Safety light curtain cascade control method and system based on bidirectional photoelectric synchronization

Abstract

The present application relates to the field of streaming communication and remote collaboration, and particularly relates to a safe light curtain cascade control method and system based on bidirectional photoelectric synchronization. The method comprises the steps of: obtaining the sampling sequence of the cascade cable and the light receiving channel; extracting the waveform gradient features in the line domain and the light domain, determining the time reference, the occurrence time and the peak gradient; normalizing the time and the gradient to obtain the time domain and the shape deviation ratio, and generating the cost index through Chebyshev distance operation; comparing the cost index with the tolerance boundary to determine the synchronization state and drive the actuator. The present application fuses the two-dimensional deviation features by using the Chebyshev boundary, effectively eliminates the time wandering caused by aging, and accurately intercepts stray light, thereby realizing high-precision time anchoring and low-delay emergency stop response.

Description

Technical Field

[0001] This invention relates to the fields of photoelectric detection and industrial safety protection, and in particular to a method and system for cascading control of safety light curtains based on bidirectional photoelectric synchronization. Background Technology

[0002] In industrial settings where multiple safety light curtains are cascaded, the transceivers of each light curtain rely on synchronization pulse signals to achieve scanning timing alignment. When this synchronization pulse signal is transmitted to the receiving end via cascaded cables, the distributed capacitance and aging attenuation of the cables cause transmission delays and waveform distortion in the synchronization level. Simultaneously, spatial diffuse reflection generates stray light interference pulses, causing the receiving end to capture multiple candidate pulses within a single listening cycle.

[0003] To identify the true synchronization pulse from the candidate pulses, existing techniques typically examine both the temporal and morphological characteristics of the pulse. The temporal characteristic reflects the deviation of the pulse's arrival time from the reference signal, while the morphological characteristic reflects the deviation of the pulse's edge steepness from the reference signal. Since stray light pulses may coincide with the reference signal in time but have significantly different waveforms, and true synchronization pulses may degrade in both time and shape due to cable aging, a single feature is insufficient for independent identification. Therefore, it is necessary to combine the temporal and morphological characteristics for a comprehensive judgment.

[0004] Existing technologies primarily employ a linear weighted summation method for feature combination. This method assigns weights to time deviation and morphological deviation, then sums them to obtain the discrimination cost. When using this method, if the time deviation of the stray light pulse is extremely small, its larger morphological deviation can be compensated for by the smaller time deviation through weight allocation, resulting in a final calculated discrimination cost below a preset threshold. This allows the stray light pulse to pass detection, triggering a safety light curtain leak alarm.

[0005] Existing technologies also employ a product-based approach for feature combination. This method directly multiplies the time deviation and morphological deviation to obtain the discrimination cost. When using this method, if the actual synchronization signal experiences moderate degradation in both time and morphology due to cable aging, the product of the two moderate deviations will expand rapidly due to nonlinear amplification and exceed a preset threshold. This causes the actual synchronization signal to be misjudged as stray light, triggering false alarms in the safety light curtain and leading to unplanned shutdowns of the controlled equipment.

[0006] Therefore, the existing technology's combination of time and morphological characteristics cannot simultaneously intercept stray light interference and tolerate signal degradation due to cable aging, making it difficult to meet the synchronous discrimination requirements of cascaded safety light curtains under complex working conditions. Summary of the Invention

[0007] To address the problem that existing safety light curtain synchronization discrimination methods cannot simultaneously handle the accurate interception of severely distorted single-dimensional signals and the reasonable tolerance of slightly aged multi-dimensional signals, this invention provides a safety light curtain cascade control method and system based on bidirectional photoelectric synchronization.

