Intelligent debugging method and system for LC filter

Abstract

The invention discloses an intelligent debugging method and system for an LC filter, and relates to the technical field of automatic control, and the method comprises the steps: generating an initialization parameter set suitable for a debugging task of the LC filter based on a meta-learning framework; constructing a quantization particle swarm hybrid optimization debugging model; and constructing a parameter change trajectory database by combining the concept of the psychological field and a knowledge distillation technology, performing time sequence modeling on the parameter change trajectory database through an LSTM network, and predicting an optimal initial guess. The LC filter does not need to depend on a large amount of historical initial data at the beginning of regulation and control, the complex program of manual debugging is deleted, the time for carrying out initialization operation on the LC filter is greatly shortened, the difficulty of local solutions in the intelligent debugging process is solved, the convergence speed is high, the debugging time is shortened, and the debugging efficiency is improved. In the debugging process, result data is quickly and efficiently obtained on the basis of combining a small amount of historical data, and the automatic intelligent debugging process of LC can be continuously optimized.

Description

Technical Field

[0001] The present invention relates to the technical field of LC filter debugging, and particularly to an intelligent debugging method and system for an LC filter. Background Art

[0002] With the development of artificial intelligence technology and automation control technology, intelligent debugging methods will be more widely applied in the field of LC filter debugging. For example, by combining more advanced technologies such as deep learning and reinforcement learning, an intelligent debugging system will be able to perform adaptive adjustment in a more complex environment, further improving the performance and application scope of the filter.

[0003] Currently, the Chinese invention patent with the application number CN201810086329.2 discloses an accurate debugging method and system for a notch filter. By performing mathematical simulation or sweep frequency test on the debugging object and establishing an accurate mathematical model of the second-order notch filter, the debugging object can be debugged. During the debugging process, if there is still a gap between the frequency characteristics of the debugging object and the index requirements, the notch depth and notch bandwidth can be appropriately adjusted according to the gap situation, gradually meeting the index requirements. During the debugging process, the next debugging can be guided based on the previous debugging results, greatly improving the debugging efficiency while achieving accurate debugging. The existing technology still has the problems that at the beginning of the LC filter regulation, it is overly dependent on historical initial data, and there are also complex procedures for manual debugging, which requires a lot of time to initialize the LC filter. There are dilemmas of local solutions in the intelligent debugging process, the convergence speed is slow, the debugging time is too long, it is impossible to quickly and efficiently obtain result data based on a small amount of historical data during debugging, and the automated intelligent debugging process of LC cannot be continuously optimized by continuously learning and accumulating to optimize the debugging process. Summary of the Invention

[0004] The technical problem solved by the present invention is that the existing technology still has the problems that at the beginning of the LC filter regulation, it is overly dependent on historical initial data, and there are also complex procedures for manual debugging, which requires a lot of time to initialize the LC filter. There are dilemmas of local solutions in the intelligent debugging process, the convergence speed is slow, the debugging time is too long, it is impossible to quickly and efficiently obtain result data based on a small amount of historical data during debugging, and the automated intelligent debugging process of LC cannot be continuously optimized by continuously learning and accumulating to optimize the debugging process.

[0005] To solve the above technical problems, the present invention provides the following technical solution: An intelligent debugging method for an LC filter, comprising the following steps: Step S1: Sampling historical LC filter debugging task data from a multi-task distribution based on a meta-learning framework, and learning the distribution characteristics of parameters through a pre-trained machine learning model to generate an initialization parameter set applicable to the LC filter debugging task; Step S2: Encode the inductance value, capacitance value, and quality factor of the LC filter into quantum states, and construct a quantum particle swarm hybrid optimization debugging model; Step S3: Combine the concept of the psychological field and knowledge distillation technology, map the high-dimensional parameter space to a low-dimensional latent space, construct a parameter change trajectory database, and perform temporal modeling on the parameter change trajectory database through an LSTM network to predict the optimal initial guess.

[0006] Preferably, the meta-learning framework adopts a two-layer optimization structure. The outer layer optimization updates the initial parameter distribution through gradient descent. The inner layer optimization simulates the gradient update process, including: minimizing the cut-off frequency error and total harmonic distortion index of the current task, and outputting the initial parameter distribution adapted to the few-shot condition.

