Radar human body falling and bed falling detection method, system and device and electronic equipment

Abstract

The invention discloses a radar human body falling and bed falling detection method, system and device and electronic equipment, and the method comprises the steps: S1, radar equipment transmits a frequency modulation continuous wave signal to a detection region, captures an echo signal reflected by the region, and converts the analog signals into digital signals containing the position, contour and motion information of a human body; s2, analyzing and processing a frame of digital signal containing the position, contour and motion information of the human body to create a frame of original radar point cloud data containing relative radar distance, radial speed, azimuth angle, pitch angle and signal intensity of a plurality of measured points; s3, screening effective radar point cloud data from each frame of original radar point cloud data; and S4, calculating a motion feature by using the multi-frame effective radar point cloud data, and detecting whether a preset motion condition is met or not. According to the invention, the calculation and equipment cost of falling and falling-off-bed detection is reduced, and the real-time performance and accuracy of falling and falling-off-bed detection are ensured by using the three-dimensional real-time point cloud features.

Description

Technical Field

[0001] This invention relates to the field of health monitoring technology, specifically to a radar method, system, device, and electronic equipment for detecting human falls and bed falls. Background Technology

[0002] With the increasing trend of population aging, the safety of elderly people living alone at home has become a major concern. Falls and bed falls are among the key threats to the life and health of elderly people living alone, and timely and effective detection of falls and bed falls is of great significance for protecting them.

[0003] Fall and bed fall detection methods based on radio frequency signals or microwave sensors used for communication such as WiFi lack the ability to distinguish human body position and contour information, and have poor anti-interference capabilities. Fall and bed fall detection methods based on sensors such as cameras and depth cameras are not suitable for indoor privacy scenarios. Millimeter-wave radar has stood out due to its advantages such as good stability, high resolution, strong anti-interference capability, and privacy protection, becoming one of the main sensor solutions for human fall and bed fall detection research.

[0004] Existing radar-based fall and bed fall detection technologies commonly employ machine learning or deep learning classification methods for fall and bed fall detection. For example, Chinese invention patent publication number CN115754967B discloses a fall detection method using deep learning networks to identify classification features, and Chinese invention patent application number 2025104843601 discloses a method for fall detection using machine learning classification. However, model-based methods such as machine learning or deep learning require storing the recognition model and demand that the computing device support machine learning or deep learning frameworks. This poses a challenge for edge embedded devices, as many edge devices often have limited memory resources and cannot support complex machine learning and deep learning inference calculations.

[0005] Another commonly used method in existing radar fall and bed fall detection technologies is to detect falls by detecting the change edge of a one-dimensional feature measured by sensors (and judging it against a specific threshold). However, methods based on change edge detection suffer from low stability, poor anti-interference ability, and limited practicality. For example, Chinese invention patent CN115343684B uses a combination of current height and average height as two conditions to detect falls, achieving an accuracy of greater than or equal to 86% and a false alarm rate of less than or equal to 20%.

[0006] Existing technologies rely solely on changes in one-dimensional features (such as a single height parameter) measured by sensors and threshold judgments for detection. This results in poor differentiation of environmental interference and non-falling human actions (such as squatting or sitting), leading to false alarms or missed detections. The limited detection dimension fails to comprehensively reflect the complex movement characteristics of falls / bed falls, resulting in low detection accuracy and failing to meet the reliability requirements of practical health monitoring. Furthermore, existing technologies lack a unified solution, making it impossible to simultaneously detect both falls and bed falls within a single method, system, or device. This necessitates the deployment of different detection schemes, increasing equipment costs and deployment complexity. Therefore, we propose a radar-based method, system, device, and electronic equipment for detecting human falls and bed falls. Summary of the Invention

[0007] The purpose of this invention is to provide a radar method, system, device, and electronic device for detecting human falls and bed falls, in order to solve the problems in the prior art.