[0008] In a first aspect, the present invention provides a safety light curtain cascade control method based on bidirectional photoelectric synchronization, which adopts the following technical solution: The safety light curtain cascade control method based on bidirectional photoelectric synchronization includes the following steps: During the listening period, discrete voltage sampling sequences of the cascaded cable synchronization channel and discrete photoelectric sampling sequences of the optical receiving channel are acquired respectively. Based on the discrete voltage sampling sequence, the line domain waveform gradient features are extracted and the line domain time base and line domain peak gradient are determined. At the same time, based on the discrete photoelectric sampling sequence, the optical domain waveform gradient features of each optical candidate point segment are extracted and the optical domain occurrence time and optical domain peak gradient are determined. The time of occurrence of the optical domain is normalized to the time reference of the linear domain to obtain the time domain deviation ratio. The peak gradient of the optical domain is normalized to the peak gradient of the linear domain to obtain the shape deviation ratio. Based on the time domain deviation ratio and the shape deviation ratio, the Chebyshev cost index is generated by Chebyshev distance calculation. The Chebyshev cost index is compared with the preset Chebyshev tolerance boundary. Based on the comparison result, the synchronization matching state is determined, and the corresponding control signal is output to directly drive the specific physical actuator to operate.

[0009] Preferably, the process of extracting line-domain waveform gradient features and determining the line-domain time base and line-domain peak gradient based on the discrete voltage sampling sequence includes: The discrete voltage sampling sequence is subjected to a differential operation to obtain a line-domain differential sequence; The maximum point in the line domain difference sequence is obtained, the timestamp corresponding to the maximum point is established as the line domain time base, and the value of the maximum point is established as the line domain peak gradient.

[0010] Preferably, the process of performing a differential operation on the discrete voltage sampling sequence to obtain a line-domain differential sequence includes: Extract the current voltage sample value and the previous voltage sample value adjacent to the current voltage sample value on the time axis from the discrete voltage sampling sequence; The current voltage sample value is subtracted from the previous voltage sample value to perform a difference operation, and the value obtained by the difference operation is used as the discrete derivative value at the current sampling time. The discrete derivative values ​​corresponding to all consecutive sampling times within the listening period are concatenated in chronological order to generate the linear domain difference sequence.

[0011] Preferably, the normalization process includes: The absolute value of the time difference between the optical domain occurrence time and the linear domain time reference is calculated and divided by the preset maximum allowable delay tolerance to obtain the time domain deviation ratio. Calculate the absolute value of the gradient difference between the peak gradient in the optical domain and the peak gradient in the linear domain, and divide it by a preset limit gradient attenuation tolerance to obtain the shape deviation ratio.

[0012] Preferably, the process of generating the Chebyshev cost index based on the temporal deviation ratio and the morphological deviation ratio through Chebyshev distance calculation includes: Determine the relationship between the temporal deviation ratio and the morphological deviation ratio; If the temporal deviation ratio is determined to be greater than or equal to the morphological deviation ratio, then the temporal deviation ratio is extracted as the Chebyshev cost index of the current optical candidate point segment; If the morphological deviation ratio is determined to be greater than the temporal deviation ratio, then the morphological deviation ratio is extracted as the Chebyshev cost index of the current optical candidate point segment.

[0013] Preferably, the process of comparing the Chebyshev cost index with a preset Chebyshev tolerance boundary includes a streaming state overwrite operation, which includes: When the listening cycle is started, the current cost record value in the current minimum cost register in the system memory is initialized to a preset full-scale extreme value; Whenever a new Chebyshev cost index is generated, a single constant-time logic comparison instruction is triggered. If the Chebyshev cost index is determined to be less than the preset Chebyshev tolerance boundary and less than the current minimum cost value stored in the current minimum cost register, then the original value in the current minimum cost register is erased, and the Chebyshev cost index is written into the current minimum cost register as the current minimum cost value. At the same time, the optical domain occurrence time of the optical candidate point segment corresponding to the Chebyshev cost index is synchronously latched as the current optimal timestamp.

[0014] Preferably, the instruction to trigger a single constant time complexity logic comparison is to compare the newly generated Chebyshev cost index with the preset Chebyshev tolerance boundary and the current minimum cost value.

[0015] Preferably, the process of directly driving the operation of a specific physical actuator by outputting the control signal corresponding to the synchronization matching state includes: When the system timer determines that the listening period is closed, it checks the current minimum cost value in the current minimum cost register; If the current minimum cost value is determined to be equal to the initial preset full-scale bias value, then a synchronization matching state representing a synchronization fault is generated; the general-purpose input / output pins are configured to output low level, and the external safety relay is de-energized through the hardware isolation circuit, physically cutting off the output signal device circuit of the safety light curtain, so as to stop the controlled machinery.

