Method of detecting based on optical signals

Abstract

This invention discloses a detection method based on optical signals. The method includes: modulating a continuous optical signal based on data; performing phase modulation on the modulated optical signal based on key information to generate a co-occurring optical pulse; splitting the co-occurring optical pulse into a first optical path signal and a second optical path signal; performing delay matching on the second optical path signal based on the transmission path of the first optical path signal in the channel; performing zero-difference interferometry measurement on the first optical path signal and the local oscillator reference light to extract the phase modulation information; determining channel evaluation parameters based on the phase modulation information; selecting a target serving base station and its corresponding target communication link from multiple base stations; performing a handover to the target communication link; sending an optical signal carrying the phase modulation information to a central measurement node through the target serving base station; and having the central measurement node perform Bell state measurement on the optical signals from two different communication parties.

Description

Technical Field

[0001] This invention relates to the field of communication technology, and more specifically, to a method for detecting optical signals. Background Technology

[0002] As fifth-generation mobile communication networks evolve towards higher speeds, lower latency, and wider connectivity, information security technologies based on key distribution have garnered widespread attention due to their theoretically provable unconditional security. Existing solutions typically employ a physically isolated parallel architecture, where a dedicated signal transmission and reception device is independently deployed alongside the operator's base station infrastructure, configured with a dedicated fiber optic link or free-space optical channel. This architecture ensures that high-power classical radio frequency signals and extremely weak light signals do not interfere with each other, thus maintaining the necessary coherence and security. While this isolation paradigm brings basic security, it also raises engineering and economic issues, directly leading to redundant construction of network infrastructure and significantly increasing operators' investment costs in spectrum, space, and hardware. More importantly, this static, point-to-point link construction model fundamentally contradicts the core advantages of dense networking, dynamic resource scheduling, and seamless user handover advocated by the network itself. Key distribution services are fixed on specific physical paths and cannot, like data services, flexibly and adaptively reconfigure resources and switch links based on user movement, network load fluctuations, and real-time changes in channel quality.

[0003] At its core, the signal, as a strong classical electromagnetic wave, is designed and optimized around power, spectral efficiency, and anti-interference capability. However, the key distribution signal operates at the single-photon or extremely weak light level, making it extremely sensitive to channel noise, especially crosstalk and nonlinear effects from strong classical signals in the same frequency band. Traditional channel isolation strategies, while effectively avoiding devastating interference between signals, also completely sever the possibility of deep integration and collaboration between the two systems at the physical and network layers. This prevents the network from fully utilizing its highly mature dense access points, efficient backhaul network, and intelligent scheduling mechanism. Summary of the Invention

[0004] To address the problems existing in the prior art, the present invention aims to provide a method for detecting optical signals that can achieve co-channel collaborative transmission of classical signals and key signals, avoid redundant infrastructure construction, adapt to the characteristics of dense networking through dynamic resource scheduling and smooth link switching, suppress channel disturbances, ensure unconditional key security, and balance the economic efficiency of network deployment with communication security and continuity.

[0005] To solve the above problems, the present invention adopts the following technical solution.

[0006] The method for detecting optical signals includes the following steps: Step 1: Provide continuous light, modulate the continuous light based on data, and perform phase modulation on the modulated light based on key information to generate a co-occurring light pulse; Step 2: Split the co-occurring optical pulse into a first optical path signal and a second optical path signal, so that the first optical path signal enters the channel, and perform delay matching on the second optical path signal based on the transmission path of the first optical path signal in the channel; Step 3: Receive the first optical path signal transmitted through the channel; provide a local oscillator reference light that has a phase correlation with the second optical path signal; perform zero-difference interferometry measurement on the first optical path signal and the local oscillator reference light to extract the phase modulation information; Step 4: Based on the extraction results of the phase modulation information, determine the channel evaluation parameters; report the channel evaluation parameters and the corresponding classical channel quality indicators; Step 5: Based on the channel evaluation parameters and the classic channel quality indicators, select the target serving base station and the corresponding target communication link from multiple base stations, and perform a handover to the target communication link; Step 6: The target serving base station sends an optical signal carrying the phase modulation information to the central measurement node; the central measurement node then performs Bell state measurement on the optical signals from the two different communication parties.

[0007] Furthermore, the method also includes: Step 11: Cut the continuous light into a sequence of light pulses with a fixed period, and inject a known and constant reference phase into each light pulse to generate a time-slotted light pulse sequence in which each light pulse carries an internal phase reference. Step 12: Within a specific starting time slot of the optical pulse sequence, the amplitude and phase of the optical pulse are changed simultaneously to generate a strong optical pulse carrying pre-distorted joint modulation information; Step 13: In the modulation time slot immediately following the specific starting time slot, the intensity noise sideband of the strong light pulse is filtered out, and a voltage corresponding to the single-photon horizontal phase shift is applied to the filtered light pulse to generate a co-generated light pulse.

[0008] Furthermore, the method also includes: Step 21: Dynamically adjust the coupling according to the reference phase, and inject different additional phase markers into the separated first optical path signal and second optical path signal; Step 22: Perform power spectrum measurements of the signal frequency bands respectively, and adjust the power of the first optical path signal and the second optical path signal independently based on the measurement results; Step 23: Construct a simulated path containing a tunable delay and a phase compensation pattern for the second optical path signal; transmit the first optical path signal through a short reference channel and perform optical correlation detection with the second optical path signal processed by the simulated path; and adjust the length of the tunable delay according to the detection results. Step 24: The second optical path signal with the adjusted delay length is coupled with the pump light, and a nonlinear optical process is used to convert specific noise frequency components into light of different wavelengths and filter them out.

[0009] Furthermore, the method also includes: Step 31: Separate the probe signal from the first optical path signal, interfere the probe signal with the test state set and measure the probability distribution, reconstruct the equivalent process matrix according to the probability distribution, and perform feedforward state transformation on the first optical path signal based on the process matrix to obtain the reconstructed first optical path signal; Step 32: Nonlinear phase locking is performed using the second optical path signal to generate the local oscillator reference light, and frequency-selective gain filtering related to the power spectrum measurement results of the signal frequency band is introduced during the locking process.

[0010] Furthermore, the method also includes: Step 33: The reconstructed first optical path signal is subjected to asymmetric beam splitting and reciprocal phase transformation with the local oscillator reference light, and the interference result is measured and phase feedback adjustment is performed. Step 34: Monitor the error signal during the nonlinear phase-locking process and obtain the error signal power; perform correlation analysis between the error signal power and the photon arrival time stamp sequence to eliminate high-noise period events, and perform digital filtering based on the error signal on the retained events.

[0011] Furthermore, the method also includes: Step 41: Process the phase modulation information to obtain phase measurement values, compare the phase measurement values ​​with the expected phase values ​​to generate a phase error sequence; combine the time information corresponding to the time-slotted optical pulse sequence, dynamically fit the unsteady coherent decay model through the sliding window maximum likelihood estimation algorithm to obtain the primary channel evaluation parameters. Step 42: Calculate the channel perturbation tensor sequence based on the update history of the process matrix, perform high-order singular value decomposition on the channel perturbation tensor sequence to extract components that match known perturbation source modes, and perform time-frequency domain cross-correlation analysis on the time derivatives of the non-steady-state coherent decay model parameters and the components to obtain secondary channel evaluation parameters.

[0012] Furthermore, the method also includes: The primary channel evaluation parameters and the secondary channel evaluation parameters are normalized to form a multidimensional channel state vector; classical channel quality indices within the same time window are obtained; a classical joint state space is constructed based on the multidimensional channel state vector and the classical channel quality indices; in the classical joint state space, a set of joint basis vectors is found using a hybrid algorithm of principal component analysis and canonical correlation analysis; the multidimensional channel state vector and the classical channel quality indices are projected onto the joint basis vectors to obtain compressed joint feature codewords; the compressed joint feature codewords and the identifiers of the joint basis vectors are encapsulated into reported data.

