Sitting posture detection method and device, millimeter wave radar equipment and intelligent health seat

Abstract

The invention provides a sitting posture detection method, and belongs to the field of human body sitting posture detection, and the method comprises the steps: calculating a scattering intensity value of each voxel in a target imaging region according to human body back posture data received by MIMO millimeter wave radar equipment, and obtaining a human body sitting posture form high-resolution image according to the scattering intensity values; feature extraction is carried out on high-resolution imaging of the human body sitting posture form, key points are selected at the tail, the middle and the top of the human body spine through a deep learning algorithm, and the spine curvature is calculated; judging that the sitting posture of the human body is abnormal when the frequency of detecting that the spine curvature exceeds the reference value reaches the set frequency; the invention further provides a sitting posture detection device. The invention also provides an MIMO millimeter wave radar device. The invention further provides an intelligent health seat. The accuracy of human body sitting posture detection can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of human sitting posture detection, and in particular to a sitting posture detection method, device, millimeter wave radar equipment, and an intelligent health chair. Background Art

[0002] With the accelerating pace of life and changes in working styles, prolonged sitting has become the norm. Office workers spend their days at their desks working on computers, while students hunch over their desks studying. Prolonged sitting has numerous adverse effects on human health, including obesity, cardiovascular disease, and musculoskeletal problems. Incorrect sitting posture can easily lead to vision loss, scoliosis, and abnormal bone growth in adolescents. By helping users maintain a correct sitting posture, posture correction technology can not only improve their physical health but also enhance their work and personal well-being. Current posture detection methods mostly use passive detection, meaning they don't emit electromagnetic wave signals but only receive images returned from the measured area or electromagnetic waves radiated, reflected, or transmitted by the target itself. Methods such as infrared detection, pressure detection, and machine vision, however, suffer from limited detection information about the target area and low accuracy.

[0003] For example, the Chinese invention patent application, "A Human Sitting Posture Detection Method Based on Channel State Information," with publication number CN109063697A, builds a data acquisition platform to collect channel state information data packets for different sitting postures. After data preprocessing, the subcarrier sequence signal is analogized to a time series signal, and a complex network is constructed using a visual graph method to extract network features. Amplitude and phase statistical features are also extracted, and after standardization, they are stored in a fingerprint database. Finally, a machine learning algorithm is used for sitting posture classification and detection. However, this method for detecting human sitting posture has problems: wireless signals are easily interfered with by metal and obstructions in the environment, resulting in fluctuations in detection accuracy, a lack of effectiveness verification in complex environments, and a lack of widespread application. Furthermore, this method involves multiple data processing steps, and there is a delay from acquisition to classification, making it difficult to meet the demand for immediate feedback and correction of sitting posture.

[0004] For example, the Chinese invention patent application, "A Device for Correcting Sitting Posture and Reminding Rest for Studying," with publication number CN119007403A, integrates an ultrasonic sensor detection module to monitor a student's sitting posture in real time. When the distance between a student's head and the pen holder is too close, the system automatically identifies the posture as poor and issues a warning sound through a speaker. This immediate feedback mechanism can effectively correct students' poor sitting habits and avoid vision loss caused by maintaining an incorrect posture for extended periods. However, the device only determines the actual sitting posture based on the distance between the head and the sensor, which lacks accuracy and makes it difficult to determine the true spinal posture of the person.

[0005] For example, the Chinese invention patent application, "A Sitting Posture Reminder Method, Device, Equipment, and Storage Medium," with publication number CN119445612A, uses an image acquisition device to capture an image of a user sitting in front of a desk, obtaining a posture image of the user. The image is then subjected to eye detection and body detection, respectively, to obtain eye position information and body position information of the user in the posture image. Based on the eye position information and body position information, a preset calculation strategy is used to calculate the distance between the user's eyes and the desk, thereby achieving sitting posture detection. This method can improve the accuracy of eye-desk distance and the timeliness and effectiveness of eye reminders. However, this method, starting from the front of the human body, determines the sitting posture only by the eyes and the distance between the human body and the desk. This method also lacks accuracy and is difficult to determine the true sitting posture of the human spine. Furthermore, the image acquisition device has extremely high lighting requirements, posing a high risk of privacy leakage. Due to the influence of lighting and obstruction by clothing, it is unable to truly detect the curvature of the human spine, making it difficult to reflect the true sitting posture of the human body.

[0006] Another example is the Chinese invention patent application, "Sitting Posture Detection and Interaction System and Method Based on Piezoresistive Array Cushion and Desk Lamp," with publication number CN119603837A. This application utilizes a piezoresistive array cushion and a smart desk lamp to perform sitting posture detection and information interaction, realizing an intelligent cushion system. This system uses the piezoresistive array cushion to accurately monitor the user's various sitting postures, and uses a visual feedback mechanism on the desk lamp to remind the user to adjust their sitting posture in real time. The system also features a sedentary reminder function, effectively managing the user's sitting posture health. However, this method uses pressure values to determine a person's sitting posture. However, the diverse sitting postures of the human spine may not correspond to the pressure values. This method is unable to calculate the curvature of the back spine in real time, and the sitting posture monitoring lacks accuracy.

[0007] Currently, health monitoring equipment needs to be worn by users, which is inconvenient to use and easy to forget. It is more convenient to integrate the sitting posture detection device into the seat. For example, the Chinese invention patent application "Intelligent Sitting Posture Correction Method and System" with publication number CN118737366A integrates multiple sensor technologies and combines them with an adaptive neural network model to monitor and analyze user sitting posture data in real time and provide personalized correction suggestions. However, the problem with this smart seat is that it is difficult to achieve multi-functional health detection through a single sensor, and multiple sensor devices are required for fusion detection, which increases the cost of the seat. Summary of the Invention

[0008] The technical problem to be solved by the present invention is: how to improve the accuracy of human sitting posture detection.

[0009] The present invention solves the above technical problems through the following technical solutions: a sitting posture detection method, the method comprising:

[0010] Based on the human back posture data received by the MIMO millimeter-wave radar device, the scattering intensity value of each voxel in the target imaging area is calculated, and high-resolution imaging of the human sitting posture is obtained based on the scattering intensity value;

[0011] Feature extraction is performed on high-resolution imaging of the human sitting posture. Key points are selected at the tail, middle, and top of the human spine using a deep learning algorithm to calculate the spinal curvature.

[0012] When the number of times that the spinal curvature is detected to exceed the reference value reaches a set number, it is determined that the sitting posture of the human body is abnormal.

