A home elevator fall detection method and system based on 4D millimeter wave radar

Abstract

This application relates to the field of elevator monitoring technology and discloses a method and system for detecting falls in home elevators based on a 4D millimeter-wave radar. The method includes acquiring a radar echo signal of any target object in the elevator in real time; determining the static point cloud and dynamic point cloud information of the target object; calculating the mean of all dynamic point clouds of the target object, and at the same time, forming a target tracking trajectory based on the dynamic point cloud, selecting a dynamic point cloud with a height value greater than the target object to solve the mean; when the means of both dynamic point clouds continuously and monotonically decrease within a preset range, and the minimum means of both dynamic point clouds are less than a first threshold, calculating the height mean of the static point cloud of the target object; when the height mean of the static point cloud of the target object is less than a second threshold for a long time, determining that the target object has fallen; when it is determined that the target object has fallen, the radar sends a person-trapped signal to the elevator master control. This application improves the detection accuracy of people falling in elevators and has better environmental adaptability.

Description

Technical Field

[0001] The present application relates to the field of elevator monitoring technology, and in particular to a method and system for detecting falls in home elevators based on 4D millimeter-wave radar. Background Art

[0002] With the arrival of an aging society, especially with the increasing number of elderly people living alone and equipped with home elevators, the consequences of falling or being trapped in the elevator are disastrous. Therefore, the safety of home elevators has become increasingly important.

[0003] Existing elevator safety monitoring systems utilize cameras and pressure sensors. However, in low-light or complex environments, camera imaging quality is poor and environmental adaptability is limited. Furthermore, camera image monitoring poses the risk of privacy leaks, which can easily cause user concerns. Pressure sensors also struggle to accurately distinguish between falls and other behaviors, resulting in a high false alarm rate and often preventing falls from being identified and effectively addressed.

[0004] With respect to the above-mentioned related technologies, the inventors have found that the existing elevator safety monitoring system has a high false alarm rate and is unable to identify falls in a timely manner. Summary of the Invention

[0005] In order to improve the detection accuracy of fall behavior and identify fall behavior in a timely manner, the present application provides a home elevator personnel fall detection method and system based on 4D millimeter wave radar.

[0006] In a first aspect, the present application provides a method for detecting falls in a home elevator based on a 4D millimeter-wave radar.

[0007] This application is achieved through the following technical solutions:

[0008] A method for detecting falls in home elevators based on 4D millimeter-wave radar comprises the following steps:

[0009] Utilize MIMO technology to acquire radar echo signals from any target object in the elevator in real time through multiple antenna channels;

[0010] Determining the three-dimensional coordinates, velocity, and energy values ​​of a static point cloud and a dynamic point cloud of the target object based on the radar echo signal;

[0011] Calculate the mean of the dynamic point clouds of all radar echo signals using the three-dimensional coordinates, velocity, and energy values ​​of the static point cloud and dynamic point cloud of the target object. At the same time, form a target tracking trajectory based on the dynamic point cloud, and select the dynamic point cloud with a height value greater than that of the target object to solve the mean value.

[0012] Determine whether the mean values ​​of the two dynamic point clouds are both continuously and monotonically decreasing within a preset range, and determine whether the minimum values ​​of the mean values ​​of the two dynamic point clouds are both less than a set first threshold;

[0013] When the means of the two dynamic point clouds are both continuously and monotonically decreasing within a preset range, and the minimum means of the two dynamic point clouds are both less than the first threshold, then calculating the height mean of the static point cloud of the target object;

[0014] Determining whether the average height of the static point cloud of the target object is lower than a set second threshold within a preset range;

[0015] If the average height of the static point cloud of the target object is lower than the second threshold within a preset range and the duration exceeds a preset time threshold, the state information of the target object is determined to be a fall.

[0016] In a preferred example, the present application may be further configured as follows: the step of determining the three-dimensional coordinates, velocity and energy value of the static point cloud and the dynamic point cloud of the target object based on the radar echo signal includes:

[0017] After performing a one-dimensional fast Fourier transform on the radar echo signal of the current frame, the mean of the one-dimensional fast Fourier transform of all chirp signals of the radar echo signal is subtracted to obtain the first signal of the dynamic point cloud;

[0018] Performing multi-frame progression and storage on the first signal of the dynamic point cloud to obtain the first signal of the static point cloud;

[0019] Performing a two-dimensional fast Fourier transform on the first signals of the acquired static point cloud and the dynamic point cloud to obtain range Doppler maps of the static point cloud and the dynamic point cloud;

[0020] Based on the acquired range Doppler maps of the static point cloud and the dynamic point cloud, combined with the radar echo signals of all antenna channels, the three-dimensional coordinates, velocity and energy values ​​of the static point cloud and the dynamic point cloud of the target object are determined.

[0021] In a preferred example, the present application can be further configured as follows: based on the acquired range Doppler maps of the static point cloud and the dynamic point cloud, combined with the radar echo signals of all antenna channels, the steps of determining the three-dimensional coordinates, velocity, and energy value of the static point cloud and the dynamic point cloud of the target object include:

[0022] Obtaining a range Doppler map of a dynamic point cloud from one of the antenna channels of the radar echo signal to perform one-dimensional CFAR detection to obtain the distance and speed of the target object;

[0023] Then, the range Doppler maps of the dynamic point clouds of all antenna channels are subjected to three-dimensional fast Fourier transform and four-dimensional fast Fourier transform along the two antenna directions to obtain the azimuth and elevation angles of the dynamic point cloud of the target object;

[0024] and performing a three-dimensional fast Fourier transform and a four-dimensional fast Fourier transform on the range Doppler maps of the static point clouds of all antenna channels along two antenna directions, respectively, to obtain the azimuth and elevation angles of the static point cloud of the target object;

[0025] The three-dimensional coordinates, velocity and energy values ​​of the static point cloud and the dynamic point cloud of the target object are determined according to the distance, velocity, azimuth and pitch angle of the target object.

