An intelligent floor system capable of recognizing fall behavior

Abstract

The application belongs to the technical field of intelligent security and protection home, and relates to a smart floor system capable of identifying falling behaviors, which comprises a smart pressure sensing floor, a data acquisition system and an upper computer, the smart pressure sensing floor is a composite structure of a solid floor and a flexible sensing fabric, the flexible sensing fabric contains 40 electrodes and 40*40 array distributed sensors, and a piezoresistive sensing layer is constructed by a graphene-carbon black mixed filler to form a conductive network; the data acquisition system collects signals through FPC flat cable, transmits data to the upper computer through a sensor router and a public network router; the upper computer adopts a "cloud-edge-end" architecture, carries a machine learning identification algorithm, can accurately identify falling and intrusion behaviors, the falling identification accuracy is more than 98%, and the response time is less than or equal to 1 second; the terminal side is associated with a cloud program to realize alarm pushing and data viewing functions. The disclosed system simultaneously integrates two core functions of old people falling monitoring and home safety protection, and meets the diversified needs of users living alone.

Description

Technical Field

[0001] This invention belongs to the field of smart home security technology, and relates to smart flooring, specifically a smart flooring system that can identify fall behavior, suitable for scenarios such as home safety monitoring for elderly people living alone and home security early warning. Background Technology

[0002] With the accelerating aging of the global population, the number of elderly people living alone continues to expand, making home safety monitoring a pressing social issue. Statistics show that the incidence of falls at home among people aged 65 and above is as high as 30%, and the proportion of injuries and deaths resulting from falls due to lack of timely medical attention is significantly increasing. Therefore, fall monitoring has become a core need in home-based elderly care.

[0003] Existing fall detection technologies are mainly divided into three categories: First, monitoring solutions based on wearable devices (such as smart bracelets and belts), which rely on continuous wearing by the user. Elderly people may refuse to use them or forget to wear them due to comfort issues, posing a risk of monitoring interruption. Second, monitoring solutions based on environmental sensors (such as infrared sensors and millimeter-wave radar). These devices have limited coverage and are easily obstructed by furniture and environmental interference, resulting in a high false alarm rate (usually exceeding 20%), making it difficult to achieve blind-spot-free monitoring throughout the house. Third, monitoring solutions based on video surveillance, although they can intuitively identify fall behavior, pose a serious risk of privacy leakage, and the equipment deployment cost is high, making them unsuitable for private spaces such as bedrooms and bathrooms.

[0004] Chinese patent CN105608839A discloses an indoor fall detection system for the elderly, employing a separate architecture for "sensing" and "positioning." Specifically, piezoelectric cables and other sensors detect falls, while independent systems such as cameras handle subsequent location and confirmation. This system requires the simultaneous deployment of piezoelectric cable sensors and camera components, resulting in a complex structure, difficult installation, and unresolved privacy concerns despite the camera's presence. Furthermore, the high cost of the entire system hinders large-scale deployment. In addition, existing technologies lack modular, expandable, and adaptable large-area monitoring solutions for different apartment layouts. Traditional array sensors are mostly custom-made, requiring redesigned circuits and structures for different indoor spaces, leading to long production cycles and high costs.

[0005] Array sensors can achieve high-resolution full-area monitoring through the collaborative work of multiple sensing units, while textile-based materials have the characteristics of flexibility, bendability, and easy laying. The flexible sensing fabric formed by combining the two can not only accurately capture pressure distribution and dynamic changes, but also achieve large-area coverage through modular splicing, providing a feasible path to solve the pain points of existing technologies.

[0006] Therefore, developing a smart floor system that features non-intrusive monitoring, strong privacy protection, modular expansion, and high accuracy has significant technological value and market potential. Summary of the Invention

[0007] To address the problems of existing fall detection technologies, such as privacy leaks, the need for user cooperation, high false alarm rates, and difficulty in large-scale deployment, this invention provides an intelligent flooring system that can identify fall behavior.

