Swimming pool automatic alarm system based on electrocardiogram monitoring

Abstract

The invention discloses a swimming pool automatic alarm system based on electrocardio monitoring, and relates to the field of electrocardio monitoring, and the system comprises a collection module which is used for collecting electrocardio data and equipment position correlation signals output by equipment worn by a swimmer in real time; the fusion module is used for carrying out de-noising processing on the collected two types of data, extracting heart rate change features and position movement features and generating a fusion feature data set; according to the method, relevant data are synchronously collected in real time and accurately denoised, key features of heart rate change and position movement are fully extracted, the abnormal linkage rule of the heart rate and the position is quantitatively analyzed, and the drowning state of the swimmer is accurately judged.

Description

Technical Field

[0001] This invention relates to the field of electrocardiogram (ECG) monitoring technology, specifically to an automatic alarm system for swimming pools based on ECG monitoring. Background Technology

[0002] Swimming pools are places for entertainment and fitness, but they also pose risks such as a large number of people, difficulty in monitoring, and difficulty in quickly detecting drowning incidents.

[0003] Patent application number 202211365158.X discloses an automatic patrol and alarm method for swimming pool safety. This method aims to address the serious shortcomings of traditional drowning monitoring systems in swimming pools. Traditional systems limit the drowning area to the bottom layer of the pool, but drowning can occur in both the bottom and upper layers. Furthermore, the algorithm for detecting drowning involves light emitted by the light-emitting module being blocked by the same object for an extended period, and the movement range is very small. Since drowning initially involves struggle and activity, by the time a person loses consciousness and becomes still, they are already in a very dangerous situation, potentially causing irreversible damage. Additionally, swimming pool water is not perfectly still; when others are swimming, the drowning person will also float with the current. Therefore, traditional drowning monitoring systems have low monitoring effectiveness and low safety levels in swimming pools.

[0004] However, relying solely on machine vision monitoring is still insufficient for a comprehensive assessment of swimming pool drowning issues.

[0005] Therefore, we propose an automatic alarm system for swimming pools based on electrocardiogram monitoring. Summary of the Invention

[0006] In view of the above-mentioned shortcomings of the existing technology, the present invention provides an automatic alarm system for swimming pools based on electrocardiogram monitoring, which can effectively solve the problems of the existing technology.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions;

[0008] This invention discloses an automatic alarm system for swimming pools based on electrocardiogram monitoring, comprising:

[0009] The acquisition module is used to acquire ECG data and device location-related signals from the device worn by the swimmer in real time; the fusion module is used to denoise the two types of data, extract heart rate change features and position movement features, and generate a fused feature dataset; the judgment module is used to receive the fused feature dataset, analyze the abnormal linkage between heart rate and position, and determine whether the swimmer is drowning.

[0010] The integrated module is used to trigger operation after a drowning situation is determined. It parses the real-time location association signal of the device to obtain the real-time coordinates of the drowning person, and simultaneously generates an alarm command containing coordinate information and transmits it to the interaction module. The analysis module is used to receive the real-time location association signals of all other swimmers in the pool who are not determined to be drowning, parse them again to obtain the real-time coordinates of all other swimmers in the pool who are not determined to be drowning, analyze the optimal entry position for pool rescue based on the real-time coordinates, and transmit it to the interaction module. The interaction module is used to receive the alarm command containing coordinates and the optimal entry position, and then transmit them to the preset control panel, displaying the panel's own position, the optimal entry position, and the coordinate information contained in the alarm command on the panel.

[0011] The acquisition module is interconnected with a fusion module via a wireless network. The fusion module is interconnected with a judgment module via a wireless network. The judgment module is interconnected with an integration module via a wireless network. The integration module is interconnected with an analysis module via a wireless network. The analysis module is interconnected with an interaction module via a wireless network.

[0012] Among them, the optimal entry point for pool rescue is any point on the edge of the pool, and the analysis module and the integration module operate synchronously in conjunction with each other, as does the interaction module.

[0013] Furthermore, the acquisition module synchronously acquires ECG data and device location-related signals based on timestamps:

[0014] ;

[0015] In the formula: For synchronization accuracy; This is the timestamp for the ECG data collection. This is the timestamp for the acquisition of location-related signals; This is the preset sampling period for the acquisition module; For synchronous calibration coefficients;

[0016] when When the preset synchronization threshold is not exceeded, synchronization is considered valid, and the corresponding ECG data and location-related signals are transmitted to the fusion module; when... If the preset synchronization threshold is exceeded, synchronization is deemed invalid, and the acquisition module immediately triggers a re-acquisition process while simultaneously initiating the synchronization calibration coefficient. The dynamic adjustment range does not exceed ±0.1 each time. If the synchronization is still determined to be invalid after 3 consecutive re-collections, a synchronization error prompt is sent to the integration module, which triggers a temporary alarm on the device and restarts the acquisition module corresponding to the invalid synchronization in the system.

