Personnel falling-into-water detection method and system based on star flash distance measurement

Abstract

The invention discloses a personnel falling-into-water detection method and system based on star flash distance measurement, and belongs to the technical field of water rescue, and the method comprises the steps: respectively configuring communication modules on water equipment and a user; reading an equipment motion signal of the water equipment, and establishing a motion state identifier; after a communication module is configured by using the motion state identifier, establishing a relative distance data set with a time identifier; activating a threshold verification channel according to the motion state identifier, and establishing a first abnormal verification result; establishing a user motion signal through an acceleration sensor arranged at a user position, and establishing a second abnormity verification result; and performing joint anomaly analysis according to the first anomaly verification result and the second anomaly verification result, and reporting a water falling detection result. According to the invention, the technical problems of high false alarm rate and insufficient detection reliability caused by environmental interference and poor scene adaptability due to the fact that people falling into water are detected only depending on single distance measurement or motion signals in the prior art are solved.

Description

Technical Field

[0001] This application relates to the field of water rescue technology, specifically to a method and system for detecting people falling into the water based on star-flash ranging. Background Technology

[0002] The increasing popularity of water-based work and recreational activities has made the prevention and emergency response to drowning accidents a crucial safety issue. Traditional drowning detection methods, such as relying on manual lookout or simple GPS positioning, suffer from long response delays, low positioning accuracy, and susceptibility to environmental interference, making it difficult to provide timely and effective rescue for those who have fallen into the water. Developing an intelligent technology capable of real-time, accurate, and automatic detection of people falling into the water and triggering a rapid response is an urgent practical need to ensure the safety of people on the water.

[0003] Existing wireless-based water-fall detection solutions mostly employ proximity sensing technologies such as Bluetooth or Wi-Fi. However, these solutions have inherent limitations in transmission latency, ranging accuracy, and anti-interference capabilities, especially in complex and turbulent water environments where false alarm rates are high. Furthermore, most solutions rely solely on single changes in distance or the immersion of wearable devices in water as the basis for judgment, failing to integrate this with the motion state of the carrying vehicle. This makes it difficult to effectively distinguish between normal human activity and genuine water-fall hazards, resulting in insufficient reliability. Therefore, a comprehensive solution integrating high-precision ranging, multi-motion information fusion analysis, and environmental adaptive judgment is urgently needed. Summary of the Invention

[0004] This application provides a method and system for detecting people falling into water based on star-flash ranging, aiming to solve the technical problems of existing technologies that rely solely on a single ranging or motion signal to detect people falling into water, which are easily affected by environmental interference, have poor scene adaptability, and thus result in high false alarm rates and insufficient detection reliability.

[0005] In view of the above problems, this application provides a method and system for detecting personnel falling into water based on star-flash ranging.

[0006] The first aspect disclosed in this application provides a method for detecting personnel falling into water based on star-flash ranging. The method includes: configuring communication modules on a water-based device and a user, wherein the communication modules are paired communication modules and connected based on star-flash technology; reading the device motion signal of the water-based device, using the device motion signal to identify the motion state, and establishing a motion state identifier; configuring the communication module using the motion state identifier, performing real-time distance measurement, and establishing a relative distance dataset with a time stamp; activating a threshold verification channel based on the motion state identifier, performing distance anomaly verification on the relative distance dataset, and establishing a first anomaly verification result; establishing a user motion signal through an accelerometer sensor located at the user's position, sending the device motion signal and the user motion signal to a joint verification network, and establishing a second anomaly verification result; performing joint anomaly analysis based on the first anomaly verification result and the second anomaly verification result, and reporting the fall-in detection result.

[0007] Another aspect of this application discloses a personnel fall-in-water detection system based on star-flash ranging. This system includes: a communication module configuration module for configuring communication modules on the water-based equipment and the user, wherein the communication modules are paired communication modules and connected based on star-flash technology; a status identifier establishment module for reading the equipment motion signal of the water-based equipment, using the equipment motion signal to identify the motion state, and establishing a motion state identifier; a distance dataset establishment module for performing real-time distance measurement after configuring the communication modules using the motion state identifier, and establishing a relative distance dataset with a time identifier; a first anomaly verification result establishment module for activating a threshold verification channel based on the motion state identifier, performing distance anomaly verification on the relative distance dataset, and establishing a first anomaly verification result; a second anomaly verification result establishment module for establishing a user motion signal through an accelerometer sensor located at the user's position, sending the equipment motion signal and the user motion signal to a joint verification network, and establishing a second anomaly verification result; and a fall-in-water detection result reporting module for performing joint anomaly analysis based on the first anomaly verification result and the second anomaly verification result, and reporting the fall-in-water detection result.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages: By employing a technical solution that uses star-based ranging to acquire relative distance datasets, simultaneously collects motion signals from both the device and the user, combines motion status identifiers, and generates abnormal results through a joint verification network and a threshold verification channel, and then reports the water-fall detection result after joint analysis, this solution solves the technical problems of existing technologies that rely solely on a single signal for detection, are susceptible to environmental interference, have poor scene adaptability, and thus suffer from high false alarm rates and insufficient reliability. This achieves multi-dimensional verification to improve detection accuracy, enhance scene adaptability and reliability, and provide effective protection for water safety.

[0009] The above description is merely an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, specific embodiments of this application are given below. Attached Figure Description

[0010] Figure 1 A flowchart illustrating a method for detecting personnel falling into water based on star-flash ranging is provided for embodiments of this application.

[0011] Figure 2 A schematic diagram of a personnel falling into water detection system based on star-flash ranging is provided for the embodiments of this application.

[0012] Explanation of reference numerals in the attached diagram: Communication module configuration module 11, status identifier establishment module 12, distance dataset establishment module 13, first anomaly verification result establishment module 14, second anomaly verification result establishment module 15, water fall detection result reporting module 16. Detailed Implementation

[0013] To further illustrate the technical means and effects adopted by this application in order to achieve the intended purpose of the invention, the following detailed description of the specific implementation methods, structures, features and effects of this application is provided in conjunction with the accompanying drawings and preferred embodiments.

[0014] The overall concept of the technical solution provided in this application is as follows: This application provides a method and system for detecting personnel falling into water based on star-flash ranging. The system acquires a dataset of the relative distance between the device and the user using star-flash technology, and simultaneously collects the motion signals of both. Combining motion status identifiers, a joint verification network is used to analyze the motion signals to obtain a second anomaly result, and a threshold verification channel is used to analyze the distance data to obtain a first anomaly result. The water fall detection result is then reported after joint analysis.

