Quantum magnetic balance current monitoring device and method based on back propagation

Abstract

The invention relates to a quantum magnetic balance current monitoring device and method based on back propagation. The method comprises the steps that a microwave signal of a microwave source passes through a microwave circulator and a power amplifier and then is transmitted to a diamond NV color center through an antenna; a to-be-measured current generates an unbalanced magnetic field at the NV color center through the proportional coil. Green laser of the laser is divided into two paths by the optical splitter, one path enters the NV color center, and the other path enters the photoelectric coupler. Under the action of microwaves, green laser and an unbalanced magnetic field, the NV color center emits red fluorescence, and the fluorescence is converted into an electric signal after the interference of the green laser is eliminated by a photoelectric coupler. The electric signal is amplified, filtered and demodulated and then is input into the data processing module. And the data processing module adjusts the frequency of the microwave source in real time through the PID-BP control circuit, so that the first-order derivative spectrum of the ODMR spectral line is kept at a zero point, and magnetic flux locking and frequency tracking are realized. And calculating a required compensation current according to the physical structure of the system and providing compensation to complete magnetic balance.

Description

Technical Field

[0001] The present invention relates to a quantum magnetic balance current monitoring device and method based on reverse propagation, belonging to the technical field of current measurement and calibration. Background Art

[0002] A diamond NV center is a special defect in a diamond crystal, consisting of a nitrogen atom replacing a carbon atom and an adjacent vacancy (a missing carbon atom). This defect has unique electronic structure and spin properties, making it a promising material for quantum computing. Its electron spin energy levels can be used to encode quantum bits, and by applying physical methods such as lasers and microwaves, the NV center's electron spin can be initialized and read out, enabling the operation of quantum logic gates.

[0003] The backpropagation algorithm is a backward propagation algorithm based on neural networks. It adjusts the weights and biases of the neural network to minimize the network's output error. In the PID-BP algorithm, the network's input signal propagates forward to the output layer, generating an output error. The backpropagation algorithm then iteratively adjusts the network's weights and biases based on the magnitude and direction of the error until the output error reaches a preset threshold or the number of iterations reaches a limit.

[0004] While traditional current monitoring devices, such as Hall-effect elements and transistors, can measure current to a certain extent, their accuracy and sensitivity are often limited by various factors. Some devices may require additional circuitry or equipment to collect, process, and transmit data, increasing system complexity and cost.

[0005] Existing technology, such as Chinese patent number "CN117054952A", discloses a current transformer based on segmented magnetic measurement of diamond nitrogen-vacancy color centers. The device includes: a laser system, a microwave system, a diamond nitrogen-vacancy color center system, and a control and processing system; wherein the laser system is used to transmit laser light to the diamond nitrogen-vacancy color center system and process detection fluorescence output by the diamond nitrogen-vacancy color center system to obtain a first electrical signal; the microwave system is used to perform microwave frequency sweeping, radiate microwaves to the diamond nitrogen-vacancy color center system, and regulate the working magnetic field of the diamond nitrogen-vacancy color center system; the diamond nitrogen-vacancy color center system is used to modulate the laser light output by the laser system with the reflected fluorescence generated by irradiation by the laser system under the regulation of the microwave system to obtain detection fluorescence; the control and processing system is used to process the first electrical signal transmitted by the laser system to obtain a corresponding first current; and adjust the laser system and the microwave system until the working magnetic field of the diamond nitrogen-vacancy color center system covers the magnetic field formed by the first current.

[0006] A problem with the prior art is that the laser light output by the laser system and the reflected fluorescence modulated within the diamond nitrogen-vacancy color center system can be susceptible to environmental noise and signal interference, thus affecting the accuracy of the detection results. The control and processing system must fine-tune the laser and microwave systems to ensure that the operating magnetic field of the diamond nitrogen-vacancy color center system overlaps the magnetic field generated by the first current. This adjustment process can be cumbersome and requires a highly precise control algorithm. Summary of the Invention

[0007] In order to solve the above problems in the prior art, the present invention proposes a quantum magnetic balance current monitoring device and method based on back propagation.

