Fall monitoring method, system and storage medium based on wifi signal recognition
By setting up wifi collection points in the fall monitoring area and building an identification model, the problems of discomfort and data processing difficulties in the existing technology are solved, and wear-free fall monitoring and accurate identification are achieved, improving user comfort and monitoring accuracy.
Patent Information
- Application Number
- CN202411167118.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-23
- Publication Date
- 2024-11-01
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing fall monitoring technologies rely on wearable devices, which leads to discomfort for users and makes it difficult to accurately extract useful information when processing large amounts of data, affecting detection accuracy.
By laying multiple wifi acquisition points in the target area, a wifi acquisition module is built to continuously collect, obtain signal data, and extract signal characteristics through Fourier transform, time domain and frequency domain analysis, combining fall event annotation and communication scenario construction Fall recognition model to achieve accurate identification and emergency response of fall events.
No need for users to wear equipment, reduce life interference, improve comfort and privacy protection, enhance environmental adaptability and accuracy, reduce false alarms and missed reports, and improve response speed.
Smart Images

Figure CN118692208B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of signal recognition, and in particular to a fall monitoring method, system and storage medium based on wifi signal recognition. Background Art
[0002] Falling is a common and serious safety problem that may lead to serious health problems such as fractures and head injuries. Therefore, an effective fall monitoring method is needed to provide timely assistance. However, there are still certain problems with the existing technology for fall monitoring. On the one hand, the fall monitoring of the existing technology usually relies on wearable devices, such as smart bracelets, belts, etc. These devices require users to actively wear them. For elderly people or patients with limited mobility, wearing these devices is not practical and may cause inconvenience or discomfort in wearing, affecting the user's quality of life; on the other hand, traditional methods often require a fixed sensor layout, which is difficult to adapt to environmental changes or the complexity of user activities, resulting in unstable monitoring effects, and face challenges when processing large amounts of data. It is difficult to accurately extract useful information from noise, affecting the detection accuracy of fall events. Summary of the invention
[0003] The present application provides a fall monitoring method, system and storage medium based on wifi signal recognition, aiming to solve the technical problems that the existing technology for fall monitoring usually relies on wearable devices, which may cause discomfort to users and affect the quality of life of users, and it is difficult to accurately extract useful information from noise when processing large amounts of data, resulting in poor detection accuracy of fall events.
[0004] The first aspect disclosed in the present application provides a fall monitoring method based on Wi-Fi signal recognition, the method comprising: extracting target area environmental information to deploy multiple Wi-Fi collection points, constructing a Wi-Fi collection module, performing continuous collection through the Wi-Fi collection module, and obtaining multiple signal data; transmitting the multiple signal data to a Wi-Fi signal processing module to obtain multiple signal features; annotating the multiple signal data based on the multiple signal features in combination with a fall event of a target user, and constructing a fall recognition model based on the annotation results and multiple communication scenarios; monitoring and identifying the target user through the fall recognition model, and generating a fall recognition result; triggering an alarm module according to the fall recognition result, and performing an emergency response based on the alarm module.
[0005] The second aspect disclosed in the present application provides a fall monitoring system based on Wi-Fi signal recognition, and the system is used for the above-mentioned fall monitoring method based on Wi-Fi signal recognition, and the system includes: a signal continuous acquisition unit, the signal continuous acquisition unit is used to extract the target area environmental information, deploy multiple Wi-Fi acquisition points, build a Wi-Fi acquisition module, and continuously acquire multiple signal data through the Wi-Fi acquisition module; a signal feature acquisition unit, the signal feature acquisition unit is used to transmit the multiple signal data to the Wi-Fi signal processing module to obtain multiple signal features; a fall recognition model construction unit, the fall recognition model construction unit is used to annotate the multiple signal data based on the multiple signal features in combination with the fall event of the target user, and construct a fall recognition model based on the annotation results in combination with multiple communication scenarios; a fall recognition result acquisition unit, the fall recognition result acquisition unit is used to monitor and identify the target user through the fall recognition model to generate a fall recognition result; an emergency response unit, the emergency response unit is used to trigger an alarm module according to the fall recognition result, and perform an emergency response based on the alarm module.
[0006] According to a third aspect disclosed in the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, any step of the first aspect disclosed in the present application is implemented.
[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages:
[0008] By monitoring fall events through WiFi signals, there is no need for users to wear additional equipment, which reduces interference with users' daily lives and improves users' comfort and privacy protection. By deploying multiple WiFi collection points in the target area and adaptively deploying them in combination with the environmental information of the target area, it can better adapt to the signal propagation characteristics in different environments. The WiFi signal processing module performs Fourier transform, time domain and frequency domain analysis on the signal data, which can extract features from complex data to achieve more accurate identification of fall events and reduce false alarms and missed alarms. Signal data is annotated based on signal features and combined with the fall events of the target users, and a fall recognition model is constructed by combining the annotated signal data and communication scenarios, so that the model can adapt to the impact of different environments on WiFi signals, improving the environmental adaptability and accuracy of the system. The alarm module is automatically triggered according to the fall recognition results, and emergency response is carried out, which improves the response speed to fall events.