[0017] Furthermore, the data denoising in the fusion module follows the following rules:

[0018] ;

[0019] In the formula: The target data after denoising; The raw data collected; These are adaptive weighting coefficients; These are the base values ​​of the original data after frequency domain low-pass filtering; The actual instantaneous frequency of the original data; This represents the median of the standard effective frequency range for the corresponding data.

[0020] Furthermore, the adaptive weighting coefficients The calculation logic is as follows:

[0021] ;

[0022] In the formula: This is the attenuation adjustment coefficient; This represents the variance of the original data within the preset statistical window. The preset variance of the standard valid data; This represents the proportion of frequency domain noise.

[0023] Furthermore, the heart rate change features extracted by the fusion module include the heart rate change rate, heart rate mutation amplitude, and heart rate abnormality duration within three consecutive sampling periods. The heart rate change rate is calculated by the ratio of the heart rate difference between adjacent sampling points to the sampling period. The heart rate mutation amplitude is the absolute value of the difference between the real-time heart rate and the average heart rate of a recent preset number of sampling periods. The heart rate abnormality duration is the product of the number of consecutive samplings in which the real-time heart rate exceeds the preset normal heart rate range and the sampling period.

[0024] The positional movement features extracted by the fusion module include real-time movement speed, movement acceleration, and trajectory continuity coefficient. The real-time movement speed is calculated by the ratio of the difference in position coordinates between adjacent sampling points to the sampling period. The movement acceleration is calculated by the ratio of the difference in speed between two adjacent sampling periods to the sampling period. The trajectory continuity coefficient is characterized by the cosine of the angle between the current sampling point and the previous two sampling points.

[0025] Furthermore, the determination module quantifies the abnormal linkage between heart rate and location using the following formula:

[0026] ;

[0027] In the formula: This is a quantified value for linkage anomalies; These are the feature weight coefficients; Real-time heart rate; This is the swimmer's preset baseline heart rate; Real-time movement speed; This is the preset minimum normal movement speed; The duration of stay at the current location; The preset judgment duration;

[0028] When A is not less than the preset abnormal threshold, the swimmer is determined to be in a state of drowning.

[0029] Furthermore, the alarm command generated by the integrated module includes a drowning hazard level indicator, the real-time coordinates of the drowning victim, and a data reliability coefficient. The hazard level indicator is based on the linkage anomaly quantification value. Sure:

[0030] when The threshold is not less than the preset abnormal threshold; if it is less than the preset first-level threshold, it is a warning level.

[0031] when The threshold value is not less than the first-level threshold, and is classified as emergency if it is less than the preset second-level threshold.

[0032] when It is classified as Special Grade if it is not less than the Level 2 threshold;

[0033] The data credibility coefficient , For efficient synchronization, This represents the signal-to-noise ratio of the denoised signal. This represents the feature extraction completeness coefficient, with values ​​ranging from 0 to 1.

[0034] when If the risk level is not less than the preset high confidence threshold, trigger the corresponding alarm according to the current risk level;

[0035] when When the data falls between a preset low confidence threshold and a preset high confidence threshold, the integrated module controls the acquisition module to increase the sampling frequency and re-acquire data, performing secondary calculations based on the new data. and The alert level will then be determined. If the calculation is repeated at least three times, the alert level will be raised by one level.

[0036] when If the level does not exceed the preset low confidence threshold, the warning level will be upgraded by one level.

[0037] Furthermore, the analysis module analyzes the optimal entry point stage, picks up candidate points at preset intervals along the edge of the pool, and then calculates the fit of each candidate point based on the following formula.

[0038] ;

[0039] In the formula: For the overall fit of candidate points; These are the weighting coefficients; The straight-line distance from the candidate point to the water area where the drowning victim was located; This represents the resistance coefficient for the corresponding water area. The population density coefficient around the candidate site; The landside accessibility coefficient for candidate points;

[0040] The candidate point with the highest fit is selected as the optimal entry point.

[0041] Furthermore, during the operation of the interactive module, in the warning level state, only the coordinates of the drowning person and the warning sign are displayed; in the emergency level state, the best entry point into the water is additionally displayed; in the special level state, the real-time location of the rescuers is overlaid and the rescuers' rescue path is recorded.