[0013] Beneficial effects: MIMO millimeter-wave radar has a higher resolution. The present invention adopts an active millimeter-wave radar imaging method. The MIMO millimeter-wave radar actively transmits FMCW waves to the human body. According to the different scattering coefficients between the measured target and the environment, more target information can be obtained, thereby achieving clearer imaging and performing three-dimensional imaging and real-time imaging. The millimeter-wave radar has good penetration, strong privacy, strong robustness, good anti-interference ability, is not affected by light, and has good all-weather working ability. According to the human body data received by the MIMO millimeter-wave radar device, the scattering intensity value of each voxel in the target imaging area is calculated, and a high-resolution imaging of the human sitting posture is obtained according to the scattering intensity value. The coordinates of three key points on the human spine are extracted through a deep learning algorithm, and then the degree of spinal curvature is calculated. Incorrect sitting posture can be quantified, thereby improving the accuracy of human sitting posture detection.

[0014] Preferably, the MIMO millimeter wave radar device transmits an FMCW signal to the back of the human body and receives posture data of the back of the human body.

[0015] Beneficial effects: The MIMO millimeter-wave radar device of the present invention transmits FMCW signals to the back of the human body and receives human posture data, detecting the human body's sitting posture from behind the human body. Combined with the backward projection imaging algorithm, it can achieve high-resolution imaging of the human back under clothing, and then can help accurately calculate the actual degree of spinal curvature of the human body, thereby improving the accuracy of human sitting posture detection.

[0016] Preferably, the process of calculating the scattering intensity value of each voxel in the target imaging area based on the human back posture data received by the MIMO millimeter wave radar device includes:

[0017] sorting the preprocessed signal of the human back posture data to obtain a sorted signal;

[0018] Performing pulse compression on the sorting signal in the distance dimension to obtain a compressed sorting signal;

[0019] The compressed sorting signal is reconstructed by matched filtering to obtain the scattering intensity value of each voxel in the target imaging area.

[0020] Beneficial effects: The present invention sorts the preprocessed signals of the human back posture data, completes the mixing and filtering processing of the receiving and transmitting signals, and obtains the sorted signals. The format of the data matrix is clearer and the generalization performance is better. The pulse compression processing is completed by performing pulse compression in the distance dimension on the sorted signals, and the time domain signal is converted to the frequency domain, thereby improving the signal-to-noise ratio and reducing the signal complexity. The scattering intensity of the target area is reconstructed by performing matched filtering on the compressed sorted signals, thereby helping to realize human back imaging.

[0021] Preferably, the preprocessing signal of the human back posture data is sorted to obtain the sorted signal, which includes:

[0022] The human body data is mixed, low-pass filtered, and amplified to obtain the preprocessed signal S pre (t), and the preprocessed signal S pre (t) is approximately expressed as S′ pre (t):

[0023] S pre (t)≈S′ pre (t) = exp(j2πf c τ)

[0024] Among them, f c is the starting frequency of the signal, τ is the time delay from the transmitted signal to the received signal, R t is a radar transmitting antenna (x t ,0,z t ) to a certain voxel (x, y, z), R r is the corresponding radar receiving antenna (x r ,0,z r ) and the Euclidean distance between the voxel (x, y, z);

[0025] For the preprocessed signal S′ pre (t) is sorted and the sorting signal S(x t ,z t ,x r ,z r ,t):

[0026]

[0027] Where ξ(x,y,z) is the scattering intensity value of each voxel.

[0028] Preferably, performing matched filtering reconstruction on the compressed sorting signal to obtain the scattering intensity value of each voxel in the target imaging area includes:

[0029] The compressed sorting signal S(x t ,z t ,x r ,z r ,k) Perform matched filtering reconstruction to obtain the estimated value of the scattering intensity of each voxel

[0030]

[0031] Use summation instead of integration to estimate the value Expressed as:

[0032]

[0033] Where k = 2πf c / c,f c is the starting frequency of the signal, c is the propagation speed of electromagnetic waves in the air, R t is a radar transmitting antenna (x t ,0,z t ) to a certain voxel (x, y, z), R r is the corresponding radar receiving antenna (x r ,0,z r ) and the Euclidean distance between the voxel (x,y,z).

[0034] Beneficial effect: The MIMO millimeter-wave radar of the present invention has a high resolution itself, and the use of summation operation instead of integration operation has little effect on the imaging resolution. Therefore, when calculating the scattering intensity of each voxel, summation operation is used instead of integration operation. Under the premise of ensuring the imaging resolution, the calculation can be simplified and the calculation speed can be accelerated.

[0035] Preferably, the spinal curvature C' is calculated as:

[0036]

[0037] Among them, A, B, and C are three key points taken at the tail, middle, and top of the human spine respectively.

[0038] Beneficial effect: The present invention extracts the coordinates of three key points in the middle of the back, namely the top of the spine, the middle of the spine, and the tail of the spine, through a deep learning algorithm, and uses the vector cross multiplication modulus length formula to calculate the curvature of the spine, thereby more accurately detecting the sitting posture of the human body.

[0039] The present invention also provides a sitting posture detection device, which includes:

[0040] The sitting posture imaging unit is used to calculate the scattering intensity value of each voxel in the target imaging area based on the human back posture data received by the MIMO millimeter wave radar device, and obtain high-resolution imaging of the human sitting posture based on the scattering intensity value;

[0041] The spinal curvature calculation unit is used to extract features from high-resolution imaging of the human sitting posture. It uses a deep learning algorithm to select key points at the tail, middle, and top of the human spine to calculate the spinal curvature.

[0042] The detection unit is used to detect that when the spinal curvature exceeds the reference value for a set number of times, it determines that the human sitting posture is abnormal.

[0043] The present invention also provides a MIMO millimeter-wave radar device, which includes a MIMO millimeter-wave radar module and a signal processing module. The MIMO millimeter-wave radar module transmits an FMCW signal to the back of a human body and receives human back posture data. The signal processing module uses the sitting posture detection method described above to process human body data. The MIMO millimeter-wave radar module includes multiple "mouth"-shaped MIMO antenna arrays arranged in an array. Each "mouth"-shaped MIMO antenna array includes 2X transmitting antenna units and 2Y receiving antenna units. The 2X transmitting antenna units are arranged in two rows of linear arrays, and each side includes X transmitting antenna units. The 2Y receiving antenna units are arranged in two columns of linear arrays, that is, each side includes Y receiving antenna units. The spacing d between adjacent transmitting antenna units and adjacent receiving antenna units is half a wavelength λ2: Among them, f c is the starting frequency of the signal, and c is the propagation speed of electromagnetic waves in the air.

