A method and system for controlling high-frequency variable optical pulse signals

Abstract

This application provides a high-frequency variable optical pulse signal control method and system, belonging to the field of signal control technology. This application acquires the ambient light intensity value of the target space and uses a wide-angle photoelectric sensor array to collect a time-series sequence of light intensity distribution. Frequency domain transformation is performed on each frame of the sequence to extract interference feature information such as frequency, amplitude, and phase. The interference feature information and the ambient light intensity value are combined to construct a time-series feature vector sequence, which is then input into a prediction model to obtain a prediction sequence containing future interference amplitude and phase. Candidate strong interference time slots are determined based on the prediction sequence, and the target strong interference time slot is selected from them. The target signal transmission time and corresponding target phase difference are calculated based on the predicted phase of the target time slot. The phase difference is used to adjust the high-frequency pulse trigger clock output by the programmable phase-locked loop to drive the optical pulse transmission circuit to transmit the signal, thereby achieving active control of the high-frequency variable optical pulse signal and improving the reliability of communication quality.

Description

Technical Field

[0001] This application belongs to the field of signal control technology, and in particular relates to a high-frequency variable optical pulse signal control method and system. Background Technology

[0002] High-frequency variable optical pulse signal control is a core technology in the fields of visible light communication and free space optical communication. By high-speed modulation of the light source to carry and transmit information, it shows broad application prospects in scenarios such as indoor wireless coverage, secure communication and vehicle networking.

[0003] Existing high-frequency variable optical pulse signal control methods mainly rely on adaptive equalization and channel coding techniques. For example, decision feedback equalizers are set at the receiver to compensate for signal distortion and inter-symbol interference, while other schemes employ strong forward error correction coding methods such as low-density parity-check codes to enhance the signal's resistance to random noise and channel fading.

[0004] However, in high-density deployment scenarios with multiple co-frequency optical signal sources, such as open-plan offices or large exhibition halls, signals from nearby light sources can cause strong coherent interference. Existing technologies struggle to effectively address this dynamically changing strong interference, lacking the ability to proactively perceive the spatiotemporal characteristics of interference sources. They typically only provide passive compensation after interference has already caused signal degradation, leading to frequent, transient drops in system communication quality and making it difficult to maintain consistently high-quality communication services. Therefore, existing technologies suffer from insufficient communication quality reliability in high-density deployment scenarios with multiple co-frequency optical signal sources. Summary of the Invention

[0005] The purpose of this application is to provide a high-frequency variable optical pulse signal control method, system, electronic device, and storage medium to solve the problem of insufficient communication quality reliability in the prior art.

[0006] To address the aforementioned technical problems, in a first aspect, this application provides a high-frequency variable optical pulse signal control method, the method comprising:

[0007] The ambient light intensity value of the target space, which includes multiple light pulse signal sources, is obtained, and the light field snapshots of the target space at multiple time points are continuously collected by a wide-angle photoelectric sensor array to obtain a time sequence of light intensity distribution.

[0008] Each frame of the light intensity distribution image in the time sequence of light intensity distribution is transformed in the frequency domain to obtain multiple frequency domain spectra. Based on multiple peak points in each frequency domain spectra, the interference feature information corresponding to each frame of light intensity distribution image is extracted. The interference feature information includes frequency, amplitude and phase.

[0009] Interference feature vectors are constructed using each interference feature information and ambient light intensity value to obtain a time-series feature vector sequence. The feature vector sequence is then input into a trained prediction model to obtain a time-series prediction feature vector sequence. Each prediction feature vector in the prediction feature vector sequence includes the amplitude and phase at future time points.

[0010] The time points corresponding to the predicted feature vectors with amplitudes greater than a preset amplitude in the predicted feature vector sequence are determined as candidate strong interference time slots, and the target strong interference time slot is determined from the candidate strong interference time slots based on the amplitude and time point.

[0011] Based on the phase of the predicted feature vector corresponding to the target strong interference time slot, calculate the target signal transmission time with a preset phase difference from the target strong interference time slot, and calculate the target phase difference based on the target signal transmission time and the preset reference clock.

[0012] The target phase difference is used as an adjustment command to the programmable phase-locked loop (PLL) to drive the PLL to adjust the phase of the output clock signal to generate a high-frequency pulse trigger clock. The high-frequency pulse trigger clock is then used to drive the optical pulse transmitting circuit to transmit a high-frequency variable optical pulse signal.

[0013] In one feasible implementation, determining the target strong interference time slot from the candidate strong interference time slots based on amplitude and time point includes:

[0014] Based on the amplitude of the candidate strong interference time slot and the time difference between the time point of the candidate strong interference time slot and the current time, the threat score of each candidate strong interference time slot is calculated. The threat score is directly proportional to the amplitude of the candidate strong interference time slot and inversely proportional to the time difference between the time point of the candidate strong interference time slot and the current time.

[0015] The candidate strong jamming slot with the highest threat score is determined as the target strong jamming slot.

[0016] In one feasible implementation, the method further includes:

[0017] Based on the amplitude of the candidate strong interference time slot, and the difference between the amplitude of the predicted feature vector corresponding to the candidate strong interference time slot and the amplitude of the previous adjacent predicted feature vector in the predicted feature vector sequence, the gradient parameter of each candidate strong interference time slot is calculated. The threat score is proportional to the gradient parameter of the candidate strong interference time slot.

[0018] In one feasible implementation, the method further includes:

[0019] Calculate the Euclidean distance between the first interference eigenvector and the other interference eigenvectors in the eigenvector sequence to obtain the Euclidean distance sequence;

[0020] If the variance of the Euclidean distance sequence is less than the preset variance, the feature vector sequence is input into the trained prediction model to obtain the time-series predicted feature vector sequence.

[0021] In one feasible implementation, each interference feature information includes at least one set of interference feature parameters, each set of interference feature parameters including frequency, amplitude and phase.

[0022] In one feasible implementation, an interference feature vector is constructed using each interference feature information and ambient light intensity value, including:

[0023] Sort the multiple sets of interference feature parameters of the interference feature information in descending order according to the amplitude;

[0024] The first N sets of interference feature parameters and ambient light intensity values ​​in the interference feature information are concatenated in a preset order to obtain the interference feature vector, where N is a positive integer.

[0025] In one feasible implementation, before inputting the feature vector sequence into a trained prediction model to obtain a temporal predicted feature vector sequence, the method includes:

[0026] Obtain a training sample set, which includes multiple training samples. Each training sample includes a historical feature vector sequence generated from historical ambient light intensity values ​​and historical light intensity distribution time series sequences, and a corresponding real feature vector sequence.

[0027] For each training sample in the training sample set, perform the following steps:

[0028] The historical feature vector sequence from each training sample is input into a pre-defined recurrent neural network model to obtain a predicted feature vector sequence.

[0029] The loss function value of the recurrent neural network model is determined based on the true feature vector sequence and the predicted feature vector sequence of each training sample.

[0030] If the loss function value does not meet the preset training stopping condition, adjust the model parameters of the recurrent neural network model to obtain an updated recurrent neural network model, and then return to input the historical feature vector sequence into the recurrent neural network model until the loss function value meets the training stopping condition, thus obtaining a trained prediction model.

[0031] Secondly, this application provides a high-frequency variable optical pulse signal control system, the system comprising:

[0032] The acquisition module is used to acquire the ambient light intensity value of the target space, which includes multiple light pulse signal sources, and to continuously collect light field snapshots of the target space at multiple time points through a wide-angle photoelectric sensor array to obtain a time sequence of light intensity distribution.

