Emergency broadcast multi-signal identification triggering method and system

Abstract

The present application relates to the technical field of ship safety communication, and particularly relates to an emergency broadcast multi-signal identification triggering method and system, comprising: acquiring original signal data of multiple heterogeneous signal sources, and acquiring a preprocessed signal sequence corresponding to each signal source according to the original signal data; acquiring signal cycle characteristics and signal dynamic characteristics of each signal source according to the preprocessed signal sequence; acquiring a single-path discrimination score of the signal source according to the signal cycle characteristics and the signal dynamic characteristics, and acquiring an effective signal source based on the single-path discrimination score; and acquiring a joint discrimination score of multiple signal sources according to the effective signal source. The present application can effectively distinguish real event signals from vibration-induced false signals under complex sea conditions and composite accident scenarios, provide more accurate response decision basis for a ship emergency broadcast system, and further optimize the situation awareness and emergency response accuracy of the ship emergency broadcast system.

Description

Technical Field

[0001] This invention relates to the field of ship safety communication technology, and in particular to an emergency broadcast multi-signal identification and triggering method and system. Background Technology

[0002] The ship's emergency broadcast system needs to receive and process trigger signals in real time from various heterogeneous signal sources, including fire alarm detectors, flood alarm sensors, emergency position beacons, and digital selective call distress signal transmitters. These signal sources differ significantly in physical interfaces, signal formats, and triggering characteristics, and are installed in different compartments of the ship.

[0003] During navigation, a ship's hull is continuously subjected to broadband mechanical vibrations generated by the operation of the main engine, propeller vibration, and wave impact. These vibrations are transmitted through the equipment mounting base to the internal signal sources, generating a flutter effect in the sensor's sensitive elements or signal transmission links. This results in vibration noise components strongly correlated with the ship's operating state being coupled into the original signal. Under high sea state conditions, the amplitude of this vibration noise component can reach or even exceed the effective signal amplitude, making it difficult for existing signal discrimination methods based on fixed thresholds or fixed filtering parameters to distinguish between real event signals and vibration-induced pseudo-signals. This leads to a significant increase in false alarm rates or the submergence of effective signals, resulting in missed detections.

[0004] Furthermore, in complex accident scenarios such as ship collisions and water ingress, multiple heterogeneous signal sources will be triggered sequentially within a short period of time, and their triggering sequence implies a causal logical relationship between the events. Existing ship emergency broadcast systems typically independently identify and respond to each signal source, lacking a multi-source temporal correlation verification mechanism based on physical causal relationships. This makes it impossible to effectively integrate multiple information sources to improve the confidence level of complex event judgments, thus affecting the system's situational awareness and response accuracy in complex scenarios. Summary of the Invention

[0005] The main objective of this invention is to provide an emergency broadcast multi-signal identification and triggering method and system, which aims to solve the technical problems mentioned in the background art.

[0006] This invention proposes an emergency broadcast multi-signal identification and triggering method, applied to ship broadcasting equipment, comprising: Acquire raw signal data from multiple heterogeneous signal sources, and obtain the preprocessed signal sequence corresponding to each signal source based on the raw signal data; The signal periodicity and dynamic characteristics of each signal source are obtained based on the preprocessed signal sequence. A single-channel discrimination score is obtained based on the signal periodicity and dynamic characteristics of the signal source, and an effective signal source is obtained based on the single-channel discrimination score. The joint discrimination score of multiple signal sources is obtained based on the effective signal source; Based on the single-path discrimination score and the joint discrimination score, a broadcast action of the corresponding level is triggered according to the preset multi-level response level conditions.

[0007] Preferably, the step of obtaining the preprocessed signal sequence corresponding to each signal source based on the original signal data includes: Based on the rising edge sampling point index where the signal amplitude first exceeds the preset silence threshold in the original signal data of each heterogeneous signal source, the relative time offset between each heterogeneous signal source is determined, and the original signal data of each channel is shifted and compensated according to the relative time offset to obtain a timestamp mark sequence carrying a unified time reference. Based on the pre-stored working frequency band parameters obtained from each of the heterogeneous signal sources, the corresponding upper and lower cutoff frequencies of the passband are determined, and the bandpass filter signal is obtained based on the aligned signal data corresponding to the timestamp mark sequence and the upper and lower cutoff frequencies of the passband. Based on the vibration acceleration signals collected by vibration reference sensors rigidly deployed at the same position as each heterogeneous signal source, the bandpass filtered signal is subjected to adaptive noise cancellation processing with the vibration acceleration signals as reference input, and the preprocessed signal sequence corresponding to each heterogeneous signal source is obtained based on the error output signal after the adaptive noise cancellation processing.

[0008] Preferably, the step of obtaining the signal periodicity characteristics and signal dynamic characteristics of each signal source based on the preprocessed signal sequence includes: Adjust the mapping relationship between the current ship navigation status parameters and the preset feature recognition window to obtain the length of the feature recognition time window corresponding to the current sampling time. The signal source type identifier is obtained according to each heterogeneous signal source, and the corresponding reference period, period deviation tolerance range and change rate reference interval are obtained from the pre-stored signal source type and reference feature mapping table. Within the time interval defined by the feature recognition time window length, zero-crossing detection is performed on the preprocessed signal sequence, and the periodic change frequency value is obtained based on the number of detected zero-crossing events. The signal periodic characteristics are obtained based on the relative deviation between the periodic change frequency value and the period allowable interval determined by the reference period and the period deviation allowable range. Within the time interval defined by the feature recognition time window length, an instantaneous rate of change sequence is obtained based on the amplitude difference between adjacent sampling points of the preprocessed signal sequence, and the instantaneous rate of change sequence is subjected to median filtering to obtain a representative rate of change. The dynamic characteristics of the signal are obtained based on the degree of deviation between the representative rate of change and the reference range of the rate of change.

[0009] Preferably, the step of obtaining a single-channel discrimination score based on the signal periodicity and dynamic characteristics of the signal source, and obtaining an effective signal source based on the single-channel discrimination score, includes: The sea state level and ship roll amplitude are obtained based on the current ship navigation status parameters. A dynamic reference offset is determined based on the sea state level and ship roll amplitude. When the sea state level is greater than or equal to a preset sea state level threshold or the ship roll amplitude is greater than or equal to a preset roll amplitude threshold, the dynamic reference offset is obtained based on a preset offset calculation function. The lower limit of the rate of change reference interval is then adjusted downward based on the dynamic reference offset to obtain a corrected rate of change reference interval. Based on the comparison between the deviation of the signal periodic characteristics and the preset periodic anomaly tolerance, the periodic anomaly marking status of the corresponding heterogeneous signal source is obtained. Based on the comparison between the degree of deviation of the signal dynamic characteristics and the dynamic anomaly tolerance determined by the modified rate of change reference interval, the dynamic anomaly labeling state of the corresponding heterogeneous signal source is obtained. Based on the periodic anomaly marking state and the dynamic anomaly marking state, the overall anomaly state of the corresponding heterogeneous signal source is determined. When the overall anomaly state is an unmarked anomaly, the heterogeneous signal source is determined as a valid signal source. Based on the deviation degree of the signal periodic characteristics and the deviation degree of the signal dynamic characteristics, the single-channel discrimination score is obtained through a preset single-channel discrimination weighting function.

[0010] Preferably, the step of obtaining the joint discrimination score of multiple signal sources based on the effective signal sources includes: Obtain a pre-stored signal causal correlation map, and based on the pre-stored signal causal correlation map, obtain the directed connection relationship between each of the effective signal sources, the causal time window length corresponding to each of the directed connection relationships, and the causal probability parameters; Based on the directed connection relationship, obtain any two valid signal sources that are connected by a directed edge, and obtain the trigger time interval between the two valid signal sources based on their respective trigger times; Based on the comparison result between the trigger time interval and the causal time window length, the causal matching state corresponding to the directed edge is generated; When the triggering time interval is less than or equal to the causal time window length, the causal matching state is determined to be a match, and the edge association score corresponding to the directed edge is obtained through a preset joint score fusion function based on the single-path discrimination score of each of the two valid signal sources and the causal probability parameter. When the triggering time interval is greater than the causal time window length, the causal matching state is determined to be mismatched, and the edge association score is set to zero; Based on the edge association scores of each directed edge corresponding to all valid signal sources and the single-path discrimination scores of each valid signal source, the joint discrimination scores of multiple signal sources are obtained through a preset graph structure score aggregation function.