[0013] Furthermore, the method also includes: Based on the aforementioned unsteady coherent decay model and its core parameters, the state fidelity evolution is simulated by numerically solving the corresponding equations to calculate the predicted remaining lifetime. Based on the secondary channel evaluation parameters from multiple base stations, the perturbation correlation matrix between base stations is calculated and harmful resonance frequencies are analyzed to generate a cooperative avoidance strategy. Based on the compressed joint feature codewords and joint basis vector indexes from each base station, an approximate nearest neighbor search matching is performed using a historical decision knowledge base to output a recommended decision. Based on the predicted remaining lifetime, the cooperative avoidance strategy, and the recommended decision, when the predicted remaining lifetime is lower than a predetermined threshold and there is a recommended target base station that conforms to the cooperative avoidance strategy, a pre-resource reservation instruction is sent to the recommended target base station, and a dual-connectivity mode with the recommended target base station is initiated. After the dual-connectivity verification period, a handover to the recommended target base station is performed based on the verification results.

[0014] Furthermore, the method also includes: Based on the time-slotted optical pulse sequence, the additional phase marker, and the spectral characteristics of the co-occurring optical pulse, the phase modulation information is modulated into an orthogonal polarization state and assigned staggered center frequencies for co-channel transmission; the composite signal transmitted via co-channel is demultiplexed and time-domain gated to extract time-overlapping pulses, and the pulses are injected into a nonlinear interferometer for Bell state projection.

[0015] Furthermore, the method also includes: performing multiple coincidence counting on the output port of the nonlinear interferometer; weighting the coincidence count based on the dual-connection verification period information; publishing the projection result; comparing and selecting key fragments based on the projection result and the transmitted basis vector information; randomly selecting a portion of the key fragments as test bits and comparing them through a classic authentication channel to estimate the error rate and correlation; when the error rate is lower than a security threshold dynamically related to the channel evaluation parameters, performing error correction and privacy amplification on the remaining key fragments to generate a shared key.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) This solution constructs a classic and signal co-channel transmission architecture through co-occurring optical pulses, avoids the redundant construction of infrastructure under the traditional physical isolation mode, reduces the investment losses of operators in spectrum, space and hardware, and achieves efficient transmission of data and key information in the same channel by relying on orthogonal polarization modulation and frequency staggered design. It is adapted to the characteristics of dense networking and dynamic scheduling, breaks the limitations of static solidification of links, and improves the utilization rate of network resources and deployment flexibility.

[0017] (2) This scheme constructs a multi-dimensional channel evaluation system, predicts the remaining channel lifetime through the non-steady-state coherent decay model, generates a collaborative avoidance strategy by combining the interference correlation analysis between base stations, accurately locates and weakens harmful resonance interference, optimizes base station selection based on historical decision knowledge base, and achieves smooth link switching by matching dual connectivity verification mechanism.

[0018] (3) This scheme optimizes signal quality through state reconstruction, nonlinear noise filtering and other technologies. Combined with dynamic security threshold determination and privacy amplification process, it generates a shared key with a minimum entropy of ≥256 bits. The key generation rate is ≥1Mbps to meet the needs of high-speed transmission. It achieves unconditional security based on mechanical principles, which can meet the needs of high-security data encryption scenarios and resist various eavesdropping and tampering attacks.

[0019] (4) This solution achieves deep collaboration with classic networks. The channel leverages mature dense access points and intelligent scheduling mechanisms, eliminating the need for additional dedicated links. Through joint state space construction and hybrid algorithm feature compression, it improves the efficiency of channel information reporting and decision-making. The link switching latency is controlled within 20ms, taking into account the advantages of low latency and wide connectivity, as well as high key security, and adapting to the communication security needs of multiple scenarios. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0021] Figure 1 This is a flowchart of the optical signal detection method of the present invention. Detailed Implementation

[0022] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0023] Please see Figure 1 The method for detecting optical signals includes the following steps: Step 1: Modulate the continuous optical signal based on the data, and perform phase modulation on the modulated optical signal based on the key information to generate a co-occurring optical pulse.

[0024] Step 1 specifically involves the following operations: Step 11: Cut the continuous light into a sequence of optical pulses with a fixed period, and inject a known and constant reference phase into each optical pulse to generate a time-slotted optical pulse sequence in which each optical pulse carries an internal phase reference. The specific operation is as follows: First, the input continuous light is converted into a fixed-period optical pulse sequence using an optical pulse cutting module. This module uses an acousto-optic modulator (AOM) as the switching element, combined with a high-precision RF driver, a synchronous clock unit, and an optical signal shaping and filtering component. The RF driver outputs an RF pulse signal with a preset repetition frequency, which is applied to the piezoelectric transducer of the AOM, causing the acousto-optic medium inside the AOM to generate a periodic refractive index grating. When continuous light is incident, it is diffracted and passes through the AOM only during the effective period of the RF pulse signal; it is blocked during the non-effective period, thus achieving pulsed cutting of continuous light. To adapt to the subsequent time-slotted modulation requirements, the repetition frequency of the RF driver signal is precisely set to 5-20 GHz, corresponding to an optical pulse period of 50-200 ps, ​​with a duty cycle controlled at 10%-30%, ensuring that the generated optical pulses have uniform width and sufficient signal energy. Simultaneously, the extinction ratio of the AOM is ≥60 dB, which effectively suppresses background light noise between pulses and avoids crosstalk between adjacent pulses. The optical signal shaping and filtering component uses a Gaussian... A narrowband filter is used to filter out pulse sidelobes and frequency spurs generated during the cutting process, making the waveform of the output optical pulse approach an ideal Gaussian pulse and improving the stability of subsequent phase modulation. The fixed period design is to achieve precise timing alignment of different modulation time slots, ensuring that data modulation and phase modulation are orderly connected in the time dimension without interference, avoiding signal crosstalk or loss of modulation information due to timing disorder. Subsequently, a known and constant reference phase is injected into each generated optical pulse. This reference phase serves as the benchmark anchor point for subsequent phase measurement and calibration. Its stability directly determines the accuracy of subsequent phase modulation and the reliability of key information extraction. Since the offset of subsequent phase modulation needs to be determined based on the benchmark phase, injecting a fixed reference phase can eliminate the interference caused by the phase fluctuation of the optical pulse itself, so that each optical pulse has an independent and traceable phase reference system. The final generated time-slotted optical pulse sequence not only realizes the segmented control of the signal in the time dimension, but also provides a basic guarantee for the phase coordination of subsequent dual modulation through the built-in phase benchmark.

[0025] Step 12: Within a specific initial time slot of the optical pulse sequence, the amplitude and phase of the optical pulses are simultaneously altered to generate a strong optical pulse carrying pre-distorted joint modulation information. The specific operation is as follows: First, a specific starting time slot in the optical pulse sequence is designated as the dedicated time slot for data modulation. This time slot designation achieves temporal isolation between the classical signal and the subsequent key signal, avoiding mutual interference caused by the two signals modulating within the same time slot. Within this starting time slot, the optical pulse undergoes simultaneous amplitude and phase modulation. Amplitude modulation carries signal strength-related information in the data, while phase modulation carries the phase coding information corresponding to the modulation symbols. This combination of dual modulation enables efficient data compression, improves spectral utilization, and meets the high-bandwidth transmission requirements of the network. Simultaneously, pre-distortion processing is performed on the modulated optical signal to compensate for subsequent distortion. Nonlinear distortion may occur during signal transmission and modulation. Since strong light pulses are easily affected by the nonlinear characteristics of optical devices in subsequent filtering, phase modulation and channel transmission, pre-distortion processing introduces phase and amplitude compensation amounts opposite to the expected distortion in advance, so that the final output strong light pulse can still maintain the integrity and accuracy of the modulation information after subsequent processing. The generated strong light pulse not only carries complete joint modulation information, but its power level is also optimized to meet the requirements of long-distance signal transmission, and reserve sufficient margin for noise filtering and intensity attenuation in subsequent modulation time slots, so as to achieve efficient data carrying and compatibility with subsequent modulation.

