An avalanche signal extraction method and system based on LMS adaptive filtering

Abstract

The application discloses an avalanche signal extraction method and system based on LMS adaptive filtering, belongs to the technical field of quantum communication, and comprises the following steps: S10, connecting instruments, generating an optical signal through a pulse laser, and triggering an avalanche photodiode (APD); S20, providing a bias voltage and a gate signal for the avalanche photodiode (APD), converting a noisy signal into a digital signal through a preamplifier and an avalanche photodiode (ADC); S30, processing a sampling signal by using a two-stage LMS adaptive filter in a field programmable gate array (FPGA); S40, filtering out high-order harmonics of the processed signal through a low-pass filter (LPF); and S50, inputting the filtered signal into a photon counting unit to obtain a photon counting result. The application optimizes filter coefficients through iteration, minimizes the difference between a convolution result and the noisy signal, accurately filters out signals of a specified frequency, maximally suppresses sinusoidal gate noise and high-order harmonics thereof, and significantly improves the signal-to-noise ratio and detection efficiency of a detector.

Description

Technical Field

[0001] This invention relates to the field of quantum communication technology, and more specifically, to a method and system for avalanche signal extraction based on LMS adaptive filtering. Background Technology

[0002] Single-photon detectors (SPDs) are highly sensitive devices specifically designed to detect single photons, and are widely used in cutting-edge fields such as quantum communication, quantum computing, and quantum measurement. Photons, as carriers of quantum information, play a crucial role in quantum technology; therefore, the performance of SPDs directly impacts research progress and application realization in these fields. SPDs not only need to operate at extremely low photon flux densities but also require precise response while maintaining low dark count rates and high detection efficiency when detecting the arrival of a single photon. With the continuous development of quantum information technology, the demand for SPDs is increasing significantly, driving the research and development of new detection technologies. Currently, the applications of SPDs mainly focus on several types, including photomultiplier tubes (PMTs), avalanche photodiodes (APDs), superconducting nanowire single-photon detectors (SNSPDs), and quantum dot detectors. Each type of detector has its unique advantages and disadvantages. PMTs offer high sensitivity but are bulky and consume a lot of power, making them less than ideal for miniaturized and low-power applications. SNSPDs boast extremely high detection efficiency and low dark count rates, making them suitable for high-performance scenarios, but their operation requires extremely low temperatures, limiting their applications. In contrast, APDs, due to their relatively simple structure, low power consumption, and ease of integration, have become one of the most widely used single-photon detectors, especially in fields like quantum key distribution (QKD) where high sensitivity and high signal-to-noise ratio are crucial, demonstrating significant advantages.

[0003] With continuous technological advancements, the performance of APDs (Active Photons) is constantly improving, particularly in terms of dark count rate, detection efficiency, and response speed, ensuring their continued crucial role in future quantum information technology applications. APDs primarily operate in two modes: linear mode and Geiger mode. In linear mode, the APD operates at a low reverse bias. In this mode, incident photons generate electron-hole pairs and are amplified through the multiplication effect, but avalanche does not occur. In contrast, in Geiger mode, the APD operates at a high reverse bias, entering an avalanche breakdown state. In this mode, a single photon can trigger an avalanche effect, generating a significant current pulse. APDs in this mode can respond quickly and detect the arrival of a single photon.

[0004] However, APDs operating in Geiger mode face a significant challenge—recovery after avalanche breakdown. After each avalanche event, the APD must be rapidly quenched by external circuitry to prevent damage and restore it to a state where it can re-detect photons. The design of the quenching circuitry has a crucial impact on the APD's recovery time and efficiency. Common quenching methods include passive and active quenching. Passive quenching is typically achieved using high-impedance resistors; while less expensive, it has a longer recovery time, affecting detection speed. In contrast, active quenching rapidly cuts off the current through external circuitry, resulting in faster recovery and higher detection efficiency, and is therefore widely used in high-performance detection systems.

[0005] In active quenching circuits, gated mode is a common technique. Gated mode applies a timing-controlled voltage gate, placing the APD (Active Photon Detector) in a high reverse bias state only within a specific time window, thus ensuring that the APD detects incident photons only at specific times. This gated method effectively reduces the dark count rate and background noise, thereby improving the signal-to-noise ratio and detection accuracy. However, the application of gated mode also introduces a significant problem: gate noise. Gate noise is a false alarm signal generated during the APD recovery process due to unstable operating timing or external environmental interference. This noise affects the performance of the detection system and may even interfere with the extraction of truly valid signals. Therefore, effectively removing gate noise, especially high-order harmonic noise, has become a key issue in improving the accuracy and reliability of single-photon detection systems.