[0011] Preferably, the preset multi-level response conditions include the following progressive response levels: The current single-channel discrimination score range, the current joint discrimination score range, and the highest priority level of the current valid signal source are obtained according to the preset response level determination mapping table. The current response level is generated based on the logical combination of the current single-path discrimination score interval, the current joint discrimination score interval, and the highest priority level. The current response level includes at least one of the following: first-level response, second-level response, third-level response, and fourth-level response. When the current response level is determined to be the first level response, the corresponding broadcast area identifier is obtained according to the signal source type identifier of the corresponding valid signal source, and the speaker in the corresponding area is triggered to play the preset first broadcast content according to the broadcast area identifier, and automatically terminated after playing for a preset first duration. When the current response level is determined to be the second level response, the ship's loudspeakers are triggered to play the preset second broadcast content according to the preset ship-wide broadcast configuration parameters, and the broadcast is played in a loop according to the preset playback interval duration and number of repetitions. When the current response level is determined to be the third level response, the ship's loudspeakers are triggered to play the preset third broadcast content according to the preset ship-wide broadcast configuration parameters, and the playback state is maintained until a manual reset command is received and the preset confirmation retention time condition is met. When the current response level is determined to be the fourth level response, the distress signal association verification result is obtained based on the time difference between the trigger time of the emergency position beacon signal and the arrival time of the digital selective call distress signal. When the distress signal association verification result is verified and the current joint discrimination score is greater than or equal to the preset distress confirmation threshold, the preset abandon ship alarm sequence is triggered to be played in a loop by all the ship's loudspeakers, and the playback state is maintained until a manual reset command is received.

[0012] Preferably, the steps to trigger a broadcast action of the corresponding level include: The broadcast priority value of each trigger signal source is obtained according to the pre-stored signal priority mapping table, and the broadcast scheduling strategy is determined according to the comparison result between the broadcast priority values. If the broadcast priority value of the current triggering signal source is higher than the broadcast priority value of the signal source that is currently broadcasting, the current broadcast is interrupted and the interruption status information is recorded, and the broadcast content corresponding to the current triggering signal source is switched. If the broadcast priority value of the current triggering signal source is lower than or equal to the broadcast priority value of the signal source that is currently broadcasting, the current triggering signal source will be added to the waiting queue. Based on the interruption status information, after the broadcast of the current trigger signal source is completed, it is determined whether the interrupted broadcast has exceeded the preset valid execution time limit. If it has not exceeded the time limit, the interrupted broadcast is resumed according to the interruption position in the interruption status information. The corresponding response suppression window duration is determined based on the broadcast priority value of each of the trigger signal sources, wherein the response suppression window duration is inversely proportional to the broadcast priority value, and repeated triggering events of the same trigger signal source are merged and suppressed within the response suppression window duration.

[0013] This invention discloses an emergency broadcast multi-signal identification and triggering system, comprising: The signal preprocessing module is used to acquire the raw signal data of multiple heterogeneous signal sources and acquire the preprocessed signal sequence corresponding to each signal source based on the raw signal data. The signal feature extraction module is used to obtain the signal periodic features and signal dynamic features of each signal source based on the preprocessed signal sequence; A single-channel discrimination module is used to obtain a single-channel discrimination score based on the signal periodicity and dynamic characteristics of the signal source, and to obtain an effective signal source based on the single-channel discrimination score; The joint discrimination module is used to obtain the joint discrimination score of multiple signal sources based on the valid signal source; The broadcast triggering module is used to trigger a broadcast action of the corresponding level according to the single-channel discrimination score and the joint discrimination score, and according to the preset multi-level response level conditions.

[0014] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of an emergency broadcast multi-signal identification triggering method.

[0015] The beneficial effects of this invention are as follows: This invention utilizes an adaptive noise cancellation method based on vibration reference signals, independently adjusting filtering parameters according to the local noise environment of each signal source, thereby effectively suppressing the interference of ship vibration noise on the original signal and obtaining a preprocessed signal sequence with a higher signal-to-noise ratio; on this basis, a signal causal correlation graph is used to verify the temporal causal correlation of multiple effective signal sources arriving sequentially within the correlation time window, thereby fusing the correlation information of multiple signals to improve the confidence of composite event judgment; furthermore, based on the single-path discrimination score and the joint discrimination score, corresponding level emergency response actions are triggered according to the hierarchical response strategy, which can significantly reduce the false alarm rate and missed alarm rate in complex sea conditions and composite accident scenarios, providing a more accurate and reliable response decision basis for the ship emergency broadcasting system, thereby optimizing the situational awareness and emergency response accuracy of the ship emergency broadcasting system. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of a method flow according to an embodiment of this application.

[0017] Figure 2 This is a schematic diagram of the system structure according to an embodiment of this application.

[0018] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0019] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0020] like Figure 1 As shown, this application provides an emergency broadcast multi-signal identification triggering method, applied to ship broadcasting equipment, including: S1. Obtain the raw signal data of multiple heterogeneous signal sources, and obtain the preprocessed signal sequence corresponding to each signal source based on the raw signal data; S2. Obtain the signal periodicity and signal dynamics of each signal source based on the preprocessed signal sequence; S3. Obtain a single-channel discrimination score based on the signal periodicity and dynamic characteristics of the signal source, and obtain an effective signal source based on the single-channel discrimination score; S4. Obtain the joint discrimination score of multiple signal sources based on the valid signal sources; S5. Based on the single-path discrimination score and the joint discrimination score, trigger the corresponding level of broadcast action according to the preset multi-level response level conditions.

[0021] As described in steps S1-S5 above, during navigation, the ship's hull is continuously subjected to broadband mechanical vibration excitation generated by the operation of the main engine, propeller vibration, and wave impact. This vibration is transmitted through the equipment mounting base to the internal components of each signal source, generating a jitter effect in the sensor's sensitive elements or signal transmission links. This results in vibration noise components strongly correlated with the ship's operating state being coupled into the original signal. Under high sea state conditions, the amplitude of this vibration noise component can reach or even exceed the effective signal amplitude, making it difficult for existing signal discrimination methods based on fixed thresholds or fixed filtering parameters to distinguish between real event signals and vibration-induced pseudo-signals. This leads to a significant increase in false alarm rates or the submergence of effective signals, resulting in missed detections. Furthermore, in complex accident scenarios such as ship collisions and water ingress, multiple heterogeneous signal sources will be triggered sequentially within a short period, and their triggering sequence implies a causal logical relationship between the events. Existing emergency broadcast systems typically independently discriminate and respond to each signal source, lacking a multi-source temporal correlation verification mechanism based on physical causal relationships. This makes it impossible to effectively integrate multiple information streams to improve the confidence level of complex event judgments. Therefore, this invention constructs a complete identification link from signal acquisition to response triggering. First, it improves the acquisition quality of each signal source by adaptive noise cancellation based on vibration reference signals. Then, it extracts the signal periodic and dynamic features of each signal source through feature recognition. On this basis, it performs initial screening of the authenticity of each signal source and generates a discrimination score for valid signal sources. Next, it uses a signal causal correlation graph to verify the correlation of multiple valid signal sources and generates a joint discrimination score. Finally, it triggers the corresponding level of emergency response according to the discrimination score and a graded response strategy, thereby solving the above-mentioned technical problems.