[0042] After receiving the rescue response status, the interaction module feeds back to the integration module in real time. The integration module adjusts the sending frequency of alarm commands according to the status of not yet departed, in rescue, and arrived. The sending frequency of alarm commands corresponding to the status of not yet departed, in rescue, and arrived decreases in a gradient.

[0043] The determination of the status of not departing, in rescue, and arrived is based on the real-time location of the rescuers. The rescuers are equipped with location sensors, and the location sensors are connected to the control panel via wireless network or Bluetooth. When the distance between the real-time location of the rescuers and the location of the control panel is less than a preset value X, it is determined that the rescue has not been triggered. When the distance between the real-time location of the rescuers and the location of the control panel is not less than the preset value X, it is determined that the rescue is in progress. When the distance between the real-time location of the rescuers and the current location of the drowning victim does not exceed the preset value Y, it is determined that the rescue has arrived. During the status determination phase, the application priority of the preset value Y is higher than the application priority of the preset value X.

[0044] Compared with the known prior art, the technical solution provided by this invention has the following beneficial effects:

[0045] This invention provides an automatic alarm system for swimming pools based on electrocardiogram (ECG) monitoring. During operation, the system collects relevant data in real time and performs precise noise reduction to fully extract key features of heart rate changes and positional movement. It quantifies and analyzes abnormal patterns in the linkage between heart rate and position to accurately determine the drowning state, effectively reducing the risk of misjudgment and missed judgment. Based on the drowning victim's real-time coordinates and the actual situation of the pool, it determines the optimal rescue entry position. It dynamically adjusts the alarm level based on data reliability, clearly displays key rescue information, and adjusts the alarm sending frequency according to the rescue response status gradient, providing clear guidance for rescuers. This significantly shortens the rescue response time, effectively improves rescue efficiency and success rate, ensures the personal safety of swimmers, and is adaptable to swimming pool scenarios of different sizes and population densities. Attached Figure Description

[0046] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0047] Figure 1 This is a schematic diagram of an automatic alarm system for swimming pools based on electrocardiogram monitoring. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0049] The present invention will be further described below with reference to embodiments.

[0050] Example:

[0051] This embodiment provides an automatic alarm system for swimming pools based on electrocardiogram monitoring, such as... Figure 1 As shown, it includes:

[0052] The data acquisition module is used to acquire electrocardiogram data and device location-related signals from the device worn by swimmers in real time.

[0053] The acquisition module synchronously acquires ECG data and device location-related signals based on timestamps:

[0054] ;

[0055] In the formula: For synchronization accuracy; This is the timestamp for the ECG data collection. This is the timestamp for the acquisition of location-related signals; This is the preset sampling period for the acquisition module; For synchronous calibration coefficients;

[0056] The above formula combines the differences in the timestamps of ECG data and location-related signals, introduces a preset sampling period and a dynamic synchronization calibration coefficient, and determines the synchronization effectiveness by quantifying the correlation between the time difference and the sampling period and calibration coefficient. The time consistency of the two types of core data ensures the reliability of subsequent data fusion.

[0057] when When the preset synchronization threshold is not exceeded, synchronization is considered valid, and the corresponding ECG data and location-related signals are transmitted to the fusion module; when... If the preset synchronization threshold is exceeded, synchronization is deemed invalid, and the acquisition module immediately triggers a re-acquisition process while simultaneously initiating the synchronization calibration coefficient. The dynamic adjustment range does not exceed ±0.1 each time. If the synchronization is still determined to be invalid after 3 consecutive re-collection, a synchronization abnormality prompt is sent to the integration module, which triggers a temporary alarm on the device and restarts the acquisition module corresponding to the synchronization invalidity in the system.

[0058] in, The initial preset range of values ​​is [0.8, 1.2]. The larger the hardware delay error of the acquisition module, the larger the value; the smaller the hardware delay error, the smaller the value.

[0059] The fusion module is used to denoise the two types of data collected, extract heart rate change features and position movement features, and generate a fused feature dataset.

[0060] Data denoising in the fusion module follows these rules:

[0061] ;

[0062] In the formula: The target data after denoising; The raw data collected; These are adaptive weighting coefficients; These are the base values ​​of the original data after frequency domain low-pass filtering; The actual instantaneous frequency of the original data; This represents the median of the standard effective frequency range for the corresponding data.