[0026] Step 13: Within the modulation time slot following a specific initial time slot, the intensity noise sidebands of the strong light pulse are filtered out, and a voltage corresponding to the single-photon horizontal phase shift is applied to the filtered light pulse to generate a co-occurring light pulse. The specific operation is as follows: First, within the modulation time slot immediately following the start time slot, intensity noise sideband filtering is performed on the strong optical pulse. After joint modulation, the strong optical pulse is accompanied by intensity noise sidebands, which are highly interfering with phase modulation. Since the signal operates at the single-photon level, it is extremely sensitive to noise. These noise sidebands can reduce the accuracy of phase shift, thus affecting the reliable encoding of key information. Therefore, filtering out the intensity noise sidebands using a dedicated filtering module can purify the signal substrate of the optical pulse, eliminate the interference of classical modulation noise on modulation, and provide a low-noise signal environment for phase modulation. Subsequently, the filtered optical pulse... A voltage corresponding to the single-photon horizontal phase shift is applied. The amplitude of this voltage is precisely calibrated to ensure that the phase shift generated by the optical pulse after application is exactly within the controllable range of the single-photon state. The key information is encoded through this phase shift. The single-photon horizontal phase modulation not only meets the security requirements of key distribution but also ensures the characteristics of the key information, preventing eavesdropping or cracking. The resulting co-existing optical pulse has its initial time slot carrying pre-distorted joint modulation information and its modulation time slot carrying key information based on the single-photon phase shift, realizing the co-existence of classical signals and key signals within the same optical pulse.

[0027] In a preferred embodiment of the present invention, step 2 is further included: splitting the co-occurring optical pulse into a first optical path signal and a second optical path signal, allowing the first optical path signal to enter the channel, and performing delay matching on the second optical path signal based on the transmission path of the first optical path signal in the channel; Step 2 further includes the following specific steps: Step 21: Dynamically adjust the coupling according to the reference phase, and inject different additional phase markers into the separated first and second optical path signals. The specific operation is as follows: First, dynamic coupling adjustment is performed based on the reference phase injected in step 11. This is achieved by using an adjustable optical coupler, such as a 3dB fiber coupler paired with a phase-sensitive feedback control unit, to split the co-existing optical pulses. The feedback control unit collects the reference phase information in the optical pulses in real time, calculates the current phase fluctuation deviation, and then dynamically adjusts the coupling coefficient of the coupler to stabilize the splitting ratio of the first and second optical path signals within a preset range, such as adjustable from 1:1 to 1:4. Furthermore, the phase loss during splitting is controlled within ±0.01π to prevent distortion of the reference phase due to the splitting operation. Subsequently, distinct additional phase markers are injected into the two split signals. These additional phase markers use a fixed phase offset setting. For example, a π / 2 phase shift is injected into the first optical path signal and a π phase shift is injected into the second optical path signal. This mark serves as a unique phase identifier for the two signals, enabling rapid differentiation between them in subsequent interferometric measurements and delay matching checks, thus avoiding signal misjudgment caused by optical path confusion. The injection of the additional phase is achieved through an electro-optic phase modulator. The modulator's driving voltage is synchronously locked with the reference phase clock signal, ensuring the stability and consistency of the additional phase mark. Furthermore, the marked phase maintains a fixed correlation with the reference phase, providing a clear phase reference for subsequent phase compensation and signal tracing. At the same time, the dynamic coupling adjustment mechanism can adapt to the optical power requirements under different channel scenarios, improving the adaptability and reliability of the dual optical path signals.

[0028] Step 22: Perform power spectrum measurements on the signal frequency bands respectively, and adjust the power of the first optical path signal and the second optical path signal independently based on the measurement results. The specific operation is as follows: First, power spectrum measurements were performed on the first and second optical path signals using a high-resolution spectrometer. The measurement frequency band was precisely aligned with the operating frequency band corresponding to the phase modulation in step 13, such as the signal band of the 1550nm communication band. The sampling frequency was adapted to the optical pulse period of 50-200ps to ensure complete capture of the power distribution characteristics within the signal frequency band, while filtering out power interference from classic signal frequency bands and extracting only the modulation-related power spectrum data. Based on the measured power spectrum results, the power of the two signals was independently adjusted using a variable optical attenuator (VOA). To accommodate subsequent channel transmission losses, the frequency band power needs to be adjusted to 0.1-1mW. The specific value is dynamically set according to the channel type, with the upper limit for fiber optic channels and the lower limit for free space channels. The second optical path signal serves as the reference optical path, and its frequency band power must be consistent with that of the first optical path signal, with the power deviation controlled within ±0.1dB. This ensures that the two signals have a balanced light intensity foundation in subsequent optical correlation detection, avoiding a decrease in detection sensitivity or an increase in delay matching error due to power differences. During the power adjustment process, the power spectrum measurement results are fed back in real time to form a closed-loop control, ensuring that the frequency band power of the two signals remains stable at the target value.

[0029] Step 23: Construct a simulated path containing a tunable delay and a phase compensation pattern for the second optical path signal. After transmitting the first optical path signal through a short reference channel, perform optical correlation detection with the second optical path signal processed by the simulated path, and adjust the length of the tunable delay based on the detection results. The specific operations are as follows: First, a simulated path including a tunable delay unit and a phase compensation pattern is constructed for the second optical path signal. The tunable delay unit uses an optical fiber delay line or an acousto-optic tunable delay device, with a delay adjustment range covering the maximum transmission delay of the target communication channel, such as 0-100ns, and an adjustment accuracy of ±1ps, to adapt to delay differences in different channel lengths and transmission environments. The phase compensation pattern is pre-stored in the control unit based on a phase distortion model of a typical channel. For different channel scenarios, such as fiber dispersion and free-space atmospheric turbulence scenarios, the corresponding compensation parameters are called, and the phase modulator applies reverse phase compensation to the second optical path signal to cancel the phase in the subsequent delay adjustment and channel simulation process. After distortion, the first optical path signal is transmitted through a short reference channel and then input to a balanced optical correlation detector along with the second optical path signal processed by the simulated path. The short reference channel can be fixed in length at 100m to simulate the basic transmission characteristics of a real channel. Next, the detector collects the interference intensity signals of the two signals. When the delays of the two signals match, the interference intensity reaches its peak; otherwise, the interference intensity attenuates. Based on the detected peak position of the interference intensity, the control unit calculates the delay deviation between the second and first optical path signals and dynamically adjusts the delay length of the tunable delay unit until the timing deviation between the two signals is controlled within ±1ps, achieving precise delay matching. This process provides a transmission delay reference for the first optical path through the short reference channel, and combined with the dynamic adjustment of the simulated path, ensures that the transmission delay of the second optical path signal is completely consistent with the transmission delay of the first optical path signal in the actual channel.