[0033] The extraction module is used to perform frequency domain transformation on each frame of light intensity distribution image in the light intensity distribution time sequence to obtain multiple frequency domain spectra, and extract the interference feature information corresponding to each frame of light intensity distribution image based on multiple peak points in each frequency domain spectra. The interference feature information includes frequency, amplitude and phase.

[0034] The prediction module is used to construct an interference feature vector with each interference feature information and ambient light intensity value to obtain a time-series feature vector sequence. The feature vector sequence is then input into the trained prediction model to obtain a time-series predicted feature vector sequence. Each predicted feature vector in the predicted feature vector sequence includes the amplitude and phase at a future time point.

[0035] The determination module is used to identify the time points corresponding to the predicted feature vectors in the predicted feature vector sequence whose amplitudes are greater than a preset amplitude as candidate strong interference time slots, and to determine the target strong interference time slot from the candidate strong interference time slots based on the amplitude and time point.

[0036] The calculation module is used to calculate the target signal transmission time with a preset phase difference from the target strong interference time slot based on the phase of the predicted feature vector corresponding to the target strong interference time slot, and to calculate the target phase difference based on the target signal transmission time and the preset reference clock.

[0037] The control module is used to apply the target phase difference as an adjustment command to the programmable phase-locked loop (PLL) to drive the PLL to adjust the phase of the output clock signal to generate a high-frequency pulse trigger clock, and use the high-frequency pulse trigger clock to drive the optical pulse transmitting circuit to transmit a high-frequency variable optical pulse signal.

[0038] Thirdly, this application provides an electronic device, comprising:

[0039] Memory, used to store computer programs;

[0040] A processor for executing a computer program to implement the steps of the high-frequency variable optical pulse signal control method as described in the first aspect above.

[0041] Fourthly, this application provides a computer-readable storage medium storing a computer program, which, when executed by a processor, can implement the steps of the high-frequency variable optical pulse signal control method described in the first aspect above.

[0042] The high-frequency variable optical pulse signal control method provided in this application uses a wide-angle photoelectric sensor array to perceive the composite optical field in a high-density deployment scenario in real time. It then utilizes frequency domain transformation and feature extraction to accurately identify the spatiotemporal characteristics of multiple co-frequency interference sources. Based on this, a spatiotemporal feature sequence incorporating interference characteristics and ambient lighting information is constructed, and a recurrent neural network model is used to predict the future dynamic evolution trend of the interference. This method can predict the occurrence time of strong interference before it actually occurs and proactively adjust the emission phase of the optical pulse for precise avoidance, fundamentally changing the lag mode of passive interference compensation in existing technologies. Therefore, this application can improve the stability and reliability of signal transmission in complex communication environments with multiple co-frequency optical signal sources, effectively ensuring continuous high-quality communication services.

[0043] Furthermore, in high-density deployment scenarios, multiple strong interferences predicted not only vary in intensity but also in urgency and duration. A multi-dimensional threat assessment model comprehensively quantifies several key factors, including the predicted intensity of interference, the proximity of occurrence time, and the rate of intensity change, to identify and prioritize the avoidance of interference events that pose the greatest risk to real-time communication quality. This avoids ignoring the communication interruption risk from other, less critical but more urgent threats by simply avoiding the strongest interference, and more accurately distinguishes and handles different forms of interference. This improves the accuracy and adaptability of interference avoidance strategies to dynamic environments, ensuring reliable and stable communication quality in complex and ever-changing interference environments. Attached Figure Description

[0044] To more clearly illustrate the technical solutions of the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0045] Figure 1 A flowchart illustrating a high-frequency variable optical pulse signal control method provided in this application embodiment;

[0046] Figure 2 This application provides a schematic diagram illustrating a specific implementation of a method for determining a target strong interference time slot.

[0047] Figure 3 This is a schematic diagram illustrating a specific implementation of a method for constructing interference feature vectors provided in an embodiment of this application;

[0048] Figure 4 A schematic diagram of a high-frequency variable optical pulse signal control system provided in this application embodiment;

[0049] Figure 5 This is a schematic diagram of the hardware structure of an electronic device provided in one embodiment of this application. Detailed Implementation

[0050] The inventive concept of this application lies in first capturing the real-time spatial pattern formed by multi-source interference through optical sensing, then decomposing the pattern into quantifiable multi-dimensional temporal features using frequency domain analysis, then using a temporal prediction model to deduce the future evolution of these features in order to predict interference, and finally directly converting the prediction results into precise phase control of the light pulse emission time, realizing a fundamental shift from passive compensation to proactive intelligent avoidance.

[0051] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0052] Figure 1 A flowchart illustrating a high-frequency variable optical pulse signal control method provided in one embodiment of this application is shown.

[0053] The core of this application is to provide a high-frequency variable optical pulse signal control method, the flowchart of one specific implementation of which is shown below. Figure 1 As shown, the method includes steps S110 to S160.

[0054] This application is primarily applied to high-density indoor environments with multiple visible light communication access points, such as open-plan offices, large conference centers, and smart factories. In these environments, multiple lighting devices acting as signal sources operate simultaneously, and their emitted optical signals at the same frequency superimpose in space. Due to frequent user movement and dynamic environmental changes, complex and unpredictable strong interference is generated at the receiving end, leading to a sharp decline in the signal quality of the communication link and making it difficult to guarantee service stability.

[0055] S110: Acquire the ambient light intensity value of the target space including multiple light pulse signal sources, and continuously collect light field snapshots of the target space at multiple time points through a wide-angle photoelectric sensor array to obtain a time sequence of light intensity distribution.

[0056] Ambient light intensity is a scalar value representing the overall background illumination intensity generated by non-communication light sources such as sunlight and ordinary lighting within a target space. A wide-angle photoelectric sensor array is a two-dimensional area array sensor device whose function is to capture spatial light information within a large field of view. A light field snapshot refers to a frame of two-dimensional image data acquired by the wide-angle photoelectric sensor array at a specific instant, represented as a pixel matrix, where the value of each pixel represents the illumination intensity at that spatial location. A light intensity distribution time series is serialized data composed of multiple consecutive light field snapshots. This sequence dynamically records the evolution of the spatial light intensity distribution over time, resulting from the interference of multiple light pulse signal sources and their interaction with environmental objects.

[0057] First, the ambient light intensity value is obtained by using a sensor specifically designed to measure slowly varying light signals, such as a PIN photodiode or an Ambient Light Sensor (ALS) chip. In a well-lit office environment, such as an indoor space under sunlight in the afternoon, the sensor measures the background light generated by the superposition of natural light and indoor lighting, and outputs a high value as the ambient light intensity value, such as 800, through an analog-to-digital converter.

[0058] Simultaneously, a high-frequency sampling clock, such as once every microsecond (e.g., once per data communication symbol rate), triggers a wide-angle photoelectric sensor array, such as a CMOS image sensor or a single-photon avalanche diode (SPAD) array, to perform a global shutter exposure to capture a single frame of light field snapshot. For example, when a user passes under the sensor with a mobile terminal, the captured 64x64 pixel grayscale image is a light field snapshot, and its pixel value matrix forms a complex interference pattern. This pattern is formed by the superposition of the local light source, two adjacent interfering light sources, and light reflected from the surface of the mobile terminal. Constructive interference areas in the image appear as bright fringes with high pixel values, while destructive interference areas appear as dark fringes with low pixel values. Over the next few microseconds, as the user continues to move, multiple consecutively captured light field snapshots show the dynamic changes in the shape, position, and contrast of these interference fringes. These consecutively captured light field snapshots are arranged according to the acquisition time sequence and finally integrated to obtain a time-series sequence of light intensity distribution.