[0022] Specifically, this invention acquires raw signal data from multiple heterogeneous signal sources, including but not limited to fire alarm detectors, water ingress alarm sensors, collision detection devices, Emergency Position Indicator Beacons (EPIRBs), and Digital Selective Call (DSC) distress signal transmitters. Each signal source differs significantly in physical interface, signal format, and triggering characteristics. This invention first preprocesses the raw signal data, including synchronous sampling, extremum removal, and adaptive filtering, generating a preprocessed signal sequence for each signal source. This preprocessed signal sequence serves as the foundation for subsequent feature extraction and discrimination. Secondly, based on this preprocessed signal sequence, the invention extracts the signal periodicity and dynamic characteristics of each signal source. The signal periodicity reflects the periodicity of the signal source's triggering mode, while the dynamic characteristics reflect the rate of change in signal amplitude. Then, based on these characteristics, the invention performs an initial screening to determine whether each signal source is a genuine and valid emergency trigger signal. A single-channel discrimination score is generated for valid signal sources, quantifying the credibility of each signal. Next, multi-source correlation verification is performed on multiple valid signal sources arriving sequentially within a preset correlation time window. A pre-stored signal causal correlation graph is used to determine the temporal causal logic relationship between the multiple signal sources, and a joint discrimination score is generated. This joint discrimination score integrates the correlation information of multiple signals, which can improve the confidence level of complex event judgment. Finally, based on the individual discrimination scores and the joint discrimination score, the corresponding level of emergency response action is triggered according to a preset graded response strategy.

[0023] Through the above steps, this invention utilizes an adaptive noise cancellation method based on vibration reference signals to independently adjust the filtering parameters of the local noise environment in which each signal source is located, effectively suppressing the interference of ship vibration noise on signal discrimination. By introducing a signal causal correlation spectrum, time-series causal correlation verification is performed on multiple effective signal sources, enabling the fusion of multiple information sources in complex accident scenarios to improve the confidence of event judgment. Through a graded response strategy, different levels of emergency response can be triggered according to signal credibility and event severity, avoiding insufficient or excessive response caused by a single threshold trigger. Compared with existing technologies that independently discriminate each signal and trigger with a fixed threshold, this invention can significantly reduce false alarm and missed alarm rates in complex sea conditions and complex accident scenarios, improving the response accuracy and reliability of the emergency broadcast system.

[0024] In one embodiment of the present invention, the step of obtaining the preprocessed signal sequence corresponding to each signal source based on the original signal data includes: S11, based on the rising edge sampling point index where the signal amplitude first exceeds the preset silence threshold in the original signal data of each heterogeneous signal source, determine the relative time offset between each heterogeneous signal source, and perform shift compensation on each original signal data according to the relative time offset to obtain a timestamp mark sequence carrying a unified time reference. S12, determine the corresponding upper and lower cutoff frequencies of the passband based on the pre-stored working frequency band parameters obtained from each heterogeneous signal source, and obtain the bandpass filter signal based on the aligned signal data corresponding to the timestamp mark sequence and the upper and lower cutoff frequencies of the passband. S13, based on the vibration acceleration signal collected by the vibration reference sensor rigidly deployed at the same position as each of the heterogeneous signal sources, the bandpass filtered signal is subjected to adaptive noise cancellation processing with the vibration acceleration signal as the reference input, and the preprocessed signal sequence corresponding to each heterogeneous signal source is obtained based on the error output signal after the adaptive noise cancellation processing.

[0025] As described in steps S11-S13 above, this invention needs to solve three key technical problems in the signal preprocessing stage: first, the clock synchronization problem between heterogeneous signal sources; second, the filter parameter adaptation problem caused by the difference in operating frequency bands of different signal sources; and third, the adaptive suppression problem of local vibration noise of the ship. Because the acquisition channels of each heterogeneous signal source are independent, and the hardware interfaces and sampling circuits of each signal source may have inherent differences in startup delay, there is a time offset between the original signal data. If time alignment is not performed, subsequent multi-source correlation verification will result in misjudgment due to timing misalignment. Therefore, this invention determines the relative time offset between each signal source by detecting the rising edge sampling point index where the amplitude of each signal first exceeds a preset silence threshold, and performs shift compensation on the original signal data accordingly.

[0026] Specifically, the preset silence threshold is set as follows: With the signal source not triggered, background noise data from each signal source is collected, and the mean and standard deviation of their amplitudes are calculated. The mean plus three times the standard deviation is taken as the preset silence threshold. This threshold ensures that rising edge detection is not falsely triggered under normal background noise, while simultaneously capturing rising edge events promptly when the signal source is triggered. The relative time offset is calculated as follows: using the signal source with the earliest rising edge among all signals as a reference, the difference between the rising edge sampling point index of each other signal source and the rising edge sampling point index of the reference signal source is calculated. This difference multiplied by the sampling period is the time offset of each signal source relative to the reference signal source. The shift compensation is implemented by cyclically shifting each original signal data according to its corresponding time offset, aligning the trigger events of each signal on the time axis, thereby obtaining a timestamp sequence carrying a unified time reference.

[0027] Secondly, this invention determines the corresponding upper and lower cutoff frequencies of the passband based on the pre-stored operating frequency band parameters obtained from each heterogeneous signal source. Different types of signal sources have different operating frequency bands. For example, the trigger signal frequency of a fire alarm detector is usually in the range of 0.5Hz to 5Hz, the signal frequency of a water ingress alarm sensor is usually in the range of 0.1Hz to 2Hz, and the transient impact signal frequency band of a collision detection device is wider, reaching 0.1Hz to 50Hz. If a uniform filtering parameter is used for all signal sources, the effective frequency components of some signal sources will be filtered out, or the out-of-band noise of other signal sources will not be effectively suppressed. Therefore, this invention establishes a mapping table between signal source types and operating frequency band parameters, queries the corresponding passband cutoff frequency according to the type identifier of each signal source, and designs a bandpass filter based on the cutoff frequency.

[0028] The bandpass filter employs a finite impulse response (FIR) structure. Its design method involves using a window function to design the coefficients of the FIR bandpass filter based on the upper and lower cutoff frequencies of the passband. The filter order is preset to a fixed value based on the transition bandwidth and stopband attenuation requirements, typically 101st order, ensuring a flat passband characteristic and sufficient stopband attenuation (typically not less than 40dB) within the cutoff frequency range. The filter's group delay is the number of sampling periods corresponding to half the order. Since all signal sources use filters with the same rate and order, the group delay is consistent and does not affect the relative timing relationship between the signals.

[0029] Furthermore, this invention utilizes vibration acceleration signals collected by vibration reference sensors rigidly deployed at the same location as each heterogeneous signal source to perform adaptive noise cancellation processing on the bandpass filtered signal, using the vibration acceleration signal as the reference input. The core principle of this adaptive noise cancellation processing is that the vibration reference sensor is installed at the same location as the signal source and rigidly fixed. The vibration acceleration signal it collects reflects the mechanical vibration environment at that location, and the vibration reference sensor does not receive trigger signals from the signal source (because emergency trigger signals are not transmitted through mechanical vibration). Therefore, the vibration acceleration signal can serve as a pure noise reference. The adaptive noise cancellation employs the Normalized Least Mean Square (NLMS) algorithm, and its iterative update formula is: ; In the formula, This represents the adaptive filter weight coefficient vector at the nth sampling time. This represents the step size factor (dimensionless, typically ranging from 0.1 to 0.5, controlling the balance between the algorithm's convergence speed and steady-state error). This represents the error output signal (i.e., the preprocessed signal sequence, which is the residual of the main signal after being processed by the adaptive filter). The vibration reference signal vector is composed of the vibration acceleration signal at the current moment and the values ​​of the previous L-1 historical sampling moments, where L is the order of the adaptive filter, typically 16 to 32. express The square of the Euclidean norm, This represents a preset small positive number (typically 0.5). This is used to prevent numerical instability caused by a denominator of zero.

[0030] The specific implementation process of adaptive noise cancellation is as follows: The bandpass-filtered signal is used as the main input, and the vibration acceleration signal is used as the reference input. The adaptive filter estimates the noise component in the main input based on the reference input, and subtracts this estimate from the main input to obtain the error output signal. Since the NLMS algorithm can automatically adjust the filter weights to minimize the output power, it can effectively cancel time-varying noise without prior knowledge of the statistical characteristics of the noise. The error output signal of this processing is the preprocessed signal sequence, which retains the effective frequency components of the signal source trigger signal while suppressing noise components related to ship vibration.