[0007] Preferably, constructing a quantum particle swarm hybrid optimization debugging model includes: The quantum particle swarm hybrid optimization debugging model is used to describe the relationship between the performance of the LC filter and the debugging parameters; Generating quantum states based on the initial parameter distribution includes: mapping the inductance value, capacitance value, and quality factor to the amplitude and phase of quantum bits respectively. The inductance value adopts a normalized logarithmic coding form, the capacitance value adopts a piecewise linear coding, and the quality factor is compressed to the interval [0, 1] through the Sigmoid function; The mathematical expression for encoding the inductance value, capacitance value, and quality factor into quantum states is: ; Where represents the superposition state of the quantum state, is the quantum bit amplitude corresponding to the inductance value, representing the relative contribution of the inductance in the quantum state, is the quantum bit amplitude corresponding to the capacitance value, representing the relative contribution of the capacitance in the quantum state, is the quantum bit amplitude corresponding to the quality factor, representing the relative contribution of the quality factor in the quantum state, is the quantum state of the inductance value, is the quantum state of the capacitance value, is the quantum state of the quality factor; Obtain the quantum parameter space, and initialize the particle positions in the quantum parameter space through the particle swarm optimization algorithm. The initialization includes: based on the preset number of iterations and fitness change rate, adjust the search direction of the particle positions through the quantum rotation gate, and dynamically adjust the inertia weight of the particle swarm. The inertia weight is used to control the speed and direction of particle search; The quantum rotation angle of the quantum rotation gate is dynamically calculated from the gradient difference between the current parameter and the global optimal parameter. The mathematical expression for the rotation angle is: ; Among them, is the rotation angle, is a function that maps two input gradients to the rotation angle, is the gradient of the current parameter, is the gradient of the global parameter.

[0008] Preferably, using the chaotic perturbation strategy to break the local optimum includes: In the convergence stagnation stage of the particle swarm algorithm, break the local optimum by injecting chaotic perturbation. After breaking the local optimum, generate a chaotic sequence using the Logistic map, and the perturbation amplitude is inversely proportional to the fitness value of the current optimal solution.

[0009] When in the iterative state, optimize the large parameter space through quantum variational optimization or quantum simulated annealing. When the preset number of iterations is reached, output the final optimization result of the quantized particle swarm hybrid optimization debugging model.

[0010] Preferably, constructing the parameter change trajectory database includes: Taking the total psychological needs of people as the output prediction parameter of the LC filter, and taking people's certain demands for physiology, environment and society as the quantum states of inductance value, capacitance value and quality factor respectively. Find the data of the missing conditions, insufficient data, expected conditions and demand data of the corresponding subject of the LC filter, and perform iterative transfer calculation based on knowledge distillation to establish the objective function of the LC filter. The mathematical expression of the objective function is: ; Among them, B is the output prediction parameter of the LC filter, is the representative function, indicating the relationship between the performance of the LC filter output by the quantized particle swarm hybrid optimization debugging model and the debugging parameters, is the total loss function, represents the learning of the debugging task data, represents the difference between two iterative calculations, represents the output prediction parameter of the nth round of iteration, represents the output prediction parameter of the (n + 1)th round of iteration, represents the Softmax function, which is used to convert the output of the model into a probability distribution, is the weight hyperparameter, which is determined by the magnitude of the backpropagation gradient of the LC filter; Store all the output prediction parameters in the iterative calculation in the parameter change trajectory database.

[0011] Preferably, the LSTM network adopts a gated recurrent unit variant, with the input being the initialization parameter set of historical debugging tasks, and the output layer predicting the optimal parameter direction for the next iteration stage through the Softmax function; Based on the learning results of the LSTM network on historical data, a similarity matching algorithm is used to quickly match similar designs and parameter changes, providing an initial guess for new debugging tasks in iterative calculations.

[0012] Preferably, a dynamic constraint processing mechanism is adopted during the process of the LSTM network performing temporal modeling on the parameter change trajectory database: Define the elastic boundary of the feasible solution. The feasible interval of the inductance value is [L_min, L_max], and the feasible interval of the capacitance value is [C_min, C_max]. When the parameter exceeds the boundary, the mirror mapping method is used to project the out-of-bounds parameter to the neighborhood feasible point of the feasible interval.