[0063] The above formula is based on the original data. It obtains the basic value through frequency domain low-pass filtering, introduces adaptive weight coefficients and the ratio of the actual instantaneous frequency to the median of the standard effective frequency range, and constructs a denoising model. This allows the denoising process to dynamically adjust the weights according to the data noise characteristics, thereby achieving the goal of both preserving the effective features in the original data and specifically filtering out interference signals, thus realizing accurate denoising of ECG data and location-related signals.

[0064] Adaptive weight coefficients The calculation logic is as follows:

[0065] ;

[0066] In the formula: This is the attenuation adjustment coefficient; This represents the variance of the original data within the preset statistical window. The preset variance of the standard valid data; The proportion of frequency domain noise;

[0067] The above formula limits the values ​​of the adaptive weight coefficients by specific operational logic, for The calculations provide support;

[0068] in, The value is set to 1.0~2.0. The higher the sensitivity of the data to noise, the larger the value; the lower the sensitivity, the smaller the value. Furthermore, because ECG data is more sensitive to noise than location-related signals, the corresponding... The value ranges from 1.5 to 2.0, corresponding to the location-related signal. The value ranges from 1.0 to 1.5, and The value of is constrained to be between 0.1 and 0.9. If the above formula yields... If the value is not between 0.1 and 0.9, then the nearest value is taken as the mean. Practical application value;

[0069] The heart rate change features extracted by the fusion module include the heart rate change rate, heart rate mutation amplitude, and heart rate abnormality duration within three consecutive sampling periods. The heart rate change rate is calculated by the ratio of the heart rate difference between adjacent sampling points to the sampling period. The heart rate mutation amplitude is the absolute value of the difference between the real-time heart rate and the average heart rate of the recent preset number of sampling periods. The heart rate abnormality duration is the product of the number of consecutive samplings in which the real-time heart rate exceeds the preset normal heart rate range and the sampling period.

[0070] The positional movement features extracted by the fusion module include real-time movement speed, movement acceleration, and trajectory continuity coefficient. Real-time movement speed is calculated by the ratio of the difference in position coordinates between adjacent sampling points to the sampling period. Movement acceleration is calculated by the ratio of the difference in speed between two adjacent sampling periods to the sampling period. The trajectory continuity coefficient is characterized by the cosine of the angle between the current sampling point and the previous two sampling points.

[0071] The determination module is used to receive the fused feature dataset, analyze the abnormal patterns of the linkage between heart rate and location, and determine whether the swimmer is drowning.

[0072] The judgment module quantifies abnormalities in the linkage between heart rate and location using the following formula:

[0073] ;

[0074] In the formula: This is a quantified value for linkage anomalies; These are the feature weight coefficients; Real-time heart rate; This is the swimmer's preset baseline heart rate; Real-time movement speed; This is the preset minimum normal movement speed; The duration of stay at the current location; The preset judgment duration;

[0075] The above formula constructs a comprehensive quantitative model by combining three core dimensions—the proportion of heart rate deviation, the degree of abnormal movement speed, and the proportion of time spent in a position—with feature weight coefficients that are dynamically adjusted according to the correlation of the swimming scenario. This model can comprehensively capture abnormal linkages between heart rate and position.

[0076] When A is not less than the preset abnormal threshold, the swimmer is determined to be in a state of drowning.

[0077] in, All are positive numbers, and their sum is 1. The stronger the correlation between abnormal heart rate and drowning status in a swimming scenario, the more likely it is to indicate a drowning state. The higher the value, the more likely the abnormal heart rate is to be caused by non-drowning factors. The smaller the value, the stronger the correlation between abnormal movement speed and drowning in a swimming scenario. The larger the value, the more likely the abnormal speed is to be caused by non-drowning factors. The smaller the value, the stronger the correlation between abnormal dwell time in a swimming scenario and drowning. The larger the value, the more likely the abnormal duration of stay is caused by non-drowning factors. The smaller the value;

[0078] The integrated module is used to trigger operation after a drowning state is determined, analyze the real-time location association signal of the device to obtain the real-time coordinates of the drowning person, and synchronously generate an alarm command containing coordinate information and transmit it to the interaction module.

[0079] The alarm command generated by the integrated module includes a drowning hazard level indicator, the real-time coordinates of the drowning victim, and a data reliability coefficient. The hazard level indicator is based on the linkage anomaly quantification value. Sure:

[0080] when The threshold is not less than the preset abnormal threshold; if it is less than the preset first-level threshold, it is a warning level.

[0081] when The threshold value is not less than the first-level threshold, and is classified as emergency if it is less than the preset second-level threshold.