[0031] It should be noted that the selection of the step size factor μ needs to balance the convergence speed and steady-state error of the algorithm: if μ is too large, the algorithm converges quickly but the steady-state residual noise is large. If the step size is too small, the steady-state error will be small, but the convergence speed will be slow. This embodiment employs a variable step size strategy: in the initial stage after detecting a signal source trigger event (e.g., the first 2 seconds), a larger step size is used (…). =0.5) for fast convergence; then switch to a smaller step size ( =0.1) to reduce steady-state error. The basis of this strategy is that after the signal source triggers an event, the vibration environment may change rapidly due to the event itself (such as the sudden change in hull vibration caused by a collision), and the filter weight coefficients need to be adjusted quickly to track the changes in noise characteristics.

[0032] In one embodiment of the present invention, the step of obtaining the signal periodicity characteristics and signal dynamic characteristics of each signal source based on the preprocessed signal sequence includes: S21, adjust the mapping relationship between the current ship navigation status parameters and the preset feature recognition window to obtain the feature recognition time window length corresponding to the current sampling time; S22, obtain the signal source type identifier according to each heterogeneous signal source, and obtain the corresponding reference period, period deviation tolerance range and change rate reference interval from the pre-stored signal source type and reference feature mapping table; S23, within the time interval defined by the feature recognition time window length, zero-crossing detection is performed on the preprocessed signal sequence, and the periodic change frequency value is obtained based on the number of detected zero-crossing events. S24, obtain the signal periodic characteristics based on the relative deviation between the periodic change frequency value and the period allowable interval determined by the reference period and the period deviation allowable range; S25, within the time interval defined by the feature recognition time window length, an instantaneous change rate sequence is obtained based on the amplitude difference between adjacent sampling points of the preprocessed signal sequence, and the instantaneous change rate sequence is subjected to median filtering to obtain a representative change rate. S26, obtain the dynamic characteristics of the signal based on the degree of deviation between the representative rate of change and the reference range of the rate of change.

[0033] As described in steps S21-S26 above, the core task of this invention in the signal feature identification stage is to extract feature quantities that can effectively distinguish between real trigger signals and interference signals from the preprocessed signal sequence. Since different types of emergency signal sources have different signal characteristic patterns—fire alarm detectors typically output periodic pulse signals, water ingress alarm sensors typically output slowly varying analog signals, and collision detection devices output transient impact signals—using a uniform feature discrimination standard for all types of signal sources would result in excessively high discrimination sensitivity for some types and excessively low sensitivity for others. Therefore, this invention introduces a signal source type and reference feature mapping table, setting differentiated reference periods and rate-of-change reference ranges for each type of signal source.

[0034] Specifically, this invention first adjusts the feature recognition time window length based on the current ship navigation state parameters. These parameters include, but are not limited to, ship speed, sea state level, and main engine speed. Under high sea state conditions, ship vibration intensifies, increasing the noise component in the signal, requiring a longer feature recognition time window to accumulate more zero-crossing event samples, thereby improving the statistical reliability of the estimated periodic frequency value. This feature recognition window adjustment mapping relationship is established as follows: experiments are conducted under different sea state conditions, recording the estimated variance of the periodic frequency value under various window lengths. The shortest window length that makes the estimated variance less than a preset upper limit (e.g., the relative standard deviation of the frequency estimate does not exceed 10%) is selected as the feature recognition time window length under that sea state condition. Typical values ​​are: under low sea states (levels 1-3)... Seconds, medium sea state (4-5) Seconds, high sea state (level 6 and above) Second.

[0035] Secondly, the reference frequency value, allowable period deviation range, and reference range of change rate corresponding to the current signal source are obtained from a pre-stored mapping table of signal source types and reference characteristics. This mapping table is pre-established based on the known operating characteristics of each type of signal source. For example, the typical triggering frequency of a fire alarm detector is 1Hz (i.e., 1 pulse per second), with an allowable deviation range of ±20%, so the allowable period range is [0.8Hz, 1.2Hz]; the typical triggering frequency of a water ingress alarm sensor is 0.5Hz, with an allowable deviation range of ±30%, so the allowable period range is [0.35Hz, 0.65Hz]. The reference range of change rate is determined based on the amplitude change characteristics of different types of signal sources. For example, the typical change rate of a fire alarm detector is [5, 50] amplitude units / second, and the typical change rate of a collision detection device is [100, 10000] amplitude units / second.

[0036] Then, within the feature recognition time window, zero-crossing detection is performed on the preprocessed signal sequence to obtain the periodic change frequency value. The specific implementation of zero-crossing detection is as follows: the median level of the preprocessed signal sequence within the feature recognition time window is calculated (taking the median of the signal amplitude within that window). The number of rising edge crossings from below the median level to above the median level, and the number of falling edge crossings from above the median level to below the median level are counted. The sum of these two counts is divided by 2 to obtain the number of zero-crossing events. Then divide by the time window length Obtain the periodic change frequency value The reason for using the median level instead of the average level is that the median level is not sensitive to abnormal amplitudes and can avoid the influence of residual pulse interference in the signal on zero-crossing detection.

[0037] Next, the absolute value of the amplitude difference between adjacent sampling points in the preprocessed signal sequence is divided by the sampling period to obtain the instantaneous rate of change sequence. Since the instantaneous rate of change sequence may contain high-frequency fluctuations caused by noise, this invention applies median filtering to the sequence to obtain a representative rate of change. The typical window length of the median filter is 5 sampling points, and its function is to filter out isolated noise spikes while preserving the signal's trend. The advantage of median filtering over mean filtering is that median filtering has a better suppression effect on impulse noise and does not blur the signal's edge features.

[0038] Finally, the signal periodicity characteristic is obtained based on the relative deviation between the periodic frequency value and the allowable periodicity range, and the signal dynamics characteristic is obtained based on the deviation between the rate of change and the reference range of the rate of change. Both the signal periodicity characteristic and the signal dynamics characteristic are dimensionless, with values ​​ranging from [0, 1]. Larger values ​​indicate that the signal characteristics are closer to the reference characteristics, meaning the signal is more likely to be a genuine trigger signal. This normalization process allows different types of signal sources to be evaluated within a unified discrimination framework.

[0039] In one embodiment of the present invention, the step of obtaining a single-channel discrimination score based on the signal periodicity characteristics and signal dynamic characteristics of the signal source, and obtaining an effective signal source based on the single-channel discrimination score, includes: S31. Obtain the sea state level value and the ship roll amplitude value according to the current ship navigation status parameters, and determine the dynamic reference offset according to the sea state level value and the ship roll amplitude value. When the sea state level value is greater than or equal to a preset sea state level threshold or the ship roll amplitude value is greater than or equal to a preset roll amplitude threshold, obtain the dynamic reference offset according to a preset offset calculation function, and adjust the lower limit value of the rate of change reference interval according to the dynamic reference offset to obtain the corrected rate of change reference interval. S32, based on the comparison result between the deviation degree of the signal periodic characteristics and the preset periodic anomaly tolerance, obtain the periodic anomaly mark state of the corresponding heterogeneous signal source; S33, based on the comparison result between the deviation degree of the signal dynamic characteristics and the dynamic anomaly tolerance determined by the modified change rate reference interval, obtain the dynamic anomaly marking state of the corresponding heterogeneous signal source. S34. Based on the periodic anomaly marking state and the dynamic anomaly marking state, determine the overall anomaly state of the corresponding heterogeneous signal source. When the overall anomaly state is an unmarked anomaly, determine the heterogeneous signal source as a valid signal source. Based on the deviation degree of the signal periodic characteristics and the deviation degree of the signal dynamic characteristics, obtain the single-channel discrimination score through a preset single-channel discrimination weighting function. ; In the formula, This represents the single-channel discrimination score (dimensionless, ranging from [0, 1], with a larger value indicating higher signal reliability). The preset weighting coefficients represent the periodic feature dimensions. This represents the preset weight coefficients for dynamic feature dimensions. This represents the normalized absolute deviation of the periodic frequency value relative to the reference frequency value. This represents the normalized absolute deviation of the rate of change relative to the median of the reference range for the rate of change.