[0016] Preferably, the process of directly driving the specific physical actuator to operate by outputting the control signal corresponding to the synchronization matching state further includes: When the system timer determines that the listening period is closed, if it is confirmed that the current minimum cost value in the current minimum cost register has been overwritten and updated, then the synchronization matching state representing normal synchronization is generated. The optical domain occurrence time of the optical candidate point segment corresponding to the last overwrite operation is extracted as the current optimal timestamp, and a synchronization confirmation command is sent to the slave light curtain through the cascaded communication interface to trigger the next round of cooperative scanning.

[0017] Secondly, this invention provides a safety light curtain cascade control system based on bidirectional photoelectric synchronization, employing the following technical solution: The data acquisition unit is used to acquire the discrete voltage sampling sequence of the cascaded cable synchronization channel and the discrete photoelectric sampling sequence of the optical receiving channel during the listening period. The feature extraction unit is used to extract line domain waveform gradient features based on the discrete voltage sampling sequence and determine the line domain time base and the line domain peak gradient. At the same time, it extracts optical domain waveform gradient features of each optical candidate point segment based on the discrete photoelectric sampling sequence and determines the optical domain occurrence time and the optical domain peak gradient. The cost synthesis unit is used to normalize the time of occurrence of the optical domain with the time reference of the linear domain to obtain the temporal deviation ratio, normalize the peak gradient of the optical domain with the peak gradient of the linear domain to obtain the morphological deviation ratio, and generate the Chebyshev cost index based on the temporal deviation ratio and the morphological deviation ratio through Chebyshev distance calculation. The physical closed-loop execution unit is used to compare the Chebyshev cost index with the preset Chebyshev tolerance boundary, determine the synchronization matching state based on the comparison result, and output the control signal corresponding to the synchronization matching state to directly drive the specific physical execution mechanism to operate.

[0018] The present invention has the following technical effects: This invention utilizes the gradient characteristics at the moment of signal transition to replace traditional amplitude characteristics, effectively overcoming the signal attenuation and trigger time wandering problems caused by the aging of long-distance cables. It also combines a sliding window and noise floor segmentation mechanism to filter out ringing noise and spatial stray light. In the multi-dimensional feature fusion stage, Chebyshev cost calculation based on the infinite norm distance is used to extract the maximum deviation parameter of a single dimension as the judgment criterion. This effectively overcomes the technical defects of traditional additive models, which are prone to missing stray light, and multiplicative models, which are prone to misjudging aging signals. It achieves reliable interception of severely distorted single-dimensional signals and reasonable tolerance for slightly degraded multi-dimensional signals. In the state determination and output stage, this method uses constant-time complexity streaming register overwrite logic, eliminating the execution time jitter caused by traditional sorting algorithms and keeping the internal processing latency of the system constant. This mechanism not only naturally filters optical crosstalk in cascaded environments and eliminates the timing scanning blind zone at the splicing of multiple light curtains, but also enables the control system to accurately deduct fixed delays to minimize the safe installation distance of the light curtains. Ultimately, by directly mapping the pin levels to the register states to physically drive the safety relays, a hardware cascade link bypassing the software middleware was constructed, achieving fail-safe control while ensuring predictable execution time in the worst case. Attached Figure Description

[0019] Figure 1 This is a flowchart of the method in the safety light curtain cascade control method based on bidirectional photoelectric synchronization provided in the embodiments of the present invention; Figure 2 This is a comparison and verification diagram of the feature fusion model determination boundary provided in the embodiments of the present invention. Detailed Implementation

[0020] This invention discloses a safety light curtain cascade control method based on bidirectional photoelectric synchronization, referring to... Figure 1 This includes steps S1-S4: S1: During the listening period, acquire the discrete voltage sampling sequence of the cascaded cable synchronization channel and the discrete photoelectric sampling sequence of the optical receiving channel respectively.