[0044] Beneficial effects: The present invention uses a MIMO radar with a two-dimensional antenna array, and a single "snap shot" can obtain all the data of a three-dimensional image. The MIMO millimeter-wave radar module includes a plurality of "mouth"-shaped MIMO antenna arrays arranged in an array. The array configuration can meet the requirements of high-resolution imaging of the human back in both lateral and longitudinal resolution, thereby improving the real-time performance of detection. The maximum angular field of view of the array can cover the target area in front of the backrest, meeting the actual needs of sitting imaging detection. In addition, the MIMO millimeter-wave radar module is composed of a plurality of "mouth"-shaped MIMO antenna arrays, which is easier to realize modularization. According to the actual resolution requirements, the "mouth"-shaped arrays can be increased or decreased to meet the user's customization needs.

[0045] Preferably, the signal processing module further includes a heart rate detection device, which includes:

[0046] A signal acquisition and preprocessing unit, used for acquiring human body data and preprocessing the human body data to obtain a preprocessed signal;

[0047] A range gate acquisition unit is used to perform distance dimension FFT processing on the pre-processed signal to determine the range gate where the chest cavity is located;

[0048] The phase extraction unit is used to extract phase change information from the preprocessed signal according to the range gate and perform phase unwrapping operation to obtain the respiratory and heartbeat signals;

[0049] A respiratory and heartbeat separation unit is used to separate the respiratory and heartbeat signals based on the ICA algorithm to obtain independent heartbeat source signals and respiratory source signals;

[0050] The heart rate detection unit is used to perform FFT processing on the heartbeat source signal, extract the heartbeat frequency and calculate the heart rate, compare the detected heart rate value with the normal heart rate information of the human body, and determine whether the heart rate is abnormal.

[0051] Preferably, the signal processing module further includes a sedentary detection device, which includes:

[0052] A signal acquisition and preprocessing unit, used for acquiring human body data and preprocessing the human body data to obtain a preprocessed signal;

[0053] A range gate acquisition unit is used to perform distance dimension FFT processing on the pre-processed signal to determine the range gate where the chest cavity is located;

[0054] A phase extraction unit, configured to extract phase change information from the preprocessed signal according to the range gate;

[0055] Micro-Doppler detection unit, used to determine whether the human sitting posture is stable based on phase change information;

[0056] The cumulative timing unit is used to start timing and record the sitting time when it is judged that the human body's sitting posture is stable. When the sitting time exceeds the set time, it is judged as sitting for a long time.

[0057] Beneficial effects: The present invention integrates human sitting posture detection, heart rate detection and sedentary detection in the signal processing module to achieve multifunctional integration of sitting posture imaging detection, human heart rate detection and sedentary detection, which helps to carry out intelligent intervention on human health in a more comprehensive and multi-dimensional manner.

[0058] The present invention also provides an intelligent health seat, including the MIMO millimeter-wave radar device. The seat also includes a body, a battery system, a wireless Bluetooth communication module and a seat management system. The MIMO millimeter-wave radar device and the battery system are both located in the backrest of the body. The backrest is made of wave-transparent and heat-dissipating materials and the rear wall is provided with ventilation holes. The wireless Bluetooth communication module is located in the MIMO millimeter-wave radar device and is used to encrypt and transmit the detection results of the signal processing module to the seat management system.

[0059] Beneficial effect: The intelligent health seat of the present invention combines millimeter-wave radar perception with AI intelligent algorithms. It only needs to integrate MIMO millimeter-wave radar equipment in the backrest of the seat body. While ensuring electromagnetic compatibility, it can realize one or more of the three detection functions of heart rate detection, sedentary detection, and sitting posture imaging detection. There is no need for multi-sensor equipment fusion detection, and the cost of the seat is greatly reduced. It has high detection accuracy, fast response speed, strong system endurance, and data can be shared, providing a comprehensive, multi-dimensional non-contact intelligent health intervention solution for office and home scenarios.

[0060] In scenarios such as office applications and home monitoring, facing different beneficiary groups, the smart health chair of the present invention can realize the combination of heart rate monitoring, sedentary detection and sitting posture imaging detection capabilities, and carry out comprehensive and multi-dimensional intelligent intervention in human health; through the non-contact, high-precision and highly adaptable technical characteristics, it solves the privacy, environmental limitations and reliability issues of traditional solutions, becoming an innovative direction in the field of smart health chairs, and helping to achieve active health management. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 This is a flow chart of the sitting posture detection method provided in Example 1 of the present invention;

[0062] Figure 2 This is a diagram showing the expected effect of the sitting posture detection method provided in Example 1 of the present invention;

[0063] Figure 3 A perspective view of a millimeter-wave radar device provided in Example 3 of the present invention;

[0064] Figure 4 A front view of the millimeter-wave radar device provided in Example 3 of the present invention;

[0065] Figure 5 A side view of the millimeter-wave radar device provided in Example 3 of the present invention;

[0066] Figure 6 A schematic diagram of a MIMO millimeter-wave radar module in a millimeter-wave radar device provided in Example 3 of the present invention;

[0067] Figure 7 A schematic diagram of a signal processing module in a millimeter-wave radar device provided in Example 3 of the present invention;

[0068] Figure 8 This is a schematic diagram of the smart health chair provided in Example 4 of the present invention;

[0069] Figure 9 A schematic diagram of another perspective of the smart health chair provided in Example 4 of the present invention;

[0070] Figure 10A schematic diagram of the use of the smart health chair provided in Example 4 of the present invention;

[0071] Figure 11 This is a diagram showing the expected interface effect of the APP module in the smart health seat provided in Example 4 of the present invention;

[0072] In the figure: 10 main body, 11 backrest, 12 seat cushion, 13 crescent-shaped armrest, 14 pneumatic telescopic rod, 15 base, 151 roller, 111 push-type hatch, 112 vent, 21 millimeter-wave radar equipment, 211 MIMO millimeter-wave radar module, 211a "mouth"-shaped MIMO antenna array, 212 signal processing module, wireless Bluetooth communication module 213, 22 battery system, 221 shell, 222 battery, 223 intelligent management chip, 31 APP module, 32 cloud module. DETAILED DESCRIPTION

[0073] To make the objectives, technical solutions, and advantages of the present invention more clearly understood, the following describes the technical solutions of the present invention clearly and completely with reference to specific embodiments and the accompanying drawings. It is obvious that the embodiments described are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by persons of ordinary skill in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0074] Example 1

[0075] See also Figure 1 This embodiment provides a sitting posture detection method, which includes the following steps:

[0076] Step 1: Based on the human back posture data received by the MIMO millimeter wave radar device, the scattering intensity value of each voxel in the target imaging area is calculated, and high-resolution imaging of the human sitting posture is obtained based on the scattering intensity value.