[0013] An intelligent debugging system for an LC filter, which is used to execute an intelligent debugging method for an LC filter, including an initialization parameter module, a quantization particle swarm hybrid optimization debugging module, an optimal prediction module, and an embedded hardware platform: The initialization parameter module is used to sample historical LC filter debugging task data from the multi-task distribution based on the meta-learning framework, learn the distribution characteristics of the parameters through a pre-trained machine learning model, and generate an initialization parameter set applicable to the LC filter debugging task; The quantization particle swarm hybrid optimization debugging module is used to encode the inductance value, capacitance value, and quality factor of the LC filter into quantum states and construct a quantization particle swarm hybrid optimization debugging model; The optimal prediction module is used to combine the concept of the psychological field and knowledge distillation technology, map the high-dimensional parameter space to the low-dimensional latent space, construct a parameter change trajectory database, perform temporal modeling on the parameter change trajectory database through the LSTM network, and predict the optimal initial guess; The embedded hardware platform is used to run the meta-learning initialization module and the quantum-inspired optimization core algorithm, dynamic parameter update interface, and temperature compensation.

[0014] Preferably, the embedded hardware platform includes: Support FPGA parallel computing when running the meta-learning initialization module and the quantum-inspired optimization core algorithm; The dynamic parameter update interface includes online adjustment of debugging parameters through the LabVIEW graphical programming environment and visual display of the frequency domain response curve and the optimization process; The temperature compensation includes integrating an NTC thermistor network, real-time monitoring of the temperature drift of the inductor and capacitor, and correcting the parameter value through a linear regression model.

[0015] Preferably, after debugging, the LC filter is subjected to index verification and evaluation: The indexes include frequency-domain performance, time-domain performance, energy efficiency index, and stability. The frequency-domain performance includes cut-off frequency error, stopband attenuation, and passband ripple. The time-domain performance includes step response rise time and overshoot rate. The energy efficiency index includes insertion loss and power loss. The stability includes temperature drift coefficient and component parameter tolerance sensitivity.

[0016] The beneficial effects of the present invention are as follows: At the beginning of the regulation of the LC filter, it does not need to rely heavily on historical initial data, and the complex procedures of manual debugging are deleted, greatly reducing the time for initializing the LC filter. It solves the dilemma of local solutions in the intelligent debugging process, and has a fast convergence speed and short debugging time. During the debugging process, based on a small amount of historical data, the result data can be obtained quickly and efficiently, and the automated intelligent debugging process of the LC can be continuously optimized by continuously learning and accumulating to optimize the debugging process. Description of the Drawings

[0017] Figure 1 It is a schematic diagram of the basic process of an intelligent debugging method for an LC filter provided by an embodiment of the present invention. Detailed Embodiments

[0018] In order to make the above objects, features, and advantages of the present invention more obvious and understandable, the following detailed description of the specific embodiments of the present invention will be made in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all embodiments.

[0019] Referring to Figure 1 , an embodiment of the present invention provides an intelligent debugging method for an LC filter, including the following steps: Step S1: Sampling historical LC filter debugging task data from a multi-task distribution based on a meta-learning framework, learning the distribution characteristics of parameters through a pre-trained machine learning model, and generating an initialization parameter set applicable to the LC filter debugging task; Step S2: Encoding the inductance value, capacitance value, and quality factor of the LC filter into quantum states, and constructing a quantum particle swarm hybrid optimization debugging model; Step S3: Combining the concept of the psychological field and knowledge distillation technology, mapping the high-dimensional parameter space to a low-dimensional latent space, constructing a parameter change trajectory database, and performing temporal modeling on the parameter change trajectory database through an LSTM network to predict the optimal initial guess.

[0020] The meta-learning framework adopts a two-layer optimization structure. The outer layer optimization updates the initialization parameter distribution through gradient descent, and the inner layer optimization simulates the gradient update process, including: minimizing the cut-off frequency error and total harmonic distortion index of the current task, and outputting the initial parameter distribution adapted to the few-shot condition.

[0021] With a small amount of sample data, meta-learning can quickly learn the debugging rules of LC filters, avoid long-term trial and error from scratch, and reduce the time and complexity of manual debugging.