[0059] S120: Perform frequency domain transformation on each frame of the light intensity distribution image in the time sequence of light intensity distribution to obtain multiple frequency domain spectra, and extract the interference feature information corresponding to each frame of light intensity distribution image based on multiple peak points in each frequency domain spectra. The interference feature information includes frequency, amplitude and phase.

[0060] A light intensity distribution image, also known as a light field snapshot, is a single frame of data in a time-series sequence of light intensity distribution. It is a two-dimensional pixel matrix, where the value of each pixel represents the light intensity at that spatial location. The image as a whole represents a complex pattern formed by the interference of multiple superimposed light signal sources. Interference feature information is a set of structured data extracted after processing a frame of light intensity distribution image. This set of data includes frequency, amplitude, and phase parameters that characterize the main components of the interference pattern.

[0061] For each frame of the light intensity distribution image in the time-series of light intensity distribution, a two-dimensional fast Fourier transform algorithm is first applied to its pixel matrix to transform the image from the spatial domain to the spatial frequency domain, resulting in a frequency domain spectrum in the form of a complex matrix. Next, peak point search is performed in this frequency domain spectrum, identifying multiple significant peak points by finding local maxima points with energy exceeding a preset energy threshold. For each identified peak point, its two-dimensional coordinates in the frequency domain spectrum are extracted as a frequency parameter, its complex magnitude as an amplitude parameter, and its complex argument as a phase parameter. Finally, the frequency, amplitude, and phase parameters corresponding to all peak points extracted from the frequency domain spectrum of a single frame of light intensity distribution image are combined to constitute the interference feature information corresponding to that frame of image.

[0062] For example, a frame of light intensity distribution image from a time-series light intensity distribution pattern is processed. For instance, the light intensity distribution image can be M×N pixels. A two-dimensional fast Fourier transform is performed on this M×N pixel matrix to generate a frequency domain spectrum, also of size M×N. In this frequency domain spectrum, in addition to the bright spot representing the DC component at the center, several discrete bright spots with significant energy are identified. These bright spots are the identified peak points, each representing a major optical signal interference component. For one peak point, its coordinates in the spectrum are (fx, fy), and the extracted frequency is the coordinate value (fx, fy); its corresponding complex value is calculated as modulus A and argument P, and the extracted amplitude is A and phase is P. Similarly, the same processing is performed on all other identified peak points to obtain their respective frequency, amplitude, and phase parameters. Finally, all groups of frequency, amplitude, and phase parameters extracted from this frame are combined to form the interference feature information corresponding to this frame of light intensity distribution image. For example, if three main peak points are identified, the final generated interference feature information can be represented as a set containing three sets of parameters: {[(fx1, fy1), A1, P1], [(fx2, fy2), A2, P2], [(fx3, fy3), A3, P3]}.

[0063] S130: Construct an interference feature vector with each interference feature information and ambient light intensity value to obtain a time-series feature vector sequence. Input the feature vector sequence into the trained prediction model to obtain a time-series prediction feature vector sequence. Each prediction feature vector in the prediction feature vector sequence includes the amplitude and phase at future time points.

[0064] An interferometric eigenvector is a fixed-length numerical vector formed by uniformly encoding the characteristics of multiple interference sources and ambient lighting information at a single time point. A temporal eigenvector sequence is a two-dimensional data matrix composed of interferometric eigenvectors from multiple consecutive time points arranged in chronological order. The trained prediction model is a recurrent neural network model that has been trained and optimized using a large amount of historical data. The predicted eigenvector sequence is the prediction result of the trained prediction model after receiving a historical eigenvector sequence, predicting the key parameters of the interference signal over a future period. Each predicted eigenvector in the sequence includes an estimate of the amplitude and phase of the interference signal at a specific future time point.

[0065] First, for each time point t, the acquired interferometric feature information is a set of N sets of frequency, amplitude, and phase parameters. To transform this into a fixed-length interferometric feature vector, the N sets of parameters are first sorted in descending order based on their amplitude values. Then, the first M sets of parameters after sorting are selected, where M is a preset integer, such as 3, and these M sets of parameters are flattened according to their sorted order, i.e., arranged sequentially into a long vector. Next, the ambient light intensity value is appended to the end of this long vector, together forming the fixed-length interferometric feature vector corresponding to time point t. Then, for prediction, K interferometric feature vectors from the current time t and the K-1 time points before it, where K is the lookback window length, are taken and combined in chronological order to form a temporal feature vector sequence. Finally, this temporal feature vector sequence is used as input and fed into a trained prediction model for forward propagation calculation, outputting a temporal predicted feature vector sequence. The prediction model can be based on a Long Short-Term Memory (LSTM) network, and this sequence contains predicted values ​​for the amplitude and phase of the interference signal at L future time points.

[0066] For example, at time point t, the interference feature information output in step S120 is {[(fx1, fy1), A1, P1], [(fx2, fy2), A2, P2], [(fx3, fy3), A3, P3]}, and the ambient light intensity value is E. At this time, the number of main interference sources M extracted is set to 2. The two sets of parameters with the largest amplitudes are expanded to obtain a subvector [fx1, fy1, A1, P1, fx2, fy2, A2, P2]. The ambient light intensity value E is then appended to the end, finally constructing the interference feature vector Vt=[fx1, fy1, A1, P1, fx2, fy2, A2, P2, E] at time point t. Assuming the review window length K is 15, the current and past 15 such interference feature vectors {Vt-14, ..., Vt} are combined into a temporal feature vector sequence. This sequence is input into the trained prediction model, and the model then outputs a time-series prediction feature vector sequence, such as a prediction feature vector sequence containing predictions for 10 future time points. Each prediction feature vector contains the prediction magnitude and prediction phase corresponding to a future time point, such as [At+1, Pt+1], [At+2, Pt+2], ..., [At+10, Pt+10].

[0067] S140: The time points corresponding to the predicted feature vectors with amplitudes greater than the preset amplitude in the predicted feature vector sequence are determined as candidate strong interference time slots, and the target strong interference time slot is determined from the candidate strong interference time slots based on the amplitude and time point.

[0068] Candidate strong interference time slots are sets of one or more future time points obtained through preliminary screening of predicted feature vector sequences, representing the times of potential interference events that may significantly impact communication in the future. Target strong interference time slots are the time points corresponding to the most pressing interference event that needs to be avoided, selected from the set of candidate strong interference time slots.

[0069] First, the predicted feature vector sequence of the time series is traversed. For each predicted feature vector in the sequence, its predicted amplitude component is extracted and compared with a preset amplitude threshold. This threshold is pre-set based on communication quality requirements and represents the lower limit of interference intensity that can significantly damage the signal. The time points corresponding to all predicted feature vectors with predicted amplitudes greater than this threshold are collectively formed into a list of candidate strong interference time slots. Next, for each time point in the candidate strong interference time slot list, a threat score is calculated. The threat score is calculated such that the score value is directly proportional to the predicted amplitude in the predicted feature vector corresponding to that time point and inversely proportional to the time difference between that time point and the current time. After calculating the threat scores of all candidate strong interference time slots, the candidate strong interference time slot with the highest threat score is determined as the final target strong interference time slot output.