[0008] Technical solution

[0009] A smart flooring system capable of recognizing fall behavior consists of three parts: a smart pressure-sensing floor, a data acquisition and transmission system, and a host computer system. These parts work together to achieve pressure sensing, data transmission, and intelligent analysis.

[0010] (a) Intelligent pressure-sensing floor

[0011] It adopts a composite structure of "flexible sensing fabric + solid floor", with the core being a modularly designed flexible sensing blanket unit, the specific structure of which is as follows:

[0012] Flexible sensing fabric: It adopts a sandwich structure, consisting of an upper electrode layer, a piezoresistive sensing layer, and a lower electrode layer from top to bottom. Each layer is fixed by contact separation assembly.

[0013] Electrode layer: A flexible composite conductive fabric is used as the substrate, and a nano-silver conductive layer is deposited on the surface through an interval coating process. The conductive layer has a thickness of 150±5nm, a line width of 50mm, and a spacing of 50mm, forming a strip electrode array. The upper electrode is the driving electrode, and the lower electrode is the receiving electrode. The two are laid perpendicularly and orthogonally to form a two-dimensional grid-like sensing array. The effective sensing area of ​​a single sensing unit is 50mm×50mm. The average surface resistance of the electrode layer is ≤2.98Ω / cm, and the resistance change rate after 1000 bends is ≤5%, exhibiting good conductivity and mechanical flexibility. The electrode layer corresponds to 40 electrodes to realize signal transmission and acquisition.

[0014] Piezoresistive sensing layer: The conductive knitted fabric is impregnated with a graphene-carbon black mixed conductive filler (mass ratio 1:3), and then dried and cured to form a stable three-dimensional conductive network; the sensing layer thickness is 2±0.2mm, the pressure response range is 0-1000Kpa, the response time is <500ms, the recovery time is <300ms, the linearity error of the resistance value with pressure change is ≤3%, and the stability is good in an environment of -10℃~60℃;

[0015] Composite Assembly: The flexible sensing fabric is encapsulated into a standard-sized (3.6m×2.8m, area 10m²) sensing blanket unit, with each unit containing 1600 sensing units (40×40 array), each corresponding to a pressure sensor; the solid floor is fixed to the surface of the sensing blanket unit in an "I" shaped arrangement, which not only physically fixes the sensing fabric but also ensures the comfort of home use; multiple sensing blanket units can be seamlessly spliced ​​through electrode interfaces to adapt to larger areas or irregularly shaped indoor spaces.

[0016] (ii) Data Acquisition System

[0017] It is responsible for the acquisition, conditioning, and wireless transmission of pressure signals. The specific configuration is as follows:

[0018] Hardware components include an array piezoresistive sensing signal acquisition and conditioning circuit, a main control MCU (model STM32F407), a wireless communication module (i.e., a Wi-Fi module supporting 802.11n / b / g protocols), and a USB TYPE-C power supply interface. The acquisition channels adopt an array design, with configurable horizontal and vertical channels (1-40 each), adapting to sensing blanket units of different sizes. The data acquisition system connects to the electrode interface of the flexible sensing fabric via an FPC cable to achieve stable signal transmission. Data cables connect to relevant devices to ensure reliable power supply and data interaction.

[0019] Signal processing flow: The pressure signals from each sensor are filtered by the anti-crosstalk conditioning circuit (low-pass filter with a cutoff frequency of 10Hz), amplified (amplification factor of 100 times), and then sent to the main control MCU for A / D conversion (sampling accuracy of 12 bits), preliminary filtering and normalization processing to eliminate environmental noise and circuit interference.

[0020] Data transmission: The acquisition frequency can be adjusted within the range of 10-100Hz (default 50Hz). The processed digital pressure data is uploaded to the sensor router via the Wi-Fi module using the TCP protocol, and then forwarded to the cloud server via the public network router. The Wi-Fi module operates at 5V (powered by the USB TYPE-C interface), has a transmission distance of ≤50m (unobstructed), a data transmission rate of ≥1Mbps, and supports disconnection and reconnection functions.