[0008] The technical solutions of the present invention are as follows:

[0009] On the one hand, the present invention provides a quantum magnetic balance current monitoring device based on reverse propagation, including a microwave source, a microwave circulator, a microwave power amplifier, a microwave antenna, a diamond NV color center, a timing controller, a laser, a dichroic mirror, a photoelectric coupler, a differential amplifier module, a lock-in amplifier, a data processing module, a proportional coil part and a spectrometer;

[0010] The laser outputs green laser light, and the diamond NV color center outputs red fluorescence light;

[0011] The microwave source output end is connected to the microwave circulator input end; the microwave circulator output end is connected to the microwave power amplifier input end;

[0012] The output end of the microwave power amplifier provides microwaves to the diamond NV color center through a microwave antenna;

[0013] The diamond NV color center outputs red fluorescence to the input end of the photoelectric coupler; the output end of the photoelectric coupler is connected to the input end of the differential amplifier module;

[0014] The output end of the differential amplifier module is connected to the input end of the lock-in amplifier; the output end of the lock-in amplifier is connected to the input end of the data processing module;

[0015] The output end of the data processing module is connected to the input end of the microwave source and the input end of the proportional coil part; the proportional coil part provides a magnetic field corresponding to the unbalanced magnetic potential for the diamond NV color center;

[0016] The output end of the timing controller is connected to the microwave source input end and the laser input end respectively, and is used to control the start-up of the microwave source and the laser;

[0017] The laser outputs green laser light to the input end of the photoelectric coupler through the beam splitter, and the laser also provides green laser light to the diamond NV color center through the beam splitter and the dichroic mirror;

[0018] The dichroic mirror is configured to allow green laser light to pass through;

[0019] The photoelectric coupler is used to eliminate the interference of the green laser signal mixed in the red fluorescence and convert the red fluorescence into an electrical signal;

[0020] The differential amplifier module is used to amplify the electrical signal output by the photoelectric coupler;

[0021] The data processing module includes a PID-BP control circuit for calculating the current to be compensated for the proportional coil part and providing a control signal for the microwave source;

[0022] The proportional coil part is used to receive the current to be measured.

[0023] As a preferred embodiment, the laser power is 100 milliwatts and the wavelength is 532 nanometers.

[0024] As a preferred embodiment, the dichroic mirror is configured to prevent the red fluorescence from returning to the original light path.

[0025] On the other hand, the present invention also provides a method for monitoring quantum magnetic balance current based on reverse propagation, comprising the following steps:

[0026] The microwaves emitted by the microwave source are sequentially input into the diamond NV color center through a microwave circulator, a microwave power amplifier, and a microwave antenna;

[0027] The current to be measured is input into the proportional coil part, and the proportional coil part provides a magnetic field corresponding to the unbalanced magnetic potential to the diamond NV color center;

[0028] The green laser emitted by the laser is split into two paths through a beam splitter. One path enters the diamond NV color center through a dichroic mirror, and the other path enters the photoelectric coupler.

[0029] The diamond NV color center generates red fluorescence under the action of the microwave, green laser and magnetic field corresponding to the unbalanced magnetic potential, and inputs the red fluorescence into the photoelectric coupler;

[0030] The photoelectric coupler eliminates the green laser signal interference mixed with the red fluorescence based on the green laser output by the spectrometer and converts the red fluorescence into an electrical signal. The electrical signal is amplified by the differential amplifier module and filtered and demodulated by the lock-in amplifier before being input into the data processing module.

[0031] The PID-BP control circuit of the data processing module outputs a control signal to the microwave source based on the modulated and demodulated electrical signal output by the phase-locked amplifier to adjust the microwave source's emission frequency, so that the electrical signal closely follows the zero-crossing point of the phase-locked demodulation curve, achieving closed-loop tracking of the center frequency. At the same time, it calculates the real-time magnetic field and the magnitude of the current to be compensated based on the physical topology and magnetic field.

[0032] Finally, the compensation current calculated by the data processing module is applied to the proportional coil to achieve magnetic balance.

[0033] As a preferred embodiment, the BP neural network structure of the PID-BP control circuit includes an input layer, a hidden layer and an output layer;

[0034] The input layer consists of four nodes: set value r(k), fundamental wave y(k), error value e(k) and bias;

[0035] The hidden layer is layers, and each layer uses neurons;

[0036] The output layer includes nodes: proportional gain K P , integral gain K I and the differential gain K D .