[0008] To achieve the above objectives, the present invention provides the following technical solution: a radar method for detecting human falls and bed falls, comprising the following steps: S1. The radar equipment transmits frequency-modulated continuous wave signals to the detection area and captures the echo signals reflected from the area. Then, it converts these analog signals into digital signals containing information about the human body's position, outline, and motion. S2. Analyze and process a frame of digital signal containing human position, outline and motion information to create a frame of raw radar point cloud data containing the relative radar distance, radial velocity, azimuth angle, elevation angle and signal strength of multiple measured points. S3. Filter out valid radar point cloud data from each frame of raw radar point cloud data; S4. Calculate motion features using multiple frames of effective radar point cloud data and check whether the preset motion conditions are met; if yes, continue to step S5; if no, return to step S1. S5. Calculate fall characteristics using multiple frames of effective radar point cloud data, and perform rule matching with preset fall and bed fall conditions to determine whether a fall or bed fall event has occurred; if yes, continue to step S6; if no, return to step S1. S6 displays fall and bed fall detection results and provides real-time visual feedback.

[0009] Preferably, step S2 specifically includes the following steps: S21. Perform initial filtering on the original parameters, selecting parameters whose distance to all points is greater than 0.05m; S22. Select the corresponding spherical coordinate system to rectangular coordinate system transformation function according to the radar installation method, and transform the human body point cloud data in the spherical coordinate system measured by the radar to the rectangular coordinate system; the installation method includes side mounting and top mounting; The coordinate system transformation function for side mounting is:

[0010] The conversion function for top mounting is:

[0011] in Indicates distance, Indicates pitch angle, Indicates azimuth. Indicates the installation height of the top-mounted radar. Ground coordinates representing directions parallel to the radar antenna plane. Ground coordinates representing directions perpendicular to the radar antenna plane; The coordinates represent the height direction.

[0012] Preferably, step S3 specifically includes the following steps: S31. Perform preliminary screening on the raw radar point cloud data of the current frame, retaining data points within the radar detection range and discarding data points outside the detection range; S32. Cluster the three-dimensional spatial coordinates processed in S31 and determine whether the number of clusters is greater than 0. If yes, retain the cluster closest to the target position in the previous frame, discard other data points, and proceed to S4. If no, return to step S31.

[0013] Preferably, step S4 specifically includes the following steps: S41. Based on the data points selected in step S3, calculate the feature parameters of the current frame. The feature parameters include the number of point clouds, radial distance, radial velocity, and three-dimensional spatial coordinates. S42. Input the motion feature parameters of the current frame into a sliding window of a set frame length to form motion features; S43. Analyze the motion features in the multi-frame sliding window and check whether the preset motion conditions are met; if yes, continue to step S5; if no, return to step S1.

[0014] Preferably, in step S5, the fall and bed fall conditions are composed of multiple threshold conditions related to multi-frame point cloud data, and the conditions are determined multiple times within a multi-frame sliding window to determine whether the fall and bed fall conditions are met; the fall and bed fall conditions are described by the following formula:

[0015]

[0016] in Indicates the first frame in a multi-frame sliding window Frame number One threshold condition value, That is, the first The first frame The first fall characteristic is greater than the first. Threshold hour, The value is 1 if it is true, otherwise the value is 0. Indicates the first to the second The product of several threshold condition values; This indicates the first to the second frame in a multi-frame sliding window. Summation of the product of frame threshold condition values; This is a threshold for fall and bed fall conditions; if the value exceeds this threshold, a fall or bed fall event is detected.

[0017] Preferably, in step S5, multi-frame point cloud data is processed and analyzed through designed fall and bed fall conditions to simultaneously detect fall and bed fall events.