[0026] In a preferred example, the present application can be further configured as follows:

[0027] Sending the status information of the target object to the elevator main control system;

[0028] When the status information of the target object is falling, the elevator main control system sends a door opening signal to the elevator door opening and closing system to control the elevator door to open.

[0029] In a preferred example, the present application can be further configured as follows:

[0030] The elevator master control system sends a distress signal to the distress device, and starts dialing an emergency call.

[0031] In a preferred example, the present application may be further configured as follows: after determining that the state information of the target object is a fall, the step further includes:

[0032] Counting the duration of the falling state of the target object;

[0033] If the duration exceeds a preset time threshold, a person-trapped signal is sent to the elevator main control system;

[0034] The elevator master control system triggers an instruction to pause the elevator operation and sends a door opening signal to the elevator door opening and closing system, calls the management personnel, and automatically dials the preset emergency contact phone number that matches the target object.

[0035] In a second aspect, the present application provides a home elevator personnel fall detection system based on 4D millimeter wave radar.

[0036] This application is achieved through the following technical solutions:

[0037] A home elevator fall detection system based on 4D millimeter wave radar, including:

[0038] The data acquisition module is used to use MIMO technology to obtain radar echo signals from any target object in the elevator in real time through multiple antenna channels;

[0039] A dynamic and static point cloud extraction module, configured to determine the three-dimensional coordinates, velocity, and energy values ​​of the static point cloud and the dynamic point cloud of the target object based on the radar echo signal;

[0040] a dynamic point cloud processing module, configured to calculate the mean of the dynamic point clouds of all radar echo signals using the three-dimensional coordinates, velocity, and energy values ​​of the static point cloud and the dynamic point cloud of the target object; and, at the same time, forming a target tracking trajectory based on the dynamic point cloud, selecting the dynamic point cloud with a height greater than that of the target object to calculate the mean value;

[0041] A dynamic point cloud analysis module is used to determine whether the mean values ​​of the two dynamic point clouds are continuously and monotonically decreasing within a preset range, and to determine whether the minimum values ​​of the mean values ​​of the two dynamic point clouds are both less than a set first threshold;

[0042] a static point cloud processing module, configured to calculate a height mean of the static point cloud of the target object when the means of the two dynamic point clouds are continuously and monotonically decreasing within a preset range and the minimum means of the two dynamic point clouds are both less than the first threshold;

[0043] A static point cloud analysis module, configured to determine whether a height average of the static point cloud of the target object is lower than a set second threshold within a preset range;

[0044] The fall detection module is configured to determine that the state information of the target object is a fall when the average height of the static point cloud of the target object is lower than the second threshold within a preset range and the duration exceeds a preset time threshold.

[0045] In a third aspect, the present application provides a computer device.

[0046] This application is achieved through the following technical solutions:

[0047] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the processor implements the steps of any one of the above-mentioned methods for detecting falls in a home elevator based on a 4D millimeter-wave radar.

[0048] In a fourth aspect, the present application provides a computer-readable storage medium.

[0049] This application is achieved through the following technical solutions:

[0050] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of any one of the above-mentioned methods for detecting a person falling in a home elevator based on a 4D millimeter-wave radar.

[0051] In a fifth aspect, the present application provides a computer program product.

[0052] This application is achieved through the following technical solutions:

[0053] A computer program product includes a computer program, which, when executed by a processor, implements the steps of any one of the above-mentioned methods for detecting people falling in a home elevator based on 4D millimeter-wave radar.

[0054] In summary, compared with the prior art, the technical solution provided by this application has at least the following beneficial effects:

[0055] Utilizing MIMO technology, radar echo signals from any target object in the elevator are acquired in real time through multiple antenna channels, so as to monitor the real-time movements of people in the elevator through 4D millimeter-wave radar technology. 4D millimeter-wave radar is not affected by changes in light and can still work stably in dim light or complex environments. It can also penetrate obstacles and can better adapt to the small and complex space in the elevator. Compared with traditional camera monitoring systems, millimeter-wave radar technology performs contactless monitoring, avoiding the risk of privacy leakage and ensuring the protection of personnel privacy. At the same time, 4D millimeter-wave radar can quickly collect target data with better real-time performance. Based on the radar echo signal, the three-dimensional coordinates, speed and energy value of the static point cloud and dynamic point cloud of the target object are determined. By combining the dynamic point cloud and static point cloud data processing, the behavioral characteristics of people in the elevator are accurately extracted from multiple dimensions, which is conducive to more accurate judgment of whether a fall has occurred. At the same time, the processing of point cloud data is more efficient, which is conducive to more rapid judgment of whether a fall has occurred in the elevator. The three-dimensional coordinates, speed and energy value of the static point cloud and dynamic point cloud of the target object are calculated. The method comprises the following steps: calculating the mean of the dynamic point cloud of the wave signal, and forming a target tracking trajectory according to the dynamic point cloud, selecting a dynamic point cloud with a larger height value than the target object tracking to solve the mean; judging whether the means of the two dynamic point clouds are continuously and monotonically decreasing within the preset range, and judging whether the minimum means of the two dynamic point clouds are both less than the set first threshold; when the means of the two dynamic point clouds are continuously and monotonically decreasing within the preset range, or the minimum means of the two dynamic point clouds are both less than the first threshold, calculating the height mean of the static point cloud of the target object; judging whether the height mean of the static point cloud of the target object is lower than the set second threshold within the preset range; if the height mean of the static point cloud of the target object is lower than the second threshold within the preset range, determining that the state information of the target object is a fall, and performing behavior analysis by fusing dynamic point cloud and static point cloud data, detecting the height of the static point cloud and the movement change of the dynamic point cloud to identify the fall behavior; thereby improving the detection accuracy of the fall behavior of people in the elevator, having better environmental adaptability, being able to identify the fall behavior in time, and facilitating timely response and processing, and having better privacy than other sensors. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 A schematic diagram of the main process of a method for detecting falls in a home elevator based on 4D millimeter-wave radar is provided as an exemplary embodiment of the present application.