[0022] Constructing a quantization particle swarm hybrid optimization debugging model includes: The quantization particle swarm hybrid optimization debugging model is used to describe the relationship between the performance of the LC filter and the debugging parameters; Generating quantum states based on the initial parameter distribution includes: mapping the inductance value, capacitance value, and quality factor to the amplitude and phase of quantum bits respectively. The inductance value adopts a normalized logarithmic coding form, the capacitance value adopts a piecewise linear coding, and the quality factor is compressed to the [0, 1] interval through the Sigmoid function; The mathematical expressions for encoding the inductance value, capacitance value, and quality factor into quantum states are: ; Among them, represents the superposition state of the quantum state, is the quantum bit amplitude corresponding to the inductance value, representing the relative contribution of the inductance in the quantum state, is the quantum bit amplitude corresponding to the capacitance value, representing the relative contribution of the capacitance in the quantum state, is the quantum bit amplitude corresponding to the quality factor, representing the relative contribution of the quality factor in the quantum state, is the quantum state of the inductance value, is the quantum state of the capacitance value, is the quantum state of the quality factor; Obtain the quantization parameter space, and initialize the particle positions in the quantization parameter space through the particle swarm optimization algorithm. The initialization includes: based on the preset number of iterations and the fitness change rate, adjust the search direction of the particle positions through the quantum rotation gate, and dynamically adjust the inertia weight of the particle swarm. The inertia weight is used to control the speed and direction of particle search; The quantum rotation angle of the quantum rotation gate is dynamically calculated from the gradient difference between the current parameter and the global optimal parameter to ensure that the search direction is adjusted as the optimal solution changes. The mathematical expression for the rotation angle is: ; Among them, is the rotation angle, is a function that maps the input of two gradients to the rotation angle, is the gradient of the current parameter, is the gradient of the global parameter.

[0023] The gradient difference reflects the gap between the current parameter and the global optimal solution, while the change in the rotation angle helps the particle conduct a more precise search within the neighborhood of the optimal solution.

[0024] Using the chaotic perturbation strategy to break through the local optimum includes: During the convergence stagnation stage of the particle swarm algorithm, break through the local optimum by injecting chaotic perturbation. After breaking through the local optimum, generate a chaotic sequence using the Logistic mapping. The perturbation amplitude is inversely proportional to the fitness value of the current optimal solution, which helps to break through the local optimum and explore new search regions.

[0025] When in the iterative state, optimize the large parameter space through quantum variational optimization or quantum simulated annealing to accelerate the search for the global optimal solution. During multiple iterations, with the dynamic adjustment of the inertia weight, the introduction of chaotic perturbation, and the application of the quantum rotation gate, the algorithm gradually approaches the optimal solution. When the preset number of iterations is reached, output the final optimization result of the quantum particle swarm hybrid optimization debugging model.

[0026] Quantum computing heuristic optimization helps to find the global optimal solution or a solution close to the optimal solution in a large parameter space. The simulated annealing algorithm escapes from the local optimal solution by introducing random perturbations and finally converges to the global optimal solution. The quantum simulated annealing algorithm utilizes the quantum superposition and interference effects in quantum computing, enabling the system to explore multiple solution spaces in parallel, thus greatly improving the convergence speed.

[0027] This method not only utilizes the efficient optimization ability of quantum computing in a high-dimensional parameter space but also combines the collective intelligence of the particle swarm and the global search ability of the chaotic perturbation mechanism, effectively avoiding the local optimal solution and accelerating the search for the global optimal solution.

[0028] Constructing the parameter change trajectory database includes: Taking the total psychological needs of people as the output prediction parameter of the LC filter, and taking people's certain demands for physiology, environment, and society as the quantum states of the inductance value, capacitance value, and quality factor respectively. Find the data of the missing conditions, insufficient data, expected conditions, and demand data of the LC filter corresponding to the main body, and conduct iterative transfer calculations based on knowledge distillation to establish the objective function of the LC filter. The mathematical expression of the objective function is: ; Among them, B is the output prediction parameter of the LC filter, is the representative function, representing the relationship between the performance and debugging parameters of the LC filter output by the quantum particle swarm hybrid optimization debugging model, is the total loss function, represents the learning of the debugging task data, Represents the difference between two iterative calculations. Represents the output prediction parameter of the n-th iteration. Represents the output prediction parameter of the (n + 1)-th iteration. Represents the Softmax function, which is used to convert the output of the model into a probability distribution. Is a weight hyperparameter determined by the magnitude of the backpropagation gradient of the LC filter. The goal of combining the concept of the psychological field with knowledge distillation for auxiliary optimization is to use the concept of the psychological field to refer to a certain correlation or interaction space between tasks and parameters. Through this mapping, we can understand and analyze the similarities and changing trends between different tasks in the high-dimensional parameter space, and then realize the guidance of parameter adjustment. By transferring the adjustment experience and rules of debugging parameters through an efficient teacher model, a relatively simplified model can debug the LC filter more quickly.