[0006] Existing inventions have some shortcomings in removing gating noise. For example, using a low-pass filter to remove gating noise often leads to distortion and amplitude attenuation of the avalanche signal during filtering, especially when the signal amplitude is small, which can easily cause missed detections and thus affect detection efficiency. While capacitor balancing methods can remove sinusoidal gating noise through capacitive balancing and differential techniques, their effectiveness is highly dependent on the capacitor matching degree. As the gating frequency increases, matching accuracy becomes difficult to guarantee, resulting in a decrease in noise suppression effectiveness. Therefore, these methods do not perform ideally under high-speed and high-sensitivity detection conditions, exhibiting low detection efficiency and high system complexity. Summary of the Invention

[0007] To overcome the shortcomings of existing technologies, an avalanche signal extraction method and system based on LMS adaptive filtering is proposed. This method uses a two-stage LMS adaptive filter to effectively reduce avalanche signals containing gating noise and higher-order harmonics, thereby significantly improving signal quality. Through iterative adjustment, the filter coefficients converge to a sine wave with the same frequency as the gating signal, thus enabling precise adjustment of the phase and amplitude of the convolution result, achieving accurate differential effects, and effectively removing gating noise and its higher-order harmonics.

[0008] The technical solution adopted by this invention to solve its technical problem is: an avalanche signal extraction method based on LMS adaptive filtering, the improvement of which is that the method includes the following steps:

[0009] S10: Connect the instrument, set the parameters and turn on the light source and power. Generate an optical signal through the pulsed laser. After being adjusted to the single-photon level by the optical attenuator, trigger the avalanche photodiode (APD).

[0010] S20: Provides bias voltage and applies gate signal to avalanche photodiode APD, and converts the avalanche signal containing gate noise into a digital signal through a preamplifier and analog-to-digital converter ADC;

[0011] S30: In a field-programmable gate array (FPGA), a two-stage LMS adaptive filter is used to process the sampled signal;

[0012] S40: Input the signal processed by the two-stage LMS adaptive filter into the low-pass filter LPF to filter out second-order and higher-order harmonics.

[0013] S50: Input the signal filtered by the low-pass filter (LPF) to the photon counting unit to obtain the photon counting result.

[0014] Furthermore, in step S30, the two-stage LMS adaptive filter includes a first-stage filter and a second-stage filter.

[0015] The first-stage filter filters out gated noise based on the gated frequency sample signal;

[0016] The second-stage filter is based on a sample signal at twice the gate frequency to filter out the first harmonic of the gate signal.

[0017] Furthermore, step S30 specifically involves the following steps:

[0018] S301: Initialize the filter tap coefficients to 0, call the sine wave sample signal that matches the gate frequency and its harmonics, and use it as the input of the first-stage filter and the second-stage filter respectively;

[0019] S302: After convolving the sinusoidal wave sample of the gated frequency with the tap coefficients of the first-stage filter, the difference is performed with the sampled signal to obtain the error signal;

[0020] S303: Adjusts the filter tap coefficients using the gradient descent method to output the first-order filtered signal;

[0021] S304: The error signal is obtained by convolving a sine wave sample with the tap coefficients of the second-stage filter at twice the gate frequency and then differentiating it with the first-order filtered signal.

[0022] S305: The filter tap coefficients are adjusted using the gradient descent method to output the second-order filtered signal.

[0023] Furthermore, in step S301, the formula for adjusting the filter tap coefficients of the LMS adaptive filter is:

[0024] w(n+1)=w(n)+2uX(n)e(n);

[0025] Where w(n) is the current tap coefficient, u is the learning rate, X(n) is the sinusoidal sample signal, and e(n) is the error signal.

[0026] Furthermore, in step S30, the formula for calculating the error signal during the filtering process is as follows:

[0027] e(n) = S(n) - y(n);

[0028] Where S(n) is the sampled signal and y(n) is the filter convolution output signal.