[0037] As a preferred embodiment, the hidden layer of the BP neural network uses a Tanh activation function.

[0038] As a preferred embodiment, the output layer of the BP neural network uses a LeakyReLU activation function.

[0039] As a preferred embodiment, the method for the PID-BP control circuit to output the control signal is:

[0040] The electrical signal output by the lock-in amplifier is used as the error value e(k), the set value r(k) is set to , the bias is set to , and the first-order derivative of the ODMR spectrum is used as the fundamental wave y(k), and input into the BP neural network input layer;

[0041] The set value r(k), fundamental wave y(k), error value e(k) and bias, as well as the set value r(k), fundamental wave y(k), error value e(k) and bias preset weight W, are passed to the hidden layer and then to the output layer;

[0042] The output of the output layer is used as the control parameter of the PID-BP control circuit:

[0043] K P =O n (k)1;

[0044] K I =On (k)2;

[0045] K D =O n (k)3;

[0046] Among them, O n (k)1, O n (k)2, O n (k)3 represents the output of the BP neural network output layer;

[0047] The output control signal is adjusted through the control algorithm:

[0048] Δu(k)=K P [e(k)-e(k-1)]+K I e(k)+K D [e(k)-2e(k-1+ek-2;

[0049] Wherein, Δu(k) represents the control signal and e(k-1) represents the previous error value.

[0050] As a preferred embodiment, the method of adjusting the BP neural network by the back propagation algorithm is:

[0051] Calculate the mean square error (MSE) between the error value e(k) and the set value r(k), and use the mean square error (MSE) as the loss function L:

[0052]

[0053] By minimizing the loss function L, the fundamental wave y(k) value approaches;

[0054] The back propagation algorithm is used to calculate the gradient corresponding to the set value r(k), fundamental wave y(k), error value e(k) and bias weight of the BP neural network input layer.

[0055] Where W represents the set of set value r(k), fundamental wave y(k), error value e(k) and bias weight;

[0056] According to the learning rate and the gradient calculated by the back propagation algorithm, the set value r(k), fundamental wave y(k), error value e(k) and bias weight of the BP neural network input layer are adjusted:

[0057]

[0058] Among them, W new Represents the adjusted weight, W old represents the weight before adjustment, η represents the learning rate, Represents the gradient descent update parameters.

[0059] As a preferred embodiment, the proportional gain K of the BP neural network output layer P , integral gain K I and the differential gain K D is the control parameter of the PID-BP control circuit;

[0060] The PID controller of the PID-BP control circuit is used to adjust the RF frequency of the microwave source:

[0061] f=f+Δu(k);

[0062] Where f represents the transmission frequency.

[0063] The present invention has the following beneficial effects:

[0064] The present invention introduces back propagation technology to adjust PID control parameters. The back propagation mechanism gives the neural network the ability to adaptively adjust weights, that is, the ability to train online. The BP neural network is introduced to dynamically adjust the PID control parameters and optimize the control performance. By adaptively optimizing the PID parameters, the controller can better cope with nonlinear and dynamically changing systems, and solves the lock-out problem of the fixed parameter PID algorithm when controlling the nonlinear quantum magnetometer. The neural network process has no testing phase, only a training process, which is like training a model all the time. It is equivalent to using the gradient descent method of the neural network to achieve parameter optimization and then control the radio frequency. By using diamond NV color centers as magnetic detectors, the 1 / f noise and current jump problems of the SQUID detector in the prior art are solved, and the measurement stability is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 It is a schematic diagram of the structure of the device of the present invention.

[0066] Reference numerals:

[0067] 1. Microwave source; 2. Microwave circulator; 3. Microwave power amplifier; 4. Magnetic field corresponding to unbalanced magnetic potential; 5. Microwave antenna; 6. Diamond NV color center; 7. Timing controller; 8. Laser; 9. Dichroic mirror; 10. Optocoupler; 11. Differential amplifier module; 12. Lock-in amplifier; 13. Data processing module; 14. Proportional coil part; 15. Spectrum splitter. DETAILED DESCRIPTION

[0068] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0069] It should be understood that the step numbers used herein are only for convenience of description and are not intended to limit the order in which the steps are to be executed.