[0029] Store all output prediction parameters in the iterative calculation in the parameter change trajectory database.

[0030] By establishing a parameter change trajectory database, intelligent decision-making can be carried out based on past debugging experience, further accelerating the debugging process. By generating reasonable initial parameters from a small number of samples, conducting global optimization, and combining historical data and intelligent prediction to optimize the debugging process, the debugging efficiency and performance can be significantly improved. And it helps to quickly find historical similar debugging cases and parameter trajectories in future debugging tasks, providing references for new debugging tasks. The parameter change trajectory database can optimize the debugging process through continuous learning and accumulation.

[0031] The LSTM network adopts a gated recurrent unit variant. The input is the initial parameter set of historical debugging tasks, and the output layer predicts the optimal parameter direction for the next iterative stage through the Softmax function. Based on the learning results of the LSTM network on historical data, similar designs and parameter changes are quickly matched through a similarity matching algorithm, providing an initial guess for new debugging tasks in iterative calculations. The similarity matching algorithm can reduce the debugging time from scratch by finding the closest historical cases.

[0032] Through multi-modal mixing, multiple influencing factors can be considered simultaneously to obtain more comprehensive and efficient debugging parameters.

[0033] Adopt a dynamic constraint processing mechanism in the process of the LSTM network performing temporal modeling on the parameter change trajectory database: Define the elastic boundary of the feasible solution. The feasible interval of the inductance value is [L_min, L_max], and the feasible interval of the capacitance value is [C_min, C_max]. When the parameter crosses the boundary, the mirror mapping method is used to project the out-of-bounds parameter to the neighboring feasible point of the feasible interval.

[0034] An intelligent debugging system for an LC filter, which is used to execute an intelligent debugging method for an LC filter, including an initialization parameter module, a quantization particle swarm hybrid optimization debugging module, an optimal prediction module, and an embedded hardware platform: The initialization parameter module is used to sample historical LC filter debugging task data from a multi-task distribution based on a meta-learning framework, learn the distribution characteristics of parameters through a pre-trained machine learning model, and generate an initialization parameter set applicable to the LC filter debugging task; The quantization particle swarm hybrid optimization debugging module is used to encode the inductance value, capacitance value, and quality factor of the LC filter into quantum states and construct a quantization particle swarm hybrid optimization debugging model; The optimal prediction module is used to combine the concept of the psychological field and knowledge distillation technology, map the high-dimensional parameter space to a low-dimensional latent space, construct a parameter change trajectory database, and perform temporal modeling on the parameter change trajectory database through an LSTM network to predict the optimal initial guess; The embedded hardware platform is used to run the meta-learning initialization module and the quantum-inspired optimization core algorithm, dynamic parameter update interface, and temperature compensation.

[0035] Deployed on the embedded hardware platform includes: Support for FPGA parallel computing when running the meta-learning initialization module and the quantum-inspired optimization core algorithm; The dynamic parameter update interface includes online adjustment of debugging parameters through the LabVIEW graphical programming environment and visual display of the frequency-domain response curve and optimization process; The temperature compensation includes integrating an NTC thermistor network, real-time monitoring of the temperature drift of inductors and capacitors, and correcting the parameter values through a linear regression model.

[0036] Perform index verification and evaluation on the debugged LC filter: The indexes include frequency-domain performance, time-domain performance, energy efficiency indexes, and stability. The frequency-domain performance includes cut-off frequency error, stop-band attenuation, and pass-band ripple. The time-domain performance includes step response rise time and overshoot rate. The energy efficiency indexes include insertion loss and power loss. The stability includes temperature drift coefficient and component parameter tolerance sensitivity.