[0077] The present invention enables the MIMO millimeter wave radar device to transmit the FMCW signal S to the back of the human body. T (t), the MIMO millimeter wave radar device receives the collected human back posture data S R (t) The MIMO millimeter-wave radar device of the present invention transmits an FMCW signal to the back of the human body and receives human posture data, detecting the human sitting posture from behind. Combined with a back-projection imaging algorithm, it can achieve high-resolution imaging of the human back under clothing, which can then help accurately calculate the actual degree of spinal curvature of the human body, improving the accuracy of human sitting posture detection. Step 1 specifically includes the following process:

[0078] Step 1.1: sort the preprocessed signal of the human back posture data to obtain the sorted signal, complete the mixing and filtering processing of the transmit and receive signals, and obtain the sorted signal. The format of the data matrix is clearer.

[0079] Human back posture data S R (t) Perform pre-processing such as mixing, low-pass filtering, and amplification to obtain the pre-processed signal S of the human back posture data pre (t).

[0080] Assume that the slope of the FMCW wave is s and the starting frequency is f c , then transmit FMCW signal S T (t), receiving the collected human back posture data S R (t) and preprocessed signal S pre (t) can be expressed as:

[0081]

[0082] S R (t) = S T (t-τ) (2)

[0083]

[0084] Preprocessed signal S pre The quadratic term exp(j2π(Sτ) in (t) 2 / 2)) and the RVP term exp(j2π(Sτt)) can be ignored, so the preprocessed signal S pre (t) is approximately expressed as S′ pre (t):

[0085] S pre (t)≈S′ pre (t) = exp(j2πf c τ) (4)

[0086] Where t is the sampling time, τ is the time delay from the MIMO millimeter-wave radar device transmitting the signal to receiving the signal, and τ can be expressed as:

[0087]

[0088] Where c is the propagation speed of electromagnetic waves in the air. Assume that the radar equipment is placed on the XOZ plane and the human body sitting imaging area is divided into many small cubes, namely a voxel, with coordinates (x, y, z). R t is a radar transmitting antenna (x t ,0,z t ) to a certain voxel (x, y, z), Rr is the corresponding radar receiving antenna (x r ,0,z r ) and the voxel (x, y, z), the Euclidean distance R t 、R r They can be expressed as:

[0089]

[0090] Then, the preprocessed signal S′ pre (t) is sorted and the sorting signal S(x t ,z t ,x r ,z r ,t):

[0091]

[0092] Wherein, ξ(x, y, z) is the scattering intensity value of each voxel, and ξ(x, y, z) is the target function that needs to be calculated and estimated in the present invention.

[0093] Step 1.2: Perform pulse compression on the sorting signal in the distance dimension to obtain the compressed sorting signal, complete the pulse compression process, convert the time domain signal to the frequency domain, improve the signal-to-noise ratio, and reduce the signal complexity. t ,z t ,x r ,z r ,t) To perform pulse compression in the distance dimension is to process the sorted signal in the spatial frequency domain, that is, k = 2πf c / c, the compressed sorting signal can be expressed as:

[0094]

[0095] Step 1.3: Perform matched filtering reconstruction on the compressed sorted signal to obtain the scattering intensity value of each voxel in the target imaging area, and obtain high-resolution imaging of the human sitting posture based on the scattering intensity value. Step 1.3 specifically includes the following processes:

[0096] Step 1.3.1, the compressed sorting signal S(x t ,z t ,x r ,z r ,k) Perform matched filtering reconstruction to obtain the estimated value of the scattering intensity of each voxel

[0097]

[0098] Step 1.3.2: Considering the fact that electromagnetic waves are transmitted and received from discrete locations in the imaging array, the integral operation in formula (10) can be replaced by a summation operation to simplify the calculation. Formula (10) can then be expressed as:

[0099]

[0100] The MIMO millimeter-wave radar of the present invention has a high resolution. The use of summation instead of integration has little effect on the imaging resolution. Therefore, when calculating the scattering intensity of each voxel, summation is used instead of integration. This simplifies the calculation and speeds up the calculation while ensuring the imaging resolution.

[0101] Step 1.3.3: According to formula (11), the set imaging area can be directly imaged with high resolution. The expected effect is that it can penetrate clothing and show the true upper body shape of the human body in a sitting position.

[0102] Step 2: Extract features from high-resolution imaging of the human sitting posture, select key points on the human spine, and calculate the spinal curvature based on a deep learning key point detection algorithm.

[0103] See also Figure 2 Based on the deep learning key point detection algorithm, the coordinates of three key points A, B, and C are taken at the tail, middle, and top of the human spine, and the spinal curvature C′ is calculated by vector product:

[0104]

[0105] in, The vector representation of the key point A at the tail of the human spine to the key point C at the top of the spine, The vector representation of the key point A at the tail of the human spine to the key point B in the middle of the spine.

[0106] The present invention uses a deep learning algorithm to extract the coordinates of three key points in the middle of the back: the top of the spine, the middle of the spine, and the tail of the spine. It uses the vector cross-multiplication modulus length formula to calculate the degree of curvature of the spine, thereby more accurately detecting the sitting posture of the human body.

[0107] Step 3: When the number of times that the spinal curvature exceeds the reference value reaches a set number, it is determined that the sitting posture of the human body is abnormal.

[0108] The reference value of spinal curvature is the clinical reference value, which is generally 20°. When the spinal curvature C′≥20° is detected, it should be recorded. When the spinal curvature C′≥20° is detected, the number of times C′ Num When the set number of times is reached, take the set number of times as 3 as an example, when the number C' Num When ≥3, the sitting posture of the human body is judged to be abnormal.