[0018] A radar-based human fall and bed fall detection system includes: Radar signal transceiver unit: responsible for transmitting frequency-modulated continuous wave signals with specific waveform parameters to the monitoring area, and receiving signals reflected back from the area, then mixing and filtering to obtain the difference frequency signal; Analog-to-digital conversion unit: Receives the analog difference frequency signal from the radar signal transceiver unit and converts it into a digital difference frequency signal; Signal preprocessing unit: Receives the digital difference frequency signal from the analog-to-digital conversion unit and performs preprocessing to generate raw radar point cloud data containing target human information; Feature extraction and analysis unit: Filters out effective point cloud data from the raw point cloud data generated by the signal preprocessing unit, and extracts motion features and fall features; Fall analysis and detection unit: Detects whether a fall or bed fall event has occurred by judging whether various fall characteristics meet the corresponding preset fall and bed fall conditions; Interactive Results Display Unit: Displays the results of the feature extraction and analysis unit and the fall analysis and detection unit in a visual form, while also enabling real-time interactive display of human body status, location, fall and bed fall events.

[0019] Preferably, it also includes a radar radio frequency front-end module and a processor module. The processor module includes an analog-to-digital conversion unit and a digital signal processing unit. The digital signal processing unit is used to process digital echo data, calculate three-dimensional point clouds and point cloud features, and output fall and bed fall detection results.

[0020] A radar-based device for detecting human falls and bed falls includes: Radar transmitting and receiving module: responsible for transmitting frequency-modulated continuous wave signals with specific waveform parameters to the monitoring area, receiving signals reflected back from the area, and then mixing and filtering to obtain the difference frequency signal; Digital signal processing module: Receives analog difference frequency signals from radar signal transceiver unit, converts them into digital difference frequency signals, and obtains the initial radar point cloud after preprocessing; Motion feature analysis module: Filters out valid data from the raw radar point cloud, calculates feature parameters, and extracts motion features and fall features from them; Fall detection module: Detects whether a fall or bed fall event has occurred by judging whether various fall characteristics meet the corresponding preset fall and bed fall conditions; Real-time interactive display module: Displays the results of the fall analysis and detection unit in a visual form, and realizes real-time interactive display of human body status, location, fall and bed fall events.

[0021] Preferably, it also includes a radar radio frequency front-end module and a processor module. The processor module includes an analog-to-digital conversion unit and a digital signal processing unit. The digital signal processing unit is used to process digital echo data, calculate three-dimensional point clouds and point cloud features, and output fall and bed fall detection results.

[0022] An electronic device includes a storage unit, a processing unit, a communication unit, and a display unit; the storage unit stores a computer-executable program, which, when executed, implements the radar-based method for detecting human falls and bed falls; the processing unit executes the executable program; the communication unit enables data interaction between the internal units of the electronic device and communication with external devices or networks; and the display unit displays the intermediate states and final output results during data processing in real time.

[0023] Compared with the prior art, the beneficial effects of the present invention are: Superior detection performance, strong stability and anti-interference capabilities: Employing 3D real-time point cloud feature analysis, compared to 1D feature change detection methods, it can comprehensively capture human position, contour, and motion trajectory, effectively resisting environmental interference (such as furniture occlusion and lighting changes), and solving the problems of poor stability and high false alarm rate of traditional methods. Through the design of "multi-frame sliding window and multi-threshold rule matching," it repeatedly verifies whether the features meet the fall / bed fall conditions, significantly improving detection accuracy and reducing false alarm rate, which is superior to traditional detection schemes that rely on only 2-3 single conditions.

[0024] Lower cost and wider hardware adaptability: It eliminates the need for pre-collecting large amounts of samples to train machine learning / deep learning models, and also avoids the high demands of complex models on hardware computing power and memory resources. It can be directly adapted to low-cost, compact edge embedded devices. The radar hardware adopts a frequency-modulated continuous wave system, simplifying signal processing and hardware solutions, making it easy to productize. Furthermore, it eliminates the need for dedicated, expensive processing equipment, significantly reducing deployment and promotion costs.

[0025] More comprehensive functionality and wider applicability: Breaking through existing technological limitations, it can simultaneously detect both falls and bed falls through a unified point cloud processing workflow and rule matching logic, without requiring additional equipment or modules. Radar technology does not involve image acquisition, protecting user privacy and adapting to private scenarios such as bedrooms and living rooms; it supports both side-mounted and top-mounted installation methods, flexibly adapting to different home environments and suitable for diverse user groups (especially elderly people living alone).