[0037] The intelligent debugging method provided in the present invention does not rely heavily on a large amount of historical initial data at the beginning of the LC filter regulation, deletes the complex procedures of manual debugging, greatly reduces the time for initializing the LC filter, solves the dilemma of local solutions in the intelligent debugging process, has a fast convergence speed, shortens the debugging time, quickly and efficiently obtains result data on the basis of combining a small amount of historical data during the debugging process, and can continuously optimize the automatic intelligent debugging process of the LC through continuous learning and accumulation to optimize the debugging process.

[0038] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media that contain computer-usable program code. Among them, the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM for short), electrically erasable programmable read-only memory (EEPROM for short), erasable programmable read-only memory (EPROM for short), programmable read-only memory (PROM for short), read-only memory (ROM for short), magnetic memory, flash memory, a magnetic disk, or an optical disc. These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured article including an instruction device, and the instruction device implements the functions specified in one process Figure 1 one process or multiple processes and / or Figure 1 boxes or multiple boxes.

[0039] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered by the scope of the claims of the present invention.

Claims

1. An intelligent debugging method for an LC filter, characterized in that: The following steps are involved: Step S1: sampling historical LC filter debugging task data from a multi-task distribution based on a meta-learning framework, learning the distribution characteristics of parameters through a pre-trained machine learning model, and generating an initialization parameter set suitable for the LC filter debugging task; Step S2: Encode the inductance value, capacitance value and quality factor of the LC filter into quantum states, and construct a quantized particle swarm hybrid optimization debugging model; Step S3: Combining the concept of psychological field and knowledge distillation technology, the high-dimensional parameter space is mapped to the low-dimensional latent space, and a parameter change trajectory database is constructed. The parameter change trajectory database is modeled in time series through the LSTM network to predict the optimal initial guess.

2. The intelligent debugging method of the LC filter according to claim 1, characterized in that: The meta-learning framework adopts a two-layer optimization structure. The outer layer optimization updates the initialization parameter distribution through gradient descent, and the inner layer optimization simulates the gradient update process including: minimizing the cutoff frequency error and total harmonic distortion index of the current task, and outputting the initial parameter distribution that adapts to the few-sample condition.

3. The intelligent debugging method of the LC filter according to claim 2, characterized in that: Building a quantized particle swarm hybrid optimization debugging model includes: The quantized particle swarm hybrid optimization debugging model is used to describe the relationship between the performance and debugging parameters of the LC filter; Generating a quantum state based on the initial parameter distribution includes: mapping the inductance value, the capacitance value and the quality factor to the amplitude phase of the quantum bit respectively, wherein the inductance value adopts a normalized logarithmic encoding form, the capacitance value adopts a piecewise linear encoding, and the quality factor is compressed to the [0,1] interval by a Sigmoid function; The mathematical expression for quantum state encoding of the inductance value, capacitance value and quality factor is: ； in, represents the superposition of quantum states, is the quantum bit amplitude corresponding to the inductance value, indicating the relative contribution of the inductance in the quantum state, is the quantum bit amplitude corresponding to the capacitance value, indicating the relative contribution of the capacitance in the quantum state, is the quantum bit amplitude corresponding to the quality factor, indicating the relative contribution of the quality factor in the quantum state, is the quantum state of the inductance value, is the quantum state of the capacitance value, is the quantum state of the quality factor; A quantized parameter space is obtained, and the particle positions in the quantized parameter space are initialized by a particle swarm optimization algorithm. The initialization includes: adjusting the search direction of the particle position by a quantum rotating gate based on a preset number of iterations and a fitness change rate, and dynamically adjusting the inertia weight of the particle swarm, wherein the inertia weight is used to control the speed and direction of the particle search; The quantum rotation angle of the quantum rotation gate is dynamically calculated by the gradient difference between the current parameter and the global optimal parameter. The mathematical expression of the rotation angle is: ； in, is the rotation angle, is a function that maps the two input gradients to rotation angles. is the gradient of the current parameter, is the gradient of the global parameter.

4. The intelligent debugging method of the LC filter according to claim 3, characterized in that: Using chaotic perturbation strategies to break local optimality includes: In the convergence stagnation stage of the particle swarm algorithm, the local optimum is broken by chaotic perturbation injection. After breaking the local optimum, the chaotic sequence is generated by Logistic mapping. The perturbation amplitude is inversely proportional to the fitness value of the current optimal solution. When in the iterative state, the large parameter space is optimized by quantum variational optimization or quantum simulated annealing, and when the preset number of iterations is reached, the final optimization result of the quantized particle swarm hybrid optimization debugging model is output.