[0070] For example, a preset amplitude threshold of 200 is set, and the predicted feature vector sequence containing predictions for 10 future time points is traversed, i.e., {[At+1, Pt+1], ..., [At+10, Pt+10]}. Assume that the predicted amplitude at future time point t+4 is 210, and the predicted amplitude at future time point t+9 is 250. Therefore, these two time points {t+4, t+9} are identified as candidate strong interference time slots. Subsequently, threat scores are calculated for these two candidate objects. For t+4, its predicted amplitude is 210, and the time difference is 4 units, resulting in a threat score of A; for t+9, its predicted amplitude is 250, and the time difference is 9 units, resulting in a threat score of B. Since the time difference plays a significant weighting role in the score calculation, even if the amplitude of t+9 is higher, its greater time distance may cause its final threat score B to be lower than the threat score A of t+4. After comparison, the candidate with the highest threat score, such as t+4, is selected as the sole target for strong interference in the time slot.

[0071] S150: Calculate the target signal transmission time with a preset phase difference from the target strong interference time slot based on the phase of the predicted feature vector corresponding to the target strong interference time slot, and calculate the target phase difference based on the target signal transmission time and the preset reference clock.

[0072] The target signal transmission time refers to the ideal future time point for transmitting the next optical pulse signal. This time point is chosen to ensure the optical pulse signal avoids predicted strong interference. The preset phase difference is a pre-defined phase angle value, which can be π radians or an odd multiple thereof, used to define the optimal phase relationship between the safe transmission time and the peak interference time. The preset reference clock is a high-frequency periodic clock signal that operates stably within the system, providing a unified time reference for all timing operations. The target phase difference directly controls the phase adjustment amount in the hardware, representing the phase offset required to shift the actual transmission time of the optical pulse from the normal time defined by the reference clock to the target signal transmission time.

[0073] First, based on the target strong interference time slot determined in step S140, the predicted phase value of the predicted feature vector corresponding to that time point is found and extracted from the predicted feature vector sequence of the timing output in step S130. Next, this predicted phase value is arithmetically calculated with a preset phase difference, such as by addition, to calculate a target phase; this target phase physically corresponds to the energy trough position of the interference signal waveform. Subsequently, based on the clock frequency of the communication system, this target phase is converted into an absolute time coordinate, which is then determined as the target signal transmission time. Finally, the calculated target signal transmission time is compared with the next trigger time of the currently preset reference clock, the time difference between the two is calculated, and this time difference is converted into a corresponding phase angle value, which is the final generated target phase difference, used to guide subsequent hardware adjustments.

[0074] For example, in step S140, the target strong interference time slot is determined as the future time point t+4. The predicted feature vector corresponding to time t+4 is found in the predicted feature vector sequence, and its predicted phase value is extracted as Pt+4. The preset phase difference is set to π radians. The purpose of choosing π radians is to make the signal transmission time of this device staggered from the predicted interference peak time by half a cycle, thereby achieving the best avoidance effect. The predicted phase Pt+4 is added to π to obtain the target phase Ptarget = Pt+4 + π. According to the high-frequency clock frequency of the system, this target phase Ptarget is converted into a specific future transmission time point, namely the target signal transmission time Ttarget. At this time, the preset reference clock inside the system is running stably, and its next naturally triggered normal transmission time is Tbase. The system calculates the time difference ΔT = Ttarget - Tbase between these two times, and converts this time difference ΔT into a phase angle value, such as Φoffset. This Φoffset is the final output target phase difference. This target phase difference will be sent to the subsequent programmable phase-locked loop, instructing it to precisely delay or advance the next clock pulse by ΔT.

[0075] S160: The target phase difference is used as an adjustment command to the programmable phase-locked loop to drive the programmable phase-locked loop to adjust the phase of the output clock signal to generate a high-frequency pulse trigger clock, and the high-frequency pulse trigger clock is used to drive the optical pulse transmitting circuit to transmit a high-frequency variable optical pulse signal.

[0076] A programmable phase-locked loop (PLL) is an electronic circuit that receives a reference clock signal and a digitized phase adjustment command, and generates an output clock signal with the same frequency as the reference clock but a precisely offset phase. A high-frequency pulse-triggered clock is the final clock signal generated by the PLL with its phase dynamically adjusted; each rising edge of this clock pulse precisely corresponds to a calculated ideal emission moment. An optical pulse emitting circuit is an electronic hardware unit used to drive a light source. It receives a high-frequency pulse-triggered clock and controls the current flowing to the light source, such as a light-emitting diode (LED) or laser diode, based on the clock's trigger, thereby generating synchronized light pulses.

[0077] First, the calculated target phase difference, a digital phase adjustment value, is sent to the phase control register of the programmable phase-locked loop (PLL) via the control bus. Upon receiving this adjustment command, the PLL uses its internal phase frequency detector and voltage-controlled oscillator to precisely adjust the phase of its output clock relative to a preset reference clock; the adjustment amount is the received target phase difference. After adjustment, the PLL outputs a high-frequency pulse trigger clock with its phase corrected in real time. Finally, this high-frequency pulse trigger clock is used as a trigger signal and sent to the optical pulse transmitting circuit. This circuit, for example, is a high-speed switch driver that instantaneously turns on upon receiving each rising edge of the trigger clock, allowing a large current pulse to flow through the light source, thereby driving the light source to emit a brief, bright light pulse perfectly synchronized with the high-frequency pulse trigger clock. This light pulse is the final output high-frequency variable optical pulse signal.

[0078] For example, the target phase difference calculated in step S150 is a value Φoffset representing the need to advance the next pulse by ΔT time. This value Φoffset is written to the control interface of the programmable phase-locked loop (PLL). The internal circuitry of the PLL responds immediately, generating an output clock pulse in the next clock cycle that is ΔT time ahead of the normal trigger point of the reference clock. This advanced pulse is the high-frequency pulse trigger clock. The rising edge signal of this trigger clock is sent to the optical pulse transmitting circuit, which then drives a high-power light-emitting diode to emit a very short-duration optical pulse. Since the transmission time of the trigger clock is based on the prediction and avoidance calculations of interference, the finally emitted optical pulse, when propagating in space, will precisely avoid the strong interference event that was previously predicted to occur at time t+4, thus ensuring that it can be clearly and accurately received by the remote receiver.

[0079] By applying the target phase difference, representing the avoidance strategy, to a programmable phase-locked loop, the phase of the high-frequency pulse trigger clock is precisely adjusted, thereby controlling the optical pulse emission circuit. Finally, using this dynamically phase-adjusted clock signal, the light source is driven to emit optical pulses at the predicted interference trough, thus achieving active and intelligent control of the time-domain characteristics of the high-frequency variable optical pulse signal emission.

[0080] Figure 2 A flowchart illustrating a method for determining a target strong interference time slot according to an embodiment of this application is shown.

[0081] In one feasible implementation, the target strong interference time slot is determined from the candidate strong interference time slots based on the amplitude and time point, such as... Figure 2 As shown, steps S210 to S220 are included.

[0082] S210: Calculate the threat score for each candidate strong interference time slot based on the amplitude of the candidate strong interference time slot and the time difference between the time point of the candidate strong interference time slot and the current time. The threat score is directly proportional to the amplitude of the candidate strong interference time slot and inversely proportional to the time difference between the time point of the candidate strong interference time slot and the current time.

[0083] The threat score is a comprehensive quantitative indicator used to assess the actual risk level posed by each candidate strong interference time slot to instant messaging. This score combines the intensity and urgency of the interference event; a higher score indicates a higher priority for avoiding the interference event. The time difference is a value representing the urgency of a future event, specifically referring to the time interval between the predicted occurrence time of the candidate strong interference time slot and the current system time.