[0009] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS
[0010] Figure 1 A schematic diagram of a fall monitoring method based on WiFi signal recognition provided in an embodiment of the present application;
[0011] Figure 2 A schematic diagram of the structure of a fall monitoring system based on WiFi signal recognition provided in an embodiment of the present application.
[0012] Explanation of the accompanying drawings: signal continuous acquisition unit 10, signal feature acquisition unit 20, fall recognition model construction unit 30, fall recognition result acquisition unit 40, emergency response unit 50. DETAILED DESCRIPTION
[0013] The embodiments of the present application provide a fall monitoring method, system and storage medium based on wifi signal recognition, thereby solving the technical problems that the prior art usually relies on wearable devices for fall monitoring, which may cause discomfort to users and affect their quality of life, and it is difficult to accurately extract useful information from noise when processing large amounts of data, resulting in poor detection accuracy of fall events.
[0014] After introducing the basic principles of the present application, various non-limiting implementation methods of the present application will be specifically introduced below in conjunction with the accompanying drawings of the specification. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.
[0015] Embodiment 1, as Figure 1 As shown, the embodiment of the present application provides a fall monitoring method based on wifi signal recognition, the method comprising:
[0016] Extract the target area environmental information, deploy multiple Wi-Fi collection points, build a Wi-Fi collection module, and perform continuous collection through the Wi-Fi collection module to obtain multiple signal data.
[0017] When a person falls in the monitoring area, his body posture and movement state will change. This change will cause the reflection and scattering characteristics of the human body to change, which will affect the signal characteristics received by the receiving end. Based on this, multiple WiFi collection points are deployed in the target area. The target area refers to a specific spatial area where a fall event needs to be monitored, such as a family room, a hospital ward, etc. The target area environmental information includes the spatial layout of the target area, the distribution of walls and furniture, etc. This information can be obtained through on-site investigation and measurement. According to the target area environmental information, multiple WiFi collection points are determined in the target area, and WiFi signal sources are deployed at multiple WiFi collection points to ensure that the WiFi signal can cover the entire target area. Based on the determined multiple WiFi collection points, combined with the WiFi collection frequency determined based on the mobile characteristics of the target user, a WiFi collection module is constructed to continuously collect WiFi transmission signals from WiFi signal sources and generate the multiple signal data.
[0018] Furthermore, extracting the target area environmental information, deploying multiple Wi-Fi collection points, constructing a Wi-Fi collection module, and continuously collecting through the Wi-Fi collection module to obtain multiple signal data, the method includes:
[0019] Traversing the target area environmental information to obtain the Wi-Fi network, and determining a plurality of Wi-Fi collection points; deploying Wi-Fi signal sources according to the plurality of Wi-Fi collection points; activating the Wi-Fi signal source to transmit signals when the target user is in the target area, and generating Wi-Fi transmission signals; determining the Wi-Fi collection frequency based on the mobility characteristics of the target user; constructing a Wi-Fi collection module according to the plurality of Wi-Fi collection points and the Wi-Fi collection frequency; and continuously collecting Wi-Fi transmission signals through the Wi-Fi collection module to generate the plurality of signal data.
[0020] The target area refers to a specific spatial area where falls need to be monitored, such as a family room, a hospital ward, etc. The target area environmental information includes the spatial layout of the target area, the distribution of walls and furniture, etc., which can be obtained through on-site investigation and measurement. Use the wifi scanning tool to scan the target area, detect the existing wifi network and signal coverage, analyze the signal strength and coverage of each location based on the wifi scanning results, and select points with strong signal strength and wide coverage as wifi collection points.
[0021] Select a suitable WiFi signal source device and install it at the corresponding location according to the determined WiFi collection points to ensure that there is a stable WiFi signal with good coverage in the target area to support subsequent signal collection and fall monitoring.
[0022] Use sensors, such as infrared sensors, cameras, etc., to detect whether the target user enters the target area. When the target user is detected to enter the target area, activate the preset Wi-Fi signal source. The activated Wi-Fi signal source transmits a Wi-Fi signal to generate a Wi-Fi transmission signal to cover the target area.
[0023] Use sensors, such as accelerometers, to obtain user movement data, extract characteristic information such as movement speed and frequency, generate movement characteristics, and automatically adjust the WiFi collection frequency according to the user's movement characteristics. For example, when the movement frequency is high, increase the collection frequency; when the movement frequency is low, reduce the collection frequency to ensure signal collection efficiency and data accuracy under different movement states.
[0024] A WiFi receiver is arranged at each WiFi collection point, and the collection frequency of the WiFi receiver is set to a certain WiFi collection frequency. A WiFi collection module is constructed to control the WiFi receiver to collect signals according to the WiFi collection frequency.