5. The intelligent debugging method of the LC filter according to claim 4, characterized in that: Constructing the parameter change trajectory database includes: The total psychological needs of people are used as the output prediction parameters of the LC filter, and people's physiological, environmental and social needs are used as the quantum states of inductance, capacitance and quality factor respectively. The missing conditions, insufficient data, expected conditions and demand data of the subject corresponding to the LC filter are found, and iterative transfer calculation is performed based on knowledge distillation to establish the objective function of the LC filter. The mathematical expression of the objective function is: ； Where B is the output prediction parameter of the LC filter, is a representative function, which indicates the relationship between the performance of the LC filter output by the quantized particle swarm hybrid optimization debugging model and the debugging parameters, is the total loss function, represents the learning of debugging task data, Represents the difference between two iterative calculations, represents the output prediction parameter of the nth iteration, represents the output prediction parameter of the n+1th iteration, Represents the Softmax function, which is used to convert the output of the model into a probability distribution. is the weight hyperparameter, which is determined by the back-propagation gradient amplitude of the LC filter; All output prediction parameters in the iterative calculation are stored in the parameter change trajectory database.

6. The intelligent debugging method of the LC filter according to claim 5, characterized in that: The LSTM network uses a gated recurrent unit variant. The input is the initialization parameter set of the historical debugging task. The output layer predicts the optimal parameter direction of the next iteration stage through the Softmax function. Based on the learning results of the LSTM network on historical data, similar designs and parameter changes are quickly matched through the similarity matching algorithm to provide initial guesses for new debugging tasks in iterative calculations.

7. The intelligent debugging method of the LC filter according to claim 6, characterized in that: The dynamic constraint processing mechanism is used in the process of LSTM network to perform time series modeling on the parameter change trajectory database: The elastic boundary of the feasible solution is defined. The feasible interval of the inductance value is [L_min, L_max], and the feasible interval of the capacitance value is [C_min, C_max]. When the parameter is out of bounds, the mirror mapping method is used to project the out-of-bounds parameter to the neighborhood feasible point of the feasible interval.

8. An intelligent debugging system for an LC filter, the system is used to perform an intelligent debugging method for an LC filter, characterized in that: Including initialization parameter module, quantized particle swarm hybrid optimization debugging module, optimal prediction module and embedded hardware platform: The initialization parameter module is used to sample historical LC filter debugging task data from a multi-task distribution based on a meta-learning framework, learn the distribution characteristics of parameters through a pre-trained machine learning model, and generate an initialization parameter set suitable for the LC filter debugging task; The quantized particle swarm hybrid optimization debugging module is used to encode the inductance value, capacitance value and quality factor of the LC filter into a quantum state, and construct a quantized particle swarm hybrid optimization debugging model; The optimal prediction module is used to combine the concept of psychological field and knowledge distillation technology, map the high-dimensional parameter space to the low-dimensional potential space, build a parameter change trajectory database, perform time series modeling on the parameter change trajectory database through the LSTM network, and predict the optimal initial guess; The embedded hardware platform is used to run the meta-learning initialization module and the quantum-inspired optimization core algorithm, dynamic parameter update interface and temperature compensation.

9. The intelligent debugging system of LC filter according to claim 8, characterized in that: The embedded hardware platform comprises: Supports FPGA parallel computing when running the meta-learning initialization module and quantum-inspired optimization core algorithm; The dynamic parameter update interface includes online adjustment of debugging parameters through the LabVIEW graphical programming environment, and visual display of frequency domain response curves and optimization progress; Temperature compensation includes an integrated NTC thermistor network to monitor the inductor and capacitor temperature drift in real time and correct the parameter values ​​through a linear regression model.

10. The intelligent debugging system of LC filter according to claim 9, characterized in that: After debugging, the LC filter is evaluated for performance verification: The indicators include frequency domain performance, time domain performance, energy efficiency indicators and stability. The frequency domain performance includes cutoff frequency error, stopband attenuation and passband ripple, the time domain performance includes step response rise time and overshoot rate, the energy efficiency indicators include insertion loss and power loss, and the stability includes temperature drift coefficient and component parameter tolerance sensitivity.

Images