[0109] MIMO millimeter-wave radar has a high resolution. The present invention adopts an active millimeter-wave radar imaging method. The MIMO millimeter-wave radar actively transmits FMCW waves to the human body. According to the different scattering coefficients between the measured target and the environment, more target information can be obtained, thereby achieving clearer imaging and performing three-dimensional imaging and real-time imaging. The millimeter-wave radar has good penetration, strong privacy, strong robustness, good anti-interference ability, is not affected by light, and has good all-weather working ability. According to the human body data received by the MIMO millimeter-wave radar device, the scattering intensity value of each voxel in the target imaging area is calculated, and a high-resolution imaging of the human sitting posture is obtained based on the scattering intensity value. The coordinates of three key points on the human spine are extracted through a deep learning algorithm, and the degree of spinal curvature is calculated. Incorrect sitting posture can be quantified, which can improve the accuracy of human sitting posture detection.

[0110] Example 2

[0111] See also Figure 7 The difference between this embodiment and embodiment 1 is that this embodiment provides a sitting posture detection device, which includes:

[0112] The sitting posture imaging unit is used to calculate the scattering intensity value of each voxel in the target imaging area based on the human back posture data received by the MIMO millimeter wave radar device, and obtain high-resolution imaging of the human sitting posture based on the scattering intensity value; the sitting posture imaging unit includes:

[0113] The signal sorting unit is used to sort the pre-processed signal of the human back posture data to obtain a sorted signal. The pre-processed signal of the human body data is sorted to obtain a sorted signal including:

[0114] The human back posture data is mixed, low-pass filtered, and amplified to obtain the preprocessed signal S pre (t), and the preprocessed signal S pre (t) is approximately expressed as S′ pre (t):

[0115] S pre (t)≈S′ pre (t) = exp(j2πf c τ)

[0116] Among them, f c is the starting frequency of the signal, τ is the time delay from the transmitted signal to the received signal, R t is a radar transmitting antenna (x t ,0,z t ) to a certain voxel (x, y, z), R r is the corresponding radar receiving antenna (xr ,0,z r ) and the Euclidean distance between the voxel (x, y, z);

[0117] For the preprocessed signal S′ pre (t) is sorted and the sorting signal S(x t ,z t ,x r ,z r ,t):

[0118]

[0119] Where ξ(x,y,z) is the scattering intensity value of each voxel.

[0120] The pulse compression unit is used to perform pulse compression on the sorting signal in the distance dimension to obtain a compressed sorting signal.

[0121] The image reconstruction unit is used to perform matched filtering reconstruction on the compressed sorting signal to obtain the scattering intensity value of each target point in the target imaging area, and obtain high-resolution imaging of the human sitting posture based on the scattering intensity value.

[0122] The compressed sorting signal is reconstructed by matched filtering to obtain the scattering intensity value of each voxel in the target imaging area, including:

[0123] The compressed sorting signal S(x t ,z t ,x r ,z r ,k) Perform matched filtering reconstruction to obtain the estimated value of the scattering intensity of each voxel

[0124]

[0125] Use summation instead of integration to estimate the value Expressed as:

[0126]

[0127] Where k = 2πf c / c,f c is the starting frequency of the signal, c is the propagation speed of electromagnetic waves in the air, R t is a radar transmitting antenna (x t ,0,z t ) to a certain voxel (x, y, z), R r is the corresponding radar receiving antenna (x r ,0,z r ) and the Euclidean distance between the voxel (x,y,z).

[0128] The spinal curvature calculation unit is used to extract features from high-resolution imaging of the human sitting posture, select key points on the human spine, and calculate the spinal curvature based on a deep learning key point detection algorithm. The spinal curvature C′ is calculated as follows:

[0129]

[0130] Among them, A, B, and C are three key points taken at the tail, middle, and top of the human spine respectively.

[0131] The detection unit is used to detect that when the spinal curvature exceeds the reference value for a set number of times, it determines that the human sitting posture is abnormal.

[0132] Example 3

[0133] The difference between this embodiment and embodiment 1 is that this embodiment provides a millimeter wave radar device 21, see Figure 3-5 The device includes a MIMO millimeter-wave radar module 211 and a signal processing module 212. The MIMO millimeter-wave radar module 211 transmits an FMCW signal to the back of the human body and receives the posture data of the back of the human body. The signal processing module 212 adopts the sitting posture detection method of Example 1 to process the human body data.

[0134] See also Figure 6 The MIMO millimeter-wave radar module 211 includes multiple "mouth"-shaped MIMO antenna arrays 211a. Each "mouth"-shaped MIMO antenna array includes 2X transmitting antenna units and 2Y receiving antenna units. The 2X transmitting antenna units are arranged in two linear arrays, with each side including X transmitting antenna units. The 2Y receiving antenna units are arranged in two linear arrays, with each side including Y receiving antenna units. The spacing d between adjacent transmitting antenna units and adjacent receiving antenna units is half a wavelength λ / 2. The wavelength λ of the electromagnetic wave signal is:

[0135]

[0136] Among them, f c is the starting frequency of the signal, f c =77×10 9 , c is the propagation speed of electromagnetic waves in the air, c = 3 × 10 8 .

[0137] Then the maximum angular field of view Fov can be expressed as:

[0138]

[0139] Formula (14) can be used to calculate Fov = ±90°, which can cover the area of the human body in front of the backrest and meet the requirements of sitting posture imaging detection.

[0140] Figure 6 The MIMO millimeter-wave radar module 211 shown includes four 2×2 "mouth"-shaped MIMO antenna arrays 211a, with X=Y=20. The "mouth"-shaped MIMO antenna arrays 211a each contain 2×20 transmitting antenna elements and 2×20 receiving antenna elements, meaning each side contains 20 antenna elements, and the antenna element spacing d is half a wavelength λ2. Using formula (13), the wavelength λ = 3.9 mm is calculated. Based on the MIMO virtual aperture principle, each "mouth"-shaped MIMO antenna array 211a can be equivalent to 40×40 virtual antenna elements. For a 2×2 "mouth"-shaped MIMO antenna array 211a, 40×40×4 virtual antenna elements are obtained, and the spacing between adjacent "mouth"-shaped MIMO antenna arrays 211a in both the horizontal and vertical directions is 20×(λ2).