[0029] Furthermore, in step S40, the low-pass filter is a digital low-pass filter, and its cutoff frequency is 2-3 times the frequency of the gate signal.

[0030] Furthermore, the filtering process discards the first 5M sampled data to avoid the impact of instability in the initial iteration.

[0031] This invention also discloses an avalanche signal extraction system based on LMS adaptive filtering, the improvement of which includes:

[0032] Photonic signal source unit: used to generate optical signals and gating signals, triggering the avalanche photodiode (APD) to generate an avalanche signal mixed with gating noise;

[0033] Data processing unit: includes a preamplifier and an analog-to-digital converter (ADC), used to amplify and convert the avalanche signal output by the avalanche photodiode (APD) into a digital signal, and input it into a programmable gate array (FPGA) for filtering.

[0034] Signal filtering unit: includes:

[0035] The LMS filtering module contains two stages of LMS adaptive filters, used to filter out gated signals and their first harmonics;

[0036] The ROM module is used to store sine wave sample signals of different frequencies;

[0037] LPF low-pass filter module, used to filter out second-order and higher harmonics;

[0038] Photon counting unit: includes a hysteresis comparator and a counter, used to determine the occurrence of photon events and count the number of photons.

[0039] In the above structure, the photonic signal source unit includes:

[0040] A signal generator is used to output gating signals;

[0041] A timer is used to adjust the synchronization between the pulsed light source and the gating signal;

[0042] A semiconductor cooling element is used to stabilize the operating temperature of an avalanche diode (APD) to a set value.

[0043] Analog low-pass filter is used to filter signals into sine waves;

[0044] Lasers and variable optical attenuators are used to generate optical signals at the single-photon level.

[0045] In the above structure, the photon counting unit includes a Schmitt trigger to suppress false counts caused by signal jitter.

[0046] The beneficial effects of this invention are: by using a two-stage LMS adaptive filter, avalanche signals containing gating noise and higher-order harmonics can be effectively reduced, thereby significantly improving the signal quality. Through iterative adjustment, the filter coefficients will converge to a sine wave with the same frequency as the gating signal, thereby enabling precise adjustment of the phase and amplitude of the convolution result, achieving accurate differential effect, and effectively removing gating noise and its higher-order harmonics. Attached Figure Description

[0047] Figure 1 This is a flowchart of an avalanche signal extraction method based on LMS adaptive filtering according to the present invention;

[0048] Figure 2 This is a block diagram of an avalanche signal extraction system based on LMS adaptive filtering according to the present invention.

[0049] Figure 3 This is a schematic diagram of an avalanche signal extraction system based on LMS adaptive filtering according to the present invention.

[0050] Figure 4 This is a schematic diagram of the signal filtering unit of an avalanche signal extraction system based on LMS adaptive filtering according to the present invention.

[0051] Figure 5 This is a flowchart of the signal filtering unit of an avalanche signal extraction system based on LMS adaptive filtering according to the present invention.

[0052] Figure 6 This is a comparison frequency domain diagram of an avalanche signal extraction system based on LMS adaptive filtering according to the present invention, using this scheme; Detailed Implementation

[0053] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0054] The following will clearly and completely describe the concept, specific structure, and technical effects of the present invention in conjunction with embodiments and accompanying drawings, so as to fully understand the purpose, features, and effects of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, not all of them. Other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are all within the scope of protection of the present invention. Furthermore, all connections / linkages involved in the patent do not simply refer to direct contact between components, but rather to the ability to form a better connection structure by adding or reducing connecting accessories according to specific implementation conditions. The various technical features in this invention can be combined interactively without contradicting each other.

[0055] See Figures 1-2 As shown, this invention provides a method for avalanche signal extraction based on LMS adaptive filtering, which includes the following steps:

[0056] S10: Connect the instrument, set the parameters and turn on the light source and power. Generate an optical signal through the pulsed laser. After being adjusted to the single-photon level by the optical attenuator, trigger the avalanche photodiode (APD).

[0057] S20: Provides bias voltage and applies gate signal to avalanche photodiode APD, and converts the avalanche signal containing gate noise into a digital signal through a preamplifier and analog-to-digital converter ADC;

[0058] S30: In a field-programmable gate array (FPGA), a two-stage LMS adaptive filter is used to process the sampled signal;

[0059] S40: Input the signal processed by the two-stage LMS adaptive filter into the low-pass filter LPF to filter out second-order and higher-order harmonics.