[0015] After introducing the basic principles of this application, various non-limiting embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0016] Example 1, as Figure 1As shown in the embodiment of this application, a method for detecting personnel falling into water based on star-flash ranging is provided. The method includes: Step S100: Configure communication modules for the watercraft and the user respectively, wherein the communication modules are paired communication modules and are connected based on star flash technology.

[0017] Specifically, waterborne equipment refers to watercraft or platforms where personnel are located, such as motorboats, sailboats, yachts, and workboats. Users refer to participants in water activities who need protection, such as drivers, passengers, athletes, and staff; users typically wear or carry a device. Paired communication modules mean that a unique, bound trust relationship has been established between two communication modules during the factory or initialization phase. After pairing, upon power-up, they automatically search for and prioritize connecting to each other, avoiding accidental connections to other unrelated devices and ensuring the exclusivity and security of communication. StarFlash technology is a new generation of short-range wireless communication technology. Compared to Bluetooth and Wi-Fi, it has significant advantages such as ultra-low latency, high-precision ranging, high reliability, and high concurrency capabilities. In this solution, its core function is to provide stable and accurate real-time distance measurement capabilities.

[0018] Specifically, a device-side StarScan communication module is installed in the main control cabin of water sports equipment such as electric surfboards and jet skis, while a portable StarScan communication module is also configured for the user. The pairing process between the two modules is triggered by the physical pairing button on the device-side module or the accompanying APP. The modules exchange device IDs and encryption keys through the StarScan protocol to complete the binding and form a unique pairing relationship. After pairing, the two modules automatically establish a StarScan link. The device-side module sends a ranging request signal to the user-side module in real time. After receiving the signal, the user-side module sends a response signal back. The real-time distance is calculated based on the signal flight time. For example, when the electric surfboard is running, the board module and the user's wrist control module complete a signal interaction every 10ms to dynamically maintain the connection and ranging function.

[0019] This step utilizes the low latency and high precision characteristics of StarFlash technology, combined with the dedicated connection of the pairing module, to avoid the problem of traditional Bluetooth signals being easily interfered with by waves and metal devices in a water environment, while ensuring the real-time and accuracy of distance data between the device and the user.

[0020] Step S200: Read the motion signal of the water equipment, use the motion signal to identify the motion state, and establish a motion state identifier.

[0021] Specifically, equipment motion signals refer to the physical parameter signals generated by water-based equipment during its movement, including acceleration, angular velocity, velocity, and displacement. These are collected in real time by sensors and reflect the equipment's motion characteristics, such as whether it is bumpy, its speed, or whether it is stationary. For example, when an electric surfboard travels in waves, its acceleration signal will exhibit high-frequency fluctuations, while the signal tends to be stable when stationary. Motion status identification is the result of standardized and tagged processing of the identified motion status, usually presented in the form of codes, symbols, or text, and used as parameters in subsequent steps.

[0022] Specifically, an inertial measurement unit (IMU) integrating accelerometers and gyroscopes is installed at the center of gravity of the water-based equipment. This sensor hardware reads the equipment's acceleration and angular velocity signals in real time. The raw signals are then fed into an edge computing module, where a Kalman filter algorithm is used to eliminate noise caused by wave impact and equipment vibration. Signal features are extracted using a sliding window method and input into a trained motion state classification model, such as a Support Vector Machine (SVM) classifier, trained using the Scikit-learn library. Training samples include raw acceleration and angular velocity data, acceleration fluctuation amplitude during turbulence, and peak frequency of the spectral spectrum. After the model outputs the current motion state of the equipment, a corresponding motion state identifier is generated and stored in the equipment's local cache for subsequent steps. The motion state classification model...

[0023] This step, by accurately identifying the movement status of the water equipment, provides crucial scenario adaptation for subsequent processes, reduces the misjudgment rate caused by changes in the equipment's own movement status, and lays the foundation for the reliability of the overall water-fall detection system.

[0024] Step S300: After configuring the communication module using the motion status identifier, perform real-time distance measurement and establish a relative distance dataset with time identifier.

[0025] Specifically, the system reads the motion status identifier and sends configuration commands to the StarScan communication module via the edge computing module: if stationary, the sampling frequency is set to low and the transmission power is reduced to lower power consumption; if traveling at high speed or in a bumpy state, the sampling frequency is increased, the transmission power is increased, and high-precision ranging algorithms, such as PDOA, are prioritized to resist interference. After configuration, the StarScan modules paired with the user end establish real-time communication through the StarScan link. The device module transmits a high-frequency signal, which is received and immediately transmitted back by the user end module. The module's built-in signal processing unit calculates the distance value based on the ToF algorithm; simultaneously, it synchronizes time via the NTP time protocol, adding a millisecond-level timestamp to the distance value. Finally, the "timestamp - distance value" is stored in the local database in the order of collection, forming a relative distance dataset with time stamps.

[0026] For example, if the motion status is identified as high-speed driving, the edge computing module sends a command to the device-side Star Flash module to adjust the sampling frequency to 50Hz, set the transmission power to 15dBm, and enable the PDOA algorithm; the module and the user's wrist control module complete a signal interaction every 20ms through the Star Flash link, calculate the distance value, synchronize the timestamp, and form a dataset fragment after continuous collection.

[0027] This step dynamically configures the communication module through motion state identifiers, achieving scene adaptation; at the same time, the relative distance dataset with time identifiers can accurately capture the temporal characteristics of distance changes, providing complete data evidence including the time dimension for distance anomaly verification.

[0028] Step S400: Activate the threshold verification channel according to the motion state identifier, perform distance anomaly verification of the relative distance dataset, and establish the first anomaly verification result.

[0029] Specifically, the threshold verification channel refers to the distance verification logic module corresponding to different motion states. Each channel is preset with parameters such as the normal distance range and fluctuation tolerance for that scenario, avoiding the problem that a single threshold cannot adapt to all motion scenarios. For example, the static scene channel adapts to distance verification when the device is parked, and the high-speed scene channel adapts to distance verification when the device is moving at high speed. The first anomaly verification result refers to the preliminary judgment result generated after anomaly verification is completed solely through distance data.

[0030] Specifically, the motion state identifier is read, and the corresponding threshold verification channel is activated based on the identifier. The channel parameters include the normal distance range in this state, the maximum allowable value for a single fluctuation, and the time window for continuous anomaly judgment. The sliding window algorithm is called through the edge computing module to extract continuous time-series distance data from the database storing the relative distance dataset, and the mean distance, instantaneous rate of change, and cumulative drift within the window are calculated. The calculated distance features are compared with the preset parameters of the currently activated channel. If the distance exceeds the normal range, the rate of change exceeds the fluctuation threshold, or the cumulative drift reaches the judgment standard, it is judged as an anomaly; otherwise, it is judged as normal. Finally, the judgment result is stored in a temporary cache to form the first anomaly verification result.