[0082] when It is classified as Special Grade if it is not less than the Level 2 threshold;

[0083] Data credibility coefficient , For efficient synchronization, This represents the signal-to-noise ratio of the denoised signal. This represents the feature extraction completeness coefficient, with values ​​ranging from 0 to 1.

[0084] when If the risk level is not less than the preset high confidence threshold, trigger the corresponding alarm according to the current risk level;

[0085] when When the data falls between a preset low confidence threshold and a preset high confidence threshold, the integrated module controls the acquisition module to increase the sampling frequency and re-acquire data, performing secondary calculations based on the new data. and The alert level will then be determined. If the calculation is repeated at least three times, the alert level will be raised by one level.

[0086] when If the warning level does not exceed the preset low confidence threshold, the warning level will be upgraded by one level.

[0087] in, The ratio of the number of synchronously valid samples within the preset statistical period to the total number of samples is taken. The ratio of the total number of heart rate change features and location movement features that were actually successfully extracted to the total number of features that should be extracted;

[0088] The analysis module is used to receive and re-analyze the real-time location association signals of all other swimmers in the pool who are not determined to be drowning, in order to obtain the real-time coordinates of all other swimmers in the pool who are not determined to be drowning, analyze the best entry position for pool rescue based on the real-time coordinates, and transmit it to the interaction module.

[0089] The analysis module analyzes the optimal entry point stage and picks up candidate points at preset intervals along the edge of the pool. The preset interval size is set according to the following: the larger the pool circumference, the larger the preset interval, and vice versa. The more swimmers the pool currently accommodates, the smaller the preset interval, and vice versa. The fit of each candidate point is then calculated based on the following formula.

[0090] ;

[0091] In the formula: For the overall fit of candidate points; These are the weighting coefficients; The straight-line distance from the candidate point to the water area where the drowning victim was located; This represents the resistance coefficient for the corresponding water area. The population density coefficient around the candidate site; The landside accessibility coefficient for candidate points;

[0092] The above formula starts from three key rescue influencing factors: waterway passage cost, personnel avoidance cost, and landside access convenience. It introduces corresponding weight coefficients and incorporates indicators such as the distance from the candidate point to the drowning person, water resistance, surrounding personnel density, and landside passage smoothness into the calculation to accurately select the optimal rescue entry point into the water, thereby improving the timeliness and safety of the rescue.

[0093] The candidate point with the highest fit is selected as the optimal entry point.

[0094] in, All three are positive numbers, and their sum is 1. The shorter the straight-line distance from the candidate point to the drowning victim's body of water, the smaller the corresponding water body's resistance coefficient. The larger the value, the lower the value. The smaller the value, the lower the density coefficient of non-drowning personnel around the candidate point, and the lower the cost of personnel avoidance during rescue. The larger the value, the lower the value. The smaller the value, the smoother the landside passage path from the candidate point to the pool passage and the higher the convenience coefficient. The larger the value, the lower the value. The smaller the value;

[0095] ∈ (0,2), the greater the water depth and the faster the water flow in the pool, the larger the value of ∈ (0,2), and vice versa.

[0096] ∈ (0,1), the higher the distribution density of non-drowning people around the candidate point, the larger its value, and vice versa;

[0097] ∈ (0,1), the smoother and unobstructed the landside passage corresponding to the candidate point and the closer it is to the current location of the rescuers, the larger its value is, and vice versa. It can be determined by using the cameras around the pool.

[0098] The interaction module is used to receive alarm commands containing coordinates and the optimal water entry position, and then transmit them to the preset control panel to display the panel's own position, the optimal water entry position, and the coordinate information contained in the alarm command on the panel.

[0099] During the interactive module's operation, in the early warning level, only the drowning victim's coordinates and warning signs are displayed; in the emergency level, the optimal entry point into the water is additionally displayed; and in the special level, the real-time location of the rescuers is overlaid, and the rescuers' rescue path is recorded.

[0100] After receiving the rescue response status (i.e., not yet dispatched, in rescue, arrived), the interaction module feeds back to the integration module in real time. The integration module adjusts the sending frequency of alarm commands according to the status of not yet dispatched, in rescue, and arrived respectively. The sending frequency of the corresponding alarm commands for not yet dispatched, in rescue, and arrived decreases in a gradient.

[0101] The determination of the status of not departing, in rescue, and arrived is based on the real-time location of the rescuers. The rescuers are equipped with location sensors, and the location sensors are connected to the control panel via wireless network or Bluetooth. When the distance between the real-time location of the rescuers and the location of the control panel is less than a preset value X, it is determined that the rescue has not been triggered. When the distance between the real-time location of the rescuers and the location of the control panel is not less than the preset value X, it is determined that the rescue is in progress. When the distance between the real-time location of the rescuers and the current location of the drowning victim does not exceed the preset value Y, it is determined that the victim has arrived. During the status determination phase, the application priority of the preset value Y is higher than the application priority of the preset value X.