[0070] It should be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, the singular forms "a", "an" and "the" are intended to include the plural forms unless the context clearly indicates otherwise.

[0071] The terms “include” and “comprising” indicate the presence of described features, integers, steps, operations, elements and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof.

[0072] The term "and / or" refers to and includes any and all possible combinations of one or more of the associated listed items.

[0073] Diamond NV centers are used to measure external magnetic fields. The measurement principle is based on the relationship between the magnetic field strength and the frequency difference between the troughs of the ODMR (Optically Detected Magnetic Resonance) spectrum. The magnetic field strength along the same axis is linearly related to the frequency difference between the two troughs. In practical applications, a magnetic measurement system based on diamond NV centers requires a microwave source to sweep the frequency range within the zero-field splitting frequency D (e.g., 2.6 GHz to 3.1 GHz). The fluorescence intensity is collected at each frequency, ultimately forming the ODMR spectrum, which is then used to determine the resonant frequency.

[0074] To improve measurement accuracy, the microwave signal is frequency modulated (FM) to avoid noise-prone frequency bands. The signal is then demodulated using phase-locked technology, resulting in the first-order derivative of the ODMR spectrum. According to the mathematical properties of first-order differentials, the zero-crossing point of the first-order derivative corresponds to the trough of the ODMR spectrum. By feeding back the changes in fluorescence intensity caused by the phase-locked offset to the microwave source, the microwave source can adjust its output frequency so that each measurement is locked to the zero-crossing point of the demodulated signal. This achieves magnetic flux locking, and locks onto the most sensitive point of the NV color center's magnetic field.

[0075] In a quantum magnetometer, when the applied radio frequency microwave frequency is equal to the Larmor frequency, the resonance signal reaches a minimum value. However, since it is difficult to directly determine the lowest point of the resonance curve in practical applications, a key feature can be used: the first-order signal (fundamental wave) value at the lowest point of the resonance curve must be zero. Based on this feature, the first-order signal value can be used as the controlled quantity, and the radio frequency frequency can be used as the controlled quantity. By adjusting the radio frequency frequency, the first-order signal value changes caused by the changes in the measured magnetic field are offset, thereby locking the first-order signal value at zero. At this point, the radio frequency frequency is equal to the Larmor frequency, so that the quantum magnetometer can track the Larmor frequency in real time and measure the magnetic field. The first-order signal is the electrical signal described in Example 2.

[0076] By exciting the NV color center with an RF microwave source, a signal indicating a change in the fluorescence intensity of the NV color center's radiation is obtained. In actual operation, the optical signal is converted into a current signal and then into a voltage signal through a current-to-voltage circuit. When the RF magnetic field is swept, a resonance curve can be plotted by collecting these voltage signals. To more quickly lock onto the Larmor frequency, the RF magnetic field is modulated and a narrowband bandpass filter is used to extract the first-order signal. Subsequently, a phase-locked amplifier circuit is designed to demodulate the first-order signal, and the data obtained can provide input for the subsequent Larmor frequency tracking algorithm. Through this series of steps, the Larmor frequency can be accurately and quickly determined.

[0077] Example 1:

[0078] See also Figure 1 The present invention provides a quantum magnetic balance current monitoring device based on reverse propagation, comprising a microwave source 1, a microwave circulator 2, a microwave power amplifier 3, a microwave antenna 5, a diamond NV color center 6, a timing controller 7, a laser 8, a dichroic mirror 9, a photoelectric coupler 10, a differential amplifier module 11, a lock-in amplifier 12, a data processing module 13, a proportional coil part 14, and a spectrometer 15;

[0079] The laser 8 outputs green laser light, and the diamond NV color center 6 outputs red fluorescence;

[0080] The output end of the microwave source 1 is connected to the input end of the microwave circulator 2; the output end of the microwave circulator 2 is connected to the input end of the microwave power amplifier 3;

[0081] The output end of the microwave power amplifier 3 provides microwaves to the diamond NV color center 6 through the microwave antenna 5;

[0082] The diamond NV color center 6 outputs red fluorescence to the input end of the photoelectric coupler 10; the output end of the photoelectric coupler 10 is connected to the input end of the differential amplifier module 11;