[0141] In the present invention, the horizontal and vertical distance resolutions of the MIMO millimeter wave radar module 211 are δ x and δ y , the formula is as follows:

[0142]

[0143] Wherein, D is the distance from the center of the MIMO millimeter wave radar module 211 to the back of the human body. In the present invention, when performing sitting posture imaging detection and heart rate detection, D≈20cm. are the physical aperture sizes of the transmitting antenna of the MIMO millimeter wave radar module 211 in the horizontal direction and the vertical direction, respectively. All are 3Nd, are the physical aperture sizes of the receiving antenna of the MIMO millimeter wave radar module 211 in the horizontal direction and the vertical direction, respectively. Both are 3Nd. The horizontal distance resolution δ of the MIMO millimeter wave radar module 211 can be calculated by formula (15) and formula (16): x , vertical distance resolution δ y They are all about 3mm, which is sufficient for high-resolution sitting posture imaging and heart rate detection of the human body.

[0144] The present invention uses a MIMO radar with a two-dimensional antenna array to obtain all the data of a three-dimensional image with a single "snap shot". The MIMO millimeter-wave radar module includes multiple "mouth"-shaped MIMO antenna arrays arranged in an array. The array configuration has a horizontal and vertical resolution of up to 3mm, meeting the needs of high-resolution imaging of the human back and improving real-time detection. The array's maximum angular field of view can reach Fov = ±90, which can cover the target area in front of the backrest, meeting the actual needs of sitting posture detection. In addition, the MIMO millimeter-wave radar module is composed of multiple "mouth"-shaped MIMO antenna arrays, which is easier to achieve modularization. According to the actual resolution requirements, the "mouth"-shaped array can be increased or decreased to meet user customization needs.

[0145] See also Figure 7 The signal processing module also includes a heart rate detection device, which includes:

[0146] A signal acquisition and preprocessing unit, used for acquiring human body data and preprocessing the human body data to obtain a preprocessed signal;

[0147] Through beamforming technology, the MIMO millimeter wave radar device transmits FMCW signal S forward T (t), gathered at the human chest, the MIMO millimeter wave radar equipment receives the collected human body data S R (t), for human body data S R (t) Perform preprocessing such as mixing, low-pass filtering, and amplification to obtain the preprocessed signal S of the human body data pre (t).

[0148] Assume that the slope of the FMCW wave is s and the starting frequency is f c , then transmit FMCW signal S T (t) Receive the collected human body data S R (t) and preprocessed signal S pre (t) can be expressed as:

[0149]

[0150] S R (t) = S T (t-τ)(18)

[0151]

[0152] Where t is the sampling time, τ is the time delay from the MIMO millimeter-wave radar device transmitting the signal to receiving the signal, and τ can be expressed as:

[0153]

[0154] Where c represents the propagation speed of electromagnetic waves in air, and d(t) represents the dynamic displacement of the human chest. Since the displacement of the chest caused by the human heartbeat is periodic, it can be assumed that both the breathing and heartbeat signals are sinusoidal signals. Then d(t) can be expressed as:

[0155] d(t)=d0+d h sin(2πf h t)+d r sin(2πf r t) (21)

[0156] Where d0 represents the static distance from the radar to the human chest, d h Represents the amplitude of human heartbeat, f h Indicates the frequency of human heartbeat (0.8-2.5Hz), d r Represents the amplitude of human breathing, f r Indicates the frequency of human breathing (0.1-0.5Hz). In general, the quadratic phase term exp(Sτ 2 / 2) can be ignored, the preprocessing signal S pre (t) can be approximately expressed as:

[0157]

[0158] The range gate acquisition unit is used to perform distance dimension FFT processing on the pre-processed signal to determine the range gate where the chest cavity is located; pre (t) Perform distance dimension FFT processing, and the frequency value at the peak of the spectrum is the intermediate frequency f IF , determine the distance gate R0 where the chest cavity is located according to Δf:

[0159]

[0160] The phase extraction unit is used to extract the phase change information Δφ(t) from the preprocessed signal according to the range gate, and perform phase unwrapping operation to obtain the respiratory and heartbeat signals.

[0161] The phase change information Δφ(t) is calculated as:

[0162]

[0163] The respiratory and heartbeat separation unit is used to separate the respiratory and heartbeat signals based on the ICA algorithm to obtain independent heartbeat source signals and respiratory source signals. The human respiratory signal and heartbeat signal are separated from the formula (24) by the independent component analysis (ICA) algorithm to obtain the independent heartbeat source S h (t):

[0164] S h (t) = A h sin(2πf h t) (25)

[0165] Among them, A h Indicates the amplitude of the separated human heartbeat signal.

[0166] The heart rate detection unit is used to perform FFT processing on the heartbeat source signal and extract the heartbeat frequency f h And calculate the heart rate HR, compare the heart rate detection value with the normal heart rate information of the human body, and determine whether the heart rate is abnormal.

[0167] Heart rate HR is calculated as:

[0168] HR=60·f h (26)

[0169] The unit of heart rate HR is bmp.

[0170] The detected heart rate signal is compared with the normal human heart rate information in the database to determine whether the current user's heart rate is abnormal.

[0171] Continue to see Figure 7 The signal processing module also includes a sedentary detection device, which includes:

[0172] The signal acquisition and preprocessing unit is used to collect human body data and preprocess the human body data to obtain a preprocessed signal; the specific process is the same as the processing process of the signal acquisition and preprocessing unit in the heart rate detection device.

[0173] The range gate acquisition unit is used to perform distance dimension FFT processing on the pre-processed signal to determine the range gate where the chest cavity is located; the specific process is the same as the processing process of the range gate acquisition unit in the heart rate detection device.

[0174] The phase extraction unit is used to extract phase change information from the preprocessed signal according to the range gate; and is used to extract phase change information Δφ(t) from the preprocessed signal according to the range gate. The phase change information Δφ(t) is calculated according to formula (14).

[0175] The micro-Doppler detection unit is used to determine whether the sitting posture of a human body is stable based on the phase change information. The specific judgment process is: if the phase change information is stable and regular, it is judged that the human body is in a sitting posture; if the change amplitude of the phase change information exceeds the stability threshold, it is judged that the human body has not yet stabilized its sitting posture; if the change amplitude of the phase change information is zero, it is judged that there is no one.

[0176] The cumulative timing unit is used to start timing and record the sitting time when it is judged that the human body's sitting posture is stable, otherwise it stops timing; when the recorded sitting time exceeds the set time, it is judged as sitting for a long time.

[0177] Existing sedentary detection methods use pressure sensors or images to detect a person and then start counting. However, the person may not have actually sat down, which can lead to inaccurate timing and affect the detection. The sedentary detection method of the present invention starts accumulating time after the micro-Doppler information of the chest cavity is detected to be stable. At this point, the person is likely to have sat down and started working or studying, which can reflect that the person has actually sat down, and the detection result is more accurate.