[0102] Among them, the best entry point for pool rescue is any point on the edge of the pool, and the analysis module and the integration module run synchronously in conjunction with each other, and the interaction module and the analysis module run synchronously in conjunction with each other.

[0103] The acquisition module is interconnected with the fusion module via a wireless network. The fusion module is interconnected with the judgment module via a wireless network. The judgment module is interconnected with the integration module via a wireless network. The integration module is interconnected with the analysis module via a wireless network. The analysis module is interconnected with the interaction module via a wireless network.

[0104] In this embodiment, the acquisition module collects ECG data and device location-related signals from the swimmer's device in real time. The fusion module then performs noise reduction on the two types of data, extracts heart rate change features and location movement features, and generates a fusion feature dataset. The judgment module receives the fusion feature dataset, analyzes the abnormal correlation between heart rate and location to determine whether the swimmer is drowning. The integration module further triggers operation after determining that the swimmer is drowning, parses the device's real-time location-related signals to obtain the drowning person's real-time coordinates, synchronously generates an alarm command containing coordinate information, and transmits it to the interaction module. The analysis module receives and reparses the real-time location-related signals of all other swimmers in the pool who are not determined to be drowning to obtain their real-time coordinates. Based on the real-time coordinates, it analyzes the optimal entry point for pool rescue and transmits it to the interaction module. Finally, the interaction module receives the alarm command containing coordinates and the optimal entry point, and then transmits it to the preset control panel, displaying the panel's own position, the optimal entry point, and the coordinate information contained in the alarm command on the panel.

[0105] In the above embodiments, the system can monitor the swimmer's electrocardiogram and location-related status in real time during operation, accurately identify drowning risks and issue warnings according to levels, simultaneously provide the best rescue entry position and key coordinate information adapted to the pool scenario, and dynamically adjust the warning rhythm based on the rescue progress, clearly present rescue-related location data, significantly shorten rescue response time, improve rescue efficiency, effectively reduce the probability of drowning injuries and deaths, and provide reliable protection for swimming safety in the pool.

[0106] The following describes an application example of the system, referring to the system in the above embodiments:

[0107] In the swimming pool of XX residential community, a swimmer wearing a special monitoring device is swimming. All modules of the system start and run according to the procedure. The acquisition module synchronously collects the swimmer's electrocardiogram data and the device's location-related signal in real time. After calculation, the synchronization accuracy does not exceed a preset threshold, indicating that synchronization is valid. The two types of data are then transmitted to the fusion module.

[0108] The fusion module first denoises the raw data to obtain clear target data. Then, it extracts features: for heart rate, it obtains the heart rate change rate, heart rate mutation amplitude, and duration of heart rate abnormality for three consecutive sampling periods; for location, it extracts real-time movement speed, movement acceleration, and trajectory continuity coefficient to form a fused feature dataset.

[0109] After receiving the dataset, the judgment module calculates the linkage anomaly quantification value. If the result reaches the preset first-level threshold but does not exceed the second-level threshold, the swimmer is judged to be drowning. The integration module immediately parses the location signal to obtain the drowning person's real-time coordinates, calculates the data credibility coefficient as high credibility, and generates an alarm command containing an emergency-level danger sign, real-time coordinates, and credibility coefficient.

[0110] The analysis module runs synchronously, determining the preset interval of candidate points based on the pool circumference and the current number of swimmers, and picking up multiple candidate points at the pool edge. Taking into account the straight-line distance of each candidate point from the drowning victim, the resistance to passage, the density of people in the surrounding area, and the ease of passage on the land side, the calculation shows that the candidate point on the west side of the pool, which is closest to the drowning victim and has fewer people in the surrounding area, has the highest suitability and is determined as the best entry point into the water.

[0111] After receiving the alarm command and the optimal entry point into the water, the interactive module synchronously transmits them to the control panel in the pool duty room. The panel clearly displays the drowning victim's coordinates, the emergency level indicator, and the optimal entry point. Upon seeing the alarm, rescuers immediately set off, and the status feedback from their position sensors changes from "not yet dispatched" to "rescuing in progress," causing the integrated module to reduce the alarm transmission frequency accordingly. When the rescuers arrive near the drowning victim, the status updates to "arrived," and the alarm transmission frequency further decreases in a gradient.