[0057] Figure 2 A dynamic and static point cloud data analysis flowchart of a home elevator personnel fall detection method based on 4D millimeter wave radar is provided as another exemplary embodiment of the present application.

[0058] Figure 3A structural block diagram of a home elevator personnel fall detection system based on 4D millimeter-wave radar is provided as another exemplary embodiment of the present application.

[0059] Figure 4 A schematic diagram of a 4D millimeter-wave radar installation for a method for detecting a person falling in a home elevator based on a 4D millimeter-wave radar is provided as an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0060] This specific embodiment is merely an explanation of the present application and is not a limitation of the present application. After reading this specification, those skilled in the art may make non-creative modifications to the present embodiment as needed, but as long as they are within the scope of the claims of the present application, they are protected by the patent law.

[0061] To make the purpose, technical solutions, and advantages of the embodiments of this application more clear, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0062] In this document, the term "and / or" simply describes a relationship between related objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A exists alone, A and B exist simultaneously, or B exists alone. Furthermore, the character " / " in this document, unless otherwise specified, generally indicates an "or" relationship between the related objects.

[0063] In order to ensure the safety of the elderly and people with mobility difficulties when riding in elevators, this application uses 4D millimeter-wave radar technology to monitor the real-time movements of people in the elevator, and integrates dynamic point cloud and static point cloud data for behavioral analysis, automatically identifying falls and handling them in a timely manner, providing a new, accurate, efficient and privacy-protected elevator safety monitoring technology, improving the accuracy and real-time performance of fall detection, and thus significantly improving the safety of home elevators.

[0064] An embodiment of the present application provides a method for detecting falls in a home elevator based on a 4D millimeter-wave radar. The main steps of the method are described as follows.

[0065] Utilize MIMO technology to acquire radar echo signals from any target object in the elevator in real time through multiple antenna channels;

[0066] Determining the three-dimensional coordinates, velocity, and energy values ​​of a static point cloud and a dynamic point cloud of the target object based on the radar echo signal;

[0067] Calculate the mean of the dynamic point clouds of all radar echo signals using the three-dimensional coordinates, velocity, and energy values ​​of the static point cloud and dynamic point cloud of the target object. At the same time, form a target tracking trajectory based on the dynamic point cloud, and select the dynamic point cloud with a height value greater than that of the target object to solve the mean value.

[0068] Determine whether the mean values ​​of the two dynamic point clouds are both continuously and monotonically decreasing within a preset range, and determine whether the minimum values ​​of the mean values ​​of the two dynamic point clouds are both less than a set first threshold;

[0069] When the means of the two dynamic point clouds are both continuously and monotonically decreasing within a preset range, and the minimum means of the two dynamic point clouds are both less than the first threshold, then calculating the height mean of the static point cloud of the target object;

[0070] Determining whether the average height of the static point cloud of the target object is lower than a set second threshold within a preset range;

[0071] If the average height of the static point cloud of the target object is lower than the second threshold within a preset range and the duration exceeds a preset time threshold, the state information of the target object is determined to be a fall.

[0072] In one embodiment, after determining that the state information of the target object is a fall, the method further includes:

[0073] Counting the duration of the falling state of the target object;

[0074] If the duration exceeds a preset time threshold, a person-trapped signal is sent to the elevator main control system;

[0075] The elevator master control system triggers an instruction to pause the elevator operation and sends a door opening signal to the elevator door opening and closing system, calls the management personnel, and automatically dials the preset emergency contact phone number that matches the target object.

[0076] In one embodiment, a method for detecting falls in a home elevator based on a 4D millimeter-wave radar further includes the following steps:

[0077] Sending the status information of the target object to the elevator main control system;

[0078] When the status information of the target object is falling, the elevator main control system sends a door opening signal to the elevator door opening and closing system to control the elevator door to open.

[0079] In one embodiment, the following steps are also included:

[0080] The elevator master control system sends a distress signal to the distress device, and starts dialing an emergency call.

[0081] The embodiments of the present application are described in further detail below with reference to the accompanying drawings.

[0082] Reference Figure 1 Utilizing MIMO technology, the 4D millimeter-wave radar collects the original echo signal of any target object in the elevator car and sends it to the fall control system controller for signal processing. The intermediate frequency signal is then sampled to obtain the original ADC time domain data, thereby obtaining the intermediate frequency ADC data (digital signal). The sampling point is set to 512, the chirp is set to 32, and the 4D millimeter-wave radar uses a four-transmitter, four-receiver radar configuration.