[0083] The output end of the differential amplifier module 11 is connected to the input end of the lock-in amplifier 12; the output end of the lock-in amplifier 12 is connected to the input end of the data processing module 13;

[0084] The output end of the data processing module 13 is connected to the input end of the microwave source 1 and the input end of the proportional coil part 14; the proportional coil part 14 provides a magnetic field 4 corresponding to the unbalanced magnetic potential for the diamond NV color center 6;

[0085] The output end of the timing controller 7 is connected to the input end of the microwave source 1 and the input end of the laser 8 respectively, and is used to control the start of the microwave source 1 and the laser 8;

[0086] The laser 8 outputs green laser light to the input end of the photoelectric coupler 10 through the beam splitter 15. The laser 8 also provides green laser light to the diamond NV color center 6 through the beam splitter 15 and the dichroic mirror 9.

[0087] The dichroic mirror 9 is configured to allow the green laser to pass through;

[0088] The photoelectric coupler 10 is used to eliminate the interference of the green laser signal mixed in the red fluorescence and convert the red fluorescence into an electrical signal;

[0089] The differential amplifier module 11 is used to amplify the electrical signal output by the photoelectric coupler 10;

[0090] The data processing module 13 includes a PID-BP control circuit for calculating the current to be compensated by the proportional coil part 14 and providing a control signal for the microwave source 1;

[0091] The proportional coil portion 14 is used to receive the current to be measured.

[0092] As a preferred embodiment, the laser 8 has a power of 100 milliwatts and a wavelength of 532 nanometers.

[0093] As a preferred embodiment, the dichroic mirror 9 is configured to prevent the red fluorescence from returning to the original light path.

[0094] The PID-BP algorithm controls the frequency of the radio frequency microwaves, achieving "real-time tracking of the electron spin resonance frequency." This is because only at specific frequencies will a sudden change in red fluorescence occur, marking the observation point. This allows the real-time magnetic field strength to be measured, resulting in real-time magnetic flux. Because magnetic flux still exists, but the magnetic field strength is very weak, quantum sensors can precisely measure this very small magnetic field strength.

[0095] The magnetic field strength B is used as the observed value. When it is small enough, I* at this time can be considered as the final current value required on the secondary side.

[0096] By observing the magnetic flux, combining the primary and secondary side ampere-turn balance with the physical topology of the sensor, the current is finally calculated.

[0097] PID (Proportional-Integral-Derivative) controllers are widely used in engineering due to their simple structure, excellent stability, and reliable operation. BP (Back Propagation of Error) neural networks, a common artificial neural network model, possess powerful nonlinear mapping and approximation capabilities. Combining a PID controller with a BP neural network enables more precise current measurement control, thereby improving the accuracy of current measurement of diamond NV color centers. The PID-BP control circuit dynamically adjusts control parameters based on real-time error feedback, thereby optimizing the measurement process.

[0098] Example 2:

[0099] The present invention also provides a method for monitoring quantum magnetic balance current based on reverse propagation, comprising the following steps:

[0100] The microwaves emitted by the microwave source 1 are sequentially input to the diamond NV color center 6 through the microwave circulator 2, the microwave power amplifier 3, and the microwave antenna 5;

[0101] The current to be measured is input into the proportional coil portion 14, and the proportional coil portion 14 provides a magnetic field 4 corresponding to the unbalanced magnetic potential to the diamond NV color center 6;

[0102] The green laser light emitted by the laser 8 is split into two paths by the beam splitter 15. One path enters the diamond NV color center 6 through the dichroic mirror 9, and the other path enters the photoelectric coupler 10.

[0103] The diamond NV color center 6 generates red fluorescence under the action of the microwave, green laser and the magnetic field 4 corresponding to the unbalanced magnetic potential, and inputs the red fluorescence into the photoelectric coupler 10;

[0104] The photoelectric coupler 10 eliminates the green laser signal interference mixed with the red fluorescence according to the green laser output by the spectrometer 15, and converts the red fluorescence into an electrical signal; the electrical signal is amplified by the differential amplifier module 11 and filtered and demodulated by the lock-in amplifier 12 before being input into the data processing module 13;

[0105] The PID-BP control circuit of the data processing module 13 outputs a control signal to the microwave source 1 according to the modulated and demodulated electrical signal output by the lock-in amplifier 12 to adjust the transmission frequency of the microwave source 1, so that the electrical signal closely follows the zero-crossing point of the phase-locked demodulation curve to achieve closed-loop tracking of the center frequency. At the same time, the real-time magnetic field is calculated, and the magnitude of the current to be compensated is calculated based on the physical topology and the magnetic field.