[0112] In summary, during operation, the system in the above embodiments collects relevant data in real time and performs precise noise reduction, fully extracts key features of heart rate changes and positional movement, quantifies and analyzes abnormal patterns in the linkage between heart rate and position, accurately determines the drowning state, effectively reduces the risk of misjudgment and missed judgment, determines the optimal rescue entry position based on the drowning person's real-time coordinates and the actual situation of the pool, dynamically adjusts the alarm level based on data reliability, synchronously and clearly displays key rescue information, adjusts the alarm sending frequency according to the rescue response status gradient, provides clear guidance for rescuers, significantly shortens the rescue response time, effectively improves rescue efficiency and success rate, ensures the personal safety of swimmers, and is adaptable to pool scenarios of different sizes and different personnel densities.

[0113] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An automatic alarm system for swimming pools based on electrocardiogram monitoring, characterized in that, include: The data acquisition module is used to acquire electrocardiogram data and device location-related signals from the device worn by swimmers in real time. The fusion module is used to denoise the two types of data collected, extract heart rate change features and position movement features, and generate a fused feature dataset. The determination module is used to receive the fused feature dataset, analyze the abnormal patterns of the linkage between heart rate and location, and determine whether the swimmer is drowning. The integrated module is used to trigger operation after a drowning state is determined, analyze the real-time location association signal of the device to obtain the real-time coordinates of the drowning person, and synchronously generate an alarm command containing coordinate information and transmit it to the interaction module. The analysis module is used to receive and re-analyze the real-time location association signals of all other swimmers in the pool who are not determined to be drowning, in order to obtain the real-time coordinates of all other swimmers in the pool who are not determined to be drowning, analyze the best entry position for pool rescue based on the real-time coordinates, and transmit it to the interaction module. The interaction module is used to receive alarm commands containing coordinates and the optimal water entry position, and then transmit them to the preset control panel to display the panel's own position, the optimal water entry position, and the coordinate information contained in the alarm command on the panel. Among them, the optimal entry point for pool rescue is any point on the edge of the pool, and the analysis module and the integration module operate synchronously, as do the interaction module and the analysis module.

2. The automatic alarm system for swimming pools based on electrocardiogram monitoring according to claim 1, characterized in that, The acquisition module synchronously acquires ECG data and device location-related signals based on timestamps. ； In the formula: For synchronization accuracy; This is the timestamp for the ECG data collection. The timestamp for the acquisition of location-related signals; This is the preset sampling period for the acquisition module; For synchronous calibration coefficients; when When the preset synchronization threshold is not exceeded, synchronization is considered valid, and the corresponding ECG data and location-related signals are transmitted to the fusion module; when... If the preset synchronization threshold is exceeded, synchronization is deemed invalid, and the acquisition module immediately triggers a re-acquisition process while simultaneously initiating the synchronization calibration coefficient. The dynamic adjustment range does not exceed ±0.1 each time. If the synchronization is still determined to be invalid after 3 consecutive re-collections, a synchronization error prompt is sent to the integration module, which triggers a temporary alarm on the device and restarts the acquisition module corresponding to the invalid synchronization in the system.

3. The automatic alarm system for swimming pools based on electrocardiogram monitoring according to claim 1, characterized in that, The data denoising in the fusion module follows the following rules: ； In the formula: The target data after denoising; The raw data collected; These are adaptive weighting coefficients; These are the base values ​​of the original data after frequency domain low-pass filtering; The actual instantaneous frequency of the original data; This represents the median of the standard effective frequency range for the corresponding data.

4. The automatic alarm system for swimming pools based on electrocardiogram monitoring according to claim 3, characterized in that, The adaptive weighting coefficient The calculation logic is as follows: ； In the formula: This is the attenuation adjustment coefficient; This represents the variance of the original data within the preset statistical window. The preset variance of the standard valid data; This represents the proportion of frequency domain noise.

5. The automatic alarm system for swimming pools based on electrocardiogram monitoring according to claim 1, characterized in that, The heart rate change features extracted by the fusion module include the heart rate change rate, heart rate mutation amplitude, and heart rate abnormality duration within three consecutive sampling periods. The heart rate change rate is calculated by the ratio of the heart rate difference between adjacent sampling points to the sampling period. The heart rate mutation amplitude is the absolute value of the difference between the real-time heart rate and the average heart rate of the recent preset number of sampling periods. The heart rate abnormality duration is the product of the number of consecutive samplings in which the real-time heart rate exceeds the preset normal heart rate range and the sampling period. The positional movement features extracted by the fusion module include real-time movement speed, movement acceleration, and trajectory continuity coefficient. The real-time movement speed is calculated by the ratio of the difference in position coordinates between adjacent sampling points to the sampling period. The movement acceleration is calculated by the ratio of the difference in speed between two adjacent sampling periods to the sampling period. The trajectory continuity coefficient is characterized by the cosine of the angle between the current sampling point and the previous two sampling points.