[0083] MIMO (Multiple-Input, Multiple-Output) technology uses multiple signal splitting to allow data to be transmitted simultaneously via multiple transmitting antennas, using different reflection or penetration paths. The receiving end then simultaneously receives the sender's echo signals through multiple receiving antennas and uses DSP to reconstruct the data, recombining the split data based on factors such as time differences to quickly restore the complete data. MIMO technology increases data transmission speeds within wireless network spectrum, improving real-time performance, while also significantly improving spectrum utilization without occupying additional spectrum. More importantly, it also increases signal reception distance.

[0084] Through high-frequency sampling and multi-channel reception, the fall detection system can obtain rich time domain data, providing an accurate and effective data basis for subsequent signal processing and fall analysis.

[0085] In one embodiment, before signal processing is performed on the radar echo signal, a high-pass filter and a low-pass filter may be used in advance to perform denoising and filtering operations to improve data quality and signal availability.

[0086] The radar data is transmitted in 100ms frames, and each frame outputs both static and dynamic point clouds. The point cloud data of the human body can be obtained through signal processing of the sampled data.

[0087] Based on the radar echo signal, the three-dimensional coordinates, velocity and energy value of the static point cloud and dynamic point cloud of the target object are determined.

[0088] In one embodiment, the step of determining the three-dimensional coordinates, velocity, and energy value of the static point cloud and the dynamic point cloud of the target object based on the radar echo signal includes:

[0089] After performing a one-dimensional fast Fourier transform on the radar echo signal of the current frame, the mean of the one-dimensional fast Fourier transform of all chirp signals of the radar echo signal is subtracted to obtain the first signal of the dynamic point cloud;

[0090] Performing multi-frame progression and storage on the first signal of the dynamic point cloud to obtain the first signal of the static point cloud;

[0091] Performing a two-dimensional fast Fourier transform on the first signals of the acquired static point cloud and the dynamic point cloud to obtain range Doppler maps of the static point cloud and the dynamic point cloud;

[0092] Based on the acquired range Doppler maps of the static point cloud and the dynamic point cloud, combined with the radar echo signals of all antenna channels, the three-dimensional coordinates, velocity and energy values ​​of the static point cloud and the dynamic point cloud of the target object are determined.

[0093] For dynamic point clouds, the one-dimensional FFT mean is stored every two frames, and is stored continuously for 32 times.

[0094] For static point clouds, the mean of the continuously stored one-dimensional FFTs of the dynamic point clouds is used as the one-dimensional FFT of the static point clouds, and is stored progressively once per frame.

[0095] The ADC data is processed with a one-dimensional FFT, the frame mean is subtracted, and then a two-dimensional FFT is performed to obtain the RD map (Range Doppler map). The frame mean is calculated by adding the energy values ​​of all chirps and dividing it by the total number of chirps to reflect the average intensity of the current frame.

[0096] In one embodiment, the step of determining the three-dimensional coordinates, velocity, and energy value of the static point cloud and the dynamic point cloud of the target object based on the acquired range Doppler maps of the static point cloud and the dynamic point cloud in combination with the radar echo signals of all antenna channels includes:

[0097] A range Doppler map of a dynamic point cloud is obtained from one of the antenna channels of the radar echo signal, and one-dimensional CFAR detection is performed to obtain the distance and velocity of the target object. One-dimensional CFAR (Constant False Alarm Rate) is an algorithm used to detect the presence of targets in radar echo signals. It can adaptively adjust the detection threshold to maintain a constant false alarm rate. It can operate in different environments and noise levels, improving the robustness of target detection. The details are as follows:

[0098] The target radar range Doppler map is divided into multiple range units, and multiple range units are selected as protection units to prevent the target signal from affecting the threshold calculation. The average background noise intensity of the area around the protection unit is calculated and used to design a constant as the threshold value (threshold factor). The signal strength of each range unit is compared with the corresponding threshold value. If it exceeds the threshold value, it is considered to be a target signal. The range index and velocity index of the target signal are calculated, and then the range and velocity of the target object are obtained based on the range resolution and velocity resolution.

[0099] Then, the range Doppler maps of the dynamic point clouds of all antenna channels are subjected to three-dimensional fast Fourier transform and four-dimensional fast Fourier transform along the two antenna directions to obtain the azimuth and elevation angles of the dynamic point cloud of the target object;

[0100] and performing a three-dimensional fast Fourier transform and a four-dimensional fast Fourier transform on the range Doppler maps of the static point clouds of all antenna channels along two antenna directions, respectively, to obtain the azimuth and elevation angles of the static point cloud of the target object;

[0101] The three-dimensional coordinates, velocity and energy values ​​of the static point cloud and the dynamic point cloud of the target object are determined according to the distance, velocity, azimuth and pitch angle of the target object.

[0102] By performing a one-dimensional CFAR on the RD map of an antenna channel, we can obtain the distance and velocity information of the target object's radar point cloud. Then, we perform a three-dimensional FFT and a four-dimensional FFT on each radar point cloud of the target object along the two antenna directions to determine the azimuth and elevation angles of the dynamic and static point clouds. Using the distance, velocity, azimuth, and elevation information of the target object's radar point cloud, we can determine the three-dimensional coordinates, velocity, and energy values ​​of the static and dynamic point clouds of the target object. The specific solution process is as follows.

[0103] Find the range gate index n corresponding to the target peak according to the distance spectrum after FFT R , calculate the target distance:

[0104] R=n R ·ΔR (1)

[0105] Where ΔR is the range resolution, defined as:

[0106]

[0107] Find the Doppler gate index n corresponding to the target peak according to the Doppler spectrum after FFT v , calculate the target speed:

[0108] v=n v ·Δv (3)

[0109] Velocity resolution (Δv) describes the minimum speed interval that the radar system can distinguish when measuring speed. The calculation formula is as follows:

[0110]

[0111] Where T c is the duration of a single Chirp, N chirpIt is the number of chirps in a frame. The static point cloud is obtained by increasing the duration of a single chirp, thereby reducing the velocity resolution and making it more sensitive to target recognition.

