Data-driven intelligent gait health detection and positioning system solution

Abstract

The invention discloses a data-driven intelligent gait health detection and positioning system solution, and belongs to the technical field of wearable rehabilitation equipment. The system comprises a three-point flexible film pressure sensor array, an IMU (inertial measurement unit), a GPS (global positioning system) positioning module, a master-slave distributed control platform and a communication module, wherein sensor data is acquired by the slave control board and then transmitted to the main control board, and the main control board realizes gait pattern recognition, fall risk early warning and user positioning based on multi-sensor fusion and a machine learning algorithm; the system adopts a hierarchical processing strategy, a first-level model is deployed on a main control board to realize real-time state recognition and low-power-consumption control, and a second-level model is deployed on a cloud to realize gait type subdivision and health assessment. According to the invention, through multi-source data fusion and edge cloud cooperative computing, the gait recognition precision and the timeliness of fall early warning are improved, the wearing comfort and the system endurance are considered at the same time, and the method is suitable for daily health monitoring of old people, rehabilitation patients and other groups.

Description

Technical Field

[0001] This invention belongs to the field of wearable rehabilitation devices and health monitoring technology, specifically involving a data-driven intelligent gait health detection and positioning system solution. Background Technology

[0002] With the accelerating aging of society, falls among the elderly have become a significant public health issue. Falls not only cause physical injuries but can also lead to prolonged bed rest, psychological distress, and other consequences, placing a heavy burden on families and society.

[0003] Currently, health monitoring devices on the market, such as smart bracelets and fall detectors, mostly rely on a single accelerometer for motion recognition. Their recognition accuracy is limited in complex environments (such as going up and down stairs or walking on slopes), and they cannot obtain key gait features such as plantar pressure distribution, making it difficult to provide risk warnings before falls.

[0004] In addition, existing gait analysis systems, such as optical motion capture and pressure force tables, are highly accurate, but they are large, expensive, and cannot be used outdoors, making them unsuitable for daily continuous monitoring. Summary of the Invention

[0005] This invention aims to overcome the shortcomings of existing technologies and provide a data-driven intelligent gait health detection and positioning system and method. Through an innovative combination of hardware architecture and software algorithms, the system achieves real-time acquisition of multi-source gait data, edge intelligent processing, and cloud-based deep analysis, thereby achieving high-precision gait recognition, early warning of fall risk, and accurate user positioning.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a data-driven intelligent gait health monitoring and positioning system, comprising:

[0007] The sensor unit includes a three-point flexible thin-film pressure sensor array, an IMU inertial measurement unit, and a GPS positioning module, all mounted on the insole.

[0008] The control panel is installed in the insole of the left and right shoes respectively, and is used to collect sensor data of the corresponding feet.

[0009] The main control board, which communicates with the slave control board, is used to receive and fuse data from multiple sensors and perform gait analysis and state recognition.

[0010] The communication module is used for data interaction with external terminals or cloud platforms;

[0011] The main control board has a built-in first-level recognition model, which is used to judge the user's motion status in real time and control the sampling frequency and data transmission strategy.

[0012] Furthermore, the three-point flexible thin-film pressure sensor array is arranged on the inner forefoot, outer forefoot, and heel areas of the insole to collect temporal data of plantar pressure distribution during the gait cycle.

[0013] Furthermore, the slave control board uses an STM32G4 series microcontroller, and the master control board uses an STM32H7 series microcontroller. The two communicate via serial port to upload data and issue commands.

[0014] Furthermore, the main control board is also used to perform the following steps:

[0015] S1. Preprocess the received sensor data and segment the gait cycle;

[0016] S2. Extract time-domain and frequency-domain features and fuse multi-source information such as pressure, attitude, and acceleration;

[0017] S3. Based on the first-level recognition model, determine whether the user is stationary, walking, running, or suspected of falling;

[0018] S4. Enter low-power mode when stationary, and increase sampling rate and trigger data upload when in motion or suspected of falling.

[0019] Furthermore, it also includes a cloud server for running the secondary recognition model, receiving multi-step continuous data uploaded by the main control board, performing gait pattern subdivision and health assessment, and generating fall warning information.

[0020] Furthermore, the secondary recognition model is constructed in the following manner:

[0021] S1. Collect plantar pressure, IMU and positioning data of multiple subjects in different sports scenarios as training set;

[0022] S2. Label the data, including gait type labels and fall event labels;

[0023] S3. Use machine learning algorithms to train the model and optimize the recognition accuracy and recall on the validation set.

[0024] Furthermore, the system is integrated into the shoe tongue and uses a detachable package. The sensor unit and the main control board are connected by a flexible circuit to ensure wearing comfort and system reliability.

[0025] Furthermore, the communication module supports multiple communication methods such as 4G / 5G, WiFi, or Bluetooth, and is used to send early warning information, location coordinates, and health data to the bound mobile terminal or monitoring platform.

[0026] Furthermore, the system also includes a fall warning rule base, which, based on historical data and real-time feature matching, enables early identification and warning of fall precursors.

[0027] Furthermore, the IMU inertial measurement unit includes a three-axis accelerometer, a three-axis gyroscope, and a three-axis magnetometer, used to collect the user's attitude angle, angular velocity, and acceleration information.

[0028] Compared with existing technologies, this invention provides a data-driven intelligent gait health detection and positioning system solution, which has the following beneficial effects:

[0029] 1. This invention employs a three-point flexible thin-film pressure sensor array combined with an IMU (Inertial Measurement Unit) to simultaneously collect multi-dimensional data such as plantar pressure distribution, attitude angle, and acceleration, overcoming the problem of insufficient recognition capability of traditional single sensors in complex scenarios. Simultaneously, a primary model is deployed on the main control board for real-time local processing and status determination, while a secondary model performs in-depth analysis and health assessment in the cloud, balancing real-time response with fine-grained recognition, effectively achieving early warning of fall risk.