[0031] This step achieves scenario-based adaptation of distance anomaly verification by binding motion state with threshold verification channel, avoiding misjudgment of a single threshold in complex scenarios; the sliding window algorithm effectively filters instantaneous interference and improves the stability of judgment.

[0032] Step S500: Establish a user motion signal by using an accelerometer set at the user's location, and send the device motion signal and the user motion signal to the joint verification network to establish a second anomaly verification result.

[0033] Specifically, the user motion signal is a motion parameter signal collected by an accelerometer installed on the user, including acceleration values, rate of change of motion acceleration, impact vibration characteristics, etc., reflecting the user's limb movement state such as normal standing, imbalance, and impact from falling into water. The second anomaly verification result refers to the anomaly judgment result generated based on the joint analysis of the device motion signal and the user motion signal.

[0034] Specifically, an accelerometer is embedded in the user's device to collect the user's acceleration signal in real time. After low-pass filtering, the signal is used to generate the user's motion signal and stored in a local cache. At the same time, the device's motion signal is read. The two types of signals are synchronously transmitted to the joint verification network in the edge computing module through a star-flash communication link. A fusion tensor is established through a registration extraction layer. A multi-dimensional abnormal feature vector is established through a feature extraction layer. A dynamic confidence judgment layer analyzes whether the fusion features conform to normal motion patterns and outputs whether it is normal or abnormal and the type of abnormality, forming a second abnormality verification result.

[0035] This step, by incorporating user motion signals and combining them with device motion signals for joint verification, overcomes the limitations of relying solely on distance data. It provides a reliable motion dimension basis for subsequent joint anomaly analysis, significantly improving the robustness of the overall water-fall detection system.

[0036] Step S600: Perform joint anomaly analysis based on the first anomaly verification result and the second anomaly verification result, and report the water fall detection result.

[0037] Specifically, the first and second anomaly verification results are read from the temporary cache, and joint analysis is performed through the edge computing module. First, the consistency of the two results is determined. If both are anomalies, a comprehensive anomaly score is calculated based on preset weights, for example, distance features and motion features are weighted at 0.4 and 0.6 respectively. If the score exceeds a threshold, it is directly determined as a fall into water. If the results are inconsistent, auxiliary parameters such as the quality of the star-flash link signal and the stability of the device's motion state are used for secondary verification. For example, if the signal quality is poor, the weight of distance anomalies is reduced, and the score is recalculated. If all are normal, it is determined as no fall into water. Finally, the determination result, including the tag, trigger time, and key features, is stored in the device's main database and synchronized to the main control system of the water equipment, completing the reporting of the fall-in-water detection result.

[0038] This step avoids the limitations of single-dimensional analysis by integrating anomaly information from both distance and motion dimensions: distance anomalies may be caused by signal interference, while motion anomalies may be caused by excessive user movements. Joint analysis, through cross-validation and auxiliary parameter correction, can effectively distinguish between real water-falling scenarios and interference scenarios. Overall, this improves the reliability of water-falling detection results and reduces equipment malfunctions or missed braking caused by misjudgments based on a single dimension.

[0039] Furthermore, sending the device motion signal and the user motion signal to the joint verification network includes: performing temporal synchronization and feature registration of the device motion signal and the user motion signal using the registration extraction layer of the joint verification network, constructing a temporally synchronized device attitude vector and a user attitude vector, and performing synchronization error analysis based on the device attitude vector and the user attitude vector to establish a fusion tensor; sending the fusion tensor to the distance change feature extraction layer, performing relative acceleration difference, instantaneous distance gradient and inertial drift index to establish a multidimensional anomaly feature vector; after initializing the dynamic confidence judgment layer using the motion state identifier, performing anomaly trigger analysis of the multidimensional anomaly feature vector to establish a second anomaly verification result.

[0040] Specifically, the registration extraction layer is the first functional layer of the joint verification network, responsible for time alignment and feature matching of device and user motion signals to ensure consistency between the two types of signals in both time and feature dimensions. The fusion tensor is a high-dimensional data structure formed by fusing the synchronized device attitude vector, user attitude vector, and synchronization error analysis results, centrally carrying the correlation information between the two types of signals. The distance change feature extraction layer is an intermediate layer of the joint verification network, extracting core feature indicators reflecting abnormal relative motion between the device and user from the fusion tensor. The dynamic confidence judgment layer is the output layer of the joint verification network; its judgment threshold can be dynamically adjusted according to the motion state identifier, used to analyze whether multi-dimensional abnormal feature vectors trigger anomalies.

[0041] Specifically, the joint validation network is deployed through an edge computing module. The joint validation network collects motion signals from devices and users in scenarios such as being stationary, driving, bumpy, and falling into water, constructing a sample dataset. After denoising and temporal synchronization preprocessing, attitude vectors and relative motion features are extracted. The network adopts a three-layer architecture: the registration and extraction layer uses a cross-correlation algorithm for temporal synchronization and PCA for feature registration; the feature extraction layer calculates indicators such as relative acceleration difference; and the dynamic confidence judgment layer uses a lightweight SVM model. The training / test set is divided into an 8:2 ratio, and the parameters are optimized using grid search. After multi-fold cross-validation tuning, the final model is obtained and adapted for edge NPU deployment. The registration extraction layer is invoked, and a cross-correlation algorithm is used to synchronize the device motion signal and the user motion signal in time. The maximum correlation coefficient between the two timestamps is calculated using a sliding window to align the signals to the same time axis. Principal component analysis is used for feature registration to filter out common features of device and user motion, constructing device and user attitude vectors. Synchronization error analysis is performed by calculating the mean residual of the two timestamps. Finally, the device attitude vector, user attitude vector, and error analysis results are integrated into a fusion tensor and stored in the network cache. The fusion tensor is passed to the distance change feature extraction layer, where the relative acceleration difference is calculated using a difference algorithm, the instantaneous distance gradient is calculated using the first derivative, and the inertial drift index is calculated using integration. These three are integrated to obtain a multidimensional anomaly feature vector. Motion state identifiers are read to initialize the dynamic confidence judgment layer: the judgment thresholds for each feature are adjusted based on the identifiers, and the multidimensional anomaly feature vector is classified. If a feature exceeds the threshold, it is judged as an anomaly, forming a second anomaly verification result.