[0040] As described in steps S31-S34 above, the reason for introducing a dynamic adjustment criteria for ship operating state parameters in this invention is that the output characteristics of sensors will systematically drift under different sea state conditions. Under high sea state conditions (sea state class 5 or higher), the ship's roll amplitude increases, and the gravitational component experienced by the sensors installed on the hull changes periodically, causing additional fluctuations in the rate of change of the sensor output signal. If a fixed rate of change reference range is used, under high sea state conditions, the rate of change of the normal signal may exceed the reference range due to additional fluctuations, leading to misjudgment as an invalid signal. Therefore, this invention, by reducing the lower limit of the rate of change reference range by a preset reference offset, can expand the permissible rate range under high sea state conditions and reduce the misjudgment rate.

[0041] Specifically, the typical preset sea state level threshold is level 5, and the typical preset roll amplitude threshold is 15 degrees. When the sea state level value is greater than or equal to the preset sea state level threshold or the ship roll amplitude value is greater than or equal to the preset roll amplitude threshold, a dynamic reference offset is obtained according to a preset offset calculation function. The input parameters of this offset calculation function include the sea state level value and the ship roll amplitude value, and the output is the dynamic reference offset, which is typically a linear function: when the sea state level increases by 1 level, the reference offset increases by 10% of the lower limit of the original interval; when the roll amplitude increases by 5 degrees, the reference offset increases by 5% of the lower limit of the original interval. The upper limit of the dynamic reference offset is set to 50% of the lower limit of the original interval to prevent excessive relaxation that could lead to a decrease in the ability to make false judgments.

[0042] It should be noted that the reason for adjusting only the reference range for the rate of change and not the allowable range for the period is that fluctuations in sensor output caused by sea state mainly affect the amplitude and rate of change of the signal, with a relatively small impact on the periodic characteristics of the signal (because the periodic characteristics are mainly determined by the triggering mechanism inside the signal source and have a weak direct correlation with the external environment). Therefore, the allowable range for the period remains unchanged, and the system adapts to changes in sea state only by dynamically adjusting the reference range for the rate of change.

[0043] Secondly, this invention obtains corresponding anomaly marker states based on the deviation degrees of signal periodic characteristics and signal dynamic characteristics, respectively. The method for obtaining the periodic anomaly marker state is as follows: when the periodic change frequency value... Deviation from reference frequency value When the relative deviation exceeds the allowable limit for periodic anomalies, the periodic anomaly flag status is set to "marked anomaly"; otherwise, it is set to "unmarked anomaly". The dynamic anomaly flag status is obtained as follows: when the distance representing the rate of change deviating from the corrected rate of change reference interval exceeds the allowable limit for dynamic anomalies, the dynamic anomaly flag status is set to "marked anomaly"; otherwise, it is set to "unmarked anomaly".

[0044] The overall anomaly status determination rule is as follows: when both the periodic anomaly labeling status and the dynamic anomaly labeling status are "unlabeled anomaly," the overall anomaly status is "unlabeled anomaly," and the signal source is determined to be a valid signal source; when either anomaly labeling status is "labeled anomaly," the overall anomaly status is "labeled anomaly," and the signal source is determined to be an invalid signal source. This dual anomaly labeling mechanism ensures that a signal must simultaneously meet the characteristic requirements of both periodic and dynamic dimensions to be determined as a valid signal, thereby improving the accuracy of authenticity determination.

[0045] When a signal source is determined to be a valid signal source, a single-channel discrimination score is obtained based on the deviation degrees of the signal's periodic characteristics and dynamic characteristics, using a preset single-channel discrimination weighting function. The preset weighting coefficient for the periodic characteristic dimension is used in the formula for calculating this discrimination score. The typical value is 0.6, which is the preset weight coefficient for the dynamic feature dimension. The typical value is 0.4. Weighting coefficient Greater than The rationale behind this setting is that the periodic characteristics of a signal are more specific than its dynamic characteristics. Different types of signal sources have significantly different trigger periods, while the rate of change has relatively low distinguishability. Therefore, the periodic characteristics carry greater weight in the discrimination score. When the signal characteristics are completely within the center value of each reference interval, Take the maximum value of 1; when the signal characteristics deviate from the edge of the reference interval, The value approaches 0. This normalized discriminant score provides a unified quantitative basis for subsequent graded responses.

[0046] In one embodiment of the present invention, the step of obtaining the joint discrimination score of multiple signal sources based on the effective signal sources includes: S41, obtain the pre-stored signal causal correlation map, and obtain the directed connection relationship between each effective signal source, the causal time window length corresponding to each directed connection relationship, and the causal probability parameter based on the pre-stored signal causal correlation map; S42, obtain any two valid signal sources with directed edge connections according to the directed connection relationship, and obtain the trigger time interval between the two valid signal sources according to the trigger time of each of the two valid signal sources; S43, Based on the comparison result between the triggering time interval and the causal time window length, generate the causal matching state corresponding to the directed edge; When the triggering time interval is less than or equal to the causal time window length, the causal matching state is determined to be a match, and the edge association score corresponding to the directed edge is obtained through a preset joint score fusion function based on the single-path discrimination score of each of the two valid signal sources and the causal probability parameter. When the triggering time interval is greater than the causal time window length, the causal matching state is determined to be mismatched, and the edge association score is set to zero; S44. Based on the edge association scores of each directed edge corresponding to all valid signal sources and the single-path discrimination scores of each valid signal source itself, the joint discrimination scores of multiple signal sources are obtained through a preset graph structure score aggregation function. ; In the formula, This represents the joint discrimination score (dimensionless). This represents the single-channel discrimination score of the first signal source. This represents the single-channel discrimination score of the second signal source. This represents the causal probability parameter (dimensionless, with a value range of [0, 1], reflecting the prior probability that the second signal source will be triggered within the causal time window after the first signal source is triggered). This represents the preset correlation enhancement coefficient (dimensionless, typically 0.5).

[0047] As described in steps S41-S44 above, the core idea of ​​this invention, which introduces a signal causal correlation graph for multi-source correlation verification, is that in complex accident scenarios such as ship collisions and flooding, the triggering of multiple emergency signal sources is not independent but rather has a physical causal relationship. For example, a ship collision may cause hull damage and flooding, thus there is a causal relationship between the collision detection device and the flooding alarm sensor; a fire may cause structural failure, thus there is a causal relationship between the fire alarm detector and the structural integrity monitoring device. By utilizing this prior knowledge of causal relationships, when multiple signal sources are triggered sequentially in time and their timing conforms to causal logic, the confidence level in judging the authenticity of the event can be enhanced.

[0048] Specifically, the signal causal relationship graph is pre-stored in the form of a directed graph. Nodes in the directed graph represent various signal sources, and directed edges represent the physical causal relationships between these sources. Each directed edge corresponds to a preset causal time window and causal probability parameter. The process of establishing the signal causal relationship graph is as follows: Based on fault tree analysis and event tree analysis in ship safety engineering, the physical causal relationships between various signal sources are identified. The causal probability parameter for each directed edge is obtained through statistical analysis of historical accident data and simulation data. The typical range of the causal time window is 30 to 300 seconds, with the specific value depending on the physical mechanism of the causal relationship. For example, the typical causal time window for a collision leading to water ingress is 120 seconds, while the typical causal time window for a fire leading to structural failure is 300 seconds.

[0049] Once any two valid signal sources connected by a directed edge are obtained, the trigger time interval is calculated based on the trigger times of each of the two valid signal sources. The trigger time interval is defined as: the trigger time of the second signal source minus the trigger time of the first signal source (where the first signal source is the starting node of the directed edge, and the second signal source is the ending node of the directed edge). When the trigger time interval is greater than zero (i.e., the first signal source triggers first) and less than or equal to the causal time window length, the causal matching state is determined to be "matched," indicating that the triggering sequence of the two signal sources conforms to causal logic.