[0106] Finally, the compensation current calculated by the data processing module 13 is compensated to the proportional coil part 14 to achieve magnetic balance.

[0107] The output of the BP neural network is the three parameters of the PID-BP control circuit, and ultimately the PID-BP control circuit is used to adjust the RF frequency.

[0108] As a preferred embodiment, the BP neural network structure of the PID-BP control circuit includes an input layer, a hidden layer and an output layer;

[0109] The input layer consists of four nodes: set value r(k), fundamental wave y(k), error value e(k) and bias;

[0110] There are 2 hidden layers, each with 8 neurons;

[0111] The output layer consists of 3 nodes: proportional gain K P , integral gain K I and the differential gain K D .

[0112] As a preferred embodiment, the hidden layer of the BP neural network uses a Tanh activation function.

[0113] As a preferred embodiment, the output layer of the BP neural network uses a LeakyReLU activation function.

[0114] As a preferred embodiment, the method for the PID-BP control circuit to output the control signal is:

[0115] The electrical signal output by the lock-in amplifier 12 is used as the error value e(k), the set value r(k) is set to 0, the bias is set to 1, and the first-order derivative of the ODMR spectrum is used as the fundamental wave y(k), and input into the BP neural network input layer;

[0116] The set value r(k), fundamental wave y(k), error value e(k) and bias, as well as the set value r(k), fundamental wave y(k), error value e(k) and bias preset weight W, are passed to the hidden layer and then to the output layer;

[0117] The output of the output layer is used as the control parameter of the PID-BP control circuit:

[0118] K P =O n (k)1;

[0119] K I =O n (k)2;

[0120] K D =On (k)3;

[0121] Among them, O n (k)1, O n (k)2, O n (k)3 represents the output of the BP neural network output layer;

[0122] The output control signal is adjusted through the control algorithm:

[0123] Δu(k)=K P [e(k)-e(k-1)]+K I e(k)+K D [e(k)-2e(k-1+ek-2;

[0124] Wherein, Δu(k) represents the control signal and e(k-1) represents the previous error value.

[0125] As a preferred embodiment, the method of adjusting the BP neural network by the back propagation algorithm is:

[0126] Calculate the mean square error (MSE) between the error value e(k) and the set value r(k), and use the mean square error (MSE) as the loss function L:

[0127]

[0128] By minimizing the loss function L, the fundamental wave y(k) value is made close to 0;

[0129] The back propagation algorithm is used to calculate the gradient corresponding to the set value r(k), fundamental wave y(k), error value e(k) and bias weight of the BP neural network input layer.

[0130] Where W represents the set of set value r(k), fundamental wave y(k), error value e(k) and bias weight;

[0131] According to the learning rate and the gradient calculated by the back propagation algorithm, the set value r(k), fundamental wave y(k), error value e(k) and bias weight of the BP neural network input layer are adjusted:

[0132]

[0133] Among them, W new Represents the adjusted weight, W old represents the weight before adjustment, η represents the learning rate, Represents the gradient descent update parameters.

[0134] Through the adjustment of the BP neural network, the changes in fluorescence intensity caused by phase-locked offset are fed back to the microwave source. The microwave source can adjust the output frequency so that each measurement is locked at the zero-crossing point of the demodulated signal. In this way, the magnetic flux is locked and locked at the most sensitive point of the NV color center to measure the magnetic field.

[0135] As a preferred embodiment, the proportional gain K of the BP neural network output layer P , integral gain K I and the differential gain K D is the control parameter of the PID-BP control circuit;

[0136] The PID controller of the PID-BP control circuit is used to adjust the radio frequency frequency of the microwave source 1:

[0137] f=f+Δu(k);

[0138] Where f represents the transmission frequency.

[0139] Through the above method, the combination of BP neural network and PID controller is used to achieve accurate monitoring of quantum magnetic balance current and dynamic adjustment of radio frequency frequency.