[0112] The three-dimensional coordinates (x, y, z) can be calculated by the distance R, azimuth angle φ and pitch angle θ of the target point cloud. The calculation formula is as follows:

[0113] x=R·cos(φ)·cos(θ) (5)

[0114] y=R·cos(φ)·sin(θ) (6)

[0115] z=R·sin(φ) (7)

[0116] The three-dimensional coordinates (x, y, z) of the dynamic point cloud and the static point cloud can be obtained by the above formula.

[0117] Finally, based on the three-dimensional coordinates and velocity, the energy values ​​of the static point cloud and dynamic point cloud of the target object can be determined.

[0118] By obtaining static point cloud and dynamic point cloud data of the human body, the static point cloud reflects the posture of the person after stabilization, while the dynamic point cloud captures the changing process of the person's movement. The combination of the two can provide richer information for the accurate identification of falling behavior and provide accurate data support for subsequent behavior analysis and fall judgment.

[0119] In one embodiment, by setting the thresholds of the CFAR and maximum value search algorithms, interference objects in the environment can be filtered out, further improving data quality and effectiveness.

[0120] Reference Figure 2 , perform fall detection on static point cloud and dynamic point cloud data of the target object, including:

[0121] Obtain dynamic point cloud data, static point cloud data and height value of the target object (i.e. tracking target data);

[0122] Based on the three-dimensional coordinates, velocity, and energy values ​​of the static and dynamic point clouds of the target object, the mean of the dynamic point clouds of all radar echo signals is calculated. At the same time, the target tracking trajectory is formed based on the dynamic point clouds, and the dynamic point clouds with a height greater than the target object are selected to solve the mean value.

[0123] Dynamic point cloud forms a height set {z1,z2,...,z M}, calculate the average value of all dynamic point clouds, the formula is:

[0124]

[0125] Among them, z represents the mean height of the dynamic point cloud, and M represents the total number of dynamic point clouds.

[0126] For the filtered dynamic point cloud, extract the height z value of all points to form a height set Calculate the average value greater than the tracking height of the target object using the formula:

[0127]

[0128] Among them, z max Represents the mean height of the dynamic point cloud, M max Indicates the number of dynamic point clouds that are larger than the tracked height value of the target object.

[0129] Determine whether the mean values ​​of the two dynamic point clouds are continuously and monotonically decreasing within a preset range, and determine whether the minimum mean values ​​of the two dynamic point clouds are both less than a set first threshold; for example, continuously determine whether the mean values ​​of all dynamic point clouds are monotonically decreasing for 10 frames and the minimum value (lowest height value) of the mean values ​​of all dynamic point clouds are less than a set first threshold, and continuously determine whether the mean value of the dynamic point cloud higher than the tracking target is monotonically decreasing for 10 frames and the minimum value (lowest height value) of the mean value of the dynamic point cloud higher than the tracking target is less than a set first threshold;

[0130] When the means of the two dynamic point clouds are continuously and monotonically decreasing within the preset range, and the minimum means of the two dynamic point clouds are both less than the first threshold, that is, it is continuously judged that the means of all dynamic point clouds are monotonically decreasing in 10 frames and their minimum value is less than the set first threshold, and it is continuously judged that the mean of the dynamic point cloud higher than the tracking target is monotonically decreasing in 10 frames and its minimum value is less than the set first threshold, the height mean of the static point cloud of the target object is calculated; for the screened static point clouds, the height z values ​​of all points are extracted to form a height set {z1, z2, ..., z N}, calculate the average value of all height values ​​z, the formula is:

[0131]

[0132] in, represents the mean height of the static point cloud, and N represents the total number of static point clouds.

[0133] When it is continuously judged that the mean of all dynamic point clouds is not monotonically decreasing in 10 frames or its minimum value is not less than the set first threshold, the mean of the dynamic point clouds of all radar echo signals continues to be calculated; and when it is continuously judged that the mean of the dynamic point cloud higher than the tracked target is not monotonically decreasing in 10 frames or its minimum value is not less than the set first threshold, the dynamic point cloud with a height value greater than the target object continues to be taken to solve the mean; until the means of both dynamic point clouds are continuously monotonically decreasing within the preset range, and the minimum means of both dynamic point clouds are less than the first threshold, the height mean of the static point cloud of the target object is calculated.

[0134] Determine whether the height mean (average) of the static point cloud of the target object is lower than a set second threshold within a preset range; for example, continuously determine whether the mean of the static point cloud is lower than the set second threshold within 600 frames;

[0135] If the average height of the static point cloud of the target object is lower than the second threshold within a preset range and the duration exceeds the preset time threshold, the state information of the target object is determined to be a fall.

[0136] When the height average of the static point cloud of the target object is not lower than the second threshold within the preset range, continue to calculate the mean of the dynamic point cloud of all radar echo signals for judgment, and continue to take the dynamic point cloud solution mean value larger than the height value of the target object for judgment.

[0137] Based on traditional radar fall detection, this application introduces a dynamic and static point cloud behavior logic analysis module. By fusing dynamic point cloud and static point cloud data, it conducts in-depth analysis of a person's motion trajectory and posture changes, thereby accurately determining whether a fall has occurred. When encountering other actions similar to fall behavior, such as suddenly losing balance or falling after leaning on an elevator, the behavior logic judgment module can distinguish these easily confused actions based on the analysis results, identify and eliminate false alarms of non-fall behaviors, reduce the false alarm rate, and ensure that the fall detection system can stably, accurately, and efficiently identify fall behaviors in complex environments.