[0026] High real-time performance and efficient feedback: Optimized radar signal parameters (large bandwidth to improve range resolution, large frequency modulation period to improve velocity resolution, and small frame period to improve temporal resolution) enable high spatiotemporal resolution point cloud data acquisition and processing, ensuring the real-time nature of the detection process. Detection results are simultaneously visualized, enabling rapid feedback on dangerous events, buying time for subsequent rescue efforts, and meeting the immediate needs of health monitoring. Attached Figure Description

[0027] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of the present invention; Figure 2 This is a scene diagram of a radar device collecting data according to the present invention. Detailed Implementation

[0028] 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 a part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention.

[0029] Please see Figure 1-2 In this embodiment of the invention, a radar method for detecting human falls and bed falls includes the following steps: S1: The radar equipment transmits frequency-modulated continuous wave signals to the detection area and captures the echo signals reflected from the area. Then, it converts these analog signals into digital signals containing information about the human body's position, outline, and motion. S2: Analyze and process a frame of digital signal containing human position, outline and motion information to create a frame of raw radar point cloud data containing the relative radar distance, radial velocity, azimuth angle, elevation angle and signal strength of multiple measured points; S3: Filter out valid radar point cloud data from each frame of raw radar point cloud data; S4: Calculate motion features using multiple frames of effective radar point cloud data and check whether the preset motion conditions are met. If yes, continue to step S5; otherwise, return to step S1. S5: Calculate fall characteristics using multiple frames of effective radar point cloud data, and perform rule matching with preset fall and bed fall conditions to determine whether a fall or bed fall event has occurred. If yes, continue to step S6; otherwise, return to step S1. S6: Displays fall and bed fall detection results and provides real-time visual feedback.

[0030] The raw parameters acquired by the radar, namely target range, azimuth, and elevation angle, are converted into three-dimensional rectangular coordinates using a function. Step S2 includes the following steps: S21: Perform initial filtering on the original parameters, selecting parameters whose distance to all points is greater than 0.05m; S22: Depending on the radar installation method (side-mounted or top-mounted), different spherical coordinate system to rectangular coordinate system transformation functions are used to transform the human body point cloud data in the spherical coordinate system measured by the radar to the rectangular coordinate system.

[0031] The coordinate system transformation function for side mounting is as follows:

[0032] The conversion function for the top-mounted device is as follows:

[0033] in Indicates distance, Indicates pitch angle, Indicates azimuth. Indicates the installation height of the top-mounted radar. Ground coordinates representing directions parallel to the radar antenna plane. Ground coordinates representing directions perpendicular to the radar antenna plane; The coordinates represent the height direction.

[0034] S23: Filter the transformed 3D coordinates again to remove noise points that are outside the detection range, and obtain initial radar point cloud data containing human features.

[0035] To filter valid point cloud data, step S3 includes the following steps: S31: Determine whether each data point in the original radar point cloud of the current frame is within the radar detection area. If yes, keep it; otherwise, discard it. S32: Cluster the three-dimensional spatial coordinates processed in S31, and determine whether the number of clusters is greater than 0. If so, retain the cluster closest to the target position in the previous frame, discard other data points, and proceed to S4. If not, return to step S31.

[0036] To extract the motion features of the target point cloud, step S4 includes the following steps: S41: Based on the data points selected in step S3, calculate the feature parameters of the current frame. These parameters include, but are not limited to, the number of point clouds, radial distance, radial velocity, and three-dimensional spatial coordinates.

[0037] S42: Input the motion feature parameters of the current frame into a sliding window of a certain frame length to form motion features.

[0038] S43: Analyze the motion features in the multi-frame sliding window and check whether the preset motion conditions are met. If the preset motion conditions are met, continue to step S5; if the preset motion conditions are not met, return to step S1 and start again.

[0039] For the fall feature in parameter calculation step S5, which includes motion features, check whether the preset fall and bed fall conditions are met.