[0140] In the embodiments of the present application, "at least one" refers to one or more, and "more" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent the existence of A alone, the existence of A and B at the same time, and the existence of B alone. Among them, A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following" and similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b and c can represent: a, b, c, a and b, a and c, b and c or a and b and c, where a, b, c can be single or multiple.

[0141] Those skilled in the art will appreciate that the various units and algorithm steps described in the embodiments disclosed herein can be implemented using a combination of electronic hardware, computer software, and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0142] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0143] In the several embodiments provided in this application, if any function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of this application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory; hereinafter referred to as: ROM), random access memory (Random Access Memory; hereinafter referred to as: RAM), magnetic disk or optical disk, and other media that can store program code.

[0144] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention's description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A quantum magnetic balance current monitoring device based on reverse propagation, characterized in that: It includes a microwave source (1), a microwave circulator (2), a microwave power amplifier (3), a microwave antenna (5), a diamond NV color center (6), a timing controller (7), a laser (8), a dichroic mirror (9), a photoelectric coupler (10), a differential amplifier module (11), a phase-locked amplifier (12), a data processing module (13), a proportional coil part (14) and a spectrometer (15); The laser (8) outputs green laser light, and the diamond NV color center (6) outputs red fluorescence light; The output end of the microwave source (1) is connected to the input end of the microwave circulator (2); the output end of the microwave circulator (2) is connected to the input end of the microwave power amplifier (3); The output end of the microwave power amplifier (3) provides microwaves to the diamond NV color center (6) through a microwave antenna (5); The diamond NV color center (6) outputs red fluorescence to the input end of the photoelectric coupler (10); the output end of the photoelectric coupler (10) is connected to the input end of the differential amplifier module (11); The output end of the differential amplifier module (11) is connected to the input end of the phase-locked amplifier (12); the output end of the phase-locked amplifier (12) is connected to the input end of the data processing module (13); The output end of the data processing module (13) is connected to the input end of the microwave source (1) and the input end of the proportional coil part (14); the proportional coil part (14) provides a magnetic field (4) corresponding to the unbalanced magnetic potential for the diamond NV color center (6); The output end of the timing controller (7) is respectively connected to the input end of the microwave source (1) and the input end of the laser (8), and is used to control the start-up of the microwave source (1) and the laser (8); The laser (8) outputs green laser light to the input end of the photoelectric coupler (10) through the beam splitter (15), and the laser (8) also provides green laser light to the diamond NV color center (6) through the beam splitter (15) and the dichroic mirror (9); The dichroic mirror (9) is configured to allow green laser light to pass through; The photoelectric coupler (10) is used to eliminate the interference of the green laser signal mixed in the red fluorescence and convert the red fluorescence into an electrical signal; The differential amplification module (11) is used to amplify the electrical signal output by the photoelectric coupler (10); The data processing module (13) comprises a PID-BP control circuit, which is used to calculate the current to be compensated for the proportional coil part (14) and to provide a control signal for the microwave source (1); The proportional coil part (14) is used for receiving the current to be measured.

2. The quantum magnetic balance current monitoring device based on reverse propagation according to claim 1 is characterized in that: The power of the laser (8) is 100 milliwatts and the wavelength is 532 nanometers.

3. The back-propagation based quantum magnetic balance current monitoring device according to claim 1, characterized in that: The dichroic mirror (9) is arranged to prevent the red fluorescence from returning to the original light path.