[0030] Step 24: Couple the second optical path signal, whose delay length has been adjusted, with the pump light. Utilize a nonlinear optical process to convert specific noise frequency components into light of different wavelengths and filter them out. The specific operation is as follows: First, the delayed-adjusted second optical path signal is efficiently coupled to the pump light input wavelength division multiplexer (WDM). The pump light uses a narrow-linewidth laser in the same wavelength band as the second optical path signal, such as a 1550nm wavelength, with a linewidth ≤10kHz and a power set to 10dBm. The pulse width is consistent with the optical pulse period of the second optical path signal to ensure efficient triggering of the nonlinear optical process after coupling. The coupled composite signal is injected into a nonlinear optical medium, such as a highly nonlinear fiber (HNLF) or a periodically polarized lithium niobate crystal (PPLN). Using a four-wave mixing (FWM) nonlinear process, specific noise frequency components carried in the second optical path signal are converted into stray light with a different wavelength than the signal, such as a 1310nm wavelength. The specific noise frequency components are mainly the 1-10GHz intensity noise sidebands introduced by modulation in step 12. Then, the conversion efficiency is controlled at ≥90% by adjusting the pump light power to ensure that the noise components can be fully converted. Subsequently, the converted composite signal is filtered by a band-stop filter. The center wavelength of the filter is precisely aligned with the wavelength of the stray light, the bandwidth is set to 0.1nm, and the extinction ratio is ≥50dB. This can efficiently filter out the noise stray light after conversion, while retaining the core frequency components and phase information of the second optical path signal. The intensity noise suppression ratio of the second optical path signal after noise filtering is improved to ≥60dB, effectively eliminating the interference of classical modulation noise and transmission noise on subsequent signal processing.

[0031] In a preferred embodiment of the present invention, step 3 is further included: receiving a first optical path signal transmitted through the channel; providing a local oscillator reference light that has a phase correlation with the second optical path signal; and performing zero-difference interferometry measurement on the first optical path signal and the local oscillator reference light to extract phase modulation information. Step 3 specifically involves the following steps: Step 31: Separate the probe signal from the first optical path signal, interfere the probe signal with the test state set and measure the probability distribution, reconstruct the equivalent process matrix based on the probability distribution, and perform a feedforward state transformation on the first optical path signal based on the process matrix to obtain the reconstructed first optical path signal. The specific operations are as follows: First, a small portion of the probe signal (10%) and the main signal (90%) are separated from the received first optical path signal using a 1:9 split-ratio fiber coupler. This is then temporarily stored in an optical buffer unit to avoid timing delays. This optical buffer unit employs a fiber ring buffer architecture, consisting of a 2×2 high-speed optical switch, a ring fiber, a low-noise optical amplifier (EDFA), and an optical isolator. The ring fiber is a single-mode fiber with a length of 100m, set according to the optical pulse period and probe signal processing time, corresponding to an optical signal transmission time of approximately 0.5μs. This satisfies the total time required for probe signal interferometry, probability distribution acquisition, and matrix reconstruction (approximately 0.3μs), ensuring no timing offset during the main signal storage period. The low-noise optical amplifier has a noise figure ≤5dB to compensate for the insertion loss of the ring fiber and optical switch, with a total loss ≤3dB to prevent main signal amplitude attenuation. The optical isolator has an isolation ≥40dB to prevent echo light within the ring cavity from interfering with the main signal state. The switching time of the optical switch is ≤1ns, ensuring accurate synchronous output after the main signal storage period, facilitating subsequent feedforward transformation. Unit timing matching; the probe signal is used to characterize state transmission impairment. The test state set is pre-stored in the control unit, containing four single-photon states under orthogonal polarization basis, such as horizontal polarization, vertical polarization, +45° polarization, and -45° polarization. After coupling with the probe signal through a polarization beam splitter, it is injected into the interferometer. The photon count after interference is collected by a single-photon detector, thereby obtaining the probability distribution of interference between different test states and probe signals. Based on this probability distribution, the equivalent process matrix is ​​reconstructed using the maximum likelihood estimation algorithm. This matrix has a dimension of 4×4, corresponding to the four test states. The matrix elements characterize the amplitude attenuation, phase shift, and decoherence effect of the channel on the state. The reconstruction accuracy is controlled within ±0.01 to ensure accurate reflection of the perturbation characteristics of the state caused by channel transmission. Subsequently, according to the reconstructed equivalent process matrix, the feedforward transformation parameters for canceling state distortion are calculated. The optical buffer unit synchronously outputs the main signal, and the corresponding state transformation is applied to the main signal of the first optical path through a high-speed electro-optic phase modulator to compensate for the phase distortion and amplitude attenuation generated in the channel transmission, and finally obtains the reconstructed first optical path signal.

[0032] Step 32: Nonlinear phase locking is performed using the second optical path signal to generate the local oscillator reference light. During the locking process, frequency-selective gain filtering related to the power spectrum measurement results of the signal frequency band is introduced. The specific operation is as follows: First, the second optical path signal processed in step 24 is input into a nonlinear phase-locked loop. This loop uses a semiconductor optical amplifier (SOA) as the nonlinear device, combined with a phase comparator, a loop filter, and a voltage-controlled oscillator. The second optical path signal serves as the reference light, coupled with the oscillating light within the loop, and then injected into the SOA. The cross-phase modulation effect of the SOA is used to achieve phase correlation between the two signals. The phase comparator acquires the phase difference between the two optical signals in real time and outputs an error voltage signal. After being filtered by the loop filter, the voltage-controlled oscillator is controlled to dynamically adjust the phase of the oscillating light, stabilizing the phase difference between the oscillating light and the second optical path signal within ±0.001π, thus achieving nonlinear phase locking and generating a signal with phase correlation. The local oscillator reference light; during the phase locking process, a frequency-selective gain filter related to the power spectrum measurement results of the signal frequency band in step 22 is introduced. The filter unit adopts a tunable active filter. According to the noise peak frequency band obtained by the power spectrum measurement, such as the modulation noise of 1-10GHz, the gain of this frequency band is suppressed by ≥30dB. At the same time, for the core frequency band of the signal, such as the vicinity of 193.5THz corresponding to 1550nm, the gain is increased by 5-10dB. Through frequency-selective modulation, the phase correlation between the local oscillator reference light and the second optical path signal is preserved, and the residual noise in the signal frequency band is effectively suppressed, so that the linewidth of the local oscillator reference light is compressed to ≤10kHz and the intensity noise suppression ratio is ≥65dB.

[0033] Step 33: After asymmetric beam splitting and reciprocal phase transformation, the reconstructed first optical path signal is interfered with the local oscillator reference light. The interference result is measured and phase feedback adjustment is performed. The specific operation is as follows: First, the reconstructed first optical path signal and the local oscillator reference light are input into an asymmetric beam splitter with a 2:8 splitting ratio. The local oscillator reference light accounts for 80% to ensure the light intensity advantage during interference and improve the interference signal-to-noise ratio, while the first optical path signal accounts for 20% to avoid the destruction of the strong light pair state. Subsequently, an electro-optic phase modulator applies a reciprocal phase transformation to the two signals respectively. The transformation parameters are derived based on the equivalent process matrix reconstructed in step 31 and are completely opposite to the phase distortion suffered by the first optical path signal in the channel. This is used to cancel the phase deviation generated during transmission and ensure that the two signals have a consistent phase before interference. The two processed signals are injected into a balanced homodyne detector. The detector converts the interference light intensity signal into an electrical signal, and its output voltage is linearly related to the phase difference between the two signals. By acquiring this voltage signal, the phase modulation information of the first optical path signal can be deduced. At the same time, a phase feedback adjustment loop is constructed to monitor the phase signal output by the detector in real time. When a phase fluctuation is detected to exceed ±0.01π, the phase of the local oscillator reference light is dynamically adjusted through a phase modulator to keep the interference process in a stable state. The phase measurement resolution can reach ±0.005π, ensuring accurate extraction of phase modulation information.