[0138] In one embodiment, before analyzing the target object point cloud data, DBSCAN clustering is performed first, which is conducive to faster analysis of the target point cloud data and can effectively distinguish the target object from noise points. The G_TRACK algorithm is then used to track the clustering results to obtain the movement trajectory of the person.

[0139] Reference Figure 1 Finally, when the target object's status information is determined to be a fall, the duration of the target object's fall state is counted. The radar works for 100ms per frame, and the duration is counted by accumulating the number of frames. If the duration exceeds the preset time threshold, a person-trapped signal is sent to the elevator main control system. The elevator main control system triggers the elevator suspension command and sends a door-opening signal to the elevator door opening and closing system, calls the management personnel, and automatically dials the preset emergency contact number that matches the target object to promptly handle elevator entrapment and fall events.

[0140] When the system detects a single person falling and remaining in the elevator for an extended period of time, such as more than five minutes, it automatically triggers the entrapment mechanism and sends a entrapment signal to the elevator's main control system. Upon receiving the entrapment signal, the elevator's main control system immediately takes appropriate emergency measures, including pausing the elevator and opening the doors to ensure the safety of the occupants. Simultaneously, the system automatically dials the emergency contact number pre-matched with the target person and notifies the relevant management personnel to rush to the scene to ensure a timely response, preventing entrapment accidents or other emergencies from delaying assistance. This entrapment mechanism minimizes elevator accidents caused by malfunctions or other emergencies, thereby enhancing elevator safety.

[0141] In one embodiment, the duration of the target object's fall state is counted; if the duration exceeds a preset time threshold, that is, when the system detects that the target object has been staying in the elevator for a long time, a person entrapment signal is sent to the elevator main control system; the elevator main control system triggers an elevator operation pause instruction and sends a door opening signal to the elevator door opening and closing system, calls the management personnel, and automatically dials the preset emergency contact number that matches the target object to promptly handle elevator entrapment and fall incidents.

[0142] In summary, a method for detecting people falling in home elevators based on 4D millimeter-wave radar utilizes MIMO technology to obtain radar echo signals from any target object in the elevator in real time, so as to monitor the real-time movements of people in the elevator through 4D millimeter-wave radar technology. 4D millimeter-wave radar is not affected by changes in light and can still work stably in dim light or complex environments. It can also penetrate obstacles and can better adapt to the narrow and complex space in the elevator. Compared with traditional camera monitoring systems, millimeter-wave radar technology performs contactless monitoring, avoids the risk of privacy leakage, and ensures the protection of people's privacy. At the same time, 4D millimeter-wave radar can quickly collect target data with better real-time performance. Based on the radar echo signal, the three-dimensional coordinates, speed and energy value of the static point cloud and dynamic point cloud of the target object are determined, and the dynamic point cloud and static point cloud data are combined to accurately extract the behavioral characteristics of people in the elevator from multiple dimensions, which is conducive to more accurate judgment of whether a fall has occurred. At the same time, the processing of point cloud data is more efficient, which is conducive to more rapid judgment of whether people in the elevator have fallen. The three-dimensional coordinates of the static point cloud and dynamic point cloud of the target object are determined. The target, speed and energy values ​​are calculated, and the mean of the dynamic point cloud of all radar echo signals is calculated. At the same time, the dynamic point cloud with a larger height value than the target object is selected to solve the mean; it is judged whether the means of the two dynamic point clouds are continuously and monotonically decreasing within a preset range, and whether the minimum means of the two dynamic point clouds are both less than the set first threshold; when the means of the two dynamic point clouds are continuously and monotonically decreasing within the preset range, or the minimum means of the two dynamic point clouds are both less than the first threshold, the height mean of the static point cloud of the target object is calculated; it is judged whether the height mean of the static point cloud of the target object is lower than the set second threshold within the preset range; if the height mean of the static point cloud of the target object is lower than the second threshold within the preset range and the duration exceeds the preset time threshold, the state information of the target object is determined to be a fall, so as to perform behavior analysis by fusing dynamic point cloud and static point cloud data, and detect the height change of the static point cloud and the movement change of the dynamic point cloud to identify the fall behavior; thereby improving the detection accuracy of the fall behavior of people in the elevator, having better environmental adaptability, being able to timely identify the fall behavior, facilitating timely response and processing, and at the same time, having better privacy.

[0143] This application detects whether the height change of a person exceeds a set threshold. If so, it enters the fall monitoring mode and continues to monitor the height change of the person to automatically identify the fall behavior; and if the person fails to return to the standing state within the set time, a trapped person signal is output to the elevator main control system.

[0144] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0145] The embodiment of the present application also provides a home elevator personnel fall detection system based on 4D millimeter wave radar, which corresponds one-to-one with the home elevator personnel fall detection method based on 4D millimeter wave radar in the above embodiment. The home elevator personnel fall detection system based on 4D millimeter wave radar includes:

[0146] The data acquisition module is used to use MIMO technology to obtain radar echo signals from any target object in the elevator in real time through multiple antenna channels;

[0147] A dynamic and static point cloud extraction module, configured to determine the three-dimensional coordinates, velocity, and energy values ​​of the static point cloud and the dynamic point cloud of the target object based on the radar echo signal;

[0148] a dynamic point cloud processing module, configured to calculate the mean of the dynamic point clouds of all radar echo signals using the three-dimensional coordinates, velocity, and energy values ​​of the static point cloud and the dynamic point cloud of the target object; and, at the same time, forming a target tracking trajectory based on the dynamic point cloud, selecting the dynamic point cloud with a height greater than that of the target object to calculate the mean value;