[0042] This step, through the registration extraction layer, controls the time deviation between the device and user motion signals to an extremely low level, significantly improving feature correlation and effectively solving the fusion error problem caused by signal asynchrony. The distance change feature extraction layer extracts multi-dimensional features that comprehensively cover various abnormal scenarios, avoiding the limitations of single feature analysis. The dynamic confidence judgment layer dynamically adjusts the threshold based on the motion state, significantly enhancing the sensitivity of anomaly response in high-speed scenarios and effectively improving the anti-interference capability in bumpy scenarios. Overall, the accuracy of the second anomaly verification result is significantly improved, and compared with the feature analysis scheme with a fixed threshold, the number of misjudgments is greatly reduced, providing a high-precision motion dimension basis for joint anomaly analysis.

[0043] Furthermore, the dynamic confidence determination layer initialization using the motion state identifier includes: performing attitude drift trend analysis based on the motion state identifier and establishing a first correction factor; reading the signal strength of the star link of the communication module and establishing a second correction factor based on the signal strength; extracting wave feature parameters from the device motion signal, calculating a wave interference index based on the wave feature parameters, and calculating a third correction factor using the wave interference index; and performing dynamic threshold initialization of the dynamic confidence determination layer based on the first correction factor, the second correction factor, and the third correction factor.

[0044] Specifically, attitude drift trend analysis refers to the process of analyzing the attitude deviation over time based on motion state identifiers to assess attitude stability. The first correction factor quantifies the impact of attitude drift trend on the judgment threshold; the more significant the drift, the larger the factor value, used to relax the corresponding threshold to avoid misjudgment. The second correction factor is a parameter generated based on signal strength; the weaker the signal, the larger the factor value, used to adjust the threshold to compensate for the impact of signal quality on data reliability. Wave characteristic parameters are equipment motion parameters used to describe the degree of wave interference, including vertical acceleration amplitude, roll / pitch angle change rate, and oscillation period. The wave interference index is a normalized exponent calculated using a weighted summation formula; the larger the value, the stronger the wave interference on the equipment motion signal. The third correction factor is a parameter generated based on the wave interference index; the stronger the interference, the larger the factor value, used to adjust the threshold to resist motion signal fluctuations caused by waves.

[0045] Specifically, the motion status identifier is read, the device motion signal is analyzed for attitude drift trend through the edge computing module, the slope of the device attitude angle change is statistically analyzed using a sliding window, and the first correction factor is generated based on the absolute value of the slope.

[0046] This method is based on the principle of inertial navigation, and calculates the slope through differential calculation. The sliding window length is set to T = 1 second, the sampling frequency to fs = 50Hz, and the attitude angle sequence within the window is... , ,..., (N=T×fs=50), the slope k is calculated using linear regression: in For timestamps. First correction factor. Generation rules: =min(1.0, |k| / ).

[0047] To preset the maximum slope, the example slope of 0.5° / s corresponds to... =0.3.

[0048] Read the signal strength of the StarSignal communication module, map the RSSI value to the 0-1 range, and convert it into a second correction factor. The lower the signal strength, the larger the factor.

[0049] Where the RSSI value range is [−100, −50] dBm, the mapping function is: For example, RSSI = -75dBm, we get =(−50−(−75)) / 50=0.5.

[0050] Wave characteristic parameters are extracted from the equipment motion signal. The vertical acceleration signal is separated by low-pass filtering and its amplitude is calculated. A peak detection algorithm is used to obtain the peak value of the roll / pitch angle change rate. A period analysis algorithm, such as an autocorrelation function, is used to determine the oscillation period. Then, wave intensity, oscillation severity, and period deviation rate are calculated using formulas, and a weighted sum is obtained to obtain the wave interference index. This index is directly used as the third correction factor; the higher the index, the larger the factor. Wave intensity is quantified by the variance of the vertical acceleration signal, oscillation severity is characterized by the peak value of the roll / pitch angle change rate, and the period deviation rate is defined by the relative deviation between the actual wave period and the reference period. The vertical acceleration amplitude variance A... V The vertical acceleration signal is separated by low-pass filtering (cutoff frequency 0.5Hz), and the variance within the window is calculated. The peak value of the roll angle change rate R... p Find the maximum absolute value after differentiating the roll angles. Period deviation rate P d Calculate the absolute deviation ratio between the actual oscillation period and the reference period. Wave Disturbance Index Calculation formula: Weight =0.4, =0.3, =0.3, after normalization Range [0, 1]. Third correction factor C3 = .

[0051] Finally, a weighted fusion algorithm is used to calculate the comprehensive correction coefficient by assigning weights of 0.3, 0.2, and 0.5 to the three factors respectively. This coefficient is then used to adjust the thresholds of features such as relative acceleration difference and instantaneous distance gradient in the dynamic confidence decision layer. The larger the comprehensive coefficient, the more lenient the threshold, thus completing the dynamic threshold initialization.

[0052] This step achieves scenario-based adaptation of the dynamic confidence judgment layer threshold through correction factors in three dimensions: attitude drift, signal strength, and wave interference. When attitude drift is significant, the threshold is appropriately widened to avoid misjudgments caused by device offset. When signal quality is poor, threshold adjustment compensates for insufficient data reliability. When wave interference is strong, the threshold is flexibly adapted to resist irregular fluctuations in motion signals. Overall, dynamic thresholds can better match the actual characteristics of different scenarios, reduce misjudgments and missed judgments caused by environmental interference or changes in device status, and improve the adaptability and accuracy of anomaly detection.

[0053] Furthermore, performing distance anomaly verification on the relative distance dataset includes: segmenting the relative distance dataset into distance sequences using the threshold verification channel, calculating the volatility, mutation rate, and continuous drift coefficient between adjacent distance samples, and establishing a local distance perturbation feature vector; configuring the dynamic threshold of the threshold verification channel based on the motion state identifier and the signal quality of the star link, and then performing a local distance perturbation feature vector comparison based on the dynamic threshold to complete the distance anomaly verification.

[0054] Specifically, the relative distance dataset is a dataset structured as "timestamp-distance value," recording the temporal changes in the relative distance between devices and users. Distance sequence segmentation refers to dividing continuous distance data into several subsequences according to time windows, facilitating local feature analysis. Volatility refers to the ratio of the magnitude of change between adjacent distance samples to the previous distance value, reflecting the relative drasticness of distance changes. Abrupt change rate refers to the ratio of the absolute difference between adjacent distance samples to a preset benchmark value, used to identify sudden jumps in distance. Continuous drift coefficient refers to the ratio of the cumulative increase or decrease in distance within a certain distance sequence to the sequence length, reflecting the continuous trend of distance change. The local distance perturbation feature vector is a multi-dimensional vector composed of the volatility, abrupt change rate, and continuous drift coefficient of a certain distance sequence, centrally representing the distance change characteristics of that local interval.