[0025] After the user enters the target area and activates the WiFi signal source, the WiFi acquisition module starts the WiFi receiver to continuously acquire WiFi transmission signals according to the set acquisition frequency to generate the multiple signal data.
[0026] The multiple signal data are transmitted to the wifi signal processing module to obtain multiple signal characteristics.
[0027] The wifi signal processing module is used to process and analyze the received signal data. The wifi signal processing module performs Fourier transform on the multiple signal data to generate signal frequency characteristics. Based on the signal frequency characteristics, time domain analysis and time domain analysis are performed on the multiple signal data to analyze the changes of the signal in time and frequency domain, and generate signal time domain analysis results and signal frequency domain analysis results.
[0028] Furthermore, the plurality of signal data are transmitted to a wifi signal processing module to obtain a plurality of signal features, the method comprising:
[0029] The wifi signal processing module performs Fourier transform on the multiple signal data to generate signal frequency characteristics; based on the signal frequency characteristics, time domain analysis is performed on the multiple signal data to generate signal time domain analysis results; based on the signal frequency characteristics, frequency domain analysis is performed on the multiple signal data to generate signal frequency domain analysis results; the multiple signal characteristics are determined according to the signal time domain analysis results and the signal frequency domain analysis results.
[0030] The wifi signal processing module is used to perform Fourier transform on multiple signal data. Fourier transform is a mathematical method that converts time domain signals into frequency domain signals. It is used to analyze the frequency components of the signal. The spectrum of the signal is obtained through Fourier transform, including the frequency components and corresponding amplitudes of the signal. The signal frequency characteristics such as characteristic frequency, main frequency components, frequency bandwidth, etc. are extracted from the spectrum.
[0031] The signal is analyzed in segments according to the preset time window. The sliding window technology is used to analyze the signal segment by segment to ensure the continuity of the time domain characteristics. The time domain characteristics of the signal, such as the mean, standard deviation, peak value, waveform shape, etc., are extracted. The changing trend and pattern of the signal in the time series are analyzed according to the time domain characteristics. The instantaneous characteristics of the signal, such as the instantaneous amplitude, instantaneous frequency, etc., are extracted to generate the signal time domain analysis results.
[0032] According to the signal spectrum, key frequency components, such as main frequency, harmonic frequency, etc., are selected as analysis features to perform frequency domain analysis. Specifically, the amplitude and phase of different frequency components in the signal are analyzed to determine the main frequency; the energy distribution of the signal in different frequency bands is calculated to determine the frequency bandwidth; the harmonic components in the signal are identified and analyzed to determine their frequency and amplitude, and the harmonic components are determined. The main frequency, frequency bandwidth, and harmonic components are integrated to generate the signal frequency domain analysis results.
[0033] The signal time domain analysis result and the signal frequency domain analysis result are combined to form a comprehensive feature vector to determine the multiple signal features.
[0034] The multiple signal data are annotated based on the multiple signal features in combination with a fall event of a target user, and a fall recognition model is constructed according to the annotation results in combination with multiple communication scenarios.
[0035] A fall event refers to an actual fall event of the target user, which can be obtained by simulating a fall situation or retrieving a historical fall event. The signal data is labeled according to the fall event of the target user. Specifically, the signal time domain analysis results are associated with the actual fall event, and the acceleration waveform change trend corresponding to the fall event is analyzed to obtain the acceleration peak of the fall event; the signal frequency domain analysis results are associated with the actual fall event, and the power spectrum density information of the fall event is analyzed to determine the frequency distribution data of the fall event. Generate a fall feature label based on the acceleration peak and frequency distribution data, label multiple signal data, and obtain the labeling results.
[0036] Multiple communication scenarios refer to communication scenarios in different environments. Different environments, such as different rooms, furniture layouts, wall materials, etc., have different effects on the propagation and reflection characteristics of wifi signals. Therefore, in practical applications, model training and tuning are required for different environments to improve the adaptability and accuracy of the model. Based on the environmental information of the target area, the environment is divided into multiple different communication scenarios, including different rooms, different environmental layouts, etc., to obtain multiple environmental scenarios. Perform wifi signal tests in each environmental scenario to obtain information such as signal strength, signal phase, and signal frequency to determine the changes in signals in different environments. Perform communication analysis based on information such as signal strength, signal phase, and signal frequency to construct multiple communication scenarios.
[0037] The annotation results are randomly divided into training data sets, validation data sets, and test data sets. According to multiple communication scenarios, the fall recognition model is trained on the training data sets, and the model parameters are adjusted to improve the recognition accuracy. The model performance is tested on the validation data sets, and necessary adjustments and optimizations are made. The final performance of the model is evaluated on the test data sets to ensure its adaptability and accuracy in different environments. Finally, a fall recognition model that meets the requirements is obtained.