[0021] In industrial environments where multiple safety light curtains are cascaded, the inherent distributed capacitance of the cascaded cables slows down the transmission speed of electrical signals and causes waveform distortion. Simultaneously, diffuse reflection from the environment generates numerous temporally random interference pulses. Using traditional single-threshold trigger level detection methods is highly susceptible to false triggering or missed triggering due to signal distortion. Therefore, this solution abandons single-point level comparison and instead performs high-frequency continuous sampling within a specific time window to fully preserve the dynamic trend of the signal.

[0022] Considering the timing requirements of multi-node cascaded communication, a dedicated time window is set each time communication is initiated. Specifically, after the master node sends a synchronization trigger signal to the cascaded cable, it starts a hardware timer to wait for the slave node to return a response signal. This waiting time window is the listening period described in this embodiment.

[0023] During the listening cycle, the weak photocurrent output by the photodiode is first converted into an analog voltage signal by a pre-amplifier circuit, while the synchronous analog voltage signal from the cascaded cable is also fed into the acquisition link. Then, it passes through a hardware-level anti-aliasing low-pass filter circuit to filter out high-frequency electromagnetic spikes generated by external frequency converters, etc.

[0024] After filtering and conditioning, the two analog voltage signals enter an analog-to-digital converter (ADC) with dual-channel synchronous sampling. The ADC continuously samples these two signals at a preset sampling frequency and directly outputs the digital sampled values ​​at the corresponding sampling times. To avoid aliasing, this sampling frequency is set to be greater than twice the highest frequency of the effective high-frequency components in the signal.

[0025] Subsequently, the digital sampled values ​​output by the analog-to-digital converter are sequentially transferred to the system memory buffer in chronological order, thereby generating the discrete voltage sampling sequence and the discrete photoelectric sampling sequence containing complete dynamic change details.

[0026] Furthermore, considering the data throughput pressure brought about by high-speed continuous sampling, the system memory cache specifically adopts a ping-pong cache structure composed of dual-port random access memory. This design ensures that when the direct memory access controller writes a new frame of sampled data, it will not overwrite the previous frame of data being read by subsequent algorithms, thereby guaranteeing the integrity and security of the discrete sampling sequence at the hardware level.

[0027] S2: Based on the discrete voltage sampling sequence, extract the linear waveform gradient features and determine the linear time base and linear peak gradient. At the same time, based on the discrete photoelectric sampling sequence, extract the optical waveform gradient features of each optical candidate point segment and determine the optical occurrence time and optical peak gradient.

[0028] Traditional synchronization determination methods typically involve directly comparing the signal amplitude with a fixed threshold. This approach is highly susceptible to trigger time slippage or even failure to trigger due to signal amplitude reduction caused by aging and attenuation in long-distance cables. In contrast, the position of the maximum slope at the moment of signal transition on the time axis is less affected by overall amplitude attenuation. Therefore, this embodiment overcomes the determination error caused by changes in physical amplitude by extracting waveform gradient features.

[0029] Specifically, the discrete voltage sampling sequence in memory is read, the difference between adjacent sample values ​​is calculated to generate a linear difference sequence, the extreme value optimization of the linear difference sequence is performed to obtain the maximum value point, and the sampling time corresponding to the obtained maximum value point is determined as the linear time reference, and its value is determined as the linear peak gradient.

[0030] Considering that the safety light curtain receiver may simultaneously receive multiple light spot signals or ambient stray light, and given that the discrete photoelectric sampling sequence contains multiple signal segments, before extracting the optical domain features, the discrete photoelectric sampling sequence is divided into multiple independent optical candidate point segments. The same difference operation and extremum optimization processing as described above are performed on each segment to extract the optical domain occurrence time and optical domain peak gradient of multiple optical candidate point segments. All multiple optical candidate point segments are retained to form an optical candidate feature set.

[0031] In an optional embodiment, considering impedance mismatch in cascaded cables or external interference causing signal ringing, multiple local extrema are generated in the line-domain differential sequence. To determine a unique line-domain time reference and filter out high-frequency noise, a sliding window of preset length is used to calculate the local mean at each position in the line-domain differential sequence. When the mean of the current window is detected to be greater than the mean of the previous window, the differential sequence is considered to be in a valid signal ramp-up phase. After being identified as being in a valid ramp-up phase, monitoring continues. When the current differential value is detected to be less than the mean of the current window or the mean of the current window is less than the mean of the previous window, the ramp-up phase is determined to end, and the sampling position corresponding to the largest differential value within the window is determined as the maximum point. If multiple valid maximum points still exist after the above processing, the largest maximum point is directly selected, its corresponding sampling time is used as the final line-domain time reference, and its value is used as the line-domain peak gradient.