[0034] Step 34: Monitor the error signal during the nonlinear phase-locking process and obtain the error signal power; perform correlation analysis between the error signal power and the photon arrival time stamp sequence to eliminate high-noise period events, and perform digital filtering based on the error signal on the retained events. The specific operations are as follows: First, the error signal of the nonlinear phase-locked loop in step 32 is monitored in real time. The power of this error signal is negatively correlated with the phase-locking stability. When the error signal power exceeds a preset threshold, such as 0.1mW, it is determined to be a high-noise period, which may be caused by sudden channel disturbances, pump light power fluctuations, etc. Simultaneously, a photon arrival timestamp sequence is acquired, output by a single-photon detector, with a time resolution of 1ps. This sequence is correlated with the error signal power time-series data to identify the photon arrival events corresponding to the high-noise period. These events are then directly removed to avoid noise interference, with a removal efficiency ≥99%. To ensure that all retained events are from low-noise, stable periods, adaptive digital filtering based on the error signal is then applied to the retained phase measurement events. The size of the filtering window is dynamically adjusted according to the power of the error signal; the lower the power of the error signal, the larger the window. For example, 0.01-0.1mW corresponds to a window size of 10-50 optical pulse cycles. Through weighted averaging within the sliding window, the influence of random noise on the phase measurement results is suppressed, reducing the standard deviation of the phase measurement value to ≤0.003π. After noise removal and digital filtering, the signal-to-noise ratio of the phase modulation information is improved to ≥30dB.

[0035] In a preferred embodiment of the present invention, step 4 is further included: determining channel evaluation parameters based on the phase modulation information; and sending the channel evaluation parameters and the classical channel quality index corresponding to the data.

[0036] Step 4 specifically involves the following operations: Step 41: Process the phase modulation information to obtain phase measurement values, compare the phase measurement values ​​with the expected phase values ​​to generate a phase error sequence; combine the time information corresponding to the time-slotted optical pulse sequence, and dynamically fit the unsteady coherent decay model using the sliding window maximum likelihood estimation algorithm to obtain the primary channel evaluation parameters. The specific operations are as follows: First, the phase modulation information extracted in step 33 is digitized. After removing high-frequency noise through low-pass filtering, the phase measurement value corresponding to each optical pulse is sampled. The sampling frequency is matched with the optical pulse period at 50-200 ps / point to ensure complete capture of the phase characteristics of each pulse. The measured value is compared point by point with the expected phase value preset during phase modulation in step 13, and the phase difference between the two is calculated to generate a phase error sequence. The amplitude of this sequence directly reflects the degree of disturbance of the phase modulation information by the channel transmission. The standard deviation of the error sequence is initially controlled within ±0.003π. Then, combined with the time information of the time-slotted optical pulse sequence in step 11, a sliding window length of 50 optical pulse periods is used. A sliding window maximum likelihood estimation algorithm is used to dynamically fit the phase error sequence and construct an unsteady coherent decay model. The model uses coherence time, decay coefficient, and phase offset baseline as parameters. Coherence time represents the duration of coherence, decay coefficient represents the rate of coherence decay, and phase offset baseline represents the inherent static phase deviation of the channel. During the fitting process, the model parameters in each window are calculated independently. By maximizing the similarity between the phase error sequence and the model prediction value, the model parameters are dynamically updated. Finally, a primary channel evaluation parameter containing coherence time, decay coefficient, phase offset baseline, and phase error standard deviation is obtained. This parameter can intuitively reflect the real-time transmission quality and basic disturbance characteristics of the channel.

[0037] The sliding window maximum likelihood estimation algorithm fits the channel evaluation parameters by minimizing the following cost function:

[0038] Where ε i Let be the phase error sequence value of the i-th time slot, μd be the time-varying mean drift term, σd be the time-varying diffusion term, Θ be the model parameter set, W be the sliding window size, and σ0² be the background noise variance.

[0039] Step 42: Calculate the channel perturbation tensor sequence based on the update history of the process matrix. Perform high-order singular value decomposition on the channel perturbation tensor sequence to extract components that match the known perturbation source modes. Perform time-frequency domain cross-correlation analysis on the time derivatives of the unsteady coherent decay model parameters and the components to obtain secondary channel evaluation parameters. The specific operations are as follows: First, based on the update history of the equivalent process matrix in step 31 (i.e., retaining the matrix data of the most recent 100 optical pulse cycles), the difference between the process matrices within adjacent cycles is calculated to construct a channel perturbation tensor sequence. This tensor has a dimension of 4×4×N, where N is the update number, i.e., 100. Tensor elements represent the perturbation changes of each component of the channel state at different times. The dynamic evolution of channel perturbation can be captured through the tensor sequence. Third-order singular value decomposition is performed on this perturbation tensor sequence. During the decomposition, the components corresponding to the first three largest singular values ​​are retained, low-amplitude noise components are removed, and feature components matching the preset known perturbation source patterns, such as fiber dispersion, free-space atmospheric turbulence, and signal crosstalk, are extracted. The amplitude of each feature component corresponds to the influence intensity of this type of perturbation source on the channel. For example, signal crosstalk... The amplitude of the corresponding characteristic components is usually between 0.01 and 0.05. Subsequently, the parameters of the unsteady coherent decay model obtained in step 41 are calculated, namely the time derivatives of the coherence time and decay coefficient. These derivatives reflect the dynamic change rate of the model parameters and indirectly characterize the intensification or weakening trend of channel disturbances. The time derivatives are cross-correlation analysis is performed with the extracted disturbance source characteristic components in the time and frequency domain. The cross-correlation window is set to 20 optical pulse periods. The correlation coefficients of the two at different frequency components are calculated. When the absolute value of the correlation coefficient is ≥0.7, the disturbance source is determined to be the main factor affecting the channel coherence. Finally, secondary channel evaluation parameters are obtained, which include the main disturbance source type, disturbance intensity, disturbance frequency, and the influence coefficient of the disturbance on coherence. These parameters can accurately locate the dynamic disturbance source of the channel.

[0040] Step 43: Normalize the primary channel evaluation parameters and secondary channel evaluation parameters to form a multi-dimensional channel state vector; obtain the classical channel quality indices within the same time window; construct a classical joint state space based on the multi-dimensional channel state vector and the classical channel quality indices; find a set of joint basis vectors in the classical joint state space using a hybrid algorithm of principal component analysis and canonical correlation analysis; project the multi-dimensional channel state vector and the classical channel quality indices onto the joint basis vectors to obtain compressed joint feature codewords; encapsulate the compressed joint feature codewords and the identifiers of the joint basis vectors into reported data. The specific operations are as follows: First, the primary and secondary channel evaluation parameters are normalized by using a linear normalization method to map all parameter values ​​to the [0,1] interval, eliminating the dimensional differences between different parameters and forming a multidimensional channel state vector with dimensions of 12-15, containing 4 primary parameters and 8-11 secondary parameters. Simultaneously, classic channel quality indicators corresponding to data within the same time window are acquired, including channel signal-to-noise ratio (SNR), bit error rate (BER), transmission delay, bandwidth utilization, and signal strength index (RSSI). A SNR ≥ 20dB and a transmission delay ≤ 1ms are the baseline thresholds for high-quality transmission in the classic channel. Based on the multidimensional channel state vector and classic channel quality indicators, a classic joint state space is constructed. The dimension of this space is the sum of the channel state vector dimension and the number of classic channel indicators, typically 17-20 dimensions. Each point in the space corresponds to the comprehensive channel state at a certain moment. In this joint state space, principal component analysis is performed... A hybrid algorithm combining principal component analysis (PCA) and canonical correlation analysis (CCI) is used to extract the joint basis vector. First, PCA is used to reduce the dimensionality of the joint state space data, retaining the top 5 principal components with a cumulative contribution rate ≥ 95%. Then, CCI is used to explore the intrinsic correlation between channel parameters and classical channel indicators, optimizing to obtain a 5-6 dimensional joint basis vector. This vector can maximize the retention of the core information of the channel's overall state. The multi-dimensional channel state vector and classical channel quality indicators are projected onto this joint basis vector to obtain a compressed joint feature codeword with a dimension compressed to 5-6. This codeword contains both the channel's transmission characteristics and disturbance information, and integrates the classical channel quality state, achieving efficient compression of channel information. Finally, the unique identifiers of the compressed joint feature codeword and the joint basis vector are encapsulated into the reported data. The encapsulation format adopts a frame structure design, with the frame header containing a timestamp and checksum to ensure the integrity and traceability of the data transmission process.