[0038] Furthermore, the method of labeling the plurality of signal data based on the plurality of signal features in combination with the fall event of the target user includes:
[0039] According to the signal time domain analysis result combined with the fall event of the target user, the acceleration waveform change trend is obtained; based on the acceleration waveform change trend, the acceleration peak is determined; according to the signal frequency domain analysis result combined with the fall event of the target user, the power spectrum density information is obtained; based on the power spectrum density information, the frequency distribution data is determined; based on the acceleration peak and the frequency distribution data, a plurality of fall feature labels are generated; the plurality of signal data are annotated by traversing the plurality of fall feature labels to generate the annotation result.
[0040] The extracted signal time domain analysis results are used to match the features related to the fall event, such as the peak value, mean value, standard deviation, etc. of the signal. Specifically, when the fall event begins, the user loses balance, causing the acceleration to increase downward, which is manifested as a sudden drop in the acceleration value; when the body collides with the ground, a large fluctuation occurs, which is manifested as a sharp rise and fluctuation in the acceleration value; after the fall, the body stops moving and the acceleration value returns to a stable state. The start, fluctuation, and end stages of the fall event are marked in the signal data, and the acceleration waveform of each fall event is extracted to form the acceleration waveform change trend.
[0041] Use signal processing algorithms, such as peak detection algorithms, to detect significant peaks in the acceleration waveform, record the time point of each peak, analyze the time distribution of the peaks, calculate the size of the acceleration peaks, which are usually higher than the acceleration values in normal activities, and calculate the time intervals between adjacent peaks to distinguish between fall events and daily activities. Acceleration peak information is determined based on the peak time point, peak size, and peak interval. During a fall, due to the rapid contact between the body and the ground, the acceleration signal will have significant peaks. The size and time of these peaks can be used to distinguish between fall events and daily activities.
[0042] Use the extracted signal frequency domain analysis results to match the features related to fall events, such as changes in the signal spectrum and frequency components. Use signal processing tools, such as the Welch method, to calculate the power spectral density of the signal, determine the energy distribution of the signal at different frequencies, and identify the energy distribution characteristics of fall events at different frequencies. The high-frequency components correspond to the collision of the body with the ground, and the low-frequency components correspond to the movement of the body. Mark the start and end points of the fall event in the signal data, and extract the power spectral density information of each fall event.
[0043] Identify the key frequency components in the power spectral density information, analyze their distribution patterns in fall events, determine the energy distribution of high-frequency and low-frequency components in fall events, identify high-frequency vibrations (such as body collision with the ground) and low-frequency movements (such as body movement), and determine frequency distribution data, including the frequency range of signal energy distribution.
[0044] The acceleration peak and frequency distribution data are combined to form a comprehensive feature set. The fall features are classified according to different feature types, such as high-frequency labels, low-frequency labels, high-acceleration labels, etc., to generate multiple fall feature labels.
[0045] Traverse all signal data, apply the generated fall feature label, match each signal data with the feature label, determine whether it meets the definition of a feature label, and label each signal data according to the matching result. The generated labeling result includes whether it meets the feature label and the type of label it meets.
[0046] Furthermore, a fall recognition model is constructed based on the annotation results and multiple communication scenarios, and the method includes:
[0047] Perform environmental identification and classification based on the target area environmental information to determine multiple environmental scenarios; perform Wi-Fi signal testing based on the multiple environmental scenarios to generate signal test results, wherein the signal test results include Wi-Fi signal strength information, Wi-Fi signal phase information, and Wi-Fi signal frequency information; perform communication analysis on the Wi-Fi signal phase information and the Wi-Fi signal frequency information based on the Wi-Fi signal strength information to construct multiple communication scenarios; perform data random division on the annotation results to determine a training data set, a verification data set, and a test data set; train and optimize the training data set and the verification data set according to the multiple communication scenarios, and evaluate them based on the test data set to generate a model evaluation result, and construct the fall recognition model.
[0048] Extract key environmental features from environmental information, such as reflective surfaces (walls, floors, ceilings), absorbing materials (furniture, curtains), interference sources (other wireless signal sources, metal objects), etc., and classify environmental layouts, such as open space, small space, complex layout, simple layout, etc., as environmental features. Classify the environmental information of the target area according to environmental features and determine multiple environmental scenarios.
[0049] Deploy Wi-Fi signal sources in various environmental scenes in the target area, transmit signals, collect signal test data in various environmental scenes, analyze the collected signal data, extract Wi-Fi signal strength information, Wi-Fi signal phase information and Wi-Fi signal frequency information, and generate signal test results.
[0050] Perform signal feature correlation analysis to analyze the relationship between signal strength, phase and frequency, for example, the relationship between strength and phase change, the impact of frequency change on signal strength, etc. Based on the signal analysis results, divide different signal propagation and reflection characteristics into multiple communication scenarios. Each communication scenario represents the signal propagation characteristics under specific environmental conditions, such as a high signal strength and low phase offset scenario, a low signal strength and high frequency interference scenario, etc.
