep neural network model.
[0031] Furthermore, in step S3, after the multi-task learning model is trained by minimizing the joint loss function, the model is updated using the Adam optimization algorithm, which specifically includes the following sub-steps:
[0032] S301. For each training sample, calculate the gradient of the joint loss function with respect to the model parameters;
[0033] S302. Update the first-order moment estimate and the second-order moment estimate according to the calculated gradient;
[0034] S303. Calculate the deviation correction of the first-order moment estimation through the first-order moment estimation, and calculate the deviation correction of the second-order moment estimation through the second-order moment estimation;
[0035] S304. Update the model parameters according to the deviation correction of the first-order moment estimation and the deviation correction of the second-order moment estimation.
[0036] The beneficial effects of the invention are:
[0037] The present invention timely captures slight changes in circuit performance through real-time monitoring, provides accurate input information for dynamic adjustment, and can simultaneously predict the optimal values of gain, phase margin and load resistance through a multi-task learning model to achieve comprehensive optimization of circuit performance. The Huber loss function is used to increase the model's robustness to outliers, so that the model can maintain stable prediction performance under different error levels. BRIEF DESCRIPTION OF THE DRAWINGS
[0038] Figure 1 A method flow chart of a predictive control method for a single-stage common-source structure and an inductive load output stage circuit provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0039] The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings, but the protection scope of the present invention is not limited to the following.
[0040] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is further described in detail in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention, that is, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations.
[0041] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative work are within the scope of protection of the present invention. It should be noted that relational terms such as the terms "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations.
[0042] Moreover, the terms "comprises", "includes" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or apparatus that includes a list of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or apparatus. In the absence of more restrictions, an element defined by the phrase "comprises a ..." does not exclude the presence of additional identical elements in the process, method, article or apparatus that includes the element.
[0043] The features and performance of the present invention are further described in detail below in conjunction with the embodiments.
[0044] Among them, Figure 1 , a predictive control method for a single-stage common source structure and an inductive load output stage circuit, comprising the following steps:
[0045] S1. Monitoring real-time performance parameters of a single-stage common-source amplifier and an inductive load output stage circuit, wherein the performance parameters include gain parameters, phase margin parameters, and load resistance parameters;
[0046] S2. construct a feature vector according to the real-time performance parameters, and construct a multi-task learning model according to the constructed feature vector;
[0047] S3. Calculate the Hunber loss of each prediction task, obtain the loss values of the gain parameter, the phase margin parameter and the load resistance parameter, calculate the joint loss function, train the multi-task learning model by minimizing the joint loss function, and obtain the trained multi-task learning model;
[0048] S4. Input new data and dynamically adjust circuit parameters through the circuit control interface according to the results predicted by the multi-task learning model;
[0049] The Hunber loss specifically includes a gain adjustment Huber loss for gain prediction, a phase compensation Huber loss for phase margin prediction, and a load modulation Huber loss for load resistance prediction.
[0050] Furthermore, as a preferred technical solution of this embodiment, the feature vector constructed according to the real-time performance parameters in step S2 may also include the amplitude, frequency, temperature, power supply voltage, etc. of the output signal. The construction of the feature vector needs to be selected according to the correlation of the task parameter characteristics, that is, the characteristics associated with the gain parameters, phase margin parameters and load resistance parameters can be used to construct a unified feature vector.
[0051] Furthermore, in step S4, dynamically adjusting the circuit parameters through the circuit control interface includes:
[0052] adjusting a control voltage of a variable gain amplifier according to the predicted gain result;
[0053] adjusting a control voltage of a voltage control capacitor according to a predicted phase margin result;
[0054] Based on the predicted load resistance result, the variable resistor connected in parallel with the inductor is adjusted.
[0055] Furthermore, the gain-adjusted Huber loss for gain prediction is specifically expressed as:
[0056]
[0057] Among them, the represents the gain-adjusted Huber loss, the G i,pred represents the prediction gain value of the i-th sample, the G i,true represents the true gain value of the i-th sample, N represents the number of samples, and δ G Represents the threshold parameter in the Huber loss function for gain prediction, which is used to balance the sensitivity of the loss function at different error levels.