6. The automatic alarm system for swimming pools based on electrocardiogram monitoring according to claim 1, characterized in that, The determination module quantifies abnormalities in the linkage between heart rate and location using the following formula: ； In the formula: This is a quantified value for linkage anomalies; These are the feature weight coefficients; Real-time heart rate; This is the swimmer's preset baseline heart rate; Real-time movement speed; This is the preset minimum normal movement speed; The duration of stay at the current location; The preset judgment duration; When A is not less than the preset abnormal threshold, the swimmer is determined to be in a state of drowning.

7. The automatic alarm system for swimming pools based on electrocardiogram monitoring according to claim 1, characterized in that, The alarm command generated by the integrated module includes a drowning hazard level indicator, the real-time coordinates of the drowning victim, and a data reliability coefficient. The hazard level indicator is based on the linkage anomaly quantification value. Sure: when The threshold is not less than the preset abnormal threshold; if it is less than the preset first-level threshold, it is a warning level. when The threshold value is not less than the first-level threshold, and is classified as emergency if it is less than the preset second-level threshold. when It is classified as Special Grade if it is not less than the Level 2 threshold; The data credibility coefficient , For efficient synchronization, This represents the signal-to-noise ratio of the denoised signal. This represents the feature extraction completeness coefficient, with values ​​ranging from 0 to 1. when If the risk level is not less than the preset high confidence threshold, trigger the corresponding alarm according to the current risk level; when When the data falls between a preset low confidence threshold and a preset high confidence threshold, the integrated module controls the acquisition module to increase the sampling frequency and re-acquire data, performing secondary calculations based on the new data. and The alert level will then be determined. If the calculation is repeated at least three times, the alert level will be raised by one level. when If the level does not exceed the preset low confidence threshold, the warning level will be upgraded by one level.

8. The automatic alarm system for swimming pools based on electrocardiogram monitoring according to claim 1, characterized in that, The analysis module analyzes the optimal entry point stage, picks up candidate points at preset intervals along the edge of the pool, and then calculates the fit of each candidate point based on the following formula. ； In the formula: For the overall fit of candidate points; These are the weighting coefficients; The straight-line distance from the candidate point to the water area where the drowning victim was located; This represents the resistance coefficient for the corresponding water area. The population density coefficient around the candidate site; The landside accessibility coefficient for candidate points; The candidate point with the highest fit is selected as the optimal entry point.

9. The automatic alarm system for swimming pools based on electrocardiogram monitoring according to claim 1, characterized in that, During the operation of the interactive module, only the coordinates of the drowning victim and the warning sign are displayed in the warning level state. In emergency situations, the optimal entry point into the water is additionally displayed; in extreme situations, the real-time location of the rescuers is overlaid, and the rescuers' rescue path is recorded. After receiving the rescue response status, the interaction module feeds back to the integration module in real time. The integration module adjusts the sending frequency of alarm commands according to the status of not yet departed, in rescue, and arrived. The sending frequency of alarm commands corresponding to the status of not yet departed, in rescue, and arrived decreases in a gradient. The determination of the status of not departing, in rescue, and arrived is based on the real-time location of the rescuers. The rescuers are equipped with location sensors, and the location sensors are connected to the control panel via wireless network or Bluetooth. When the distance between the real-time location of the rescuers and the location of the control panel is less than a preset value X, it is determined that the rescue has not been triggered. When the distance between the real-time location of the rescuers and the location of the control panel is not less than the preset value X, it is determined that the rescue is in progress. When the distance between the real-time location of the rescuers and the current location of the drowning victim does not exceed the preset value Y, it is determined that the rescue has arrived. During the status determination phase, the application priority of the preset value Y is higher than the application priority of the preset value X.

10. The automatic alarm system for swimming pools based on electrocardiogram monitoring according to claim 1, characterized in that, The acquisition module is interconnected with a fusion module via a wireless network. The fusion module is interconnected with a judgment module via a wireless network. The judgment module is interconnected with an integration module via a wireless network. The integration module is interconnected with an analysis module via a wireless network. The analysis module is interconnected with an interaction module via a wireless network.

Images