[0051] Set the division ratio of the training dataset, validation dataset, and test dataset. For example, the training dataset accounts for 70%, the validation dataset accounts for 15%, and the test dataset accounts for 15%. Based on the division ratio, use the random division method to divide the annotation results into the training dataset, validation dataset, and test dataset to ensure that the ratio of fall events (positive samples) and non-fall events (negative samples) in each dataset is relatively balanced.
[0052] Match the samples in the training data set to different communication scenarios so that the model can learn the features in different scenarios. Select appropriate machine learning algorithms, such as support vector machines, decision trees, neural networks, etc., for training the fall recognition model. Use the training data set and communication scenarios to train the model. Adjust model parameters, such as learning rate, regularization parameters, etc., to improve the prediction performance of the model. Tune the model through methods such as cross-validation to optimize the performance of the model and reduce overfitting and underfitting. Use the validation data set to test the model and evaluate the performance of the model in different communication scenarios. Based on the results of the validation data set, adjust the model's hyperparameters and optimize the generalization ability of the model. Use the test data set to conduct a final evaluation of the trained model, such as accuracy, to evaluate the overall performance of the model, analyze the performance of the model in different communication scenarios, identify the strengths and weaknesses of the model, and ensure the reliability and accuracy of the model in practical applications. Based on the model evaluation results of the test data set, generate the final fall recognition model for real-time monitoring and identification of fall events in different communication scenarios.
[0053] The target user is monitored and identified through the fall identification model to generate a fall identification result; an alarm module is triggered according to the fall identification result, and an emergency response is performed based on the alarm module.
[0054] The multiple signal data obtained from the wifi acquisition module are transmitted to the fall recognition model in real time to monitor and identify the target user. The model generates a fall recognition result based on the input signal data. The recognition results can be divided into two types. The first recognition result determines that the user has a fall event and assigns a value of 1. In this case, the alarm module is triggered for emergency response. The alarm module can be configured to send a text message notification to the preset emergency contact to inform the target user of the fall event; or it can automatically dial an emergency number, such as contacting emergency services or family members, so that rescue measures can be taken quickly. The second recognition result determines that the user has not had a fall event and assigns a value of 0. In this case, the user status is continuously monitored and the wifi acquisition frequency is updated according to the monitoring data.
[0055] By real-time monitoring and identifying fall events, and quickly triggering alarms and emergency response mechanisms when events occur, it ensures that users can receive timely help after falling.
[0056] Furthermore, the target user is monitored and identified by the fall identification model, a fall identification result is generated, and an alarm module is triggered according to the fall identification result. The method includes:
[0057] The WiFi transmission signal is synchronized to the fall recognition model to monitor and identify the target user, and a fall recognition result is obtained, wherein the fall recognition result is the first recognition result or the second recognition result; when the fall recognition result is the first recognition result, an activation instruction is generated in combination with the target area environmental information, the fall recognition result is assigned a value of 1, and the alarm module is triggered; when the fall recognition result is the second recognition result, a continuous monitoring instruction is generated, and the continuous monitoring duration and continuous monitoring dynamic information are retrieved according to the continuous monitoring instruction for analysis, the WiFi acquisition frequency is updated according to the analysis result, the fall recognition result is assigned a value of 0, and the fall recognition result is generated.
[0058] The collected WiFi transmission signal is input into the fall recognition model. The fall recognition model analyzes the input WiFi signal data, identifies whether a fall event occurs, and outputs a fall recognition result. The fall recognition result includes a first recognition result or a second recognition result. The first recognition result is recognition that a fall has occurred, and the second recognition result is recognition that no fall has occurred.
[0059] When the fall recognition result is the first recognition result, an activation instruction is generated in combination with the environmental information of the target area, such as ground conditions, surrounding obstacles, etc. The environmental information helps determine the severity of the fall event and the response strategy. For example, falling on a cement floor is different from falling on a bed or sofa. The alarm response level is divided in this way and added to the activation instruction. The fall recognition result is assigned a value of 1, indicating that a fall event has been confirmed. This value is used to trigger the alarm module. According to the generated activation instruction, the alarm module is triggered. For example, a fall alarm can be issued in the form of a sound alarm, SMS notification, phone alarm, etc. to ensure that the fall event is handled in a timely manner.
[0060] When the fall recognition result is the second recognition result, a continuous monitoring instruction is generated, which requires the system to continue to monitor the state of the target area. If no fall event is detected for a long time, check whether the user is in a dynamic state. Specifically, according to the needs, set the monitoring duration, such as 1 hour or 30 minutes, continue to collect signals within the monitoring duration, obtain continuous monitoring dynamic information, and analyze whether the target user is in an active state based on the continuous monitoring dynamic information. For example, by analyzing the signal change pattern to determine whether the user is moving, if the user is in a static state for a long time, the signal collection frequency can be reduced to save resources and reduce interference. The fall recognition result is assigned a value of 0, indicating that no fall event is identified, and the final fall recognition result is generated.