[0149] A dynamic point cloud analysis module is used to determine whether the mean values ​​of the two dynamic point clouds are continuously and monotonically decreasing within a preset range, and to determine whether the minimum values ​​of the mean values ​​of the two dynamic point clouds are both less than a set first threshold;

[0150] a static point cloud processing module, configured to calculate a height mean of the static point cloud of the target object when the means of the two dynamic point clouds are continuously and monotonically decreasing within a preset range and the minimum means of the two dynamic point clouds are both less than the first threshold;

[0151] A static point cloud analysis module, configured to determine whether a height average of the static point cloud of the target object is lower than a set second threshold within a preset range;

[0152] The fall detection module is configured to determine that the state information of the target object is a fall when the average height of the static point cloud of the target object is lower than the second threshold within a preset range and the duration exceeds a preset time threshold.

[0153] Reference Figure 3 ,A home elevator personnel fall detection system based on 4D millimeter-wave radar acquires the radar echo signal from any target object in the elevator in real time by setting a radar RF module and a radar MCU module.

[0154] The radar RF module uses a 4D millimeter-wave radar sensor and four transmit and four receive antennas with MIMO technology to acquire real-time echo signals from people inside the elevator. After receiving the echo signals, the transmit and receive antennas convert the intermediate frequency signals into ADC data and transmit them to the MCU chip for processing.

[0155] The radar MCU module (fall system controller) integrates FFT and CFAR algorithms, efficiently processing radar echo signals and extracting information such as a person's three-dimensional coordinates, speed, and energy. Through this data processing algorithm, the system can calculate a person's duration of stay and detect falls.

[0156] A home elevator personnel fall detection system based on 4D millimeter wave radar then sends the personnel status information to the elevator main control system through a communication module (such as 485) and a 485 communication protocol.

[0157] The elevator's main control system is electrically connected to the radar MCU module, the elevator's door opening and closing system, and the emergency call device. It receives trigger signals and issues control commands. When a fall is detected, the elevator's main control system sends a door-opening signal to the elevator's door opening and closing system and transmits a distress signal to the emergency call device, initiating an emergency call.

[0158] When the radar MCU module detects that a person has fallen and maintained this condition for a certain period of time, the fall system will output a trapped person signal, control the emergency call device to link with the elevator door opening and closing system, automatically open the elevator door, and make an emergency call to the relevant personnel.

[0159] Reference Figure 4 A 4D millimeter-wave radar-based home elevator fall detection system can be installed at the top of the elevator doorway at a 30-degree angle, ensuring wide coverage and enabling real-time monitoring of occupant behavior. The radar includes a radar RF module and a radar MCU module. Table 1 lists the radar configuration parameters.

[0160] Table 1

[0161]

[0162] Initialize the radar.

[0163] If no one enters the elevator, no personal information will be detected and no distress signal will be output.

[0164] When a person enters the elevator, the radar detects the target object, transmits a millimeter-wave signal, and receives the reflected signal, obtaining the radar echo signal of the person in the elevator and monitoring the changes in the radar echo signal in real time. The radar MCU module extracts and analyzes the received radar echo signal to ensure efficient and accurate fall detection.

[0165] If the person leaves within the set threshold time and the set altitude change threshold is not reached, no distress signal will be output.

[0166] Once a person falls or stays for a long time, the fall system will promptly notify the elevator main control system and take emergency measures, such as stopping the elevator and opening the elevator door, and automatically calling the emergency contact and notifying the management staff, greatly improving the safety of home elevators.

[0167] For the specific limitations of a home elevator personnel fall detection system based on 4D millimeter-wave radar, please refer to the limitations of a home elevator personnel fall detection method based on 4D millimeter-wave radar above, which will not be repeated here.

[0168] Each module in the aforementioned 4D millimeter-wave radar-based home elevator fall detection system can be implemented in whole or in part through software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in hardware form, or can be stored in a computer device's memory in software form, allowing the processor to call and execute the corresponding operations of each module.

[0169] In one embodiment, a computer device is provided, which may be a server. The computer device includes a processor, a memory, a network interface, and a database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, it implements any of the above-mentioned methods for detecting falls in home elevators based on 4D millimeter-wave radar.

[0170] In one embodiment, a computer-readable storage medium is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, any one of the above-mentioned methods for detecting a person falling in a home elevator based on a 4D millimeter-wave radar is implemented.

[0171] In one embodiment, a computer program product is provided, which includes a computer program, and when the computer program is executed by a processor, it implements any of the above-mentioned methods for detecting people falling in a home elevator based on 4D millimeter wave radar.

[0172] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium, including a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the various embodiments of the present application. When the computer program is executed, it may include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0173] Those skilled in the art will clearly understand that for the sake of convenience and brevity in description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above.

Claims

1. A method for detecting falls in home elevators based on 4D millimeter-wave radar, characterized in that: The following steps are included: Utilize MIMO technology to acquire radar echo signals from any target object in the elevator in real time through multiple antenna channels; Determining the three-dimensional coordinates, velocity, and energy values ​​of a static point cloud and a dynamic point cloud of the target object based on the radar echo signal; Calculate the mean of the dynamic point clouds of all radar echo signals using the three-dimensional coordinates, velocity, and energy values ​​of the static point cloud and dynamic point cloud of the target object. At the same time, form a target tracking trajectory based on the dynamic point cloud, and select the dynamic point cloud with a height value greater than that of the target object to solve the mean value. Determine whether the mean values ​​of the two dynamic point clouds are both continuously and monotonically decreasing within a preset range, and determine whether the minimum values ​​of the mean values ​​of the two dynamic point clouds are both less than a set first threshold; When the means of the two dynamic point clouds are both continuously and monotonically decreasing within a preset range, and the minimum means of the two dynamic point clouds are both less than the first threshold, then calculating the height mean of the static point cloud of the target object; Determining whether the average height of the static point cloud of the target object is lower than a set second threshold within a preset range; If the average height of the static point cloud of the target object is lower than the second threshold within a preset range and the duration exceeds a preset time threshold, the state information of the target object is determined to be a fall.