[0032] In another optional embodiment, the segmentation processing method of the discrete photoelectric sampling sequence is as follows: Multiple sampling points are collected during a silent period without light signal input, and their statistical average value is calculated as the noise floor benchmark. The positions of the noise floor benchmark with amplitudes lower than a preset multiple in the discrete photoelectric sampling sequence are determined as segmentation points, thereby dividing them into multiple independent light candidate point segments; wherein, the preset multiple is greater than 1, preferably 1.5.

[0033] S3: Normalize the time of occurrence of the optical domain with the time reference of the linear domain to obtain the time domain deviation ratio, normalize the peak gradient of the optical domain with the peak gradient of the linear domain to obtain the shape deviation ratio, and generate the Chebyshev cost index based on the time domain deviation ratio and the shape deviation ratio through Chebyshev distance calculation.

[0034] In the signal synchronization determination of the safety light curtain, time deviation and signal waveform distortion need to be comprehensively evaluated. Traditional feature addition models, when fusing multi-dimensional features, are prone to masking high deviations in one dimension due to low deviations in another, leading to the false passage of abnormal signals. Feature multiplication models, on the other hand, tend to nonlinearly amplify slight uniform attenuation in both dimensions, resulting in excessive penalty for normal aging signals. This step employs a cost synthesis method based on the infinite norm distance, extracting the maximum deviation parameter in a single dimension as the basis for safety determination. This approach intercepts severely excessive signals in a single dimension while maintaining tolerance for natural degradation in both dimensions.

[0035] Specifically, the process first iterates through the set of optical candidate features, calculating the absolute difference between the optical domain occurrence time and the linear domain time reference for each optical candidate point segment, as well as the absolute difference between the optical domain peak gradient and the linear domain peak gradient for each optical candidate point segment. Then, the absolute value of the time difference is normalized by dividing it by a pre-defined maximum allowable delay tolerance, yielding the temporal deviation ratio for each optical candidate point segment. Simultaneously, the absolute value of the gradient difference is normalized by dividing it by a preset limit gradient attenuation tolerance, yielding the morphological deviation ratio for each optical candidate point segment. Finally, the magnitudes of these two deviation ratios are compared, and the maximum value is extracted as the Chebyshev cost index for the current optical candidate point segment. Extracting the maximum deviation accurately reflects the degree of signal degradation in the worst dimension, avoiding the risk of missed detection caused by feature averaging in the additive model and eliminating the false alarm problem caused by feature coupling amplification in the multiplicative model. This generates an objective cost index reflecting the synchronization quality of each optical candidate point segment in the set.

[0036] In an optional embodiment, the maximum permissible delay tolerance and the limiting gradient attenuation tolerance are not fixed empirical values, but are obtained based on physical models and operating condition calibrations. Specifically, the limiting gradient attenuation tolerance is based on an equivalent RC low-pass filter model composed of the parasitic distributed capacitance and conductor resistance of the cascaded cable. It is obtained by using a standard step signal excitation and measuring the rise slope test method to obtain the actual gradient attenuation of the cable with the maximum permissible physical length, and then calibrated by combining the aging derating factor in the functional safety design specification. The maximum permissible delay tolerance is based on the delay accumulation principle in the worst-case execution time analysis of the real-time system. Specifically, the maximum permissible delay tolerance is calculated by accumulating the cable transmission delay, the inherent group delay of the photoelectric sensor, the inherent processing delay of the microcontroller system, the jitter error of the system master clock, and the safety time margin required by the functional safety standard.

[0037] In another optional embodiment, to prevent systemic failure, before performing the above normalization operation, it is mandatory to check whether the maximum permissible delay tolerance and the limiting gradient decay tolerance used as the denominator are strictly greater than zero. If the maximum permissible delay tolerance or the limiting gradient decay tolerance is detected to be less than or equal to zero, it is determined that the factory parameter configuration is incorrect or the read-only memory data has been flipped due to strong electromagnetic interference. At this time, the current operation cycle will be immediately terminated, a preset safety abnormality interrupt will be triggered, and the external physical actuator will be controlled to enter a safe shutdown state.