[0055] Specifically, the threshold verification channel is invoked, and the edge computing module performs distance sequence segmentation on the relative distance dataset, employing a sliding window algorithm. For example, the window length is set to 1 second with a step size of 0.5 seconds, dividing the continuous "timestamp-distance value" data into several local subsequences. For each subsequence, features are calculated using statistical analysis algorithms. The volatility is calculated using the ratio of the difference between adjacent distances to the previous value, the mutation rate is calculated using the ratio of the absolute difference between adjacent distances to the baseline value, and the continuous drift coefficient is calculated using the ratio of the cumulative distance change in the subsequence to the window duration. These three factors are integrated into a local distance perturbation feature vector. Subsequently, the motion status identifier and the star link signal quality are read, and the dynamic threshold of the threshold verification channel is configured according to preset mapping rules, such as tightening the mutation rate threshold at high speeds and relaxing the volatility threshold when the signal is weak. The values ​​of each dimension of the local distance perturbation feature vector are compared with the corresponding dynamic thresholds. If any dimension value exceeds the threshold, the distance change corresponding to that subsequence is determined to be abnormal, completing the distance anomaly verification.

[0056] Regarding the threshold verification channel, the typical motion states of water equipment are first analyzed, including stationary, low-speed, high-speed, and turbulent conditions. Relative distance datasets during normal operation in each state are collected. Time-series analysis is performed on the data, and statistical methods, such as mean ± 3σ, are used to determine the normal distance range for each state. The maximum instantaneous fluctuation value is calculated to set the fluctuation tolerance, and the anomaly judgment delay is set in combination with response requirements. The threshold dynamic correction rules are established by associating the starburst signal quality parameters. Through verification with measured data, the parameters are iteratively optimized to form a threshold verification channel adapted to each state.

[0057] This step, through distance sequence segmentation and multi-dimensional feature extraction, can accurately capture the distance change characteristics of local intervals, avoiding the limitations of single distance value analysis. By combining motion state and signal quality to configure dynamic thresholds, the judgment criteria can adapt to the distance change patterns of different scenarios, reducing misjudgments caused by environmental interference or differences in equipment status. Overall, it improves the pertinence and reliability of distance anomaly verification, providing a solid basis for the first anomaly verification result.

[0058] Furthermore, reporting the water-fall detection results also includes: executing a three-level braking response for the water-based equipment, wherein the three-level braking response includes a power output reduction response within a first preset distance, an engine power-off response within a second preset distance, and a rescue coordinate transmission response within a third preset distance; and handling the water-fall anomaly based on the three-level braking response.

[0059] Specifically, the three-level braking response is a safety protection system that executes actions in stages based on the real-time distance between the equipment and the person in the water. It balances braking safety and rescue timeliness through a gradient response. The first preset distance refers to the distance threshold that triggers a power output reduction response, at which point the priority is to slow the equipment's movement away. The power output reduction response refers to controlling the power system of the watercraft to reduce its output power, thus slowing the equipment's speed. The second preset distance is the distance threshold that triggers an engine power cut-off response, at which point the equipment's power must be completely stopped to prevent further movement away. The third preset distance is the distance threshold that triggers a rescue coordinate transmission response, at which point precise location information must be transmitted to the outside world to accelerate the rescue.

[0060] Specifically, after reporting the water-fall detection result, the system reads the relative distance between the device and the user transmitted in real time by the StarScan communication module and compares it with the preset three-level distance thresholds. If the distance is within the first preset distance, the main control system sends a command to the power control system via the CAN bus to control the motor driver to reduce the output power, executing a power output reduction response. If the distance reaches the second preset distance, the main control system triggers a relay to cut off the engine power supply circuit, executing an engine power-off response, and simultaneously activates an audible and visual alarm via the vehicle buzzer and LED lights. If the distance exceeds the third preset distance, the main control system calls the GPS module to obtain latitude and longitude coordinates and sends the coordinates and personnel water-fall alarm information to the preset rescue platform and the user's associated mobile phone via the communication module, executing a rescue coordinate transmission response. Simultaneously, the water-fall anomaly handling module is activated in conjunction with the system. If the device is equipped with an automatic inflatable life raft, the gas generator is triggered to inflate it; the event log is recorded to local storage, completing the entire emergency response process.

[0061] This step, through a three-level braking response and gradient actions, achieves progressive protection of deceleration, stopping, and calling for help, avoiding potential equipment loss of control or secondary dangers caused by a single emergency braking. Simultaneously, the response intensity is dynamically adjusted according to distance, balancing equipment safety and rescue efficiency. The rescue coordinate transmission response ensures accurate transmission of the water-fall location, providing crucial information for rapid rescue. The coordinated water-fall anomaly handling further enhances the comprehensiveness of emergency response, reducing the overall risk of injury or death in water-fall accidents and strengthening the safety protection capabilities of water sports equipment.

[0062] Furthermore, the three-level braking response of the watercraft includes: when the water-fall detection result meets the third preset distance, real-time coordinate reporting is performed, wherein the real-time coordinate reporting is the watercraft coordinate reporting at the current time node; the trajectory backtracking of the watercraft's water-fall time node and the current time node is performed, and a guidance marker is established using the trajectory backtracking result; the guidance marker is synchronously sent as auxiliary information of the real-time coordinates.

[0063] Specifically, guidance markers refer to auxiliary rescue information generated based on trajectory backtracking results, including the direction and distance of the point of entry into the water relative to the current location, and the movement trend of the equipment. This auxiliary information supplements the guidance markers as real-time coordinates, helping rescuers to more accurately locate the point of entry into the water and the trajectory of the equipment.

[0064] Specifically, when the water-fall detection result meets the third preset distance, the main control system of the water-based equipment calls the positioning module, such as the Beidou dual-mode positioning module, to obtain the equipment's latitude and longitude coordinates at the current time point. The communication module then executes real-time coordinate reporting, sending the coordinate data to the preset rescue platform. Historical positioning records of the water-fall time point and the coordinates of the current time point are retrieved from the local database. For example, trajectory backtracking is performed based on time-series coordinate point line fitting, generating the equipment's movement trajectory from the moment of waterfall to the present, including key turning points and velocity change segments. Based on the trajectory backtracking results, the azimuth angle, straight-line distance, and trajectory curvature of the water-fall point relative to the current position are calculated and integrated into a guidance marker. For example, if the water-fall point is located 30° northwest of the current position, with a straight-line distance of 80 meters, the trajectory is a straight line. This guidance marker, as auxiliary information, is synchronously sent to the rescue platform and associated terminals via the same communication link along with the real-time coordinates.