[0041] In a preferred embodiment of the present invention, step 5 is further included: selecting a target serving base station and its corresponding target communication link from multiple base stations based on channel evaluation parameters and classic channel quality indicators, and performing a handover to the target communication link; Step 5 specifically involves the following operations: Step 51: Based on the unsteady-state coherent decay model and its core parameters, the state-fidelity evolution is simulated by numerically solving the corresponding equations, and the remaining lifetime is predicted. The specific operations are as follows: First, the unsteady coherent decay model and core parameters obtained from step 41 are extracted, including coherence time, decay coefficient, and phase offset baseline. These parameters directly determine the evolution of the state in the channel. Based on the density matrix evolution equation in mechanics, a fourth-order Runge-Kutta numerical solution method is adopted, using the current state fidelity as the initial condition. This current state fidelity is the state fidelity of the first optical path signal after reconstruction in step 31, with an initial value ≥95%. The parameters of the unsteady coherent decay model are input to simulate the decay process of the state fidelity over time. The time step of the numerical solution is set to 1ms, and the solution range covers 0-100ms to ensure accurate capture of the fidelity decay. The trend is set with a fidelity threshold of 90%. Below this threshold, the security and reliability of key distribution will decrease significantly. By fitting the fidelity decay curve, the time point when the fidelity drops from the current value to 90% is located. The difference between this time point and the current time is the predicted remaining lifetime of the channel. During the solution process, the evolution equation is dynamically corrected by combining the perturbation intensity in the secondary evaluation parameters extracted in step 42. If the perturbation intensity is ≥0.05, the decay coefficient is amplified by 1.2 times to adapt to the scenario where the fidelity decays faster when the perturbation intensifies, ensuring the accuracy of the predicted remaining lifetime. The final output of the predicted remaining lifetime accuracy is controlled within ±1ms.

[0042] Step 52: Based on the secondary channel evaluation parameters from multiple base stations, calculate the inter-base station disturbance correlation matrix and analyze the harmful resonance frequencies to generate a cooperative avoidance strategy. The specific operations are as follows: First, secondary channel evaluation parameters are collected from multiple candidate base stations within the current communication range, typically 3-5 candidate base stations, covering a 1km radius around the user's movement trajectory. The focus is on extracting core information such as the type of disturbance source, disturbance intensity, and disturbance frequency for each base station. Based on this information, a disturbance correlation matrix is ​​calculated between any two base stations. The matrix dimension is the number of candidate base stations multiplied by the number of candidate base stations, and the matrix elements are the correlation coefficients of the disturbance frequencies between the two base stations. The calculation uses a 20ms time window and employs the Pearson correlation coefficient algorithm. The correlation coefficient ranges from -1 to 1; the closer the absolute value is to 1, the stronger the correlation between the disturbance frequencies of the two base stations. Subsequently, eigenvalue decomposition is performed on the correlation matrix, extracting eigenvalues ​​≥ 0.8. The frequency components corresponding to the feature vectors are identified as harmful resonance frequencies. At these frequencies, the disturbances from multiple base stations will superimpose, causing a sharp drop in state fidelity. For example, when the correlation coefficient of the crosstalk frequencies of two base stations is 0.9, the intensity of the superimposed disturbance will increase to 1.5 times that of a single base station. Based on the identified harmful resonance frequencies, a collaborative avoidance strategy is generated. For each candidate base station, the operating frequency of its channel is adjusted, for example, by shifting it by ±50MHz to avoid the harmful resonance frequency. At the same time, the transmit power of the base station is optimized, reducing the transmit power of base stations near the resonance frequency by 1-2dBm to weaken the disturbance superposition effect. After the strategy is generated, candidate base stations that meet the avoidance requirements are marked, and base stations with the risk of disturbance superposition are excluded.

[0043] Step 53: Based on the compressed joint feature codewords and joint basis vector indexes from each base station, perform approximate nearest neighbor search matching using the historical decision knowledge base, and output the recommendation decision. The specific operations are as follows: First, a historical decision-making knowledge base is constructed, storing historical data from the past 1000 link handovers. This includes compressed joint feature codewords, joint basis vector indices, candidate base station information, final selection results, and post-handover channel quality feedback, such as state fidelity, classic channel bit error rate, and classic channel signal-to-noise ratio. For each historical data point, a corresponding weight is assigned: historical data with excellent post-handover channel quality has a weight of 0.8, while data with good post-handover channel quality has a weight of 0.5. The criteria for excellent channel quality are: fidelity ≥ 93% and bit error rate ≤ 10%. -7The standard for judging good channel quality is a fidelity of 90%-93%. Generally, values ​​below this standard are weighted at 0.2 to highlight the reference value of high-quality decision-making experience. Subsequently, the compressed joint feature codewords and joint basis vector indices of each candidate base station are input. An approximate nearest neighbor search is performed using the KD-tree algorithm to calculate the Euclidean distance between the current codeword and all historical codewords in the knowledge base. The top 5 historical data with the smallest distances are selected (e.g., a distance threshold ≤ 0.1). The recommendation results of these 5 historical data are then weighted and voted on, with the weights corresponding to the historical data. The base station with the highest number of votes is the initial recommendation. If multiple historical data points recommend the same base station, and the base station has a high-quality feedback rate of ≥70% in its historical feedback, it is directly identified as the recommended target base station. If the base station with the highest number of votes has a high-quality feedback rate of <50%, it is removed, and the remaining four historical data points are re-voted to ensure the reliability of the recommendation decision. The final recommendation decision output includes the recommended target base station identifier, the corresponding target communication link, and the recommendation confidence level. The target communication link is a combination configuration of channel and classic channel. The recommendation confidence level ranges from 0 to 1, and a confidence level ≥0.7 is considered a highly reliable recommendation.