[0021] Anti-crosstalk design: The acquisition circuit adopts a differential input method, and each channel is independently configured with an isolation resistor (10kΩ) and a filter capacitor (0.1μF) to effectively suppress crosstalk between adjacent channels, with a crosstalk suppression ratio ≥60dB.

[0022] (III) Host Computer System

[0023] Employing a "cloud-edge-device" collaborative architecture, it achieves data storage, intelligent analysis, and result delivery, including two working modes: cloud guardian mode and cloud guard mode.

[0024] The "cloud-edge-device" collaborative architecture consists of:

[0025] Edge side: Deploy a local gateway (model ESP32-C3) to preprocess the collected data (including signal noise reduction, outlier removal, and basic feature extraction), with a preprocessing latency of ≤100ms, reducing the computing pressure on the cloud;

[0026] On the cloud side: Deploy a cloud server (configured with CPU ≥ 8 cores, memory ≥ 16GB, and storage ≥ 1TB), equipped with a multimodal behavior recognition algorithm based on machine learning. The algorithm is trained on a dataset of 100,000+ samples (including behavioral data such as standing, walking, and falling from different weights and ages). The cloud server is connected to the cloud mini-program service to achieve data synchronization and interaction.

[0027] Terminal side: including WeChat mini program and bound mobile terminal (phone), supporting functions such as data viewing, mode switching, alarm reception;

[0028] Working mode and core algorithm (derived from experimental data statistics):

[0029] Mode 1: Cloud Guardian Mode (Home Monitoring for the Elderly) Core Functions: Recognizes standing, walking, falls, and provides early warnings for prolonged falls. The specific algorithm process is as follows:

[0030] S1: Floor Calibration: In unmanned mode, collect 10-30 frames (default 20 frames) of pressure data from each sensor, calculate the average value as the baseline value, and subtract the corresponding baseline value from all subsequent data to eliminate the influence of the floor's own weight;

[0031] S2: Feature extraction: Real-time acquisition of pressure data, calculation of feature parameters such as pressure coefficient of variation (CV value), pressure distribution area, pressure center trajectory, and long side length of pressure region;

[0032] S3: Behavioral Judgment

[0033] Standing determination: Real-time pressure coefficient of variation > set threshold lower limit (the threshold lower limit is the minimum CV value when a person walks after calibration, default 0.3);

[0034] Walking determination: The real-time pressure variation coefficient is within the threshold range, and the pressure center trajectory changes continuously (displacement ≥ 5cm / frame).

[0035] Fall detection: Real-time pressure coefficient of variation < set upper threshold (the upper threshold is the maximum CV value when the human body is lying flat after calibration, the default is 0.1);

[0036] S4: Dynamic threshold adjustment:

[0037] Weight-appropriate: When the weight is greater than 80kg, the overall threshold range increases by 5%;

[0038] Walking adaptation: When continuous changes in the pressure center trajectory are detected, the threshold range decreases by 10-20%;

[0039] Region adaptation: For every 1cm the length of the long side of the pressure zone exceeds the maximum value of the long side when standing, the threshold increases by 2.5%; for every 1cm decrease, the threshold decreases by 2.5%.

[0040] S5: Alarm Mechanism: After a fall is detected, continuous monitoring is performed for 1 minute. If no change in pressure distribution is detected (indicating inability to get up), an alarm message is immediately pushed through the cloud mini-program service, and the bound mobile phone number is dialed (up to 3 contacts, dialed in order of priority).

[0041] Mode 2: Cloud Guardian Mode (Home Security Early Warning) Core Function: Monitors home intrusions when individuals living alone are away. The specific algorithm process is as follows:

[0042] Step 1: Floor Calibration: The calibration method is the same as that of Cloud Guardian Mode, eliminating the influence of floor weight;

[0043] Step 2: Intrusion Detection: If any sensor pressure value is detected after calibration to be >0.5KPa (ambient noise threshold) and the duration is >3 seconds, it is determined to be an intrusion by personnel;

[0044] Step 3: Alarm Mechanism: Immediately push intrusion alerts via cloud mini-program service and call the bound mobile phone number. At the same time, record the intrusion time and pressure distribution trajectory and store them in the cloud (retained for 30 days).