[0065] This step ensures that rescuers can obtain the current location of the equipment by reporting real-time coordinates, providing basic positioning for the rescue; trajectory backtracking and guidance markers supplement the relative relationship and movement path between the point of fall into the water and the equipment, helping rescuers quickly deduce the location of the point of fall into the water and avoid expanding the search area due to relying solely on the current coordinates; the synchronous transmission of auxiliary information and real-time coordinates significantly improves the accuracy of rescue positioning, shortens the rescue response time, and increases the probability of rescuing the person who fell into the water.

[0066] Furthermore, the guidance identifier is established using the trajectory backtracking results, including: establishing a device attitude vector and a set of relative distance change rates based on the trajectory backtracking results; using the device attitude vector and the set of relative distance change rates to perform inversion fitting of the user's water entry angle, water entry speed, and drift direction to establish a set of dynamic feature parameters for falling into the water; and using the set of dynamic feature parameters for falling into the water to establish the guidance identifier.

[0067] Specifically, the device attitude vector, extracted from trajectory backtracking results, is a multi-dimensional vector describing the device's motion attitude at various time points, such as 3-axis acceleration, 3-axis angular velocity, and heading angle, quantifying the device's tilt, turning, and other states. The relative distance change rate set, extracted from the relative distance dataset, is a set of distance changes over time, reflecting the speed and trend of distance changes between the device and the user. The water entry dynamic feature parameter set consists of parameters such as the user's entry angle, entry velocity, and drift direction, comprehensively describing the water entry process and subsequent motion state.

[0068] Specifically, from the trajectory backtracking results, the edge computing module extracts the device's 3-axis acceleration, 3-axis angular velocity, and heading angle at key time points to construct the device's attitude vector. Simultaneously, distance values ​​at corresponding time points are extracted from the relative distance dataset, and the rate of change of distance between adjacent time points is calculated using a difference algorithm, forming a set of relative distance change rates. An inversion fitting algorithm, such as curve fitting based on the least squares method, is then used as a benchmark. Based on the device's attitude vector and the set of relative distance change rates, the spatial angle (entry angle) and velocity (entry velocity) at the moment of the user's fall into the water are calculated. Combined with water flow direction data derived from the device's onboard water flow sensors and the trend characteristics of the velocity set, the drift direction after the user falls into the water is inverted. These three factors are integrated into a set of dynamic feature parameters for the fall into the water. Finally, the data formatting module converts the parameter set into intuitive rescue guidance information, forming guidance signs.

[0069] This step transforms the equipment trajectory and distance change data into dynamic features of the user falling into the water through inversion fitting, upgrading the guidance signage from simple location information to multi-dimensional guidance that includes movement trends, which is more in line with the characteristics of users who may drift with the water flow in actual rescue scenarios. The introduction of the dynamic feature parameter set of falling into the water helps rescuers predict changes in the user's position, narrows the search range, improves the accuracy and practicality of rescue guidance, and provides more effective decision support for rapid rescue.

[0070] Furthermore, before performing joint anomaly analysis based on the first anomaly verification result and the second anomaly verification result, the process includes: activating the star-flash link waveform residual analyzer, performing a secondary fitting correction of the relative distance dataset, wherein the secondary fitting correction establishes a waveform residual vector by calculating the time offset of the ranging signal echo envelope, the signal attenuation rate, and the multipath interference ratio; using the waveform residual vector to correct the time synchronization deviation of the relative distance dataset, and correcting the first anomaly verification result based on the correction result.

[0071] Specifically, the star-flash link waveform residual analyzer is a module used to analyze the waveform deviation of star-flash ranging signals. It extracts characteristic parameters of the echo signal to evaluate and correct the accuracy of the ranging data. The multipath interference ratio (MIR) is the ratio of the reflected signal intensity to the direct signal intensity, reflecting the degree of interference from environmental reflections on the signal waveform. The waveform residual vector is a multi-dimensional vector composed of time offset, signal attenuation rate, and MIR, which centrally quantifies the waveform deviation characteristics of the ranging signal. Time synchronization deviation refers to the deviation between the centralized timestamp of the relative distance data and the actual ranging time, originating from signal transmission delay or clock asynchrony.

[0072] Specifically, before performing joint anomaly analysis, the edge computing module activates the StarSpark link waveform residual analyzer. This analyzer retrieves the original ranging signal waveform, including transmitted and echoed waves, from the signal buffer of the StarSpark communication module corresponding to the relative distance dataset. The echo envelope is extracted using Hilbert transform, and the time offset (the difference between the actual propagation time and the theoretical value) is calculated. A logarithmic attenuation model is used to fit the signal strength variation with distance to obtain the signal attenuation rate. Power spectrum analysis separates the direct and reflected signals, and the multipath interference ratio is calculated. These three parameters are integrated into a waveform residual vector, and then a quadratic curve is fitted using the least squares method based on this vector to correct the timestamps of each sample in the relative distance dataset and adjust the corresponding distance values. The distance anomaly verification is then re-executed using the corrected relative distance dataset to correct the original first anomaly verification result.

[0073] This step achieves precise correction of relative distance data through waveform residual analysis, effectively eliminating ranging errors caused by time synchronization deviation, signal attenuation, and multipath interference, thus significantly improving the accuracy of distance data. Based on the corrected data, the first anomaly verification result is corrected, reducing misjudgments caused by interference from the original signal, providing a more reliable distance dimension basis for subsequent joint anomaly analysis, and improving the overall anti-interference capability and judgment accuracy of the water fall detection system.

[0074] Furthermore, reporting the water-fall detection results also includes: performing a startup test on the backup communication module; if the backup communication module starts successfully, connecting the backup communication module and the communication module of the water equipment, and performing a detachment verification of the user communication module; updating the water-fall detection results based on the detachment verification feedback.

[0075] Specifically, the backup communication module refers to the main communication module, such as redundant communication devices like Bluetooth BLE or LoRa modules that are activated when the StarScan module fails, to ensure the continuity of communication between the device and the user.

[0076] Specifically, after reporting the water-fall detection results, the main control system of the water equipment immediately sends a start command to the backup communication module. It confirms normal power supply to the module through a voltage detection circuit and calls a signal self-test function to check if its transmission power and reception sensitivity meet the standards, completing the start-up test. If the test passes, the main control system establishes a connection between the backup communication module and the main communication module via the UART interface, synchronizing the equipment identification and communication protocol. The backup module sends encrypted verification signals to the user communication module at a preset frequency and monitors the response signal through the receiving window: if no response is received or the response signal strength drops sharply, it is determined to be a detachment; if a response containing the correct verification code is received and the signal is stable, it is determined not to be a detachment, and a detachment verification feedback is generated. The main control system updates the water-fall detection results based on the feedback.