[0084] For each candidate in the strong interference time slot list, firstly, obtain its corresponding prediction amplitude value in the prediction feature vector sequence, and calculate the time difference between its predicted occurrence time and the current time recorded by the current system clock. Then, calculate the threat score of the candidate through a preset scoring function; the design principle of this function is that the threat score is positively correlated with the prediction amplitude, that is, the stronger the interference, the higher the score, and negatively correlated with the calculated time difference, that is, the closer the event, the higher the score. This calculation process will be performed on all candidate strong interference time slots in the list one by one, and finally generate a corresponding threat score for each candidate. The preset scoring function can be expressed as shown in formula (1):

[0085] (1)

[0086] in, Indicates the first Threat scores for each candidate strong interference slot. Indicates the first The predicted amplitude of each candidate strong interference time slot, Indicates the first The time difference between the predicted occurrence time of each candidate strong interference time slot and the current time. and These are preset normal weight coefficients, used to adjust the importance of amplitude and time difference in the overall score.

[0087] For example, the candidate strong interference time slot list is {t+4, t+9}, with corresponding predicted amplitudes of 210 and 250, respectively. Assume weighting coefficients... and All are set to 10. For time point t+4, the time difference is 4 time units, and the threat score calculated according to formula (1) is 25.2; for time point t+9, the time difference is 9 time units, and the threat score calculated is 25.0.

[0088] S220: The candidate strong jamming slot with the highest threat score is determined as the target strong jamming slot.

[0089] After calculating the threat scores for all candidate strong interference time slots, these scores are compared. The maximum value is found by iterating through all calculated threat scores. Finally, the candidate strong interference time slot corresponding to this maximum threat score is determined as the target strong interference time slot, and its time point information is output. For example, after comparing two threat scores of 25.2 and 25.0, 25.2 is determined to be the maximum value. Therefore, the candidate strong interference time slot with the highest threat score, i.e., t+4, is determined as the final target strong interference time slot, and its time point information is output.

[0090] In one feasible implementation, the method further includes:

[0091] Based on the amplitude of the candidate strong interference time slot, and the difference between the amplitude of the predicted feature vector corresponding to the candidate strong interference time slot and the amplitude of the previous adjacent predicted feature vector in the predicted feature vector sequence, the gradient parameter of each candidate strong interference time slot is calculated. The threat score is proportional to the gradient parameter of the candidate strong interference time slot.

[0092] The gradient parameter is a value used to quantify the instantaneity of the interference peak. It represents the rate at which the interference energy increases. The larger the value, the steeper the interference peak and the more likely it is to be instantaneous.

[0093] For each candidate in the list of strong interference time slots, in addition to obtaining its predicted amplitude and calculating the time difference, it is also necessary to find the time point immediately preceding the candidate's time point from the predicted feature vector sequence of the time series, and extract the predicted amplitude value of the preceding time point. Subsequently, by subtracting the predicted amplitude value of the preceding adjacent time point from the predicted amplitude value of the current strong interference time slot, the difference between the two is calculated, and this difference is determined as the gradient parameter of the candidate. Finally, when calculating the threat score, this gradient parameter will be incorporated into the scoring function as a new positive correlation factor. For example, the threat scoring function can also be as shown in formula (2):

[0094] (2)

[0095] in, Indicates the first Threat scores for each candidate strong interference slot. Indicates the first The predicted amplitude of each candidate strong interference time slot, Indicates the first The time difference between the predicted occurrence time of each candidate strong interference time slot and the current time. and These are preset normal coefficient weighting factors, used to adjust the importance of amplitude and time difference in the overall score. Indicates the first The gradient parameters of the candidate strong interference time slots, , It is a preset positive constant weight coefficient used to adjust the importance of the gradient parameter in the overall score.

[0096] For example, the candidate strong interference time slot list is {t+4, t+9}, with corresponding predicted amplitudes of 210 and 250, respectively. Assume weighting coefficients... , , The values ​​are set to 10, 10, and 0.1 respectively. First, the gradient parameters are calculated: for t+4, the magnitude is 210. Assuming the predicted magnitude of the previous time point t+3 is 150, the gradient parameter G1 = 60; for t+9, the magnitude is 250. Assuming the predicted magnitude of the previous time point t+8 is 240, the gradient parameter G2 = 10. Then, the total threat score is calculated according to formula (2): for t+4, the threat score S1 ≈ 31.22; for t+9, the threat score S2 ≈ 25.98.

[0097] In this embodiment, multiple strong interferences predicted in a high-density deployment scenario not only vary in intensity but also in urgency and duration. A multi-dimensional threat assessment model comprehensively quantifies several key factors, including the predicted intensity of the interference, the proximity of its occurrence time, and the rate of intensity change, to identify and prioritize the avoidance of interference events that pose the greatest risk to real-time communication quality. This avoids ignoring the communication interruption risk caused by other, less serious but more urgent threats by simply avoiding the strongest interference, and more accurately distinguishes and handles different forms of interference. This improves the accuracy of the interference avoidance strategy and its adaptability to dynamic environments, ensuring reliable and stable communication quality in complex and ever-changing interference environments.

[0098] In one feasible implementation, the method further includes:

[0099] Calculate the Euclidean distance between the first interference eigenvector and the other interference eigenvectors in the eigenvector sequence to obtain the Euclidean distance sequence.

[0100] Euclidean distance is a numerical value used to quantify the degree of difference between two multidimensional vectors. It calculates the straight-line distance between the two vectors in multidimensional space; the larger the distance value, the greater the difference in the channel states described by the two vectors. An Euclidean distance sequence is a one-dimensional time series composed of multiple Euclidean distance values. Each value in the sequence represents the degree of deviation of the interference eigenvector at a certain time point from the state of the starting point of the sequence.

[0101] First, from the temporal feature vector sequence of the input model, the first interferometric feature vector is selected as the reference vector. Then, a loop calculation process is used to sequentially extract all other interferometric feature vectors from the second to the last in the sequence, and calculate the Euclidean distance between each of them and the previously selected reference vector. The specific method for calculating the Euclidean distance is to subtract the values ​​of corresponding dimensions of the two vectors one by one, sum the squares, and then take the square root of the result. This loop calculation process generates a corresponding Euclidean distance value for each non-initial vector in the sequence. Finally, all the calculated Euclidean distance values ​​are arranged according to the temporal order of their corresponding vectors in the original sequence, forming an Euclidean distance sequence. For example, the temporal feature vector sequence to be processed is {Vt-14, ..., Vt}. The first vector Vt-14 is selected as the reference vector. Then, the Euclidean distance D1 between Vt-13 and Vt-14, the Euclidean distance D2 between Vt-12 and Vt-14 are calculated sequentially until the Euclidean distance D14 between Vt and Vt-14 is calculated. Finally, a sequence of Euclidean distances {D1, D2, ..., D14} containing 14 values ​​is obtained.

[0102] If the variance of the Euclidean distance sequence is less than the preset variance, the feature vector sequence is input into the trained prediction model to obtain the time-series predicted feature vector sequence.

[0103] Variance is a statistical indicator used to measure the dispersion or volatility of a data sequence. The smaller the variance value, the more stable the fluctuation of the data sequence. Preset variance is a pre-set threshold used to define the level of stability of channel state changes. When the actual calculated variance is less than this threshold, the recent changes in the channel are considered to be regular and predictable.