[0040] Through extensive testing, we identified key parameters that distinguish falls and bed falls from other actions, such as height change, maximum current height, and current speed change, while also determining the threshold ranges for these parameters. Considering the data differences caused by different installation methods, we designed tests for both side-mounted and top-mounted installations, employing different threshold judgment criteria.

[0041] Taking a side-mounted radar as an example: The radar equipment is placed vertically on a support 2m above the ground. The movement range of the target human body is controlled within a semicircle with a radius of 3m centered on the radar. Data on various actions such as walking, squatting, sitting, lying down, falling, turning over in bed, and falling out of bed are collected and tested.

[0042] The effective point cloud data in the filtered multi-frame sliding window is processed to determine the height change value of each frame. Maximum height Maximum radial velocity Do the parameters meet the threshold? If three consecutive frames simultaneously meet the threshold condition, it is determined that a fall or bed fall event has occurred.

[0043] A radar-based human fall and bed fall detection system includes: Radar signal transceiver unit: responsible for transmitting frequency-modulated continuous wave signals with specific waveform parameters to the monitoring area, and receiving signals reflected back from the area, then mixing and filtering to obtain the difference frequency signal; Analog-to-digital conversion unit: Receives the analog difference frequency signal from the radar signal transceiver unit and converts it into a digital difference frequency signal; Signal preprocessing unit: Receives the digital difference frequency signal from the analog-to-digital conversion unit and performs preprocessing to generate raw radar point cloud data containing target human information; Feature extraction and analysis unit: Filters out effective point cloud data from the raw point cloud data generated by the signal preprocessing unit, and extracts motion features and fall features; Fall analysis and detection unit: Detects whether a fall or bed fall event has occurred by judging whether various fall characteristics meet the corresponding preset fall and bed fall conditions; Interactive Results Display Unit: Displays the results of the feature extraction and analysis unit and the fall analysis and detection unit in a visual form, while also enabling real-time interactive display of human body status, location, fall and bed fall events.

[0044] Preferably, it also includes a radar radio frequency front-end module and a processor module. The processor module includes an analog-to-digital conversion unit and a digital signal processing unit. The digital signal processing unit is used to process digital echo data, calculate three-dimensional point clouds and point cloud features, and output fall and bed fall detection results.

[0045] A radar-based device for detecting human falls and bed falls includes: Radar transmitting and receiving module: responsible for transmitting frequency-modulated continuous wave signals with specific waveform parameters to the monitoring area, receiving signals reflected back from the area, and then mixing and filtering to obtain the difference frequency signal; Digital signal processing module: Receives analog difference frequency signals from radar signal transceiver unit, converts them into digital difference frequency signals, and obtains the initial radar point cloud after preprocessing; Motion feature analysis module: Filters out valid data from the raw radar point cloud, calculates feature parameters, and extracts motion features and fall features from them; Fall detection module: Detects whether a fall or bed fall event has occurred by judging whether various fall characteristics meet the corresponding preset fall and bed fall conditions; Real-time interactive display module: Displays the results of the fall analysis and detection unit in a visual form, and realizes real-time interactive display of human body status, location, fall and bed fall events.

[0046] Preferably, it also includes a radar radio frequency front-end module and a processor module. The processor module includes an analog-to-digital conversion unit and a digital signal processing unit. The digital signal processing unit is used to process digital echo data, calculate three-dimensional point clouds and point cloud features, and output fall and bed fall detection results.

[0047] An electronic device includes a storage unit, a processing unit, a communication unit, and a display unit; the storage unit stores a computer-executable program, which, when executed, implements the radar-based method for detecting human falls and bed falls; the processing unit executes the executable program; the communication unit enables data interaction between the internal units of the electronic device and communication with external devices or networks; and the display unit displays the intermediate states and final output results during data processing in real time.

[0048] The working principle of this invention is: Signal acquisition and conversion: The radar equipment transmits frequency-modulated continuous wave signals to cover the monitoring area, receives the echo signals reflected by the human body, and converts the analog signals into digital signals containing information on the human body's position, outline, and movement, providing raw data support for subsequent analysis.