[0060] S50: Input the signal filtered by the low-pass filter (LPF) to the photon counting unit to obtain the photon counting result.

[0061] In this invention, a two-stage LMS adaptive filter effectively reduces avalanche signals containing gating noise and higher-order harmonics, thereby significantly improving signal quality. The adaptive nature of the LMS algorithm allows the filter to dynamically adjust, automatically optimize the filtering effect, accurately remove unwanted noise, and improve the signal-to-noise ratio. Compared with traditional filtering methods, the LMS adaptive filter minimizes signal distortion and amplitude attenuation during the filtering process, ensuring the integrity of the avalanche signal. This is particularly important for single-photon signals requiring precise detection. Through iterative adjustment, the filter coefficients converge to a sine wave with the same frequency as the gating signal, thereby precisely adjusting the phase and amplitude of the convolution result to achieve accurate differential effects and effectively remove gating noise and its higher-order harmonics. The adaptive filtering mechanism and higher-order harmonic removal capability of this method enable the system to extract effective signals more accurately, thereby improving the detector's detection efficiency and sensitivity, especially in low photon flux density applications, providing higher detection accuracy. By effectively suppressing noise and higher-order harmonics, the probability of false alarms is reduced, thereby improving the reliability of the detection results, making it suitable for high-precision quantum communication, quantum computing, and other fields. This method performs signal processing in real time using a field-programmable gate array (FPGA), which has high flexibility and parallel processing capabilities, enabling the entire process to respond quickly and optimize signal processing in real time, thereby improving the system's real-time performance and response speed.

[0062] Furthermore, in step S30, the two-stage LMS adaptive filter includes a first-stage filter and a second-stage filter.

[0063] The first-stage filter filters out gated noise based on the gated frequency sample signal;

[0064] The second-stage filter is based on a sample signal at twice the gate frequency to filter out the first harmonic of the gate signal.

[0065] The first-stage filter focuses on noise suppression at the gate frequency, effectively removing noise components with the same frequency as the gate signal. This is because gate noise is usually close to the gate signal frequency, and filtering can directly remove this noise, preserving the original form of the gate signal. Harmonics in the gate signal often interfere with the accurate extraction of the target signal; the second-stage filter specifically filters out these harmonics, significantly improving signal purity. Utilizing the characteristics of the Least Mean Square (LMS) adaptive filter, the filter dynamically adjusts its coefficients according to changes in the input signal, ensuring that the filtering process can be optimized according to different noise conditions.

[0066] Furthermore, step S30 specifically involves the following steps:

[0067] S301: Initialize the filter tap coefficients to 0, call the sine wave sample signal that matches the gate frequency and its harmonics, and use it as the input of the first-stage filter and the second-stage filter respectively;

[0068] S302: After convolving the sinusoidal wave sample of the gated frequency with the tap coefficients of the first-stage filter, the difference is performed with the sampled signal to obtain the error signal;

[0069] S303: Adjusts the filter tap coefficients using the gradient descent method to output the first-order filtered signal;

[0070] S304: The error signal is obtained by convolving a sine wave sample with the tap coefficients of the second-stage filter at twice the gate frequency and then differentiating it with the first-order filtered signal.

[0071] S305: The filter tap coefficients are adjusted using the gradient descent method to output the second-order filtered signal.

[0072] This scheme employs two-stage filtering to more accurately remove or reduce noise associated with the gating frequency and its harmonics, thereby improving signal quality. This is crucial for communication systems, audio processing, or any scenario requiring high-precision signals. Using gradient descent to adjust the filter tap coefficients means the filter can adaptively adjust its parameters to best match the characteristics of the input signal. This adaptability allows the filter to remain effective as signal characteristics change over time. By precisely adjusting the filter tap coefficients, signal distortion during processing can be reduced. This is essential for maintaining the original characteristics and integrity of the signal.

[0073] like Figure 3 As shown, this invention provides an avalanche signal extraction system based on LMS adaptive filtering, comprising:

[0074] Photonic signal source unit: used to generate optical signals and gating signals, triggering the avalanche photodiode (APD) to generate an avalanche signal mixed with gating noise;

[0075] Data processing unit: includes a preamplifier and an analog-to-digital converter (ADC), used to amplify and convert the avalanche signal output by the avalanche photodiode (APD) into a digital signal, and input it into a programmable gate array (FPGA) for filtering.