[0038] S4: Compare the Chebyshev cost index with the preset Chebyshev tolerance boundary, determine the synchronization matching state based on the comparison result, and output the control signal corresponding to the synchronization matching state to directly drive the specific physical actuator to run.

[0039] Functional safety standards require that the worst-case execution time of the control program must be predictable and fixed. Traditional light curtains often use array caching and sorting operations when processing multiple light pulses, and their execution time fluctuates with the increase of stray light, making it difficult to meet high-level safety certification requirements. Therefore, this step uses streaming comparison logic to replace dynamic sorting, making the computation time strictly fixed, and constructing a hardware-level safety output path directly related to the underlying physical relays.

[0040] Specifically, when the listening cycle starts, the minimum cost value in memory is pre-initialized to a preset full-bias extreme value. During window listening, for each Chebyshev cost index, a single comparison instruction is executed with constant time complexity to determine whether the Chebyshev cost index is simultaneously less than a preset Chebyshev tolerance boundary and the current minimum cost value recorded in the register. If both conditions are met, the old value in the register is erased and overwritten with the Chebyshev cost index, while simultaneously latching the optical domain occurrence time of the optical candidate point segment corresponding to the Chebyshev cost index as the current optimal timestamp. In this streaming overwrite process, cascaded crosstalk light with phase shift or amplitude attenuation and a large maximum deviation value can be naturally filtered out, retaining only the effective direct light with a small cost to achieve optical crosstalk immunity. The preset Chebyshev tolerance boundary is set to 0.6, and the full-bias extreme value is set to 1.

[0041] When the listening window closes, the moment of occurrence of the optical domain corresponding to the minimum cost exponent retained in the register is extracted and used as the timing reference with the highest synchronization quality and the most stable transmission delay. This is used to align the scanning timing of the master and slave light curtains and eliminate timing scanning blind spots at the splicing points. At the same time, the algorithm's processing time is kept constant through a single overwrite operation with constant time complexity, thereby accurately deducting the fixed internal delay and minimizing the installation safety distance of the light curtain while ensuring safety.

[0042] When the listening window closes, the final state of the register is read directly and converted into a drive signal for the safety light curtain via a hardware cascaded control link. If the read register value is still equal to the initial full-scale bias, it is determined that the master and slave light curtains have lost synchronization. The general-purpose input / output pins are configured to output a low level, and the external safety relay is de-energized via a hardware isolation circuit, physically cutting off the output signal device circuit of the safety light curtain and stopping the controlled machinery. Conversely, if the register value is less than the full-scale bias, it is determined that a valid synchronization signal has been successfully captured. The general-purpose input / output pins are kept high to maintain the operation of the light curtain, and a synchronization confirmation command is sent to the slave light curtain via the cascaded communication interface to trigger the next round of cooperative scanning.

[0043] To demonstrate the effectiveness of the solution, relevant experiments were conducted. Below are the images obtained from the experiments: Figure 2 This is a comparison and verification diagram of the boundary determination for the feature fusion model. The horizontal axis represents the temporal deviation ratio, and the vertical axis represents the morphological deviation ratio. In the diagram, the dotted line represents the traditional product fusion envelope, the long dashed line represents the traditional linear summation envelope, the thick solid line represents the Chebyshev tolerance boundary of this invention, the shaded area enclosed by the thick solid line represents the rigid safe receiving area of ​​this invention, the hollow circle marks the two-dimensional attenuated true signal, and the cross mark marks the one-dimensional heavily distorted stray light. As shown in the diagram, the two-dimensional attenuated true signal located within the safe receiving area of ​​this invention is correctly received, but located outside the traditional product fusion envelope, it will be misjudged; the one-dimensional heavily distorted stray light located outside the tolerance boundary of this invention is accurately intercepted, but located inside the traditional linear summation envelope, it will be missed.