[0049] Point cloud data generation and optimization: The digital signal is processed to generate raw radar point cloud data containing range, radial velocity, azimuth, elevation angle, and signal strength. The clean initial point cloud data is obtained through three steps: "range filtering (retaining data points > 0.05m), coordinate transformation (side-mounted / top-mounted adaptation of the corresponding spherical coordinate system to rectangular coordinate system transformation function), and noise filtering". Then, the effective point cloud is filtered through "range filtering, cluster analysis, and target tracking" (retaining the cluster closest to the target in the previous frame) to eliminate interference data.

[0050] Feature extraction and condition judgment: Based on multiple frames of effective point cloud data, the feature parameters such as the number of point clouds, radial distance, and three-dimensional coordinates are calculated, and a motion feature sequence is constructed through a sliding window. First, it is judged whether the preset motion conditions are met (preliminary screening of suspected dangerous actions), and then the fall features (such as height change value, speed change value, etc.) are calculated. The rule matching formula of multiple threshold combinations is used to quantify whether a fall / bed fall event has occurred, without relying on machine learning models or template libraries.

[0051] Results visualization feedback: The test results are displayed in real time, and the status, location and event warnings of the human body are presented simultaneously to achieve dynamic monitoring.

[0052] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A radar method for detecting human falls and bed falls, characterized in that, Includes the following steps: S1. The radar equipment transmits frequency-modulated continuous wave signals to the detection area and captures the echo signals reflected from the area. Then, it converts these analog signals into digital signals containing information about the human body's position, outline, and motion. S2. Analyze and process a frame of digital signal containing human position, outline and motion information to create a frame of raw radar point cloud data containing the relative radar distance, radial velocity, azimuth angle, elevation angle and signal strength of multiple measured points. S3. Filter out valid radar point cloud data from each frame of raw radar point cloud data; S4. Calculate motion features using multiple frames of effective radar point cloud data, and detect whether the preset motion conditions are met; if so, continue to step S5. If not, return to step S1; S5. Calculate fall characteristics using multiple frames of effective radar point cloud data, and perform rule matching with preset fall and bed fall conditions to determine whether a fall or bed fall event has occurred; if yes, continue to step S6; if no, return to step S1. S6 displays fall and bed fall detection results and provides real-time visual feedback.

2. The radar method for detecting human falls and bed falls according to claim 1, characterized in that, S2 specifically includes the following steps: S21. Perform initial filtering on the original parameters, selecting parameters whose distance to all points is greater than 0.05m; S22. Select the corresponding spherical coordinate system to rectangular coordinate system transformation function according to the radar installation method, and transform the human body point cloud data in the spherical coordinate system measured by the radar to the rectangular coordinate system; the installation method includes side mounting and top mounting; The coordinate system transformation function for side mounting is: ; The conversion function for top mounting is: ; in Indicates distance, Indicates pitch angle, Indicates azimuth. Indicates the installation height of the top-mounted radar. Ground coordinates representing directions parallel to the radar antenna plane. Ground coordinates representing directions perpendicular to the radar antenna plane; The coordinates represent the height direction.

3. The radar method for detecting human falls and bed falls according to claim 1, characterized in that, S3 specifically includes the following steps: S31. Perform preliminary screening on the raw radar point cloud data of the current frame, retaining data points within the radar detection range and discarding data points outside the detection range; S32. Cluster the three-dimensional spatial coordinates processed in S31 and determine whether the number of clusters is greater than 0. If yes, retain the cluster closest to the target position in the previous frame, discard other data points, and proceed to S4. If no, return to step S31.

4. The radar method for detecting human falls and bed falls according to claim 1, characterized in that, S4 specifically includes the following steps: S41. Based on the data points selected in step S3, calculate the feature parameters of the current frame. The feature parameters include the number of point clouds, radial distance, radial velocity, and three-dimensional spatial coordinates. S42. Input the motion feature parameters of the current frame into a sliding window of a set frame length to form motion features; S43. Analyze the motion features in the multi-frame sliding window and check whether the preset motion conditions are met; if yes, continue to step S5; if no, return to step S1.