[0178] It should be noted that if the signal processing module integrates both the heart rate detection device and the sedentary detection device, the signal acquisition and preprocessing unit, the distance gate acquisition unit and the phase extraction unit can be shared by the sedentary detection device and the heart rate detection device. Of course, the signal acquisition and preprocessing unit, the distance gate acquisition unit and the phase extraction unit can also be separately provided in the sedentary detection device.

[0179] Example 4

[0180] The difference between this embodiment and embodiment 3 is that this embodiment provides a smart health seat, including the millimeter wave radar device of embodiment 3, see Figure 8 and Figure 9 The seat also includes a body 10, a battery system 22, a wireless Bluetooth communication module 213 and a seat management system. The millimeter-wave radar device 21 and the battery system 22 are both located in the backrest 11 of the body 10. The thickness of the backrest 11 is appropriately widened. The millimeter-wave radar device 21 is embedded in the middle and upper layer of the backrest 11, and the battery system 22 is embedded in the lower part of the backrest 11. A push-type hatch 111 is provided on the backrest 11. By providing a push-type hatch on the backrest of the seat, the placement and maintenance of the battery system 22 are facilitated. The backrest 11 is made of wave-transmitting and heat-dissipating materials and the rear wall is provided with a vent 112 to facilitate the propagation of electromagnetic wave signals and system heat dissipation. The backrest 11 can also be integrated with a central control touch screen. The wireless Bluetooth communication module 213 is located in the millimeter-wave radar device 21 and is used to encrypt and transmit the detection results of the signal processing module 212 to the seat management system, and the transmission is stable and secure.

[0181] It should be noted that the present invention takes into account that heart rate detection and sedentary prolongation detection require the use of beamforming technology to focus the beam on the chest. Therefore, when the millimeter-wave radar device 21 is integrated into the smart health chair, it is embedded in the middle and upper part of the backrest 11 to achieve more accurate heart rate detection and sedentary prolongation detection. When detecting sitting posture, in order to better cover the human back, it is best to embed the millimeter-wave radar device 21 in the middle or upper part of the backrest 11.

[0182] Continue to see Figure 8 The chair's main body 10 is ergonomically designed to provide users with support while sitting and leaning, while also enhancing comfort. The main body 10 also includes a seat cushion 12, crescent-shaped armrests 13, and a base 15. The crescent-shaped armrests 13 are located on either side of the seat cushion 12. The base 15 is mounted to the bottom of the seat cushion 12 via pneumatic telescopic rods 14. Rollers 151 are also mounted on the four corners of the base 15. The seat cushion 12 is made of a soft, resilient material, providing users with a sense of comfort and security.

[0183] Continue to see Figure 5 The battery system 22 is high-performance, replaceable, and rechargeable, and includes a housing 221, a battery 222, and an intelligent management chip 223. The housing 221 features an ingenious quick-release structure. Through a push-type hatch on the back of the seatback, the battery can be easily removed from the seatback and replaced with a new one. This housing structure design not only allows users to quickly replace the battery when it is exhausted, maintaining the normal use of the seat, but also facilitates separate maintenance or charging of the battery. Battery 222 is located inside the housing 221. Battery 222 uses lithium-ion technology, which has high energy density and a high number of charge and discharge cycles. Compared to traditional batteries of the same volume, it can store 30%-50% more electricity, providing a lightweight and compact power supply for the system, enabling it to operate for a long time, helping to reduce the frequency of battery replacement and increase the overall service life of the product. Smart management chip 223, located on the side of housing 221, monitors battery voltage, current, temperature, and remaining charge in real time. Upon detecting an abnormality such as overcharging, over-discharging, overheating, or a short circuit, the chip activates multiple protection mechanisms within microseconds, including circuit disconnection and adjusting charging current and voltage to ensure user safety. The intelligent chair system of this invention requires no electrical connection, making it safe and convenient to use.

[0184] See also Figure 10 The seat management system includes an APP module 31 and a cloud module 32. Among them, the APP module 31 is an APP for smart health seats applied to millimeter wave radar (can be applied to mobile phones, tablets, smart watches, etc.), which is connected to the millimeter radar device 21 through the wireless Bluetooth communication module 213, and receives the detection results transmitted by the signal processing module 212 in real time, and promptly feeds back to the user by sending corresponding message prompts; at the same time, the user can also adjust the working frequency of the millimeter wave radar device 21 through the APP module interface to achieve multi-dimensional and intelligent intervention in the user's health.

[0185] For example, in the present invention, when it is judged that the current sitting posture of the human body is abnormal, the signal processing module 213 transmits the detection result to the mobile phone APP through the wireless Bluetooth communication module 213. When the recorded sitting time exceeds the set time value, it is judged as sedentary, and the signal processing module 213 transmits the detection result to the mobile phone APP through the wireless Bluetooth communication module 213. When it is judged that the current user's heart rate is abnormal, the signal processing module 213 transmits the detection result to the mobile phone APP through the wireless Bluetooth communication module 213. The expected effect is as follows: Figure 11 As shown, if T ≥ 1.5h is detected, a "Sedentary reminder!" message will be pushed; if the spinal curvature C ≥ 20° is detected, a "Incorrect sitting posture reminder!" message will be pushed; if an abnormal heart rate is detected, a "Heart rate abnormality reminder!" message will be pushed and an alarm will be issued immediately.

[0186] The cloud module 32 is used for a specific user group, such as a business or a family, to upload the results of human heart rate detection, sedentary detection and sitting posture imaging detection to the group data cloud, so that family members can share and monitor each other's health data. Parents or children can view the home health data of children and the elderly in real time, and take timely measures in case of emergencies to improve the family happiness index and safety index, and realize the sharing and mutual monitoring of health data.

[0187] The intelligent health seat of the present invention combines millimeter-wave radar perception with AI intelligent algorithms. It only requires the integration of MIMO millimeter-wave radar equipment in the backrest of the seat body. While ensuring electromagnetic compatibility, it can realize one or more of the three detection functions of heart rate detection, sedentary detection, and sitting posture imaging detection. There is no need for multi-sensor equipment fusion detection, which greatly reduces the cost of the seat. It has high detection accuracy, fast response speed, strong system endurance, and data sharing, providing a comprehensive, multi-dimensional non-contact intelligent health intervention solution for office and home scenarios.