[0104] First, variance is calculated on the Euclidean distance sequence generated in the previous step. The variance result is then compared with a pre-set variance threshold based on experimental data or system requirements. Only if the calculated variance is less than this pre-set threshold is the current historical channel state considered sufficiently stable and suitable for prediction, and the subsequent steps proceed, inputting the complete time-series feature vector sequence into the trained prediction model. If the calculated variance is greater than or equal to the pre-set variance threshold, it indicates that recent channel changes are too drastic or irregular, and the prediction process is terminated. For example, suppose the pre-set variance threshold is 0.1. In one scenario, due to slow user movement, the variance of the calculated Euclidean distance sequence {D1, ..., D14} is 0.05. Since 0.05 is less than 0.1, the channel is considered stable, and the original 15-point time-series feature vector sequence is fed into the prediction model. In another scenario, due to severe obstruction of the optical path, the calculated variance is 0.3. Since 0.3 is greater than 0.1, the channel is determined to be unstable, so the prediction is terminated and the sequence is not input into the prediction model.

[0105] In one feasible implementation, each interference feature information includes at least one set of interference feature parameters, each set of interference feature parameters including frequency, amplitude and phase.

[0106] Interference characteristic parameters are a set of quantitative values ​​used to describe the physical properties of the interference components of a single optical signal. These values ​​include the frequency characterizing the spatial distribution characteristics of the interference component, the amplitude characterizing its energy strength, and the phase characterizing its waveform relative position.

[0107] In step S120, by searching for peak points in the frequency domain spectrum, one or more significant peak points may be identified, each corresponding to a major optical signal interference component. Therefore, to fully describe the complex interference situation at the current moment, the interference feature information ultimately generated by S120 is a set containing data extracted from all identified peak points. Specifically, this interference feature information consists of at least one set of interference feature parameters; in a simple scenario with only one major interference source, this information may contain only one set of interference feature parameters. For example, at a certain moment, since the receiver is simultaneously within the coverage area of ​​the local light source and two other nearby interfering light sources, step S120 identifies three independent peak points in its frequency domain spectrum, extracts their respective frequencies, amplitudes, and phases, and thus obtains three sets of interference feature parameters, such as the first set being [(fx1, fy1), A1, P1], the second set being [(fx2, fy2), A2, P2], and the third set being [(fx3, fy3), A3, P3]. At this point, the final output of the interference feature information is a set containing these three sets of interference feature parameters.

[0108] Figure 3 A flowchart illustrating a method for constructing interference feature vectors according to an embodiment of this application is shown.

[0109] In one feasible implementation, an interference feature vector is constructed using each interference feature information and the ambient light intensity value, such as... Figure 3 As shown, the method includes steps S310 to S320.

[0110] S310: Sort multiple sets of interference feature parameters in descending order based on amplitude.

[0111] First, each set of interference feature parameters in the interference feature information set is traversed, and the amplitude value is extracted as the sorting key. Then, a standard sorting algorithm, such as quicksort or mergesort, is used to sort these multiple sets of interference feature parameters in descending order. After sorting, the originally unordered parameter set is reorganized into an ordered list. For example, if the interference feature information output by S120 is {[(fx2, fy2), 180.0, P2], [(fx1, fy1), 250.0, P1], [(fx3, fy3), 155.0, P3]}, then based on the amplitude values ​​of 250.0, 180.0, and 155.0, these three sets of parameters are rearranged into an ordered list: [[(fx1, fy1), 250.0, P1], [(fx2, fy2), 180.0, P2], [(fx3, fy3), 155.0, P3]].

[0112] S320: The first N sets of interference feature parameters and ambient light intensity values ​​in the interference feature information are concatenated in a preset order to obtain the interference feature vector, where N is a positive integer.

[0113] First, based on a preset positive integer N, the first N sets of interference feature parameters are selected from the obtained ordered parameter list. Next, these N sets of parameters are flattened, that is, the frequency, amplitude, and phase values ​​within each set are arranged in a fixed internal order, for example, frequency first, then amplitude, then phase. Then, these N flattened parameters are concatenated end-to-end into a long vector according to their order in the ordered list. Finally, the ambient light intensity value is appended as an independent value to the end of this long vector. Through this concatenation process, regardless of the number of originally detected interference sources, an interference feature vector with the same fixed length and structured order can be generated. Assuming the preset positive integer N is 2, meaning only the two strongest interference sources are considered, the first two sets of parameters are selected from the sorted list: [(fx1, fy1), 250.0, P1] and [(fx2, fy2), 180.0, P2]. The two sets of parameters are then concatenated into a long vector: [fx1, fy1, 250.0, P1, fx2, fy2, 180.0, P2]. Finally, the ambient light intensity value of 800 is appended to the end of this vector to obtain the final interference eigenvector: [fx1, fy1, 250.0, P1, fx2, fy2, 180.0, P2, 800].

[0114] In one feasible implementation, before inputting the feature vector sequence into a trained prediction model to obtain a temporal predicted feature vector sequence, the method includes:

[0115] Obtain a training sample set, which includes multiple training samples. Each training sample includes a historical feature vector sequence generated from historical ambient light intensity values ​​and historical light intensity distribution time series sequences, and a corresponding real feature vector sequence.

[0116] First, a database is constructed by continuously collecting time-series sequences of light intensity distribution and ambient light intensity values ​​over a long period of time in a laboratory or actual deployment environment across multiple scenarios. Then, the feature extraction and vector construction processes described in S120 and S130 are performed on all the raw data in this database, transforming it into a sufficiently long baseline feature vector sequence encompassing various operating conditions. Finally, a sliding window technique is used to segment a large number of training samples from this baseline feature vector sequence. Specifically, a historical window length K and a future prediction length L are set. Starting from the beginning of the sequence, segments of length K are sequentially extracted as historical feature vector sequences, and simultaneously, the immediately following segments of length L are extracted as their corresponding true feature vector sequences. This process slides continuously along the baseline feature vector sequence until the end of the sequence, thereby generating a training sample set containing multiple training samples.

[0117] For each training sample in the training sample set, perform the following steps:

[0118] The historical feature vector sequence from each training sample is input into a pre-defined recurrent neural network model to obtain a predicted feature vector sequence. Based on the true and predicted feature vector sequences of each training sample, the loss function value of the recurrent neural network model is determined. If the loss function value does not meet the pre-defined training stopping condition, the model parameters of the recurrent neural network model are adjusted to obtain an updated recurrent neural network model. The process is repeated until the loss function value meets the training stopping condition, resulting in a trained prediction model.

[0119] First, the historical feature vector sequence of a training sample from the training sample set is input into an initialized recurrent neural network model based on a long short-term memory network. The model calculates a predicted feature vector sequence with the same length as the true feature vector sequence through forward propagation. Next, a loss function, such as the mean squared error function, is used to calculate the difference between the predicted and true feature vector sequences point by point. The differences at all points are summed or averaged to obtain a total loss function value. Then, it is determined whether the loss function value meets the preset training stopping condition. If not, the backpropagation process is initiated. A backpropagation algorithm over time can be used to calculate the gradient of the loss with respect to each parameter within the model based on the loss function value. Based on the gradient, optimizers such as Adam and SGD are used to fine-tune the model parameters so that the model can make more accurate predictions when faced with the same input in the next iteration. This process is repeated for all samples in the training sample set until the model's loss function value converges to a sufficiently low level, satisfying the training stopping condition. The resulting model is the trained prediction model.

[0120] This embodiment uses a wide-angle photoelectric sensor array to perceive the composite light field in a high-density deployment scenario in real time, and accurately identifies the spatiotemporal characteristics of multiple co-frequency interference sources by utilizing frequency domain transformation and feature extraction. Based on this, a spatiotemporal feature sequence incorporating interference characteristics and ambient lighting information is constructed, and a recurrent neural network model is used to predict the future dynamic evolution trend of the interference. This allows for the prediction of the occurrence time of strong interference before it actually happens, and proactive adjustment of the emission phase of the light pulse for precise avoidance, fundamentally changing the lag mode of passive interference compensation in existing technologies. Therefore, this application can improve the stability and reliability of signal transmission in complex communication environments with multiple co-frequency optical signal sources, effectively ensuring continuous high-quality communication services.