[0076] Signal filtering unit: includes:

[0077] The LMS filtering module contains two stages of LMS adaptive filters, used to filter out gated signals and their first harmonics;

[0078] The ROM module is used to store sine wave sample signals of different frequencies;

[0079] LPF digital low-pass filter module, used to filter out second-order and higher harmonics;

[0080] Photon counting unit: Includes a hysteresis comparator and a counter, used to determine the occurrence of photon events and count photons. The hysteresis comparator prevents erroneous counting caused by minute signal jitter, and the counter counts the number of photons generated per second and sends the count to the host computer.

[0081] The photon counting unit may also include a Schmitt trigger to suppress false counting caused by signal jitter.

[0082] Reference Figure 4 and Figure 5 As shown, in the LMS filtering module, two M-bit empty arrays W1(n) and W2(n) are first initialized as the first and second filters, with all initial tap coefficients set to 0. A ROM module is established to store gating frequency samples of different frequencies, i.e., sine wave samples, with each sample containing one cycle of the sine wave at that frequency. In this embodiment, 16 samples are used. Sine wave samples corresponding to the gating frequency and its twice-multiple frequency are selected from the ROM and used as input signals X1(n) and Xn(n), respectively. The selected frequencies are 600 MHz (gating frequency) and 1200 MHz (twice the gating frequency). These samples are read cyclically and convolved with the filter coefficients W1(n) and W2(n), respectively.

[0083] In the first-stage LMS adaptive filtering, X1(n) is convolved with W1(n), and the convolution result is differentially divided with the sampled input signal of the ADC to obtain the error signal e1(n). Based on the value of the error signal, the tap coefficients of W1(n) are adjusted to decrease the error signal in its negative gradient direction. The filter coefficients W1(n) are initialized as empty filter tap coefficients with all zeros. The inputs are sinusoidally gated samples X1(n) and a noisy signal S(n).

[0084] Calculate the convolution output signal y(n) of the filter and adjust the tap coefficients according to the error signal e(n).

[0085] The formula used is: the error signal formula: e(n)=S(n)-y(n); where S(n) is the sampled signal and y(n) is the filter convolution output signal.

[0086] The gradient descent update formula is: w(n+1)=w(n)+2uX(n)e(n); where w(n) is the current tap coefficient, u is the learning rate (0.01 in this example), X(n) is the sinusoidal sample signal, and e(n) is the error signal.

[0087] Furthermore, in the initial iteration phase, the first 5M data points are discarded because the filter coefficients are not yet stable. Once the system stabilizes, the result e1(n) after first-order LMS adaptive filtering is obtained, thereby eliminating gating noise.

[0088] In the second-stage LMS adaptive filtering, e1(n) is used as the input signal for the second-order LMS. After convolving X2(n) with W2(n), the difference between X2(n) and e1(n) is obtained to yield the error result e2(n). Similarly, in the initial iteration phase, since the coefficients of the second-order LMS filter are not yet stable, the first 5M data points are discarded. After stabilization, the result e2(n) after second-order LMS adaptive filtering is obtained, thus eliminating the first-order harmonics generated by gating noise. The second-order filtered signal e2(n) is then output to the next module, the digital low-pass filter (LPF).

[0089] In the LPF digital low-pass filter module, a digital low-pass filter with a cutoff frequency 2.5 times that of the gate signal is used to filter out second-order and higher-order harmonic components. In this example, the cutoff frequency of the low-pass digital filter selected is 1500MHz. The processed result is as follows... Figure 5 As shown, since the avalanche signal is mainly concentrated in the range of 0 to 1 GHz, and the frequency distribution gradually decreases with increasing frequency, the second-order and higher-order harmonic components have very little impact on the avalanche signal. Therefore, higher-order harmonics can be effectively filtered out without affecting the avalanche signal.

[0090] The preferred embodiments of the present invention have been described in detail, but the present invention is not limited to the embodiments described. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention, and these equivalent modifications or substitutions are all included within the scope defined by the claims of this application.