[0045] Data security: Cloud data is stored using AES-256 encryption, and the transmission process uses TLS1.2 encryption. User identity is verified through mobile phone number and verification code to ensure data privacy and security.

[0046] Beneficial effects

[0047] Compared to conventional array-type electronic thin-film sensors, this invention integrates the sensor within the floor, eliminating the need for cameras or wearable devices. This achieves truly seamless and concealed monitoring, completely avoiding privacy issues and increasing user acceptance. The core flexible sensing fabric adopts a standard modular design, allowing for flexible adaptation to rooms of different sizes and shapes through splicing. This greatly simplifies the installation process, reduces damage to existing decorations, and facilitates large-scale production and cost control. Based on a high-density two-dimensional pressure sensing array, it can capture subtle pressure distribution and changes. Combined with multi-dimensional feature extraction and dynamic threshold adjustment algorithms, it significantly improves the accuracy of behavior recognition (especially fall detection), effectively reducing false alarms and missed alarms. It adopts common communication protocols (such as Wi-Fi) and power supply interfaces (such as USB Type-C), facilitating integration with existing home networks (via public network routers and sensor routers). The sensing material is specially designed with excellent flexibility, bending resistance, and environmental stability, ensuring long-term reliability. The Wi-Fi module supports mainstream routing protocols, and the cloud-based mini-program service is compatible with iOS and Android systems. The USB Type-C power interface is compatible with common chargers, ensuring strong system compatibility. High stability and long lifespan are achieved through the use of nano-silver conductive material in the electrode layer (corresponding to 40 electrodes), offering excellent bending performance. The sensing layer uses graphene-carbon black composite filler, ensuring high stability. Tests show it can operate continuously for ≥30,000 hours, meeting the needs of long-term home use. This system integrates two core functions: fall monitoring for the elderly and home safety protection, meeting the diverse needs of users living alone and enhancing the product's practical value. Attached Figure Description

[0048] Figure 1 Schematic diagram of the overall structure of the intelligent floor system;

[0049] Figure 2 Schematic diagram of data processing in the host computer system;

[0050] Figure 3 Sensitivity characteristics of intelligent sensing flooring;

[0051] Figure 4 Cloud-based mini-program function flowchart;

[0052] Figure 5 Example 1: Product schematic diagram;

[0053] The components are labeled as follows: 1. Solid floor, 2. FPC cable, 3. Data acquisition system, 4. Data cable, 5. Host computer, 6. Flexible sensing fabric, 7. 40-channel electrode, 8. Public network router, 9. Cloud mini-program service, 10. Sensor router, 11. Pressure sensor. Detailed Implementation

[0054] The present invention will be described in detail below with reference to embodiments to enable those skilled in the art to better understand the invention, but the invention is not limited to the following embodiments. Furthermore, unless specific techniques or conditions are specified in the embodiments, they are all performed according to conventional techniques or conditions in the art.

[0055] Example 1

[0056] Standard apartment type (10m²) 2 Bedroom) Deployment Plan

[0057] Deployment of smart pressure-sensing floors:

[0058] Select a standard-sized (3.6m×2.8m) flexible sensing blanket unit and lay it flat on the bedroom floor. The edges of the sensing blanket are fixed with tape to prevent displacement. Arrange Shengxiang flooring in a "I" shape on the surface of the sensing blanket. The seams of the flooring are staggered from the edges of the sensing blanket unit to avoid affecting the sensitivity of the sensing unit.