[0058] Furthermore, the phase compensation Huber loss for phase margin prediction is specifically expressed as:
[0059]
[0060] Among them, the represents the phase compensation Huber loss, the φ i,pred represents the predicted phase margin value of the i-th sample, the φ i,truerepresents the true phase margin value of the i-th sample, N represents the number of samples, and δ G Represents the threshold parameter in the Huber loss function for phase margin prediction, which is used to balance the sensitivity of the loss function at different error levels.
[0061] Furthermore, the load modulation Huber loss predicted for the load resistance is specifically expressed as:
[0062]
[0063] Among them, the represents the load modulation Huber loss, the R i,pred represents the predicted load resistance value of the ith sample, the R i,true represents the true load resistance value of the i-th sample, N represents the number of samples, and δ G Represents the threshold parameter in the Huber loss function for load modulation prediction, which is used to balance the sensitivity of the loss function at different error levels.
[0064] Specifically, the Huber loss is defined by the Huber loss function, which is specifically:
[0065]
[0066] The hyperparameter δ controls the behavior of the loss function when the error is small and large.
[0067] Furthermore, the joint loss function is calculated based on weighting, which is specifically expressed as:
[0068]
[0069] Among them, the represents the joint loss function, represents the gain-adjusted Huber loss, represents the phase compensation Huber loss, represents the load modulation Huber loss, the ω G ,ω φ and ω R Represent the weight coefficients for gain, phase margin, and load resistance tasks respectively.
[0070] Furthermore, the multi-task learning model includes three task-specific layers, including a gain adjustment task layer, a phase compensation layer and a load modulation layer.
[0071] Furthermore, in step S1, the gain parameter is calculated by comparing the amplitude of the input signal and the output signal, the phase margin parameter is calculated by measuring the phase delay of the output signal relative to the input signal using a network analyzer, and the load resistance parameter is obtained by directly measuring the load resistance value in parallel with the inductor using a resistance measuring instrument.
[0072] Furthermore, the multi-task learning model is constructed through a deep neural network model.
[0073] Furthermore, in step S3, after the multi-task learning model is trained by minimizing the joint loss function, the model is updated using the Adam optimization algorithm, which specifically includes the following sub-steps:
[0074] S301. For each training sample, calculate the gradient of the joint loss function with respect to the model parameters;
[0075] S302. Update the first-order moment estimate and the second-order moment estimate according to the calculated gradient;
[0076] S303. Calculate the deviation correction of the first-order moment estimation through the first-order moment estimation, and calculate the deviation correction of the second-order moment estimation through the second-order moment estimation;
[0077] S304. Update the model parameters according to the deviation correction of the first-order moment estimation and the deviation correction of the second-order moment estimation.
[0078] Specifically, for the adam optimization solution described in the above embodiment, wherein:
[0079] The gradient of the calculated model parameters is expressed as:
[0080] The updated first-order moment estimate is expressed as:
[0081] The updated second-order moment estimate is expressed as:
[0082] The bias correction for computing the first moment estimate is expressed as:
[0083] The bias correction for computing the second-order moment estimate is expressed as:
[0084] Updating model parameters is expressed as:
[0085] Among them, the represents the gradient of the type parameter, the v t and trepresents the momentum term for updating the first-order moment estimate and the adjustment term for updating the second-order moment estimate, wherein β1 and β2 are momentum parameters between 0 and 1, used to balance the previous first-order moment / second-order moment estimate and the current gradient, It represents the bias correction of the first-order moment estimate, which is used to correct the bias of the first-order moment estimate and accurately reflect the average value of the long-term gradient. represents the deviation correction of the second-order moment estimation, the α represents the step size of the control model parameter update, the ∈ represents a numerical stability constant for maintaining numerical stability, the Θ t represents the model parameters, the Θ t+1 Represents the updated model parameters
[0086] Furthermore, as a preferred implementation scheme, a single-stage common source structure and an inductive load output stage circuit are proposed. The prediction compensation method of a single-stage common source structure and an inductive load output stage circuit proposed in the above embodiment can be applied to the above circuit, wherein the circuit includes: an input signal source, a common source amplifier, an output inductive load, an output buffer, a variable gain control circuit, a voltage control capacitor, an output signal detection circuit, a phase margin detection circuit and a closed-loop control circuit digital signal processor, the gate end of the common source amplifier is connected to the input signal source to control the conductive state of the transistor, the drain end of the common source amplifier is connected to the output buffer through the output inductive load to form a signal amplification path, and the source of the common source amplifier The pole is grounded, the variable gain control circuit is used to adjust the input signal gain of the common source amplifier according to the instruction of the closed-loop control circuit digital signal processor, the voltage-controlled capacitor is used to adjust the phase of the output signal according to the instruction of the closed-loop control circuit, the output signal detection circuit and the phase margin detection circuit are used to monitor the amplitude and phase of the output signal in real time, and feed the information back to the closed-loop control circuit, the closed-loop control circuit is used to adjust the variable gain control circuit and the voltage-controlled capacitor according to the feedback signal, and perform real-time control, the closed-loop control circuit digital signal processor is used to make a comprehensive judgment and output a control instruction according to the detected signal amplitude and phase information, or the ambient temperature information provided by the temperature sensor.