Claims

1. A method for avalanche signal extraction based on LMS adaptive filtering, characterized in that, The method includes the following steps: S10: Connect the instrument, set the parameters and turn on the light source and power. Generate an optical signal through the pulsed laser. After being adjusted to the single-photon level by the optical attenuator, trigger the avalanche photodiode (APD). S20: Provides bias voltage and applies gate signal to avalanche photodiode APD, and converts the avalanche signal containing gate noise into a digital signal through a preamplifier and analog-to-digital converter ADC; S30: In a field-programmable gate array (FPGA), a two-stage LMS adaptive filter is used to process the sampled signal; in step S30, the two-stage LMS adaptive filter includes a first-stage filter and a second-stage filter; The first-stage filter filters out gated noise based on the gated frequency sample signal; The second-stage filter is based on a sample signal at twice the gate frequency, filtering out the first harmonic of the gate signal; The specific steps of step S30 are as follows: S301: Initialize the filter tap coefficients to 0, call the sine wave sample signal that matches the gate frequency and its harmonics, and use it as the input of the first-stage filter and the second-stage filter respectively; S302: After convolving the sinusoidal wave sample of the gated frequency with the tap coefficients of the first-stage filter, the difference is performed with the sampled signal to obtain the error signal; S303: Adjusts the filter tap coefficients using the gradient descent method to output the first-order filtered signal; S304: The error signal is obtained by convolving a sine wave sample with the tap coefficients of the second-stage filter at twice the gate frequency and then differentiating it with the first-order filtered signal. S305: The filter tap coefficients are adjusted using the gradient descent method to output the second-order filtered signal; S40: Input the signal processed by the two-stage LMS adaptive filter into the low-pass filter LPF to filter out second-order and higher-order harmonics. S50: Input the signal filtered by the low-pass filter (LPF) to the photon counting unit to obtain the photon counting result.

2. The avalanche signal extraction method based on LMS adaptive filtering according to claim 1, characterized in that, In step S301, the formula for adjusting the filter tap coefficients of the LMS adaptive filter is as follows: ； Where w(n) is the current tap coefficient, u is the learning rate, X(n) is the sinusoidal sample signal, and e(n) is the error signal.

3. The avalanche signal extraction method based on LMS adaptive filtering according to claim 1, characterized in that, In step S30, the formula for calculating the error signal during the filtering process is as follows: ； Where S(n) is the sampled signal and y(n) is the filter convolution output signal.

4. The avalanche signal extraction method based on LMS adaptive filtering according to claim 1, characterized in that, In step S40, the low-pass filter is a digital low-pass filter, and its cutoff frequency is 2-3 times the frequency of the gate signal.

5. The avalanche signal extraction method based on LMS adaptive filtering according to claim 1, characterized in that, In step S30, the first 5M sampled data are discarded during signal processing to avoid the effects of instability in the initial iteration.

6. An avalanche signal extraction system based on LMS adaptive filtering, applicable to the method described in claim 1, characterized in that, include: Photonic signal source unit: used to generate optical signals and gating signals, triggering the avalanche photodiode (APD) to generate an avalanche signal mixed with gating noise; Data processing unit: includes a preamplifier and an analog-to-digital converter (ADC), used to amplify and convert the avalanche signal output by the avalanche photodiode (APD) into a digital signal, and input it into a programmable gate array (FPGA) for filtering. Signal filtering unit: includes: The LMS filtering module contains two stages of LMS adaptive filters, used to filter out gated signals and their first harmonics; The ROM module is used to store sine wave sample signals of different frequencies; LPF digital low-pass filter module, used to filter out second-order and higher harmonics; Photon counting unit: includes a hysteresis comparator and a counter, used to determine the occurrence of photon events and count the number of photons.

7. The avalanche signal extraction system based on LMS adaptive filtering according to claim 6, characterized in that, The photon signal source unit includes: A signal generator is used to output gating signals; A timer is used to adjust the synchronization between the pulsed light source and the gating signal; A semiconductor cooling element is used to stabilize the operating temperature of an avalanche diode (APD) to a set value. Analog low-pass filter is used to filter signals into sine waves; Lasers and variable optical attenuators are used to generate optical signals at the single-photon level.

8. The avalanche signal extraction system based on LMS adaptive filtering according to claim 6, characterized in that, The photon counting unit includes a Schmitt trigger to suppress false counts caused by signal jitter.

Images