[0121] Based on the same concept, this application provides a high-frequency variable optical pulse signal control system, which is described below in conjunction with... Figure 4The high-frequency variable optical pulse signal control system provided in the embodiments of this application will be described in detail.

[0122] Figure 4 This is a structural block diagram of a high-frequency variable optical pulse signal control system illustrated in an embodiment of this application. Figure 4 As shown, the system may include:

[0123] The acquisition module 410 is used to acquire the ambient light intensity value of the target space including multiple light pulse signal sources, and continuously collect light field snapshots of the target space at multiple time points through a wide-angle photoelectric sensor array to obtain a time sequence of light intensity distribution.

[0124] The extraction module 420 is used to perform frequency domain transformation on each frame of light intensity distribution image in the light intensity distribution time sequence to obtain multiple frequency domain spectra, and extract the interference feature information corresponding to each frame of light intensity distribution image based on multiple peak points in each frequency domain spectra. The interference feature information includes frequency, amplitude and phase.

[0125] The prediction module 430 is used to construct an interference feature vector with each interference feature information and the ambient light intensity value to obtain a time-series feature vector sequence, and input the feature vector sequence into the trained prediction model to obtain a time-series prediction feature vector sequence. Each prediction feature vector in the prediction feature vector sequence includes the amplitude and phase at a future time point.

[0126] The determination module 440 is used to determine the time point corresponding to the predicted feature vector with an amplitude greater than a preset amplitude in the predicted feature vector sequence as a candidate strong interference time slot, and determine the target strong interference time slot in the candidate strong interference time slot according to the amplitude and time point.

[0127] The calculation module 450 is used to calculate the target signal transmission time with a preset phase difference from the target strong interference time slot based on the phase of the predicted feature vector corresponding to the target strong interference time slot, and to calculate the target phase difference based on the target signal transmission time and a preset reference clock.

[0128] The control module 460 is used to apply the target phase difference as an adjustment command to the programmable phase-locked loop, so as to drive the programmable phase-locked loop to adjust the phase of the output clock signal to generate a high-frequency pulse trigger clock, and use the high-frequency pulse trigger clock to drive the optical pulse transmitting circuit to transmit a high-frequency variable optical pulse signal.

[0129] In one embodiment, the determining module 440 is specifically used to calculate the threat score of each candidate strong interference time slot based on the amplitude of the candidate strong interference time slot and the time difference between the time point of the candidate strong interference time slot and the current time. The threat score is directly proportional to the amplitude of the candidate strong interference time slot and inversely proportional to the time difference between the time point of the candidate strong interference time slot and the current time. The candidate strong interference time slot with the highest threat score is determined as the target strong interference time slot.

[0130] In one embodiment, the determining module 440 is further configured to calculate the gradient parameter of each candidate strong interference time slot based on the magnitude of the candidate strong interference time slot and the difference between the magnitude of the predicted feature vector corresponding to the candidate strong interference time slot and the magnitude of the previous adjacent predicted feature vector in the predicted feature vector sequence, and the threat score is proportional to the gradient parameter of the candidate strong interference time slot.

[0131] In one embodiment, the prediction module 430 is further configured to calculate the Euclidean distance between the first interference feature vector and other interference feature vectors in the feature vector sequence to obtain an Euclidean distance sequence; if the variance of the Euclidean distance sequence is less than a preset variance, the feature vector sequence is input into the trained prediction model to obtain a time-series predicted feature vector sequence.

[0132] In one embodiment, each interference feature information includes at least one set of interference feature parameters, and each set of interference feature parameters includes frequency, amplitude, and phase.

[0133] In one embodiment, the prediction module 430 is specifically used to sort multiple sets of interference feature parameters of the interference feature information in descending order according to the amplitude; the first N sets of interference feature parameters in the interference feature information and the ambient light intensity value are concatenated in a preset order to obtain the interference feature vector, where N is a positive integer.

[0134] In one embodiment, before inputting the feature vector sequence into the trained prediction model to obtain the time-series predicted feature vector sequence, the prediction module 430 is further configured to acquire a training sample set, which includes multiple training samples. Each training sample includes a historical feature vector sequence generated from historical ambient light intensity values ​​and historical light intensity distribution time-series sequences, and a corresponding true feature vector sequence. For each training sample in the training sample set, the following steps are performed: inputting the historical feature vector sequence in each training sample into a preset recurrent neural network model to obtain a predicted feature vector sequence; determining the loss function value of the recurrent neural network model based on the true feature vector sequence and the predicted feature vector sequence of each training sample; if the loss function value does not meet the preset training stopping condition, adjusting the model parameters of the recurrent neural network model to obtain an updated recurrent neural network model, and returning to input the historical feature vector sequence into the recurrent neural network model until the loss function value meets the training stopping condition to obtain the trained prediction model.

[0135] The high-frequency variable optical pulse signal control system of this application embodiment is used to implement the aforementioned high-frequency variable optical pulse signal control method. Therefore, the specific implementation of the high-frequency variable optical pulse signal control system can be found in the embodiment section of the high-frequency variable optical pulse signal control method above. The specific implementation can be referred to the description of the corresponding embodiments, which will not be repeated here.

[0136] Figure 5 A schematic diagram of the hardware structure of an electronic device provided in one embodiment of this application is shown.

[0137] The electronic device may include a processor 510 and a memory 520 storing computer program instructions.

[0138] Specifically, the processor 510 may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0139] Memory 520 may include mass storage for data or instructions. For example, and not limitingly, memory 520 may include a hard disk drive (HDD), floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, memory 520 may include removable or non-removable (or fixed) media. Where appropriate, memory 520 may be internal or external to the integrated gateway disaster recovery device. In a particular embodiment, memory 520 is non-volatile solid-state memory.

[0140] Memory may include read-only memory (ROM), random access memory (RAM), disk storage media devices, optical storage media devices, flash memory devices, and electrical, optical, or other physical / tangible memory storage devices. Therefore, typically, memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., memory devices) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the method according to the first aspect of this disclosure.

[0141] The processor 510 reads and executes computer program instructions stored in the memory 520 to implement any of the high-frequency variable optical pulse signal control methods in the above embodiments.

[0142] In one example, the electronic device may also include a communication interface 530 and a bus 540. Wherein, such as Figure 5 As shown, the processor 510, memory 520, and communication interface 530 are connected through bus 540 and complete communication with each other.

[0143] The communication interface 530 is mainly used to realize communication between various modules, devices, units and / or equipment in the embodiments of this application.

[0144] Bus 540 includes hardware, software, or both, that couples components of an online data traffic metering device together. For example, and not limitingly, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, bus 540 may include one or more buses. Although specific buses are described and illustrated in embodiments of this application, any suitable bus or interconnect is contemplated herein.

[0145] The electronic device can execute the high-frequency variable optical pulse signal control method in the embodiments of this application, thereby realizing the high-frequency variable optical pulse signal control method described in conjunction with the accompanying drawings.

[0146] Furthermore, in conjunction with the high-frequency variable optical pulse signal control method in the above embodiments, this application embodiment can provide a computer-readable storage medium for implementation. This computer-readable storage medium stores computer program instructions; when executed by a processor, these computer program instructions implement any of the high-frequency variable optical pulse signal control methods in the above embodiments.