[0061] Furthermore, when the fall recognition result is the first recognition result, an activation instruction is generated in combination with the target area environment information, and the method includes:
[0062] When the fall recognition result is the first recognition result, the fall posture information of the target user is generated according to the analysis based on the WiFi transmission signal; the fall alarm response level is divided based on the fall posture information combined with the target area environmental information; and the fall alarm response level is added to the activation instruction.
[0063] When the fall recognition result is the first recognition result, it means that the target user has fallen. After the fall recognition, relevant information is extracted from the existing WiFi signal data, including signal strength, signal phase, signal frequency, etc., and the time series data of the WiFi signal is analyzed to identify the change pattern of the signal during the fall, such as sudden changes or fluctuations in signal strength, analyze the spectral characteristics of the signal, and identify the frequency domain changes that may be caused by the fall event, such as the enhancement of high-frequency components in the spectrum.
[0064] Based on the results of the Wi-Fi signal analysis and the acceleration change information, the user's posture during the fall is determined. For example, the user's falling posture, such as leaning forward or lying on the side, can be inferred by analyzing the signal fluctuation pattern. In addition, the signal pattern during the fall can be analyzed to identify whether there is an obvious collision or violent movement and determine the collision pattern. The falling posture and collision pattern are integrated to form the falling posture information.
[0065] The impact of ground materials is analyzed based on the environmental information of the target area. Specifically, hard floors, such as concrete floors and tile floors, will cause more serious injuries when falling. Therefore, falls in such environments usually require a higher response level. Soft floors, such as beds and sofas, will reduce the impact of falling and have a relatively low risk of injury.
[0066] The impact of the falling posture is analyzed based on the falling posture information. Specifically, landing on the head may cause serious injury, so the highest level of alarm response is required; although landing on the knees or buttocks may also cause injuries, the risk is usually lower than landing on the head.
[0067] According to the results of the impact analysis, the alarm response levels are divided. Specifically, for high-risk fall events, such as hard ground and head landing, it is classified as a high response level, and the alarm module needs to be triggered immediately for emergency response, such as sending text messages to notify relevant personnel and calling emergency numbers; for medium-risk fall events, such as hard ground, knees or buttocks landing, it is classified as a medium response level, and a warning signal needs to be generated. Health checks or monitoring are recommended, and further observation may be required; for low-risk fall events, such as soft ground and non-serious landing methods, it is classified as a low response level. It only needs to record the event and continue to monitor the status of the target user. If there are no further abnormalities, the low response can be maintained.
[0068] The activation instruction is used to trigger the alarm module or other response mechanism. Depending on the severity of the fall event, the activation instruction can include different response operations, such as sending a notification, making a call, starting an emergency procedure, etc. According to the determined fall alarm response level, this information is integrated into the activation instruction. In this way, the activation instruction contains detailed information about the severity of the event and indicates the operations that the system needs to perform.
[0069] In summary, the fall monitoring method based on wifi signal recognition provided by the embodiment of the present application has the following technical effects:
[0070] By monitoring fall events through WiFi signals, there is no need for users to wear additional equipment, which reduces interference with users' daily lives and improves users' comfort and privacy protection. By deploying multiple WiFi collection points in the target area and adaptively deploying them in combination with the environmental information of the target area, it can better adapt to the signal propagation characteristics in different environments. The WiFi signal processing module performs Fourier transform, time domain and frequency domain analysis on the signal data, which can extract features from complex data to achieve more accurate identification of fall events and reduce false alarms and missed alarms. Signal data is annotated based on signal features and combined with the fall events of the target users, and a fall recognition model is constructed by combining the annotated signal data and communication scenarios, so that the model can adapt to the impact of different environments on WiFi signals, improving the environmental adaptability and accuracy of the system. The alarm module is automatically triggered according to the fall recognition results, and emergency response is carried out, which improves the response speed to fall events.
[0071] Embodiment 2 is based on the same inventive concept as the fall monitoring method based on wifi signal recognition in the above embodiment. Figure 2 As shown, an embodiment of the present application provides a fall monitoring system based on wifi signal recognition, the system comprising:
[0072] A signal continuous acquisition unit 10, the signal continuous acquisition unit 10 is used to extract the target area environmental information, deploy multiple wifi acquisition points, build a wifi acquisition module, and continuously acquire data through the wifi acquisition module to obtain multiple signal data; a signal feature acquisition unit 20, the signal feature acquisition unit 20 is used to transmit the multiple signal data to the wifi signal processing module to obtain multiple signal features; a fall recognition model construction unit 30, the fall recognition model construction unit 30 is used to annotate the multiple signal data based on the multiple signal features in combination with the fall event of the target user, and construct a fall recognition model based on the annotation results and multiple communication scenarios; a fall recognition result acquisition unit 40, the fall recognition result acquisition unit 40 is used to monitor and identify the target user through the fall recognition model to generate a fall recognition result; an emergency response unit 50, the emergency response unit 50 is used to trigger the alarm module according to the fall recognition result, and perform an emergency response based on the alarm module.