[0087] Further, the input signal source output is connected to the gate terminal of the common source amplifier, the closed-loop control circuit receives feedback signals from the output signal detection circuit and the phase margin detection circuit, and dynamically adjusts the variable gain control circuit and the voltage-controlled capacitor, the output inductive load is connected between the drain terminal of the common source amplifier and the output buffer, and is used to manage the output current and the resonant frequency, the output buffer receives the signal from the common source amplifier, and provides additional amplification or buffering to drive the inductive load, the variable gain control circuit is connected to the gate terminal of the common source amplifier, and dynamically adjusts the gain of the input signal, wherein the voltage-controlled attenuator and / or variable resistor are connected in parallel, and the voltage-controlled capacitor and the voltage-controlled capacitor are connected in parallel. The output signal detection circuit is located between the output end of the output buffer and the ground, and is used to dynamically adjust the phase characteristics of the output signal. The output signal detection circuit samples the signal amplitude from the output end of the output buffer through the sampling resistor, and feeds the information back to the closed-loop control circuit. The phase margin detection circuit measures the phase of the output signal and compares it with the phase of the input signal, and outputs the phase margin information. The closed-loop control circuit receives feedback signals from the output signal detection circuit and the phase margin detection circuit, and dynamically adjusts the variable gain control circuit and the voltage control capacitor. The digital signal processor receives the signal of the closed-loop control circuit, and outputs a control instruction according to a preset algorithm to adjust the variable gain control circuit and the voltage control capacitor.
[0088] The above-mentioned embodiment adopts real-time monitoring of the performance parameters of the single-stage common-source amplifier and the inductive load output stage circuit, and uses a deep learning model to predict the optimal adjustment values of the gain, phase margin and load resistance to achieve accurate and dynamic adjustment. Specifically, the above-mentioned embodiment adopts the Huber loss function to perform customized design for different prediction tasks, improves the robustness of the model to outliers, and ensures the sensitivity balance at different error levels. In addition, through the dynamic parameter update mechanism of the Adam optimization algorithm, the model can quickly respond to changes in circuit performance and adjust circuit parameters in real time, thereby significantly improving the stability and reliability of the circuit. In summary, the above-mentioned embodiment realizes intelligent and automated compensation control of the single-stage common-source amplifier and the inductive load output stage circuit, optimizes the circuit design process, and improves the prediction accuracy and adjustment efficiency of circuit performance.
[0089] The above is only a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the form disclosed herein, and should not be regarded as excluding other embodiments, but can be used in various other combinations, modifications and environments, and can be modified within the scope of the concept described herein through the above teachings or the technology or knowledge of the relevant field. The changes and modifications made by those skilled in the art shall not deviate from the spirit and scope of the present invention, and shall be within the scope of protection of the claims attached to the present invention.
Claims
1. A predictive compensation method for a single-stage common source structure and an inductive load output stage circuit, characterized in that: The following steps are involved: S1. Monitoring real-time performance parameters of a single-stage common-source amplifier and an inductive load output stage circuit, wherein the performance parameters include gain parameters, phase margin parameters, and load resistance parameters; S2. construct a feature vector according to the real-time performance parameters, and construct a multi-task learning model according to the constructed feature vector; S3. Calculate the Hunber loss of each prediction task, obtain the loss values of the gain parameter, the phase margin parameter and the load resistance parameter, calculate the joint loss function, train the multi-task learning model by minimizing the joint loss function, and obtain the trained multi-task learning model; S4. Input new data, and dynamically adjust circuit parameters through the circuit control interface according to the prediction results of the multi-task learning model; wherein the Hunber loss specifically includes a gain adjustment Huber loss for gain prediction, a phase compensation Huber loss for phase margin prediction, and a load modulation Huber loss for load resistance prediction.