[0147] It should be clarified that this application is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of this application is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of this application.

[0148] The functional blocks shown in the above-described structural diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this application are programs or code segments used to perform the required tasks. Programs or code segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried on a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc. Code segments can be downloaded via computer networks such as the Internet, intranets, etc.

[0149] It should also be noted that the exemplary embodiments mentioned in this application describe methods or systems based on a series of steps or apparatus. However, this application is not limited to the order of the above steps; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0150] The aspects of this application have been described above with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It should be understood that each block in the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that these instructions, executable via the processor of the computer or other programmable data processing apparatus, enable the implementation of the functions / actions specified in one or more blocks of the flowchart illustrations and / or block diagrams. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field-programmable logic circuit. It is also understood that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can also be implemented by dedicated hardware performing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0151] The above provides a detailed description of a high-frequency variable optical pulse signal control method, system, electronic device, and storage medium provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are merely for the purpose of helping to understand the method and its core ideas. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A method of controlling high frequency variable optical pulse signals, characterized by, The method comprises: acquiring an ambient light intensity value of a target space comprising a plurality of light pulse signal sources, and continuously collecting light field snapshots of the target space at multiple time points through a wide-angle photoelectric sensor array to obtain a light intensity distribution time sequence; performing frequency domain transformation on each light intensity distribution image in the light intensity distribution time sequence to obtain a plurality of frequency domain maps, and extracting interference characteristic information corresponding to each light intensity distribution image based on a plurality of peak points in each frequency domain map, wherein the interference characteristic information comprises frequency, amplitude and phase; constructing an interference characteristic vector using each interference characteristic information and the ambient light intensity value to obtain a time sequence characteristic vector sequence, and inputting the characteristic vector sequence into a trained prediction model to obtain a time sequence prediction characteristic vector sequence, wherein each prediction characteristic vector in the prediction characteristic vector sequence comprises amplitude and phase at a future time point; determining a time point corresponding to a prediction characteristic vector with an amplitude greater than a preset amplitude in the prediction characteristic vector sequence as a candidate strong interference time slot, and determining a target strong interference time slot in the candidate strong interference time slot according to the amplitude and the time point; calculating a target signal transmission time point having a preset phase difference with the target strong interference time slot according to a phase of the prediction characteristic vector corresponding to the target strong interference time slot, and calculating a target phase difference according to the target signal transmission time point and a preset reference clock; using the target phase difference as an adjustment instruction to act on a programmable phase-locked loop to drive the programmable phase-locked loop to adjust the phase of an output clock signal to generate a high-frequency pulse trigger clock, and using the high-frequency pulse trigger clock to drive a light pulse emission circuit to emit a high-frequency variable light pulse signal.

2. The method of claim 1, wherein, The method further comprises: calculating a threat score of each candidate strong interference time slot according to the amplitude of the candidate strong interference time slot and a time difference between the time point of the candidate strong interference time slot and a current time point, wherein the threat score is proportional to the amplitude of the candidate strong interference time slot, and the threat score is inversely proportional to the time difference between the time point of the candidate strong interference time slot and the current time point; determining the candidate strong interference time slot with the highest threat score as the target strong interference time slot.

3. The method of claim 2, wherein, The method further comprises: calculating a gradient parameter of each candidate strong interference time slot according to a difference between the amplitude of the prediction characteristic vector corresponding to the candidate strong interference time slot and the amplitude of a previous adjacent prediction characteristic vector in the prediction characteristic vector sequence, wherein the threat score is proportional to the gradient parameter of the candidate strong interference time slot.

4. The method of claim 1, wherein, The method further comprises: calculating the Euclidean distance between a first interference characteristic vector and other interference characteristic vectors in the characteristic vector sequence to obtain a Euclidean distance sequence; in a case where the variance of the Euclidean distance sequence is less than a preset variance, inputting the characteristic vector sequence into a trained prediction model to obtain a time sequence prediction characteristic vector sequence.

5. The method of claim 1, wherein, Each interference characteristic information comprises at least one set of interference characteristic parameters, and each set of interference characteristic parameters comprises frequency, amplitude and phase.

6. The method of claim 5, wherein, The interference feature information and the ambient light intensity value are used to construct an interference feature vector, which includes: The multiple sets of interference feature parameters in the interference feature information are sorted in descending order according to the amplitudes; The first N sets of interference feature parameters in the interference feature information and the ambient light intensity value are spliced in a preset order to obtain the interference feature vector, where N is a positive integer.

7. The method of claim 1, wherein, Before the feature vector sequence is input into the trained prediction model to obtain a time-series prediction feature vector sequence, the method includes: obtaining a training sample set, the training sample set including multiple training samples, each training sample including a historical feature vector sequence generated by a historical ambient light intensity value and a historical light intensity distribution time-series sequence, and a corresponding real feature vector sequence; for each training sample in the training sample set, the following steps are performed respectively: inputting the historical feature vector sequence in each training sample into a preset recurrent neural network model to obtain a prediction feature vector sequence; determining a loss function value of the recurrent neural network model according to the real feature vector sequence and the prediction feature vector sequence of each training sample; if the loss function value does not satisfy a preset training stop condition, adjusting the model parameters of the recurrent neural network model to obtain an updated recurrent neural network model, and returning to input the historical feature vector sequence into the recurrent neural network model until the loss function value satisfies the training stop condition, thereby obtaining the trained prediction model.

8. A high frequency variable optical pulse signal control system, characterized by, includes: an acquisition module configured to acquire an ambient light intensity value of a target space including multiple light pulse signal sources, and continuously collect light field snapshots of the target space at multiple time points through a wide-angle photoelectric sensor array to obtain a light intensity distribution time-series sequence; an extraction module configured to perform frequency domain transformation on each light intensity distribution image in the light intensity distribution time-series sequence to obtain multiple frequency domain spectra, and extract interference feature information corresponding to each light intensity distribution image based on multiple peak points in each frequency domain spectrum, the interference feature information including frequency, amplitude and phase; a prediction module configured to construct an interference feature vector using each interference feature information and the ambient light intensity value, obtain a time-series feature vector sequence, and input the feature vector sequence into a trained prediction model to obtain a time-series prediction feature vector sequence, each prediction feature vector in the prediction feature vector sequence including amplitude and phase at a future time point; a determination module configured to determine a time point corresponding to a prediction feature vector with an amplitude greater than a preset amplitude in the prediction feature vector sequence as a candidate strong interference time slot, and determine a target strong interference time slot according to the amplitude and the time point in the candidate strong interference time slot; a calculation module configured to calculate a target signal transmission time based on a phase of the prediction feature vector corresponding to the target strong interference time slot, and calculate a target phase difference based on the target signal transmission time and a preset reference clock. A control module is configured to use the target phase difference as an adjustment instruction to act on a programmable phase-locked loop to drive the programmable phase-locked loop to adjust the phase of an output clock signal to generate a high-frequency pulse trigger clock, and use the high-frequency pulse trigger clock to drive a light pulse emitting circuit to emit a high-frequency variable light pulse signal.

9. An electronic device, comprising: The application further provides a computer readable storage medium having a computer program stored therein, wherein the computer program is executable by a processor to implement the high-frequency variable light pulse signal control method according to any one of claims 1 to 7. The application further provides a computer readable storage medium having a computer program stored therein, wherein the computer program is executable by a processor to implement the high-frequency variable light pulse signal control method according to any one of claims 1 to 7. The application further provides a computer readable storage medium having a computer program stored therein, wherein the computer program is executable by a processor to implement the high-frequency variable light pulse signal control method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, ​

Images