[0073] Furthermore, the system further includes a signal data generating module to perform the following operation steps:
[0074] Traversing the target area environmental information to obtain the Wi-Fi network, and determining a plurality of Wi-Fi collection points; deploying Wi-Fi signal sources according to the plurality of Wi-Fi collection points; activating the Wi-Fi signal source to transmit signals when the target user is in the target area, and generating Wi-Fi transmission signals; determining the Wi-Fi collection frequency based on the mobility characteristics of the target user; constructing a Wi-Fi collection module according to the plurality of Wi-Fi collection points and the Wi-Fi collection frequency; and continuously collecting Wi-Fi transmission signals through the Wi-Fi collection module to generate the plurality of signal data.
[0075] Furthermore, the system further includes a signal feature determination module to perform the following operation steps:
[0076] The wifi signal processing module performs Fourier transform on the multiple signal data to generate signal frequency characteristics; based on the signal frequency characteristics, time domain analysis is performed on the multiple signal data to generate signal time domain analysis results; based on the signal frequency characteristics, frequency domain analysis is performed on the multiple signal data to generate signal frequency domain analysis results; the multiple signal characteristics are determined according to the signal time domain analysis results and the signal frequency domain analysis results.
[0077] Furthermore, the system further includes a labeling result generating module to perform the following operation steps:
[0078] According to the signal time domain analysis result combined with the fall event of the target user, the acceleration waveform change trend is obtained; based on the acceleration waveform change trend, the acceleration peak is determined; according to the signal frequency domain analysis result combined with the fall event of the target user, the power spectrum density information is obtained; based on the power spectrum density information, the frequency distribution data is determined; based on the acceleration peak and the frequency distribution data, a plurality of fall feature labels are generated; the plurality of signal data are annotated by traversing the plurality of fall feature labels to generate the annotation result.
[0079] Furthermore, the system further includes a fall recognition model building module to perform the following operation steps:
[0080] Perform environmental identification and classification based on the target area environmental information to determine multiple environmental scenarios; perform Wi-Fi signal testing based on the multiple environmental scenarios to generate signal test results, wherein the signal test results include Wi-Fi signal strength information, Wi-Fi signal phase information, and Wi-Fi signal frequency information; perform communication analysis on the Wi-Fi signal phase information and the Wi-Fi signal frequency information based on the Wi-Fi signal strength information to construct multiple communication scenarios; perform random data division on the annotation results to determine a training data set, a verification data set, and a test data set;
[0081] The training data set and the verification data set are trained and tuned according to the multiple communication scenarios, and are evaluated according to the test data set to generate a model evaluation result and construct the fall recognition model.
[0082] Furthermore, the system further includes a fall recognition result generation module to perform the following operation steps:
[0083] The WiFi transmission signal is synchronized to the fall recognition model to monitor and identify the target user, and a fall recognition result is obtained, wherein the fall recognition result is the first recognition result or the second recognition result; when the fall recognition result is the first recognition result, an activation instruction is generated in combination with the target area environmental information, the fall recognition result is assigned a value of 1, and the alarm module is triggered; when the fall recognition result is the second recognition result, a continuous monitoring instruction is generated, and the continuous monitoring duration and continuous monitoring dynamic information are retrieved according to the continuous monitoring instruction for analysis, the WiFi acquisition frequency is updated according to the analysis result, the fall recognition result is assigned a value of 0, and the fall recognition result is generated.
[0084] Furthermore, the system further includes an activation instruction generation module to perform the following operation steps:
[0085] When the fall recognition result is the first recognition result, analyzing the WiFi transmission signal to generate the fall posture information of the target user; dividing the fall alarm response level based on the fall posture information combined with the target area environment information;
[0086] The fall alarm response level is added to the activation instructions.
[0087] Through the above detailed description of the fall monitoring method based on wifi signal recognition in this specification, those skilled in the art can clearly understand the fall monitoring system based on wifi signal recognition in this embodiment. Since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.
[0088] Embodiment three, based on the same inventive concept as an intelligent drug screening method in the aforementioned embodiment, the present application also provides a storage medium, on which a computer program is stored, and when the computer program is executed, the steps of any one of the methods described in the aforementioned embodiment one are implemented.