2. The predictive compensation method for a single-stage common source structure and an inductive load output stage circuit as claimed in claim 1, characterized in that: In step S4, dynamically adjusting the circuit parameters through the circuit control interface includes: adjusting a control voltage of a variable gain amplifier according to the predicted gain result; adjusting a control voltage of a voltage control capacitor according to a predicted phase margin result; Based on the predicted load resistance result, the variable resistor connected in parallel with the inductor is adjusted.
3. The predictive compensation method for a single-stage common source structure and an inductive load output stage circuit as claimed in claim 1, characterized in that: The gain-adjusted Huber loss for gain prediction is specifically expressed as: Among them, the represents the gain-adjusted Huber loss, the G i,pred represents the prediction gain value of the i-th sample, the G i,true represents the true gain value of the i-th sample, N represents the number of samples, and δ G Represents the threshold parameter in the Huber loss function for gain prediction, which is used to balance the sensitivity of the loss function at different error levels.
4. The predictive compensation method for a single-stage common source structure and an inductive load output stage circuit as claimed in claim 1, characterized in that: The phase compensation Huber loss for phase margin prediction is specifically expressed as: Among them, the represents the phase compensation Huber loss, the φ i,pred represents the predicted phase margin value of the i-th sample, the φ i,true represents the true phase margin value of the i-th sample, N represents the number of samples, and δ G Represents the threshold parameter in the Huber loss function for phase margin prediction, which is used to balance the sensitivity of the loss function at different error levels.
5. The predictive compensation method for a single-stage common source structure and an inductive load output stage circuit as claimed in claim 1, characterized in that: The load modulation Huber loss predicted for the load resistance is specifically expressed as: Among them, the represents the load modulation Huber loss, the R i,pred represents the predicted load resistance value of the ith sample, the R i,true represents the true load resistance value of the i-th sample, N represents the number of samples, and δ G Represents the threshold parameter in the Huber loss function for load modulation prediction, which is used to balance the sensitivity of the loss function at different error levels.
6. The predictive compensation method for a single-stage common source structure and an inductive load output stage circuit as claimed in claim 1, characterized in that: The joint loss function is calculated based on weighting, which is specifically expressed as: Among them, the represents the joint loss function, represents the gain-adjusted Huber loss, represents the phase compensation Huber loss, represents the load modulation Huber loss, the ω G ,ω φ and ω R Represent the weight coefficients for gain, phase margin, and load resistance tasks respectively.
7. The predictive compensation method for a single-stage common source structure and an inductive load output stage circuit as claimed in claim 1, characterized in that: The multi-task learning model includes three task-specific layers, including a gain adjustment task layer, a phase compensation layer, and a load modulation layer.
8. The predictive compensation method for a single-stage common source structure and an inductive load output stage circuit as claimed in claim 1, characterized in that: In step S1, the gain parameter is calculated by comparing the amplitude of the input signal and the output signal, the phase margin parameter is calculated by measuring the phase delay of the output signal relative to the input signal using a network analyzer, and the load resistance parameter is obtained by directly measuring the load resistance value connected in parallel with the inductor using a resistance measuring instrument.
9. The predictive compensation method for a single-stage common source structure and an inductive load output stage circuit as claimed in claim 1, characterized in that: The multi-task learning model is constructed through a deep neural network model.
10. The predictive compensation method for a single-stage common source structure and an inductive load output stage circuit according to claim 1, characterized in that: In step S3, after the multi-task learning model is trained by minimizing the joint loss function, the model is updated using the Adam optimization algorithm, which specifically includes the following sub-steps: S301. For each training sample, calculate the gradient of the joint loss function with respect to the model parameters; S302. Update the first-order moment estimate and the second-order moment estimate according to the calculated gradient; S303. Calculate the deviation correction of the first-order moment estimation through the first-order moment estimation, and calculate the deviation correction of the second-order moment estimation through the second-order moment estimation; S304. Update the model parameters according to the deviation correction of the first-order moment estimation and the deviation correction of the second-order moment estimation.
Citation Information
Cited By
Signal optimization method of ADAM algorithm assisted multi-stage logarithmic amplifier
CN120687748A