[0188] In scenarios such as office applications and home monitoring, facing different beneficiary groups, the smart health chair of the present invention can realize the combination of heart rate monitoring, sedentary detection and sitting posture imaging detection capabilities, and carry out comprehensive and multi-dimensional intelligent intervention in human health; through the non-contact, high-precision and highly adaptable technical characteristics, it solves the privacy, environmental limitations and reliability issues of traditional solutions, becoming an innovative direction in the field of smart health chairs, and helping to achieve active health management.

[0189] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A sitting posture detection method, characterized in that: Methods include: Based on the human back posture data received by the MIMO millimeter-wave radar device, the scattering intensity value of each voxel in the target imaging area is calculated, and high-resolution imaging of the human sitting posture is obtained based on the scattering intensity value; Feature extraction is performed on high-resolution imaging of the human sitting posture. Key points are selected at the tail, middle, and top of the human spine using a deep learning algorithm to calculate the spinal curvature. When the number of times that the spinal curvature is detected to exceed the reference value reaches a set number, it is determined that the sitting posture of the human body is abnormal.

2. The sitting posture detection method according to claim 1, wherein: The MIMO millimeter-wave radar device transmits FMCW signals to the back of the human body and receives the posture data of the back of the human body.

3. The sitting posture detection method according to claim 1 or 2, characterized in that: The process of calculating the scattering intensity value of each voxel in the target imaging area based on the human back posture data received by the MIMO millimeter wave radar device includes: sorting the preprocessed signal of the human back posture data to obtain a sorted signal; Performing pulse compression on the sorting signal in the distance dimension to obtain a compressed sorting signal; The compressed sorting signal is reconstructed by matched filtering to obtain the scattering intensity value of each voxel in the target imaging area.

4. The sitting posture detection method according to claim 3, wherein: The compressed sorting signal is reconstructed by matched filtering to obtain the scattering intensity value of each voxel in the target imaging area, including: The compressed sorting signal S(x t ,z t ,x r ,z t ,k) Perform matched filtering reconstruction to obtain the estimated value of the scattering intensity of each voxel Use summation instead of integration to estimate the value Expressed as: Where k = 2πf c / c,f c is the starting frequency of the signal, c is the propagation speed of electromagnetic waves in the air, R t is a radar transmitting antenna (x t ,0,z t ) to a certain voxel (x, y, z), R r is the corresponding radar receiving antenna (x r ,0,z r ) and the Euclidean distance between the voxel (x,y,z).

5. The sitting posture detection method according to claim 1, wherein: The calculation method of spinal curvature C' is: Among them, A, B, and C are three key points taken at the tail, middle, and top of the human spine respectively.

6. Sitting posture detection device, characterized in that: The device includes: The sitting posture imaging unit is used to calculate the scattering intensity value of each voxel in the target imaging area based on the human back posture data received by the MIMO millimeter wave radar device, and obtain high-resolution imaging of the human sitting posture based on the scattering intensity value; The spinal curvature calculation unit is used to extract features from high-resolution imaging of the human sitting posture. It uses a deep learning algorithm to select key points at the tail, middle, and top of the human spine to calculate the spinal curvature. The detection unit is used to detect that when the spinal curvature exceeds the reference value for a set number of times, it determines that the human sitting posture is abnormal.

7. MIMO millimeter-wave radar equipment, characterized by: The device includes a MIMO millimeter-wave radar module and a signal processing module. The MIMO millimeter-wave radar module transmits an FMCW signal to the back of a human body and receives human back posture data. The signal processing module adopts the sitting posture detection method described in any one of claims 1 to 5 to process human body data; the MIMO millimeter-wave radar module includes multiple array-arranged "mouth"-shaped MIMO antenna arrays, each "mouth"-shaped MIMO antenna array includes 2X transmitting antenna units and 2Y receiving antenna units, the 2X transmitting antenna units are arranged in two rows of linear arrays and each side contains X transmitting antenna units, the 2Y receiving antenna units are arranged in two columns of linear arrays, that is, each side contains Y receiving antenna units, and the spacing d between adjacent transmitting antenna units and adjacent receiving antenna units is half a wavelength λ / 2: Among them, f c is the starting frequency of the signal, and c is the propagation speed of electromagnetic waves in the air.

8. The MIMO millimeter wave radar device according to claim 7, characterized in that: The signal processing module also includes a heart rate detection device, which includes: A signal acquisition and preprocessing unit, used for acquiring human body data and preprocessing the human body data to obtain a preprocessed signal; A range gate acquisition unit is used to perform distance dimension FFT processing on the pre-processed signal to determine the range gate where the chest cavity is located; The phase extraction unit is used to extract phase change information from the preprocessed signal according to the range gate and perform phase unwrapping operation to obtain the respiratory and heartbeat signals; A respiratory and heartbeat separation unit is used to separate the respiratory and heartbeat signals based on the ICA algorithm to obtain independent heartbeat source signals and respiratory source signals; The heart rate detection unit is used to perform FFT processing on the heartbeat source signal, extract the heartbeat frequency and calculate the heart rate, compare the detected heart rate value with the normal heart rate information of the human body, and determine whether the heart rate is abnormal.

9. The MIMO millimeter-wave radar device according to claim 7 or 8, characterized in that: The signal processing module also includes a sedentary detection device, which includes: A signal acquisition and preprocessing unit, used for acquiring human body data and preprocessing the human body data to obtain a preprocessed signal; A range gate acquisition unit is used to perform distance dimension FFT processing on the pre-processed signal to determine the range gate where the chest cavity is located; A phase extraction unit, configured to extract phase change information from the preprocessed signal according to the range gate; The micro-Doppler detection unit is used to determine whether the human body's sitting posture is stable based on the phase change information. The cumulative timing unit is used to start timing and record the sitting time when it is judged that the human body's sitting posture is stable. When the sitting time exceeds the set time, it is judged as sitting for a long time.

10. A smart health seat comprising the MIMO millimeter-wave radar device according to any one of claims 7 to 9, characterized in that: The seat also includes a main body, a battery system, a wireless Bluetooth communication module and a seat management system. The MIMO millimeter-wave radar device and the battery system are both located in the backrest of the main body. The backrest is made of wave-transparent and heat-dissipating materials and has vents on the back wall. The wireless Bluetooth communication module is located in the MIMO millimeter-wave radar device and is used to encrypt and transmit the detection results of the signal processing module to the seat management system.

Images