[0044] This invention also discloses a safety light curtain cascade control system based on bidirectional photoelectric synchronization, specifically including: The data acquisition unit is used to acquire the discrete voltage sampling sequence of the cascaded cable synchronization channel and the discrete photoelectric sampling sequence of the optical receiving channel during the listening period. The feature extraction unit is used to extract line domain waveform gradient features based on the discrete voltage sampling sequence and determine the line domain time base and the line domain peak gradient. At the same time, it extracts optical domain waveform gradient features of each optical candidate point segment based on the discrete photoelectric sampling sequence and determines the optical domain occurrence time and the optical domain peak gradient. The cost synthesis unit is used to normalize the time of occurrence of the optical domain with the time reference of the linear domain to obtain the temporal deviation ratio, normalize the peak gradient of the optical domain with the peak gradient of the linear domain to obtain the morphological deviation ratio, and generate the Chebyshev cost index based on the temporal deviation ratio and the morphological deviation ratio through Chebyshev distance calculation. The physical closed-loop execution unit is used to compare the Chebyshev cost index with the preset Chebyshev tolerance boundary, determine the synchronization matching state based on the comparison result, and output the control signal corresponding to the synchronization matching state to directly drive the specific physical execution mechanism to operate.

Claims

1. A safety light curtain cascade control method based on bidirectional photoelectric synchronization, characterized in that, include: During the listening period, discrete voltage sampling sequences of the cascaded cable synchronization channel and discrete photoelectric sampling sequences of the optical receiving channel are acquired respectively. Based on the discrete voltage sampling sequence, the line domain waveform gradient features are extracted and the line domain time base and line domain peak gradient are determined. At the same time, based on the discrete photoelectric sampling sequence, the optical domain waveform gradient features of each optical candidate point segment are extracted and the optical domain occurrence time and optical domain peak gradient are determined. The time of occurrence of the optical domain is normalized to the time reference of the linear domain to obtain the time domain deviation ratio. The peak gradient of the optical domain is normalized to the peak gradient of the linear domain to obtain the shape deviation ratio. Based on the time domain deviation ratio and the shape deviation ratio, the Chebyshev cost index is generated by Chebyshev distance calculation. The Chebyshev cost index is compared with the preset Chebyshev tolerance boundary. Based on the comparison result, the synchronization matching state is determined, and the corresponding control signal is output to directly drive the specific physical actuator to operate.

2. The safety light curtain cascade control method based on bidirectional photoelectric synchronization according to claim 1, characterized in that, The process of extracting line-domain waveform gradient features and determining the line-domain time base and line-domain peak gradient based on the discrete voltage sampling sequence includes: The discrete voltage sampling sequence is subjected to a differential operation to obtain a line-domain differential sequence; The maximum point in the line domain difference sequence is obtained, the timestamp corresponding to the maximum point is established as the line domain time base, and the value of the maximum point is established as the line domain peak gradient.

3. The safety light curtain cascade control method based on bidirectional photoelectric synchronization according to claim 2, characterized in that, The process of performing a differential operation on the discrete voltage sampling sequence to obtain a line-domain differential sequence includes: Extract the current voltage sample value and the previous voltage sample value adjacent to the current voltage sample value on the time axis from the discrete voltage sampling sequence; The current voltage sample value is subtracted from the previous voltage sample value to perform a difference operation, and the value obtained by the difference operation is used as the discrete derivative value at the current sampling time. The discrete derivative values ​​corresponding to all consecutive sampling times within the listening period are concatenated in chronological order to generate the linear domain difference sequence.

4. The safety light curtain cascade control method based on bidirectional photoelectric synchronization according to claim 1, characterized in that, The normalization process includes: The absolute value of the time difference between the optical domain occurrence time and the linear domain time reference is calculated and divided by the preset maximum allowable delay tolerance to obtain the time domain deviation ratio. Calculate the absolute value of the gradient difference between the peak gradient in the optical domain and the peak gradient in the linear domain, and divide it by a preset limit gradient attenuation tolerance to obtain the shape deviation ratio.