[0044] Step 54: Based on the predicted remaining lifetime, the cooperative avoidance strategy, and the recommendation decision, when the predicted remaining lifetime is lower than a predetermined threshold and there is a recommended target base station that meets the cooperative avoidance strategy, a pre-resource reservation instruction is sent to the recommended target base station, and a dual-connectivity mode with the recommended target base station is initiated; after the dual-connectivity verification period, a handover to the recommended target base station is performed based on the verification result. The specific operations are as follows: First, a predetermined threshold for the predicted remaining lifetime is set to 50ms. This threshold is based on the typical latency and channel attenuation characteristics of network handover, allowing sufficient time for handover preparation and verification. When the predicted remaining lifetime output in step 51 is less than 50ms, and the recommended target base station output in step 53 conforms to the cooperative avoidance strategy generated in step 52 (i.e., no harmful resonant frequency superposition and disturbance intensity ≤0.03), the link handover process is initiated. The control unit sends a pre-resource reservation instruction to the recommended target base station. The instruction includes channel bandwidth, reservation of ≥100MHz classic channel resources, reservation of a link with ≥10Mbps bandwidth and ≤1ms latency, and phase calibration parameters to ensure that the target base station completes resource configuration and parameter adaptation in advance. Subsequently, dual-connectivity mode is initiated, and the user equipment simultaneously establishes communication links with the current serving base station and the recommended target base station. The target base station synchronously receives and verifies the signal, undertaking the main data transmission task, to achieve a smooth transition of data transmission. The dual-connection verification period is set to 20ms. During the verification period, the transmission latency of the target link is monitored in real time and must be ≤1ms. If all indicators meet the requirements, the verification is deemed successful, and link switching is performed to completely migrate the data transmission task to the target communication link. If any indicator fails to meet the requirements during the verification period, the dual-connection mode is immediately terminated, and the system returns to the current serving base station. The search and matching logic in step 53 is called again to select the next target base station from the remaining recommended base stations. After the switch is completed, the channel quality of the target link is continuously monitored. If there are no abnormalities within 10ms, the resources occupied by the current base station are released, and the entire link switching process is completed. If an abnormality occurs, an emergency switching mechanism is activated, and the system falls back to the most recently verified link to ensure communication continuity.

[0045] In a preferred embodiment of the present invention, step 6 is further included: sending an optical signal carrying phase modulation information to the central measurement node through the target serving base station; and the central measurement node performing Bell state measurement on the optical signals from the two different communication parties.

[0046] The aforementioned Bell state measurement of optical signals from two different communicating parties by the central measurement node includes: the estimated success probability of Bell state projection is... P success =1 / 4[1+exp(-△t / T2).cos(△Φ)] Where Δt is the arrival time difference between the two optical pulses, T2 is the decoherence time, and ΔΦ is the relative phase difference between the two signals.

[0047] Step 6 specifically involves the following operations: Step 61: Based on the time-slotted optical pulse sequence, the added phase marker, and the spectral characteristics of the co-occurring optical pulses, the phase modulation information is modulated into orthogonal polarization states and assigned staggered center frequencies for cochannel transmission. The specific operations are as follows: First, using the time-slotted optical pulse sequence in step 11 as the timing reference, and combining the mutually exclusive additional phase markers injected in step 21, the unique timing and phase identifier of each optical pulse are clarified, providing a dual identification basis for subsequent signal demultiplexing. Based on the spectral characteristics of the co-occurring optical pulse, its core parameters are a center wavelength of 1550nm and a spectral width of 20nm. The phase modulation information is modulated into horizontal and vertical polarization states respectively by a polarization beam splitter. The two polarization states are orthogonal to each other and can be transmitted independently in the same physical channel without interference. The horizontal polarization state carries the information of communication party A, and the vertical polarization state carries the information of communication party B. At the same time, in order to further improve the signal isolation, the optical signals of the two polarization states are assigned staggered center frequencies, and the frequency offset is set to 50MHz. This ensures that even if polarization crosstalk occurs during channel transmission, signal separation can be achieved through frequency filtering. Through the dual design of orthogonal polarization modulation and frequency staggering, the phase modulation information can be transmitted in the same optical fiber channel, which avoids the cost of additional channel construction and ensures the independence of the two signals. The state fidelity loss during transmission is controlled to ≤3%.

[0048] Step 62: Demultiplex and time-domain gating are performed on the composite signal transmitted via the common channel to extract time-overlapping pulses. The pulses are then injected into a nonlinear interferometer for Bell state projection. The specific operations are as follows: First, the composite signal transmitted via a common channel to the central measurement node is demultiplexed. Using a tunable optical filter, two optical signals corresponding to the horizontal and vertical polarization states are separated according to a preset staggered center frequency. The filtering bandwidth is set to 10MHz to ensure that only the frequency components corresponding to the phase modulation information are extracted, filtering out channel noise and frequency spurious signals. Subsequently, based on the time-slotted timing information from step 11 and the additional phase marker from step 21, time-domain gating is performed through a high-speed optical switch to retain only the time-overlapping optical pulses in the two signals. The gating window and the optical pulse period are matched to 50-200ps. By comparing the additional phase marker with the time slot timing, asynchronous pulses caused by transmission delay differences are eliminated to ensure that the two pulses entering the interferometer are completely aligned in the time dimension, with the time deviation controlled within ±1ps. The filtered synchronous pulses are injected into a nonlinear interferometer. This interferometer uses a periodically polarized lithium niobate crystal as its core device. By adjusting the crystal temperature, the intensity of the nonlinear optical effect is controlled, causing the two signals to interact within the interferometer and achieving the initial projection of the Bell state.

[0049] Step 63: Perform multiple coincidence counting on the output port of the nonlinear interferometer, weight the coincidence count based on the dual-connection verification period information, publish the projection result, and the communicating parties compare and select based on the projection result and the transmitted basis vector information to extract the key fragment. The specific operation is as follows: First, single-photon detectors are deployed at the four output ports of the nonlinear interferometer. Multiple coincidence counting is performed on the projected signal, with the counting window set to 10 ps. Only synchronous photon events at different ports within the same time slot are recorded to distinguish the four Bell states. Based on the corresponding detection results, the coincidence count results are weighted according to the channel quality information during the dual-connectivity verification period in step 54. If the state fidelity of the target link is ≥92% and the classical channel bit error rate is ≤10 during the verification period, the result is considered as follows: -7 The coincidence count weight for this period is assigned as 1.0; if the fidelity is between 90% and 92%, the weight is assigned as 0.8; if it is below the above standard, the weight is assigned as 0.5. By weighting, the counting noise of low-quality channel periods is suppressed, and the accuracy of Bell state recognition is improved. The central measurement node publishes the Bell state projection results, only publishing the projection type and not disclosing the specific state information. The two communicating parties compare the projection results with the basis information when they send the information, and then select the phase modulation information corresponding to the state with consistent basis, and convert it into a binary key fragment. During the comparison process, the basis consistency rate is controlled at ≥50% to ensure that the extracted key fragment has sufficient length and reliability.

[0050] A portion of the key fragments is randomly selected as test bits and compared through a classic authentication channel to estimate the error rate and correlation. When the error rate is lower than a security threshold dynamically related to the channel evaluation parameters, the remaining key fragments are corrected and privacy amplified to generate a shared key. The specific operation is as follows: First, both communicating parties randomly select 20% of the extracted key fragments as test bits and compare them through a classic authentication channel. The classic authentication channel uses the SHA-256 hash algorithm to encrypt the transmission of the test bits to ensure that the information is not eavesdropped or tampered with during the comparison process. Based on the comparison results, the error rate and correlation of the key fragments are estimated. The error rate is the proportion of inconsistent test bits, and the correlation is the sequence correlation coefficient of consistent bits. The security threshold is dynamically set according to the primary channel evaluation parameters in step 41. For example, when the channel coherence time is ≥1... When the coherence time is 0ms and the decay coefficient is ≤0.02, the security threshold is set to 5%. When the coherence time is <5ms and the decay coefficient is >0.05, the security threshold is lowered to 3% to ensure that the security threshold matches the actual transmission quality of the channel. When the estimated error rate is lower than this dynamic security threshold, low-density parity-check code (LDPC) is used to correct the remaining 80% of the key fragment. Then, privacy amplification is performed through the Toeplitz matrix to compress the length of the corrected key fragment to 60% of the original length, eliminating possible eavesdropping information and residual errors, and finally generating a shared key. The security of the shared key is guaranteed by mechanical principles. After privacy amplification, the minimum entropy of the key is ≥256 bits, which can meet the high-security data encryption requirements in communication, and the key generation rate is ≥1Mbps, which is suitable for high-speed network transmission scenarios.