[0089] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A fall monitoring method based on wifi signal recognition, characterized in that: The method comprises: Extract the target area environmental information, deploy multiple Wi-Fi collection points, build a Wi-Fi collection module, and continuously collect data through the Wi-Fi collection module to obtain multiple signal data; Transmitting the plurality of signal data to a wifi signal processing module to obtain a plurality of signal features; Annotating the multiple signal data based on the multiple signal features in combination with a fall event of the target user, and building a fall recognition model according to the annotation results in combination with multiple communication scenarios; Monitoring and identifying the target user through the fall recognition model to generate a fall recognition result; triggering an alarm module according to the fall recognition result, and performing an emergency response based on the alarm module; Traversing the target area environmental information to obtain the Wi-Fi network and determine multiple Wi-Fi collection points; Deploy WiFi signal sources according to the multiple WiFi collection points; When the target user is in the target area, the Wi-Fi signal source is activated to transmit a signal and generate a Wi-Fi transmission signal; Determine the wifi acquisition frequency based on the target user's mobility characteristics; Constructing a wifi acquisition module according to the multiple wifi acquisition points and the wifi acquisition frequency; Continuously collect the WiFi transmission signals through the WiFi collection module to generate the multiple signal data; Performing Fourier transform on the plurality of signal data through the wifi signal processing module to generate signal frequency characteristics; Performing time domain analysis on the plurality of signal data based on the signal frequency characteristics to generate a signal time domain analysis result; Perform frequency domain analysis on the plurality of signal data based on the signal frequency characteristics to generate a signal frequency domain analysis result; Determine the plurality of signal features according to the signal time domain analysis result and the signal frequency domain analysis result; Obtaining the acceleration waveform change trend according to the signal time domain analysis result and the fall event of the target user; Determining an acceleration peak value based on a change trend of the acceleration waveform; Obtaining power spectrum density information based on the signal frequency domain analysis result and the fall event of the target user; determining frequency distribution data based on the power spectral density information; generating a plurality of fall feature labels based on the acceleration peak value and the frequency distribution data; Traversing the multiple fall feature tags to annotate the multiple signal data and generate the annotated result; Performing environmental identification and classification based on the target area environmental information to determine multiple environmental scenes; Perform a Wi-Fi signal test based on the multiple environmental scenarios to generate a signal test result, wherein the signal test result includes Wi-Fi signal strength information, Wi-Fi signal phase information, and Wi-Fi signal frequency information; Based on the wifi signal strength information, communication analysis is performed on the wifi signal phase information and the wifi signal frequency information to construct multiple communication scenarios; Randomly divide the labeled results into a training data set, a validation data set, and a test data set; The training data set and the verification data set are trained and tuned according to the multiple communication scenarios, and are evaluated according to the test data set to generate a model evaluation result and construct the fall recognition model.
2. The fall monitoring method based on wifi signal recognition as claimed in claim 1, characterized in that: The target user is monitored and identified by the fall identification model, a fall identification result is generated, and an alarm module is triggered according to the fall identification result. The method includes: Synchronize the WiFi transmission signal to the fall recognition model to monitor and identify the target user, and obtain a fall recognition result, where the fall recognition result is the first recognition result or the second recognition result; When the fall recognition result is the first recognition result, an activation instruction is generated in combination with the target area environment information, a value of 1 is assigned to the fall recognition result, and the alarm module is triggered; When the fall recognition result is the second recognition result, a continuous monitoring instruction is generated, and the continuous monitoring duration and continuous monitoring dynamic information are retrieved and analyzed according to the continuous monitoring instruction. The wifi acquisition frequency is updated according to the analysis result, and the fall recognition result is assigned a value of 0 to generate the fall recognition result.
3. The fall monitoring method based on wifi signal recognition as claimed in claim 2, characterized in that: When the fall recognition result is the first recognition result, an activation instruction is generated in combination with the target area environment information, the method comprising: When the fall recognition result is the first recognition result, analyzing the WiFi transmission signal to generate the fall posture information of the target user; Dividing the fall alarm response level based on the fall posture information and the target area environment information; The fall alarm response level is added to the activation instructions.
4. A fall monitoring system based on wifi signal recognition, characterized in that: A system for implementing the fall monitoring method based on wifi signal recognition according to any one of claims 1 to 3, comprising: A signal continuous acquisition unit, which is used to extract the target area environmental information, deploy multiple wifi acquisition points, build a wifi acquisition module, and perform continuous acquisition through the wifi acquisition module to obtain multiple signal data; A signal feature acquisition unit, the signal feature acquisition unit is used to transmit the multiple signal data to a wifi signal processing module to obtain multiple signal features; A fall recognition model building unit, the fall recognition model building unit is used to annotate the multiple signal data based on the multiple signal features in combination with the fall event of the target user, and build a fall recognition model according to the annotated results in combination with multiple communication scenarios; A fall recognition result acquisition unit, the fall recognition result acquisition unit is used to monitor and identify the target user through the fall recognition model to generate a fall recognition result; An emergency response unit is used to trigger an alarm module according to the fall recognition result and perform an emergency response based on the alarm module.
5. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the fall monitoring method based on wifi signal recognition described in any one of claims 1 to 3 are implemented.