4. A quantum magnetic balance current monitoring method based on reverse propagation, characterized in that: The following steps are involved: The microwaves emitted by the microwave source (1) are sequentially input into the diamond NV color center (6) through a microwave circulator (2), a microwave power amplifier (3), and a microwave antenna (5); The current to be measured is input into the proportional coil part (14), and a magnetic field (4) corresponding to the unbalanced magnetic potential is provided to the diamond NV color center (6) through the proportional coil part (14); The green laser light emitted by the laser (8) is divided into two paths through a beam splitter (15), one path enters the diamond NV color center (6) through a dichroic mirror (9), and the other path enters the photoelectric coupler (10); The diamond NV color center (6) generates red fluorescence under the action of the microwave, the green laser and the magnetic field (4) corresponding to the unbalanced magnetic potential, and inputs the red fluorescence into the photoelectric coupler (10); The photoelectric coupler (10) eliminates the interference of the green laser signal mixed in the red fluorescence according to the green laser output by the spectrometer (15), and converts the red fluorescence into an electrical signal; the electrical signal is amplified by the differential amplifier module (11) and filtered and demodulated by the lock-in amplifier (12), and then input into the data processing module (13); The PID-BP control circuit of the data processing module (13) outputs a control signal to the microwave source (1) to adjust the emission frequency of the microwave source (1) according to the modulated and demodulated electrical signal output by the phase-locked amplifier (12), so that the electrical signal closely follows the zero-crossing point of the phase-locked demodulation curve to achieve closed-loop tracking of the center frequency, and simultaneously calculates the real-time magnetic field and calculates the magnitude of the current to be compensated according to the physical topological structure and the magnetic field; Finally, the compensation current calculated by the data processing module (13) is compensated to the proportional coil part (14) to achieve magnetic balance.

5. The method for monitoring quantum magnetic balance current based on reverse propagation according to claim 4, characterized in that: The BP neural network structure of the PID-BP control circuit includes an input layer, a hidden layer and an output layer; The input layer consists of four nodes: set value r(k), fundamental wave y(k), error value e(k) and bias; There are 2 hidden layers, each with 8 neurons; The output layer consists of 3 nodes: proportional gain K P , integral gain K I And the differential gain K D .

6. The method for monitoring quantum magnetic balance current based on reverse propagation according to claim 5, characterized in that: The hidden layer of the BP neural network uses the Tanh activation function.

7. The method for monitoring quantum magnetic balance current based on reverse propagation according to claim 5, characterized in that: The output layer of the BP neural network uses the LeakyReLU activation function.

8. The method for monitoring quantum magnetic balance current based on reverse propagation according to claim 5, characterized in that: The method for the PID-BP control circuit to output a control signal is: The electric signal output by the lock-in amplifier (12) is used as an error value e(k), the set value r(k) is set to 0, the bias is set to 1, and the first-order derivative of the ODMR spectrum is used as a fundamental wave y(k), and inputted into the BP neural network input layer; The set value r(k), fundamental wave y(k), error value e(k) and bias, as well as the set value r(k), fundamental wave y(k), error value e(k) and bias preset weight W, are passed to the hidden layer and then to the output layer; The output of the output layer is used as the control parameter of the PID-BP control circuit: K P =O n (k)1; K I =O n (k)2; K D =O n (k)3; Among them, O n (k) 1. O n (k)2, O n (k)3 represents the output of the output layer of the BP neural network; The output control signal is adjusted through the control algorithm: Δu(k)=K P [e(k)-e(k-1)]+K I e(k)+K D [e(k)-2e(k-1)+e(k-2)]; Among them, Δu(k) represents the control signal, and e(k-1) represents the previous error value.

9. The method for monitoring quantum magnetic balance current based on reverse propagation according to claim 8, characterized in that: The method of adjusting the BP neural network through the back propagation algorithm is: Calculate the mean square error (MSE) between the error value e(k) and the set value r(k), and use the mean square error (MSE) as the loss function L: By minimizing the loss function L, the fundamental wave y(k) value is made close to 0; The back propagation algorithm is used to calculate the set value r(k), fundamental wave y(k), error value e(k) and the gradient corresponding to the bias weight of the BP neural network input layer. Wherein, W represents the set of set value r(k), fundamental wave y(k), error value e(k) and bias weight; According to the learning rate and the gradient calculated by the back propagation algorithm, the set value r(k), base wave y(k), error value e(k) and bias weight of the BP neural network input layer are adjusted: Among them, W new represents the adjusted weight, W old represents the weight before adjustment, η represents the learning rate, Represents the gradient descent update parameters.

10. The method for monitoring quantum magnetic balance current based on reverse propagation according to claim 9, characterized in that: The proportional gain K of the output layer of the BP neural network P , integral gain K I And the differential gain K D is the control parameter of the PID-BP control circuit; The PID controller of the PID-BP control circuit is used to adjust the radio frequency frequency of the microwave source (1): f=f+Δu(k); Wherein, f represents the transmission frequency.

Images