2. The method for detecting falls in a home elevator based on 4D millimeter wave radar according to claim 1, characterized in that: The step of determining the three-dimensional coordinates, velocity and energy value of the static point cloud and the dynamic point cloud of the target object based on the radar echo signal includes: After performing a one-dimensional fast Fourier transform on the radar echo signal of the current frame, the mean of the one-dimensional fast Fourier transform of all chirp signals of the radar echo signal is subtracted to obtain the first signal of the dynamic point cloud; Performing multi-frame progression and storage on the first signal of the dynamic point cloud to obtain the first signal of the static point cloud; Performing a two-dimensional fast Fourier transform on the first signals of the acquired static point cloud and the dynamic point cloud to obtain range Doppler maps of the static point cloud and the dynamic point cloud; Based on the acquired range Doppler maps of the static point cloud and the dynamic point cloud, combined with the radar echo signals of all antenna channels, the three-dimensional coordinates, velocity and energy values ​​of the static point cloud and the dynamic point cloud of the target object are determined.

3. The method for detecting falls in a home elevator based on 4D millimeter wave radar according to claim 2, characterized in that: The steps of determining the three-dimensional coordinates, velocity and energy value of the static point cloud and the dynamic point cloud of the target object based on the obtained range Doppler map of the static point cloud and the dynamic point cloud and combining the radar echo signals of all antenna channels include: Obtaining a range Doppler map of a dynamic point cloud from one of the antenna channels of the radar echo signal to perform one-dimensional CFAR detection to obtain the distance and speed of the target object; Then, the range Doppler maps of the dynamic point clouds of all antenna channels are subjected to three-dimensional fast Fourier transform and four-dimensional fast Fourier transform along the two antenna directions to obtain the azimuth and elevation angles of the dynamic point cloud of the target object; and performing a three-dimensional fast Fourier transform and a four-dimensional fast Fourier transform on the range Doppler maps of the static point clouds of all antenna channels along two antenna directions, respectively, to obtain the azimuth and elevation angles of the static point cloud of the target object; The three-dimensional coordinates, velocity and energy values ​​of the static point cloud and the dynamic point cloud of the target object are determined according to the distance, velocity, azimuth and pitch angle of the target object.

4. The method for detecting falls in a home elevator based on a 4D millimeter-wave radar according to any one of claims 1 to 3, characterized in that: Also includes the following steps, Sending the status information of the target object to the elevator main control system; When the status information of the target object is falling, the elevator main control system sends a door opening signal to the elevator door opening and closing system to control the elevator door to open.

5. The method for detecting falls in a home elevator based on 4D millimeter-wave radar according to claim 4, characterized in that: Also includes the following steps, The elevator master control system sends a distress signal to the distress device, and starts dialing an emergency call.

6. The method for detecting falls in a home elevator based on 4D millimeter wave radar according to claim 1, characterized in that: After determining that the state information of the target object is a fall, the method further includes: Counting the duration of the falling state of the target object; If the duration exceeds a preset time threshold, a person-trapped signal is sent to the elevator main control system; The elevator master control system triggers an instruction to pause the elevator operation and sends a door opening signal to the elevator door opening and closing system, calls the management personnel, and automatically dials the preset emergency contact phone number that matches the target object.

7. A home elevator fall detection system based on 4D millimeter wave radar, characterized in that: include, The data acquisition module is used to use MIMO technology to obtain radar echo signals from any target object in the elevator in real time through multiple antenna channels; A dynamic and static point cloud extraction module, configured to determine the three-dimensional coordinates, velocity, and energy values ​​of the static point cloud and the dynamic point cloud of the target object based on the radar echo signal; a dynamic point cloud processing module, configured to calculate the mean of the dynamic point clouds of all radar echo signals using the three-dimensional coordinates, velocity, and energy values ​​of the static point cloud and the dynamic point cloud of the target object; and, at the same time, forming a target tracking trajectory based on the dynamic point cloud, selecting the dynamic point cloud with a height greater than that of the target object to calculate the mean value; A dynamic point cloud analysis module is used to determine whether the mean values ​​of the two dynamic point clouds are continuously and monotonically decreasing within a preset range, and to determine whether the minimum values ​​of the mean values ​​of the two dynamic point clouds are both less than a set first threshold; a static point cloud processing module, configured to calculate a height mean of the static point cloud of the target object when the means of the two dynamic point clouds are continuously and monotonically decreasing within a preset range and the minimum means of the two dynamic point clouds are both less than the first threshold; A static point cloud analysis module, configured to determine whether a height average of the static point cloud of the target object is lower than a set second threshold within a preset range; The fall detection module is configured to determine that the state information of the target object is a fall when the average height of the static point cloud of the target object is lower than the second threshold within a preset range and the duration exceeds a preset time threshold.

8. A computer device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, characterized in that The invention comprises a computer program, which implements the steps of the method according to any one of claims 1 to 6 when the computer program is executed by a processor.

Images