[0077] This step ensures communication redundancy in extreme cases through the startup detection of the backup communication module, avoiding verification interruptions caused by main module failure; the drop verification effectively distinguishes between user water drop and module accidental drop scenarios, reducing false alarms caused by equipment failure; by providing feedback and updating results, the water drop detection results are made more consistent with the actual situation, improving the robustness and decision accuracy of the system.

[0078] In summary, the personnel falling into water detection method based on star-flash ranging provided in this application has the following technical effects: 1. By integrating relative distance data from star-flash ranging with device and user motion signals, multi-dimensional verification avoids misjudgment based on single data points. Combining motion state adaptation thresholds with joint analysis, it accurately identifies water-fall scenarios, solving the problems of environmental interference and poor scenario adaptation in traditional detection, and providing reliable detection basis for water safety.

[0079] 2. A three-stage braking response achieves progressive protection of "deceleration-stop-distress call," balancing equipment safety and rescue efficiency. Power regulation, power-off braking, and rescue coordinates are linked to avoid the risks of relying on isolated emergency operations. Combined with anomaly handling, this enhances emergency response capabilities and reduces the risk of injury or death in water accidents.

[0080] 3. A backup communication module ensures communication redundancy, preventing verification interruptions due to main module failure. The detachment verification distinguishes between water immersion and module detachment, reducing false alarms caused by equipment failure. Feedback updates improve system robustness and decision-making accuracy, ensuring test results accurately reflect real-world scenarios.

[0081] Example 2, based on the same inventive concept as the personnel falling into water detection method based on star-flash ranging in the foregoing examples, such as... Figure 2As shown in the figure, this application provides a personnel fall-in-water detection system based on star-flash ranging. The system includes: a communication module configuration module 11, used to configure communication modules on the water equipment and the user respectively, wherein the communication modules are paired communication modules and are connected based on star-flash technology; a status identifier establishment module 12, used to read the equipment motion signal of the water equipment, use the equipment motion signal to identify the motion status, and establish a motion status identifier; a distance dataset establishment module 13, used to configure the communication module using the motion status identifier, perform real-time distance measurement, and establish a relative distance dataset with time identifier; a first anomaly verification result establishment module 14, used to activate a threshold verification channel according to the motion status identifier, perform distance anomaly verification of the relative distance dataset, and establish a first anomaly verification result; a second anomaly verification result establishment module 15, used to establish a user motion signal through an accelerometer set at the user's location, send the equipment motion signal and the user motion signal to a joint verification network, and establish a second anomaly verification result; and a fall-in-water detection result reporting module 16, used to perform joint anomaly analysis based on the first anomaly verification result and the second anomaly verification result, and report the fall-in-water detection result.

[0082] Furthermore, the second anomaly verification result establishment module 15 is also used to perform the following steps: using the registration extraction layer of the joint verification network to perform temporal synchronization and feature registration of the device motion signal and the user motion signal, constructing the temporally synchronized device attitude vector and user attitude vector, and performing synchronization error analysis based on the device attitude vector and the user attitude vector to establish a fusion tensor; sending the fusion tensor to the distance change feature extraction layer, performing relative acceleration difference, instantaneous distance gradient and inertial drift index to establish a multidimensional anomaly feature vector; after initializing the dynamic confidence judgment layer using the motion state identifier, performing anomaly trigger analysis of the multidimensional anomaly feature vector to establish the second anomaly verification result.

[0083] Furthermore, the second anomaly verification result establishment module 15 is also used to perform the following steps: perform attitude drift trend analysis based on the motion state identifier and establish a first correction factor; read the signal strength of the star link of the communication module and establish a second correction factor based on the signal strength; extract wave feature parameters from the device motion signal, calculate the wave interference index based on the wave feature parameters, and calculate a third correction factor using the wave interference index; and perform dynamic threshold initialization of the dynamic confidence determination layer based on the first correction factor, the second correction factor, and the third correction factor.

[0084] Furthermore, the first anomaly verification result establishment module 14 is also used to perform the following steps: segment the relative distance dataset into distance sequences using the threshold verification channel, calculate the volatility, mutation rate and continuous drift coefficient between adjacent distance samples, and establish a local distance perturbation feature vector; configure the dynamic threshold of the threshold verification channel according to the motion state identifier and the signal quality of the star link, and then perform a local distance perturbation feature vector comparison according to the dynamic threshold to complete the distance anomaly verification.

[0085] Furthermore, the water-fall detection result reporting module 16 is also used to perform the following steps: execute a three-level braking response for the water equipment, the three-level braking response including a power output reduction response within a first preset distance, an engine power-off response within a second preset distance, and a rescue coordinate transmission response within a third preset distance; and perform water-fall anomaly processing based on the three-level braking response.

[0086] Furthermore, the water-fall detection result reporting module 16 is also used to perform the following steps: when the water-fall detection result meets the third preset distance, perform real-time coordinate reporting, wherein the real-time coordinate reporting is the coordinate reporting of the water equipment at the current time node; perform trajectory backtracking of the water equipment's water-fall time node and the current time node, and establish a guidance marker using the trajectory backtracking result; and synchronously send the guidance marker as auxiliary information of the real-time coordinates.

[0087] Furthermore, the water-fall detection result reporting module 16 is also used to perform the following steps: establish a device attitude vector and a set of relative distance change rates based on the trajectory backtracking results; use the device attitude vector and the set of relative distance change rates to perform inversion fitting of the user's water entry angle, water entry speed and drift direction to establish a set of water-fall dynamic feature parameters; and use the set of water-fall dynamic feature parameters to establish the guidance identifier.

[0088] Furthermore, the water-fall detection result reporting module 16 is also used to perform the following steps: activate the star-flash link waveform residual analyzer, perform secondary fitting correction of the relative distance dataset, wherein the secondary fitting correction establishes a waveform residual vector by calculating the time offset of the ranging signal echo envelope, the signal attenuation rate and the multipath interference ratio; use the waveform residual vector to correct the time synchronization deviation of the relative distance dataset, and correct the first anomaly verification result according to the correction result.

[0089] Furthermore, the water-fall detection result reporting module 16 is also used to perform the following steps: perform a startup detection of the backup communication module; if the backup communication module starts successfully, connect the backup communication module and the communication module of the water equipment, and perform a detachment verification of the user communication module; update the water-fall detection result based on the detachment verification feedback.