[0050] When the causal matching status is "matched," the edge correlation score is obtained through a joint score fusion function based on the individual discrimination scores and causal probability parameters of the two valid signal sources. This fusion function means that when there is a causal correlation between the two signal sources and their triggering timing conforms to causal logic, the joint discrimination score is higher than the maximum value of the individual discrimination scores. Specifically, the larger of the individual discrimination scores of the two signal sources is taken as the base score, and the product of the causal probability parameter, the smaller discrimination score, and the correlation enhancement coefficient is added to form the correlation enhancement term. The correlation enhancement coefficient is then used. The typical value is 0.5, and the physical meaning of this value is: when the causal relationship is completely determined ( Furthermore, when the discrimination scores of the two signals are equal, the joint discrimination score is 50% higher than the single-channel discrimination score. This improvement reflects the gain effect of multi-source correlation verification on the discrimination confidence.

[0051] When the triggering time interval is greater than the causal time window length or the triggering time interval is negative (i.e., the second signal source triggers before the first signal source), the causal matching state is determined to be "mismatched", and the edge correlation score is set to zero. At this time, there is no causal correlation enhancement between the two signal sources, and the joint discrimination score is the maximum value of the single-path discrimination scores of each signal source.

[0052] When more than two valid signal sources arrive sequentially within the associated time window, causal correlation matching verification is performed on each combination of valid signal sources. The edge correlation score for each matching group is calculated, and the final joint discrimination score is obtained through a graph structure score aggregation function. The graph structure score aggregation function is implemented as follows: the individual discrimination score of each valid signal source is used as the initial score of the node, and the edge correlation score of each matched directed edge is used as the edge weight. Score aggregation is performed on the graph structure through a message passing mechanism. For signal source combinations without causal correlation, the joint discrimination score is the maximum value of the individual discrimination scores of each signal source.

[0053] This invention also includes a noise-correlated edge exclusion mechanism: within a preset statistical period, if the cumulative number of causal triggering events corresponding to a certain directed edge is lower than a preset minimum causal triggering frequency threshold, then the directed edge is marked as noise-correlated and excluded from the correlation verification. The purpose of this mechanism is that some directed edges in the signal causal correlation graph may never or rarely experience causal triggering in actual operation; if retained in the correlation verification, they may generate false joint discrimination score enhancements.

[0054] The causal time window is dynamically adjusted based on the current ship navigation status parameters: when the sea state level is greater than or equal to a preset high sea state threshold, the causal time window is expanded by a preset amount to cover a wider range of temporal correlations. This expansion is based on the fact that under high sea state conditions, the delay in the signal acquisition link increases, and the occurrence and transmission of events may be accelerated or delayed due to severe sea conditions, requiring a wider causal time window to accommodate this temporal uncertainty.

[0055] In one embodiment of the present invention, the preset multi-level response level conditions include the following progressive response levels, including: S51, obtain the current single-channel discrimination score interval, the current joint discrimination score interval, and the highest priority level of the current valid signal source according to the preset response level determination mapping table; S52, generate a current response level based on the logical combination between the current single-path discrimination score interval, the current joint discrimination score interval, and the highest priority level, wherein the current response level includes at least one of the first-level response, the second-level response, the third-level response, and the fourth-level response; When the current response level is determined to be the first level response, the corresponding broadcast area identifier is obtained according to the signal source type identifier of the corresponding valid signal source, and the speaker in the corresponding area is triggered to play the preset first broadcast content according to the broadcast area identifier, and automatically terminated after playing for a preset first duration. When the current response level is determined to be the second level response, the ship's loudspeakers are triggered to play the preset second broadcast content according to the preset ship-wide broadcast configuration parameters, and the broadcast is played in a loop according to the preset playback interval duration and number of repetitions. When the current response level is determined to be the third level response, the ship's loudspeakers are triggered to play the preset third broadcast content according to the preset ship-wide broadcast configuration parameters, and the playback state is maintained until a manual reset command is received and the preset confirmation retention time condition is met. When the current response level is determined to be the fourth level response, the distress signal association verification result is obtained based on the time difference between the trigger time of the emergency position beacon signal and the arrival time of the digital selective call distress signal. When the distress signal association verification result is verified and the current joint discrimination score is greater than or equal to the preset distress confirmation threshold, the preset abandon ship alarm sequence is triggered to be played in a loop by all the ship's loudspeakers, and the playback state is maintained until a manual reset command is received.

[0056] As described in steps S51-S52 above, this embodiment of the invention discloses a specific method for triggering corresponding levels of emergency response actions based on single-channel discrimination scores and joint discrimination scores. After obtaining the single-channel discrimination scores of each valid signal source and the joint discrimination scores of multiple signal sources, it is necessary to trigger an appropriate level of emergency response based on signal reliability and event severity. If only a single threshold is used for triggering, the following situations may occur: when the threshold is set too high, some real emergency events may not be responded to because the discrimination score does not reach the threshold (missed reporting); when the threshold is set too low, noise interference may trigger unnecessary broadcasts (false alarms). Therefore, the graded response strategy of this invention automatically selects the response level based on the discrimination score and signal source priority, achieving a balance between response sensitivity and accuracy.

[0057] Specifically, the current response levels include the following four progressive response tiers: Level 1 Response (Information Level), Level 2 Response (Warning Level), Level 3 Response (Emergency Level), and Level 4 Response (Abandon Ship Level). Typical parameter configurations for each response level are as follows: Level 1 response typically lasts for 30 seconds at a normal conversation level; Level 2 response typically lasts for 60 seconds, repeats 3 times, and is at a warning level; Level 3 response continues until manually reset by watch personnel and the retention time is confirmed; Level 4 response continuously plays the abandon ship warning sequence specified by the International Maritime Organization (IMO) until manually reset, with no automatic termination option.

[0058] Typical values ​​for the first and second score intervals and each recognition threshold are as follows: The first score interval is [0.3, 0.6], indicating that the signal has a certain degree of reliability but is not certain; the second score interval is [0.5, 0.8], indicating that the signal has a high degree of reliability; the first recognition threshold is 0.7, the second recognition threshold is 0.85, the third recognition threshold is 0.9, and the fourth recognition threshold is 0.95. These parameter values ​​can be adjusted through factory calibration and actual shipboard testing.

[0059] The typical configuration of signal source priorities is as follows: emergency position beacons and abandon ship command signal sources are the first priority (highest), digital selective call distress signals are the second priority, fire alarm detectors are the third priority, water ingress alarm sensors are the third priority, collision detection devices are the fourth priority, and other auxiliary monitoring signal sources are the fifth priority (lowest).

[0060] The response level determination logic is as follows: When the single-channel discrimination score is in the first score range and the signal source priority is higher than or equal to the fifth priority, a first-level response (indication level) is triggered; when the single-channel discrimination score is greater than or equal to the first identification threshold and the signal source priority is higher than or equal to the third priority, or when the combined discrimination score is in the second score range, a second-level response (warning level) is triggered; when the combined discrimination score is greater than or equal to the second identification threshold, or when the single-channel discrimination score is greater than or equal to the third identification threshold and the signal source priority is lower than or equal to the second priority, a third-level response (emergency level) is triggered; when the emergency beacon signal is triggered and the DSC distress signal arrives within the preset distress association time window, and the combined discrimination score is greater than or equal to the fourth identification threshold, a fourth-level response (abandon ship level) is triggered.

[0061] In one embodiment of the present invention, the step of triggering a broadcast action of the corresponding level includes: S61, obtain the broadcast priority value of each trigger signal source according to the pre-stored signal priority mapping table, and determine the broadcast scheduling strategy according to the comparison result between the broadcast priority values. If the broadcast priority value of the current triggering signal source is higher than the broadcast priority value of the signal source that is currently broadcasting, the current broadcast is interrupted and the interruption status information is recorded, and the broadcast content corresponding to the current triggering signal source is switched. If the broadcast priority value of the current triggering signal source is lower than or equal to the broadcast priority value of the signal source that is currently broadcasting, the current triggering signal source will be added to the waiting queue. S62, based on the interruption status information, after the broadcast of the current trigger signal source is completed, determine whether the interrupted broadcast has exceeded the preset valid execution time limit. If it has not exceeded the time limit, resume the execution of the interrupted broadcast according to the interruption position in the interruption status information. S63, determine the corresponding response suppression window duration according to the broadcast priority value of each of the trigger signal sources, wherein the response suppression window duration is inversely proportional to the broadcast priority value, and within the response suppression window duration, repeated triggering events of the same trigger signal source are merged and suppressed.