[0059] Installation of the data acquisition and transmission system:

[0060] The data acquisition and transmission module is fixed in the corner of the bedroom wall (≥30cm from the ground) and connected to the electrode interface of the sensing blanket via an FPC flexible cable (all 40 horizontal and 40 vertical channels are enabled); the module is powered by a 5V / 2A charger connected to a USB TYPE-C interface, and the Wi-Fi module is connected to the home router (2.4GHz band) to establish a TCP long connection with the cloud server;

[0061] Host computer system configuration:

[0062] Users can search for and register the mini-program on WeChat, and bind their mobile phone number (supporting 1 primary contact + 2 backup contacts); enter the device addition page, scan the QR code on the data collection and transmission module to complete the device binding, and set the monitoring mode to "cloud guardian mode" with a monitoring period of 24 hours;

[0063] System calibration and testing:

[0064] After the device is paired, it automatically enters calibration mode and collects 20 frames of pressure data under unattended conditions to complete the baseline value calculation. Three test subjects weighing 60kg, 75kg, and 90kg were invited to conduct standing, walking, and fall (simulated fall, no actual injury) tests respectively.

[0065] Standing position: The system has a 100% recognition accuracy and a response time of ≤300ms;

[0066] Walking status: The system's recognition accuracy is 100%, with no false alarms;

[0067] Falling status: The system has an accuracy rate of 98.5%, a response time of ≤1 second, and will automatically dial the main contact's mobile phone number if the person does not get up within 1 minute. The alarm information will also be pushed to the mini program.

[0068] Cloud Guardian Mode Test:

[0069] After a user switches to "Cloud Guardian Mode" in the mini-program and leaves home, they invite a tester into their bedroom. Once the system detects a change in stress, it immediately sends an intrusion alert to the user's phone and dials the main contact's number. The intrusion trajectory data is stored in the cloud and can be viewed through the mini-program.

[0070] The core of the intelligent floor system is a multi-layered flexible sensing blanket with a sandwich structure. Both the upper and lower electrode layers utilize nano-silver conductive knitted fabric, with multi-channel, mutually insulated strip electrodes formed on a polyester base fabric through a precision hot-pressing process. The strip electrodes of the upper and lower layers are arranged spatially orthogonally, their intersections forming an independent pressure sensing unit. In this embodiment, the spacing between the orthogonal electrodes is designed to be 0.5 cm, thereby forming a high-density dot matrix of 1600 sensing units within the monitoring area. This resolution is sufficient to accurately capture the subtle distribution contours and dynamic changes of foot pressure. Between the two electrode layers is a graphene-carbon black functional fabric serving as a piezoresistive sensing layer. This fabric uses polyester fiber as a matrix, and through impregnation-drying or in-situ synthesis processes, highly conductive graphene nanosheets and carbon black particles are used as mixed conductive fillers, firmly and uniformly loaded onto the fiber surface and interwoven network, forming a stable three-dimensional conductive network. When pressure is applied to the sensing unit, the contact state and conductive path density of the fibers inside the fabric change, resulting in a sensitive and reversible change in its macroscopic resistance value. To ensure signal stability and product durability, the aforementioned sandwich structure is encapsulated into a standard-sized sensing carpet, and then Shengxiang flooring is securely fixed to the sensing carpet in an "I" shape. This "I"-shaped fixing method ensures that there is no relative displacement between the sensing layer and the upper and lower electrode layers during use, avoiding signal drift or damage caused by interlayer slippage, thus guaranteeing long-term reliability.

[0071] The array-type data acquisition circuit system connects to each electrode of the sensing blanket via FPC flexible cables, cyclically scanning all sensing units at a frequency of 50Hz to read their real-time resistance values ​​and converting them into digital pressure signals using a built-in algorithm. The processed pressure distribution data is uploaded to a cloud server via an integrated Wi-Fi module. A cloud-deployed machine learning algorithm analyzes the continuous data stream in real time, extracting various feature parameters such as the pressure center trajectory, pressure distribution area, and the rate of change of pressure values ​​over time to achieve accurate gait recognition and fall detection. This model, trained on a large dataset containing both normal activities and fall scenarios, exhibits excellent generalization ability. Finally, the system's judgment results and alarm information are pushed to the user's WeChat mini-program and linked mobile phone number. In actual deployment, the system demonstrated excellent performance, recognizing the basic behavioral states (standing, walking, and falling) of adults weighing 40-90kg in simulated tests, with a fall recognition accuracy exceeding 98% and a second-level response time, promptly sending notifications upon detecting a fall. If a person is detected falling and unable to get up for one minute, an alarm is immediately triggered by calling the caregiver's mobile phone (or through voice call / vibration alert, etc.). This fully verifies the technical feasibility and practical value of the system.