[0090] The above description is merely a preferred embodiment of this application and is not intended to limit this application in any way. Although this application has been disclosed above with reference to preferred embodiments, it is not intended to limit this application. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the technical solution of this application. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of this application without departing from the scope of the technical solution of this application shall still fall within the scope of the technical solution of this application.

Claims

1. A method for detecting personnel falling into water based on star-flash ranging, characterized in that, The method includes: Communication modules are configured on the water equipment and the user respectively, wherein the communication modules are paired communication modules and the communication modules are connected based on star flash technology; Read the motion signals of the water equipment, use the motion signals to identify the motion status, and establish a motion status identifier; After configuring the communication module using the motion status identifier, real-time distance measurement is performed to establish a relative distance dataset with time identifiers; Activate the threshold verification channel based on the motion state identifier, perform distance anomaly verification on the relative distance dataset, and establish the first anomaly verification result; The user motion signal is established by an accelerometer set at the user's location, and the device motion signal and the user motion signal are sent to the joint verification network to establish a second anomaly verification result; Based on the first anomaly verification result and the second anomaly verification result, a joint anomaly analysis is performed, and the water fall detection result is reported.

2. The method for detecting personnel falling into water based on star-flash ranging as described in claim 1, characterized in that, Sending the device motion signal and the user motion signal to the joint authentication network includes: The registration extraction layer of the joint verification network is used to perform temporal synchronization and feature registration of the device motion signal and the user motion signal, construct the temporally synchronized device attitude vector and user attitude vector, and perform synchronization error analysis based on the device attitude vector and the user attitude vector to establish a fusion tensor; The fused tensor is sent to the distance change feature extraction layer to perform relative acceleration difference, instantaneous distance gradient and inertial drift index to establish a multidimensional anomaly feature vector; After initializing the dynamic confidence judgment layer using the motion state identifier, perform anomaly trigger analysis of the multidimensional anomaly feature vector to establish a second anomaly verification result.

3. The method for detecting personnel falling into water based on star-flash ranging as described in claim 2, characterized in that, Initialization of the dynamic confidence determination layer using the motion state identifier includes: Based on the motion state identifier, perform attitude drift trend analysis and establish a first correction factor; Read the signal strength of the star flash link of the communication module, and establish a second correction factor based on the signal strength; Extract wave characteristic parameters from the equipment motion signal, calculate the wave interference index based on the wave characteristic parameters, and use the wave interference index to calculate the third correction factor. Dynamic threshold initialization of the dynamic confidence decision layer is performed based on the first correction factor, the second correction factor, and the third correction factor.

4. The method for detecting personnel falling into water based on star-flash ranging as described in claim 1, characterized in that, Perform distance anomaly validation on the relative distance dataset, including: The threshold verification channel is used to segment the relative distance dataset into distance sequences, and the volatility, mutation rate and continuous drift coefficient between adjacent distance samples are calculated to establish a local distance perturbation feature vector. After configuring the dynamic threshold of the threshold verification channel based on the motion state identifier and the signal quality of the star link, the local distance perturbation feature vector comparison is performed based on the dynamic threshold to complete the distance anomaly verification.

5. The method for detecting personnel falling into water based on star-flash ranging as described in claim 1, characterized in that, The report of the water-related test results also includes: The system executes a three-level braking response for the watercraft, which includes a power output reduction response within a first preset distance, an engine power cut-off response within a second preset distance, and a rescue coordinate transmission response within a third preset distance. The abnormal water-falling response is handled according to the three-level braking response.

6. The method for detecting personnel falling into water based on star-flash ranging as described in claim 5, characterized in that, Perform a three-level braking response for the watercraft, including: When the water-fall detection result meets the third preset distance, real-time coordinate reporting is executed, wherein the real-time coordinate reporting is the coordinate reporting of the water equipment at the current time node; Perform trajectory backtracking of the time point when the water equipment fell into the water and the current time point, and use the trajectory backtracking results to establish guidance markers; The guidance identifier is sent synchronously as auxiliary information for real-time coordinates.

7. The method for detecting personnel falling into water based on star-flash ranging as described in claim 6, characterized in that, Establish guidance markers using trajectory backtracking results, including: Based on the trajectory backtracking results, establish the device attitude vector and the set of relative distance change rates; Using the device attitude vector and the set of relative distance change rates, the user's water entry angle, water entry velocity, and drift direction are inverted and fitted to establish a set of dynamic feature parameters for falling into the water; The guidance identifier is established using the set of dynamic feature parameters of the falling water.

8. The method for detecting personnel falling into water based on star-flash ranging as described in claim 1, characterized in that, Before performing joint anomaly analysis based on the first anomaly verification result and the second anomaly verification result, the following is included: Activate the star link waveform residual analyzer and perform a second-order fitting correction of the relative distance dataset. The second-order fitting correction establishes the waveform residual vector by calculating the time offset of the ranging signal echo envelope, the signal attenuation rate, and the multipath interference ratio. The time synchronization deviation of the relative distance dataset is corrected using the waveform residual vector, and the first anomaly verification result is corrected based on the correction result.

9. The method for detecting personnel falling into water based on star-flash ranging as described in claim 1, characterized in that, The report of the water-related test results also includes: Perform startup detection on the backup communication module; If the backup communication module starts successfully, connect the backup communication module and the communication module of the water equipment, and perform a detachment verification of the user communication module; Update the water-fall detection results based on the feedback from the detachment verification.

10. A personnel fall-in-water detection system based on star-flash ranging, characterized in that, The system is used to perform the personnel fall-in-water detection method based on star-flash ranging as described in any one of claims 1 to 9, the system comprising: A communication module configuration module is used to configure communication modules for both the watercraft and the user, wherein the communication modules are paired communication modules and are connected based on star-flash technology; The status identification module is used to read the equipment motion signals of the water equipment, use the equipment motion signals to identify the motion status, and establish a motion status identification. The distance dataset establishment module is used to configure the communication module using the motion state identifier, perform real-time distance measurement, and establish a relative distance dataset with time identifier; The first anomaly verification result establishment module is used to activate the threshold verification channel according to the motion state identifier, perform distance anomaly verification of the relative distance dataset, and establish the first anomaly verification result; The second anomaly verification result establishment module is used to establish a user motion signal through an accelerometer set at the user's location, and send the device motion signal and the user motion signal to the joint verification network to establish a second anomaly verification result; The water fall detection result reporting module is used to perform joint anomaly analysis based on the first anomaly verification result and the second anomaly verification result, and report the water fall detection result.

Images