[0062] As described in steps S61-S63 above, this invention, based on graded response triggering, further provides a priority arbitration scheduling mechanism for simultaneous triggering of multiple signal sources. In complex ship accident scenarios, multiple signal sources may trigger emergency responses simultaneously, with each signal source corresponding to a different response level and broadcast content. Without effective scheduling management, broadcast content may conflict or switch frequently, preventing crew members from obtaining clear emergency instructions. Therefore, this invention, through priority arbitration and waiting queue mechanisms, ensures that high-priority responses are executed promptly, while guaranteeing that interrupted low-priority responses can be restored after the high-priority responses have ended.

[0063] Specifically, the interruption point is recorded as follows: For replayable response content, the current loop count and playback progress within the current loop are recorded (expressed as a time offset); for one-time playback response content, the time elapsed during playback is recorded. A typical effective execution time limit is 300 seconds (i.e., 5 minutes). If the interrupted response has exceeded this time limit since its triggering, playback will not be resumed and will be terminated directly. This mechanism is based on the fact that after the effective execution time limit has expired, the interrupted response content may no longer be timely, and resuming playback might interfere with the ongoing emergency response.

[0064] The duration of the response suppression window corresponds to the signal source priority as follows: 5 seconds for the first priority signal source, 10 seconds for the second priority, 20 seconds for the third priority, 30 seconds for the fourth priority, and 60 seconds for the fifth priority. Higher priority signal sources have shorter response suppression windows to ensure rapid response. Within the response suppression window duration, repeated response actions triggered by the same signal source are merged and suppressed, extending the remaining duration of the currently executing response action to its original duration. This mechanism prevents repeated interruptions and restarts of broadcast content caused by sensor jitter.

[0065] This invention also includes a queue timeout forced scheduling mechanism: a queue timeout duration is set for signal sources in the waiting queue with a priority higher than or equal to a preset high-priority threshold. If execution has not started by the end of the queue timeout period, the currently executing low-priority response is forcibly interrupted, and the response content corresponding to the high-priority signal source is switched. The high-priority threshold is set to the second priority, the queue timeout duration for first-priority signal sources is 30 seconds, and the queue timeout duration for second-priority signal sources is 60 seconds. Signal sources in the waiting queue are sorted from highest to lowest priority, and signal sources with the same priority are sorted according to their arrival time.

[0066] like Figure 2 As shown, the present invention also provides an emergency broadcast multi-signal identification triggering system, comprising: The signal preprocessing module is used to acquire the raw signal data of multiple heterogeneous signal sources and acquire the preprocessed signal sequence corresponding to each signal source based on the raw signal data. The signal feature extraction module is used to obtain the signal periodic features and signal dynamic features of each signal source based on the preprocessed signal sequence; A single-channel discrimination module is used to obtain a single-channel discrimination score based on the signal periodicity and dynamic characteristics of the signal source, and to obtain an effective signal source based on the single-channel discrimination score; The joint discrimination module is used to obtain the joint discrimination score of multiple signal sources based on the valid signal source; The broadcast triggering module is used to trigger a broadcast action of the corresponding level according to the single-channel discrimination score and the joint discrimination score, and according to the preset multi-level response level conditions.

[0067] The present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of an emergency broadcast multi-signal identification triggering method.

[0068] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of an emergency broadcast multi-signal identification triggering method.

[0069] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, apparatus, article, or method. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.

[0070] The above description is merely a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. An emergency broadcast multi-signal identification triggering method, applied to ship broadcasting equipment, characterized in that, include: Acquire raw signal data from multiple heterogeneous signal sources, and obtain the preprocessed signal sequence corresponding to each signal source based on the raw signal data; Adjust the mapping relationship between the current ship navigation status parameters and the preset feature recognition window to obtain the length of the feature recognition time window corresponding to the current sampling time. The signal source type identifier is obtained according to each heterogeneous signal source, and the corresponding reference period, period deviation tolerance range and change rate reference interval are obtained from the pre-stored signal source type and reference feature mapping table. Within the time interval defined by the feature recognition time window length, zero-crossing detection is performed on the preprocessed signal sequence, and the periodic change frequency value is obtained based on the number of detected zero-crossing events. The signal periodic characteristics are obtained based on the relative deviation between the periodic variation frequency value and the period allowable interval determined by the reference period and the period deviation allowable range; Within the time interval defined by the feature recognition time window length, an instantaneous rate of change sequence is obtained based on the amplitude difference between adjacent sampling points of the preprocessed signal sequence, and the instantaneous rate of change sequence is subjected to median filtering to obtain a representative rate of change. Based on the degree of deviation between the representative rate of change and the reference range of the rate of change, and based on the degree of deviation, the dynamic characteristics of the signal are obtained; A single-channel discrimination score is obtained based on the signal periodicity and dynamic characteristics of the signal source, and an effective signal source is obtained based on the single-channel discrimination score. Obtain a pre-stored signal causal correlation map, and based on the pre-stored signal causal correlation map, obtain the directed connection relationship between each of the effective signal sources, the causal time window length corresponding to each of the directed connection relationships, and the causal probability parameters; Based on the directed connection relationship, obtain any two valid signal sources that are connected by a directed edge, and obtain the trigger time interval between the two valid signal sources based on their respective trigger times; Based on the comparison result between the trigger time interval and the causal time window length, the causal matching state corresponding to the directed edge is generated; When the triggering time interval is less than or equal to the causal time window length, the causal matching state is determined to be a match, and the edge association score corresponding to the directed edge is obtained through a preset joint score fusion function based on the single-path discrimination score of each of the two valid signal sources and the causal probability parameter. When the triggering time interval is greater than the causal time window length, the causal matching state is determined to be mismatched, and the edge association score is set to zero; Based on the edge association scores of each directed edge corresponding to all valid signal sources and the single-path discrimination scores of each valid signal source, the joint discrimination scores of multiple signal sources are obtained through a preset graph structure score aggregation function. Based on the single-path discrimination score and the joint discrimination score, a broadcast action of the corresponding level is triggered according to the preset multi-level response level conditions; The step of obtaining a single-channel discrimination score based on the signal periodicity and dynamic characteristics of the signal source, and obtaining an effective signal source based on the single-channel discrimination score, includes: The sea state level and ship roll amplitude are obtained based on the current ship navigation status parameters. A dynamic reference offset is determined based on the sea state level and ship roll amplitude. When the sea state level is greater than or equal to a preset sea state level threshold or the ship roll amplitude is greater than or equal to a preset roll amplitude threshold, the dynamic reference offset is obtained based on a preset offset calculation function. The lower limit of the rate of change reference interval is then adjusted downward based on the dynamic reference offset to obtain a corrected rate of change reference interval. Based on the comparison between the deviation of the signal periodic characteristics and the preset periodic anomaly tolerance, the periodic anomaly marking status of the corresponding heterogeneous signal source is obtained. Based on the comparison between the degree of deviation of the signal dynamic characteristics and the dynamic anomaly tolerance determined by the modified rate of change reference interval, the dynamic anomaly labeling state of the corresponding heterogeneous signal source is obtained. Based on the periodic anomaly marking state and the dynamic anomaly marking state, the overall anomaly state of the corresponding heterogeneous signal source is determined. When the overall anomaly state is an unmarked anomaly, the heterogeneous signal source is determined as a valid signal source. Based on the deviation degree of the signal periodic characteristics and the deviation degree of the signal dynamic characteristics, the single-channel discrimination score is obtained through a preset single-channel discrimination weighting function.