[0072] Example 2

[0073] Large apartment (30m²) 2 Living Room Deployment Plan

[0074] Deployment of smart pressure-sensing floors:

[0075] Three standard-sized flexible sensing blanket units are selected and seamlessly spliced ​​together through electrode docking interfaces (resistance change rate at splicing point ≤3%) to form a large sensing area of ​​3.6m × 8.4m; after splicing, the sensing blanket units are laid flat on the living room floor, and the edges are fixed by buckles; solid wood flooring (adapted to the composite structure of this invention) is laid, and the flooring arrangement direction is consistent with the electrode direction of the sensing units;

[0076] Installation of the data acquisition and transmission system:

[0077] The system adopts a distributed layout of one main acquisition module and two slave acquisition modules. The main module is fixed in the central corner of the living room, and the slave modules are fixed in the two side corners respectively. The slave modules are connected to the main module via RS485 bus. The main module communicates with the cloud server via Wi-Fi. The acquisition frequency is set to 80Hz to meet the high-resolution monitoring requirements of large areas.

[0078] System testing:

[0079] The tester conducted a fall simulation test in different areas of the living room (joints, edge areas, and center areas). The system's recognition accuracy was 98.2%, with no signal blind spots at the joints, and the alarm mechanism was triggered normally.

[0080] Example 3

[0081] Cloud Guardian Mode Security Test

[0082] System Configuration:

[0083] Users set up "Cloud Guardian Mode", enable intrusion warning function, and set sensitivity to "High" (pressure detection threshold 0.3KPA).

[0084] Intrusion simulation test:

[0085] Test 1: A pet cat (5kg) entered the monitoring area. The system detected the pressure change but determined it to be a non-human intrusion (pressure area < set threshold), and no alarm was triggered.

[0086] Test 2: When a stranger (65kg) enters the monitoring area, the system immediately sends an intrusion alert and calls the user, recording the intrusion time (accurate to the second) and movement trajectory. The test accuracy is 100%.

[0087] The manual defines a public network router as a standard network device that connects the local area network (LAN) of the smart floor system (networked via sensor routers or home routers) to the public internet. Its core function is to enable remote cross-network data transmission from the indoor site to the cloud server, serving as the key network hub connecting the "edge" to the "cloud" in the entire system's "cloud-edge-device" architecture.

[0088] This invention combines flexible textile materials, array sensing technology, wireless communication, and intelligent algorithms to construct an efficient, reliable, and easy-to-use intelligent floor monitoring system. Those skilled in the art should understand that, without departing from the concept of this invention, various modifications and substitutions can be made to the materials (e.g., electrodes using other metal nanowires), structural parameters (e.g., electrode spacing, sensing unit density), and algorithm details (e.g., feature selection, threshold setting) in the above embodiments, and all such modifications fall within the scope of protection claimed by this invention.

[0089] The above description is merely a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A smart flooring system capable of recognizing fall behavior, characterized in that, Includes an intelligent pressure-sensing floor, a data acquisition system (3), and a host computer (5), wherein: The intelligent pressure sensing floor is a composite structure, including a flexible sensing fabric (6) and a solid floor (1) laid on its surface. Furthermore, the solid floor (1) is fixed to the surface of the flexible sensing fabric (6) in an "I" shaped arrangement. The flexible sensing fabric (6) is a modularly designed sensing unit, which includes an upper electrode layer, a lower electrode layer and a piezoresistive sensing layer between them; the electrode strips of the upper electrode layer and the lower electrode layer are arranged perpendicularly and orthogonally to form a two-dimensional sensing array, and the intersection of the array constitutes a pressure sensor (11). The data acquisition system (3) is electrically connected to the flexible sensing fabric (6) for acquiring and processing the signal of the pressure sensor (11) and transmitting the processed pressure data to the host computer (5) via the network. Furthermore, the data acquisition system (3) is connected to the flexible sensing fabric (6) via the FPC cable (2) and powered by the USB TYPE-C interface. The host computer (5) is used to receive and analyze the pressure data. The host computer (5) includes at least a first working mode and a second working mode. In the first working mode, the host computer (5) is used to identify the user's behavior state based on the pressure data and trigger an alarm when a fall is detected. In the second working mode, the host computer (5) is used to detect personnel intrusion behavior based on the pressure data and trigger an alarm.