5. The radar method for detecting human falls and bed falls according to claim 1, characterized in that, In S5, the fall and bed fall conditions are composed of multiple threshold conditions related to multi-frame point cloud data. The conditions for fall and bed fall are determined multiple times in a multi-frame sliding window. The conditions for falls and bed falls are described by the following formula: ; ； in Indicates the first frame in a multi-frame sliding window Frame number One threshold condition value, That is, the first The first frame The first fall characteristic is greater than the first. Threshold hour, The value is 1 if it is true, otherwise the value is 0. Indicates the first to the second The product of several threshold condition values; This indicates the first to the second frame in a multi-frame sliding window. Summation of the product of frame threshold condition values; This is a threshold for fall and bed fall conditions; if the value exceeds this threshold, a fall or bed fall event is detected. In S5, multi-frame point cloud data is processed and analyzed through designed fall and bed fall conditions, and fall and bed fall event detection is realized simultaneously.

6. A radar-based system for detecting human falls and bed falls, characterized in that, include: Radar signal transceiver unit: responsible for transmitting frequency-modulated continuous wave signals with specific waveform parameters to the monitoring area, and receiving signals reflected back from the area, then mixing and filtering to obtain the difference frequency signal; Analog-to-digital conversion unit: Receives the analog difference frequency signal from the radar signal transceiver unit and converts it into a digital difference frequency signal; Signal preprocessing unit: Receives the digital difference frequency signal from the analog-to-digital conversion unit and performs preprocessing to generate raw radar point cloud data containing target human information; Feature extraction and analysis unit: Filters out effective point cloud data from the raw point cloud data generated by the signal preprocessing unit, and extracts motion features and fall features; Fall analysis and detection unit: Detects whether a fall or bed fall event has occurred by judging whether various fall characteristics meet the corresponding preset fall and bed fall conditions; Interactive Results Display Unit: Displays the results of the feature extraction and analysis unit and the fall analysis and detection unit in a visual form, while also enabling real-time interactive display of human body status, location, fall and bed fall events.

7. A radar-based human fall and bed fall detection system according to claim 6, characterized in that, It also includes a radar radio frequency front-end module and a processor module. The processor module includes an analog-to-digital conversion unit and a digital signal processing unit. The digital signal processing unit is used to process digital echo data, calculate three-dimensional point clouds and point cloud features, and output fall and bed fall detection results.

8. A radar-based device for detecting human falls and bed falls, characterized in that, include: Radar transmitting and receiving module: responsible for transmitting frequency-modulated continuous wave signals with specific waveform parameters to the monitoring area, receiving signals reflected back from the area, and then mixing and filtering to obtain the difference frequency signal; Digital signal processing module: Receives analog difference frequency signals from radar signal transceiver unit, converts them into digital difference frequency signals, and obtains the initial radar point cloud after preprocessing; Motion feature analysis module: Filters out valid data from the raw radar point cloud, calculates feature parameters, and extracts motion features and fall features from them; Fall detection module: Detects whether a fall or bed fall event has occurred by judging whether various fall characteristics meet the corresponding preset fall and bed fall conditions; Real-time interactive display module: Displays the results of the fall analysis and detection unit in a visual form, and realizes real-time interactive display of human body status, location, fall and bed fall events.

9. A radar-based human fall and bed fall detection device according to claim 8, characterized in that, It also includes a radar radio frequency front-end module and a processor module. The processor module includes an analog-to-digital conversion unit and a digital signal processing unit. The digital signal processing unit is used to process digital echo data, calculate three-dimensional point clouds and point cloud features, and output fall and bed fall detection results.

10. An electronic device, characterized in that, It includes a storage unit, a processing unit, a communication unit, and a display unit; the storage unit is used to store a computer-executable program, which, when executed, implements the radar-based human fall and bed fall detection method as described in any one of claims 1-5; the processing unit is used to execute the executable program; the communication unit is used to realize data interaction between the various units within the electronic device, as well as communication with external devices or networks; The display unit is used to display the intermediate states and final output results during the data processing process in real time.

Images