5. The safety light curtain cascade control method based on bidirectional photoelectric synchronization according to claim 1, characterized in that, The process of generating the Chebyshev cost index based on the temporal deviation ratio and the morphological deviation ratio through Chebyshev distance calculation includes: Determine the relationship between the temporal deviation ratio and the morphological deviation ratio; If the temporal deviation ratio is determined to be greater than or equal to the morphological deviation ratio, then the temporal deviation ratio is extracted as the Chebyshev cost index of the current optical candidate point segment; If the morphological deviation ratio is determined to be greater than the temporal deviation ratio, then the morphological deviation ratio is extracted as the Chebyshev cost index of the current optical candidate point segment.

6. The safety light curtain cascade control method based on bidirectional photoelectric synchronization according to claim 1, characterized in that, The process of comparing the Chebyshev cost index with a preset Chebyshev tolerance boundary includes a streaming state overwrite operation, which includes: When the listening cycle is started, the current cost record value in the current minimum cost register in the system memory is initialized to a preset full-scale extreme value; Whenever a new Chebyshev cost index is generated, a single constant-time logic comparison instruction is triggered. If the Chebyshev cost index is determined to be less than the preset Chebyshev tolerance boundary and less than the current minimum cost value stored in the current minimum cost register, then the original value in the current minimum cost register is erased, and the Chebyshev cost index is written into the current minimum cost register as the current minimum cost value. At the same time, the optical domain occurrence time of the optical candidate point segment corresponding to the Chebyshev cost index is synchronously latched as the current optimal timestamp.

7. The safety light curtain cascade control method based on bidirectional photoelectric synchronization according to claim 6, characterized in that, The instruction to trigger a single constant time complexity logic comparison is to compare the newly generated Chebyshev cost index with the preset Chebyshev tolerance boundary and the current minimum cost value.

8. The safety light curtain cascade control method based on bidirectional photoelectric synchronization according to claim 6, characterized in that, The process by which the output control signal corresponding to the synchronization matching state directly drives the operation of the specific physical actuator includes: When the system timer determines that the listening period is closed, it checks the current minimum cost value in the current minimum cost register; If the current minimum cost value is determined to be equal to the initial preset full-scale bias value, then a synchronization matching state representing a synchronization fault is generated; the general-purpose input / output pins are configured to output low level, and the external safety relay is de-energized through the hardware isolation circuit, physically cutting off the output signal device circuit of the safety light curtain, so as to stop the controlled machinery.

9. The safety light curtain cascade control method based on bidirectional photoelectric synchronization according to claim 6, characterized in that, The process of directly driving the operation of a specific physical actuator by outputting the control signal corresponding to the synchronization matching state also includes: When the system timer determines that the listening period is closed, if it is confirmed that the current minimum cost value in the current minimum cost register has been overwritten and updated, then the synchronization matching state representing normal synchronization is generated. The optical domain occurrence time of the optical candidate point segment corresponding to the last overwrite operation is extracted as the current optimal timestamp, and a synchronization confirmation command is sent to the slave light curtain through the cascaded communication interface to trigger the next round of cooperative scanning.

10. A safety light curtain cascade control system based on bidirectional photoelectric synchronization, characterized in that, include: The data acquisition unit is used to acquire the discrete voltage sampling sequence of the cascaded cable synchronization channel and the discrete photoelectric sampling sequence of the optical receiving channel during the listening period. The feature extraction unit is used to extract line domain waveform gradient features based on the discrete voltage sampling sequence and determine the line domain time base and the line domain peak gradient. At the same time, it extracts optical domain waveform gradient features of each optical candidate point segment based on the discrete photoelectric sampling sequence and determines the optical domain occurrence time and the optical domain peak gradient. The cost synthesis unit is used to normalize the time of occurrence of the optical domain with the time reference of the linear domain to obtain the temporal deviation ratio, normalize the peak gradient of the optical domain with the peak gradient of the linear domain to obtain the morphological deviation ratio, and generate the Chebyshev cost index based on the temporal deviation ratio and the morphological deviation ratio through Chebyshev distance calculation. The physical closed-loop execution unit is used to compare the Chebyshev cost index with the preset Chebyshev tolerance boundary, determine the synchronization matching state based on the comparison result, and output the control signal corresponding to the synchronization matching state to directly drive the specific physical execution mechanism to operate.

Images