2. The intelligent floor system capable of recognizing fall behavior according to claim 1, characterized in that: The flexible sensing fabric (6) is a multi-layer composite structure, and each layer is fixed by contact separation assembly; the piezoresistive sensing layer is based on conductive fabric and impregnated with a mixed conductive filler of graphene and carbon black.

3. The intelligent floor system capable of recognizing fall behavior according to claim 2, characterized in that: The mass ratio of graphene to carbon black is 1:3; the thickness of the piezoresistive sensing layer is 2±0.2mm, the pressure response range is 0-1000kPa, the response time is <500ms, and the recovery time is <300ms.

4. The intelligent floor system capable of recognizing fall behavior according to claim 1, characterized in that: The flexible sensing fabric (6) is a standardized rectangular sensing blanket unit. Each sensing blanket unit contains multiple pressure sensors (11) arranged in an M×N array. Multiple sensing blanket units can be electrically connected and physically spliced ​​through an electrode interface, where M and N are both positive integers.

5. The intelligent floor system capable of recognizing fall behavior according to claim 4, characterized in that: The standard size of the sensing blanket unit is 3.6m×2.8m, containing 40 driving electrodes and 40 receiving electrodes to form a 40×40 sensor array. The effective sensing area of ​​a single pressure sensor (11) is 50mm×50mm.

6. The intelligent floor system capable of recognizing fall behavior according to claim 1, characterized in that: Both the upper and lower electrode layers are based on flexible composite conductive fabric with a nano-silver conductive layer on the surface; the average surface resistance of the electrode layers is ≤2.98Ω / cm, and the resistance change rate after 1000 bends is ≤5%.

7. The intelligent floor system capable of recognizing fall behavior according to claim 1, characterized in that: The data acquisition system (3) includes a main control MCU, an array piezoresistive sensor signal conditioning circuit and a wireless communication module; the array piezoresistive sensor signal conditioning circuit includes a low-pass filter and an amplifier; the main control MCU is used to perform analog-to-digital conversion and preprocessing on the conditioned signal; the wireless communication module is used to send data to the host computer (5).

8. The intelligent floor system capable of recognizing fall behavior according to claim 7, characterized in that: The number of acquisition channels of the data acquisition system (3) is configurable, and the acquisition frequency is adjustable in the range of 10-100Hz; the cutoff frequency of the array piezoresistive sensing signal conditioning circuit is 10Hz, and the amplification factor is 100 times; the analog-to-digital conversion sampling accuracy of the main control MCU is 12 bits.

9. The intelligent floor system capable of recognizing fall behavior according to claim 7, characterized in that: The data acquisition system (3) adopts a differential input method, and each acquisition channel is equipped with an independent isolation resistor and a filter capacitor; the wireless communication module is a Wi-Fi module that supports the 802.11n / b / g protocol.

10. The intelligent floor system capable of recognizing fall behavior according to claim 1, characterized in that: The host computer (5) adopts a "cloud-edge-device" collaborative architecture, including: The local gateway on the edge side is used to perform real-time preprocessing of the received stress data; A cloud server, equipped with a machine learning-based behavior recognition model, is used to perform intelligent analysis in the first working mode and / or the second working mode; The terminal application is used for users to configure the system, view data, and receive alarm information; further, the terminal application is a mobile terminal applet that supports binding multiple contacts; the data storage and transmission process of the cloud server adopts encryption measures.

Images