[0051] The above are merely preferred embodiments of the present invention; however, the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and its improved concept, should be covered within the scope of protection of the present invention.

Claims

1. A method of detecting based on an optical signal, characterized in that, The method includes: Step 1: Modulate the continuous optical signal based on the data, and perform phase modulation on the modulated optical signal based on the key information to generate a co-occurring optical pulse; Step 2: Split the co-occurring optical pulse into a first optical path signal and a second optical path signal, so that the first optical path signal enters the channel, and perform delay matching on the second optical path signal based on the transmission path of the first optical path signal in the channel; Step 3: Receive the first optical path signal transmitted through the channel; provide a local oscillator reference light that has a phase correlation with the second optical path signal; perform zero-difference interferometry measurement on the first optical path signal and the local oscillator reference light to extract the phase modulation information; Step 4: Determine channel evaluation parameters based on the phase modulation information; send the channel evaluation parameters and the corresponding classical channel quality indices for the data; Step 5: Based on the channel evaluation parameters and the classic channel quality indicators, select the target serving base station and the corresponding target communication link from multiple base stations, and perform a handover to the target communication link; Step 6: The target serving base station sends an optical signal carrying the phase modulation information to the central measurement node; the central measurement node then performs Bell state measurement on the optical signals from the two different communication parties.

2. The method of detecting an optical signal according to claim 1, wherein The method further includes: Step 11: Cut the continuous light into a sequence of light pulses with a fixed period, and inject a known and constant reference phase into each light pulse to generate a time-slotted light pulse sequence in which each light pulse carries an internal phase reference. Step 12: Within a specific starting time slot of the optical pulse sequence, the amplitude and phase of the optical pulse are changed simultaneously to generate a strong optical pulse carrying pre-distorted joint modulation information; Step 13: In the modulation time slot following the specific starting time slot, the intensity noise sideband of the strong light pulse is filtered out, and a voltage corresponding to the single-photon horizontal phase shift is applied to the filtered light pulse to generate a co-generated light pulse.

3. The method of detecting an optical signal according to claim 2, wherein The method further includes: Step 21: Dynamically adjust the coupling according to the reference phase, and inject different additional phase markers into the separated first optical path signal and second optical path signal; Step 22: Perform power spectrum measurements of the signal frequency bands respectively, and adjust the power of the first optical path signal and the second optical path signal independently based on the measurement results; Step 23: Construct a simulated path containing a tunable delay and a phase compensation pattern for the second optical path signal; transmit the first optical path signal through a short reference channel and perform optical correlation detection with the second optical path signal processed by the simulated path; and adjust the length of the tunable delay according to the detection results. Step 24: The second optical path signal with the adjusted delay length is coupled with the pump light, and a nonlinear optical process is used to convert specific noise frequency components into light of different wavelengths and filter them out.

4. The method of detecting an optical signal according to claim 3, wherein The method further includes: Step 31: Separate the probe signal from the first optical path signal, interfere the probe signal with the test state set and measure the probability distribution, reconstruct the equivalent process matrix according to the probability distribution, and perform a feedforward transformation on the first optical path signal based on the process matrix to obtain the reconstructed first optical path signal; Step 32: Nonlinear phase locking is performed using the second optical path signal to generate the local oscillator reference light, and frequency-selective gain filtering related to the power spectrum measurement results of the signal frequency band is introduced during the locking process.

5. The method of detecting an optical signal according to claim 4, wherein The method further includes: Step 33: The reconstructed first optical path signal is subjected to asymmetric beam splitting and reciprocal phase transformation with the local oscillator reference light, and the interference result is measured and phase feedback adjustment is performed. Step 34: Monitor the error signal during the nonlinear phase-locking process and obtain the error signal power; perform correlation analysis between the error signal power and the photon arrival time stamp sequence to eliminate high-noise period events, and perform digital filtering based on the error signal on the retained events.

6. The method of detecting an optical signal according to claim 5, wherein, The method further includes: Step 41: Process the phase modulation information to obtain phase measurement values, compare the phase measurement values ​​with the expected phase values ​​to generate a phase error sequence; combine the time information corresponding to the time-slotted optical pulse sequence, dynamically fit the unsteady coherent decay model through the sliding window maximum likelihood estimation algorithm to obtain the primary channel evaluation parameters. Step 42: Calculate the channel perturbation tensor sequence based on the update history of the process matrix, perform high-order singular value decomposition on the channel perturbation tensor sequence to extract components that match the known perturbation source mode, and perform time-frequency domain cross-correlation analysis on the time derivatives of the non-steady-state coherent decay model parameters and the components to obtain secondary channel evaluation parameters.

7. The method of detecting an optical signal according to claim 6, wherein The method further includes: The primary channel evaluation parameters and the secondary channel evaluation parameters are normalized to form a multidimensional channel state vector; classical channel quality indices within the same time window are obtained; a classical joint state space is constructed based on the multidimensional channel state vector and the classical channel quality indices; in the classical joint state space, a set of joint basis vectors is found using a hybrid algorithm of principal component analysis and canonical correlation analysis; the multidimensional channel state vector and the classical channel quality indices are projected onto the joint basis vectors to obtain compressed joint feature codewords; the compressed joint feature codewords and the identifiers of the joint basis vectors are encapsulated into reported data.

8. The method of detecting an optical signal according to claim 7, wherein, The method further includes: Based on the aforementioned unsteady coherent decay model and its core parameters, the state fidelity evolution is simulated by numerically solving the corresponding equations, and the remaining lifetime is calculated and predicted. Based on the secondary channel evaluation parameters from multiple base stations, the inter-base station disturbance correlation matrix is ​​calculated and harmful resonance frequencies are analyzed to generate a cooperative avoidance strategy. Based on the compressed joint feature codewords and joint basis vector indexes from each base station, an approximate nearest neighbor search matching is performed using the historical decision knowledge base, and a recommendation decision is output. Based on the predicted remaining lifetime, the cooperative avoidance strategy, and the recommendation decision, when the predicted remaining lifetime is lower than a predetermined threshold and there is a recommended target base station that meets the cooperative avoidance strategy, a pre-resource reservation instruction is sent to the recommended target base station, and a dual-connection mode with the recommended target base station is initiated; after the dual-connection verification period, a handover to the recommended target base station is performed based on the verification result.

9. The method of detecting an optical signal according to claim 8, wherein, The method further includes: Based on the time-slotted optical pulse sequence, the additional phase marker, and the spectral characteristics of the co-occurring optical pulse, the phase modulation information is modulated into an orthogonal polarization state and assigned staggered center frequencies for cochannel transmission; The composite signal transmitted via a common channel is demultiplexed and time-domain gated to extract time-overlapping pulses, which are then injected into a nonlinear interferometer for Bell state projection.

10. The method of detecting an optical signal according to claim 9, wherein, The method further includes: Multiple coincidence counting is performed on the output port of the nonlinear interferometer. The coincidence count is weighted based on the dual-connection verification period information to obtain the projection result. The two communicating parties compare and select based on the projection result and the sent basis vector information to extract the key fragment. A portion of the key fragments are randomly selected as test bits and compared through a classic authentication channel to estimate the error rate and correlation. When the error rate is lower than a security threshold dynamically related to the channel evaluation parameters, the remaining key fragments are corrected and privacy amplified to generate a shared key.

Images