2. The emergency broadcast multi-signal identification triggering method according to claim 1, characterized in that, The step of obtaining the preprocessed signal sequence corresponding to each signal source based on the original signal data includes: Based on the rising edge sampling point index where the signal amplitude first exceeds the preset silence threshold in the original signal data of each heterogeneous signal source, the relative time offset between each heterogeneous signal source is determined, and the original signal data of each channel is shifted and compensated according to the relative time offset to obtain a timestamp mark sequence carrying a unified time reference. Based on the pre-stored working frequency band parameters obtained from each of the heterogeneous signal sources, the corresponding upper and lower cutoff frequencies of the passband are determined, and the bandpass filter signal is obtained based on the aligned signal data corresponding to the timestamp mark sequence and the upper and lower cutoff frequencies of the passband. Based on the vibration acceleration signals collected by vibration reference sensors rigidly deployed at the same position as each of the heterogeneous signal sources, the bandpass filtered signal is subjected to adaptive noise cancellation processing with the vibration acceleration signals as reference input, and the preprocessed signal sequence corresponding to each of the heterogeneous signal sources is obtained based on the error output signal after the adaptive noise cancellation processing.

3. The emergency broadcast multi-signal identification triggering method according to claim 1, characterized in that, The preset multi-level response conditions include the following progressive response levels: The current single-channel discrimination score range, the current joint discrimination score range, and the highest priority level of the current valid signal source are obtained according to the preset response level determination mapping table. The current response level is generated based on the logical combination of the current single-path discrimination score interval, the current joint discrimination score interval, and the highest priority level. The current response level includes at least one of the following: first-level response, second-level response, third-level response, and fourth-level response. When the current response level is determined to be the first level response, the corresponding broadcast area identifier is obtained according to the signal source type identifier of the corresponding valid signal source, and the speaker in the corresponding area is triggered to play the preset first broadcast content according to the broadcast area identifier, and automatically terminated after playing for a preset first duration. When the current response level is determined to be the second level response, the ship's loudspeakers are triggered to play the preset second broadcast content according to the preset ship-wide broadcast configuration parameters, and the broadcast is played in a loop according to the preset playback interval duration and number of repetitions. When the current response level is determined to be the third level response, the ship's loudspeakers are triggered to play the preset third broadcast content according to the preset ship-wide broadcast configuration parameters, and the playback state is maintained until a manual reset command is received and the preset confirmation retention time condition is met. When the current response level is determined to be the fourth level response, the distress signal association verification result is obtained based on the time difference between the trigger time of the emergency position beacon signal and the arrival time of the digital selective call distress signal. When the distress signal association verification result is verified and the current joint discrimination score is greater than or equal to the preset distress confirmation threshold, the preset abandon ship alarm sequence is triggered to be played in a loop by all the ship's loudspeakers, and the playback state is maintained until a manual reset command is received.

4. The emergency broadcast multi-signal identification triggering method according to claim 1, characterized in that, The steps to trigger a broadcast action of the corresponding level include: The broadcast priority value of each trigger signal source is obtained according to the pre-stored signal priority mapping table, and the broadcast scheduling strategy is determined according to the comparison result between the broadcast priority values. If the broadcast priority value of the current triggering signal source is higher than the broadcast priority value of the signal source that is currently broadcasting, the current broadcast is interrupted and the interruption status information is recorded, and the broadcast content corresponding to the current triggering signal source is switched. If the broadcast priority value of the current triggering signal source is lower than or equal to the broadcast priority value of the signal source that is currently broadcasting, the current triggering signal source will be added to the waiting queue. Based on the interruption status information, after the broadcast of the current trigger signal source is completed, it is determined whether the interrupted broadcast has exceeded the preset valid execution time limit. If it has not exceeded the time limit, the interrupted broadcast is resumed according to the interruption position in the interruption status information. The corresponding response suppression window duration is determined based on the broadcast priority value of each of the trigger signal sources, wherein the response suppression window duration is inversely proportional to the broadcast priority value, and repeated triggering events of the same trigger signal source are merged and suppressed within the response suppression window duration.

5. An emergency broadcast multi-signal identification triggering system, characterized in that, include: The signal preprocessing module is used to acquire the raw signal data of multiple heterogeneous signal sources and acquire the preprocessed signal sequence corresponding to each signal source based on the raw signal data. The signal feature extraction module is used to adjust the mapping relationship between the current ship navigation status parameters and the preset feature recognition window to obtain the feature recognition time window length corresponding to the current sampling time. The signal source type identifier is obtained according to each heterogeneous signal source, and the corresponding reference period, period deviation tolerance range and change rate reference interval are obtained from the pre-stored signal source type and reference feature mapping table. Within the time interval defined by the feature recognition time window length, zero-crossing detection is performed on the preprocessed signal sequence, and the periodic change frequency value is obtained based on the number of detected zero-crossing events. The signal periodic characteristics are obtained based on the relative deviation between the periodic variation frequency value and the period allowable interval determined by the reference period and the period deviation allowable range; Within the time interval defined by the feature recognition time window length, an instantaneous rate of change sequence is obtained based on the amplitude difference between adjacent sampling points of the preprocessed signal sequence, and the instantaneous rate of change sequence is subjected to median filtering to obtain a representative rate of change. Based on the degree of deviation between the representative rate of change and the reference range of the rate of change, and based on the degree of deviation, the dynamic characteristics of the signal are obtained; A single-channel discrimination module is used to obtain a single-channel discrimination score based on the signal periodicity and dynamic characteristics of the signal source, and to obtain an effective signal source based on the single-channel discrimination score; The joint discrimination module is used to obtain a pre-stored signal causal correlation map, and based on the pre-stored signal causal correlation map, obtain the directed connection relationship between each of the effective signal sources, the causal time window length corresponding to each directed connection relationship, and the causal probability parameter. Based on the directed connection relationship, obtain any two valid signal sources that are connected by a directed edge, and obtain the trigger time interval between the two valid signal sources based on their respective trigger times; Based on the comparison result between the trigger time interval and the causal time window length, the causal matching state corresponding to the directed edge is generated; When the triggering time interval is less than or equal to the causal time window length, the causal matching state is determined to be a match, and the edge association score corresponding to the directed edge is obtained through a preset joint score fusion function based on the single-path discrimination score of each of the two valid signal sources and the causal probability parameter. When the triggering time interval is greater than the causal time window length, the causal matching state is determined to be mismatched, and the edge association score is set to zero; Based on the edge association scores of each directed edge corresponding to all valid signal sources and the single-path discrimination scores of each valid signal source, the joint discrimination scores of multiple signal sources are obtained through a preset graph structure score aggregation function. The broadcast triggering module is used to trigger a broadcast action of the corresponding level according to the single-channel discrimination score and the joint discrimination score, and according to the preset multi-level response level conditions. The step of obtaining a single-channel discrimination score based on the signal periodicity and dynamic characteristics of the signal source, and obtaining an effective signal source based on the single-channel discrimination score, includes: The sea state level and ship roll amplitude are obtained based on the current ship navigation status parameters. A dynamic reference offset is determined based on the sea state level and ship roll amplitude. When the sea state level is greater than or equal to a preset sea state level threshold or the ship roll amplitude is greater than or equal to a preset roll amplitude threshold, the dynamic reference offset is obtained based on a preset offset calculation function. The lower limit of the rate of change reference interval is then adjusted downward based on the dynamic reference offset to obtain a corrected rate of change reference interval. Based on the comparison between the deviation of the signal periodic characteristics and the preset periodic anomaly tolerance, the periodic anomaly marking status of the corresponding heterogeneous signal source is obtained. Based on the comparison between the degree of deviation of the signal dynamic characteristics and the dynamic anomaly tolerance determined by the modified rate of change reference interval, the dynamic anomaly labeling state of the corresponding heterogeneous signal source is obtained. Based on the periodic anomaly marking state and the dynamic anomaly marking state, the overall anomaly state of the corresponding heterogeneous signal source is determined. When the overall anomaly state is an unmarked anomaly, the heterogeneous signal source is determined as a valid signal source. Based on the deviation degree of the signal periodic characteristics and the deviation degree of the signal dynamic characteristics, the single-channel discrimination score is obtained through a preset single-channel discrimination weighting function.

6. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 4.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 4.

Images