[0030] 2. The system of this invention features intelligent power consumption management and adaptive sampling mechanisms, extending the device's battery life and making it suitable for continuous daily monitoring. The main control board can dynamically adjust the sensor sampling frequency and data transmission strategy according to the user's movement status. For example, it enters a low-power mode when stationary and increases the sampling rate and triggers uploading when in motion or suspected of falling, thereby significantly reducing overall energy consumption while ensuring data validity.

[0031] 3. The system of this invention adopts a wearable-friendly design, taking into account comfort, reliability, and practicality. The hardware is integrated into the shoe tongue, using detachable packaging and flexible circuit connection, which does not affect the user's normal walking and wearing experience; it supports multiple communication methods such as 4G / 5G, WiFi, and Bluetooth, and can realize real-time push of early warning information, location data, and health reports, which facilitates remote monitoring and timely intervention by guardians.

[0032] In summary, this system, through collaborative innovation of software and hardware, achieves high-precision, low-power, and wearable health monitoring and positioning functions. It is suitable for daily gait monitoring and fall risk prevention for groups such as the elderly and rehabilitation patients, and has significant social significance and application value. Attached Figure Description

[0033] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention, but do not constitute a limitation thereof. In the drawings:

[0034] Figure 1 This is a schematic diagram of the control system flow of the data-driven intelligent gait health monitoring and positioning system in an embodiment of the present invention;

[0035] Figure 2 This is a schematic diagram of the data acquisition module of the master-slave distributed control platform in an embodiment of the present invention;

[0036] Figure 3 This is a system schematic diagram of the machine learning control algorithm in an embodiment of the present invention;

[0037] Figure 4 This is a schematic diagram illustrating the process of gait mode switching and fall warning rule determination in an embodiment of the present invention. Detailed Implementation

[0038] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0039] Please see Figures 1-4 This embodiment provides a data-driven intelligent gait health detection and positioning system, which is suitable for groups such as the elderly and rehabilitation patients. It is used for real-time gait monitoring, fall warning and location positioning. The system adopts a master-slave distributed architecture, integrates multi-sensor data, and achieves efficient and accurate status recognition and health assessment through edge-cloud collaborative computing.

[0040] The hardware components of this system include sensor units, slave control boards, main control boards, communication modules, and cloud servers.

[0041] The sensor unit includes:

[0042] A three-point flexible thin-film pressure sensor array is arranged on the inner forefoot, outer forefoot, and heel areas of the insole to collect temporal distribution data of plantar pressure during the gait cycle.

[0043] An IMU (Inertial Measurement Unit) includes a three-axis accelerometer, a three-axis gyroscope, and a three-axis magnetometer, used to collect the user's attitude angle, angular velocity, and acceleration information in real time.

[0044] The GPS positioning module is used to obtain the user's real-time location coordinates;

[0045] The control board uses an STM32G4 series microcontroller, which is built into the left and right insoles respectively. It is responsible for collecting sensor data from the corresponding foot and uploading the data to the main control board via serial communication.

[0046] The main control board uses an STM32H7 series microcontroller, which is located in the shoe tongue and communicates with the slave control board to receive, fuse and process data from multiple sources of sensors. The main control board has a built-in first-level recognition model that supports real-time motion state judgment and adaptive control logic.

[0047] The communication module supports multiple communication methods such as 4G / 5G, WiFi and Bluetooth, and is used to send processed data, early warning information and location coordinates to the bound mobile terminal or remote monitoring platform.

[0048] The cloud server deploys a secondary recognition model, receives multi-step continuous data uploaded by the main control board, and performs gait pattern segmentation, health assessment, and fall risk analysis.

[0049] The workflow of this system is as follows: After the system starts up, it will run according to the following steps:

[0050] Step 1: Data Collection and Upload;

[0051] The control board collects plantar pressure, IMU, and positioning data at a set frequency and uploads them to the main control board in real time via serial port.

[0052] Step 2: Data preprocessing and feature extraction;

[0053] The main control board performs preprocessing such as filtering and noise reduction on the received sensor data, and performs gait period segmentation; then, it extracts time-domain features (such as mean, variance, and peak value) and frequency-domain features (such as spectral energy), and fuses multi-source information such as pressure, attitude, and acceleration to form a feature vector;

[0054] Step 3: First-level identification and adaptive control;

[0055] The main control board calls the built-in first-level recognition model to determine the user's current state based on the feature vector, including being stationary, walking, running, or suspected of falling.

[0056] If the system is identified as being in a stationary state, it enters a low-power mode, reduces the sampling frequency, and pauses data upload.

[0057] If movement or a suspected fall is detected, the sensor sampling rate is increased, and the data is uploaded to the cloud via the communication module.

[0058] Step 4: In-depth cloud-based analysis and early warning generation;

[0059] The cloud server receives the uploaded multi-step continuous data, calls the secondary recognition model to subdivide the gait pattern (such as going up and down stairs, walking on ramps, etc.) and conduct health assessment; combined with historical data and real-time features, it matches the fall warning rule library. If a fall warning sign is identified, a warning message is generated and sent to the monitoring terminal along with the user's location.

[0060] Step 5: Model Update and Optimization;

[0061] The system continuously collects labeled data generated in actual use and regularly performs incremental training on the first and second level models to improve recognition accuracy and environmental adaptability.

[0062] Example 1: Model Training and Validation;

[0063] To construct the primary and secondary recognition models, this embodiment organized 10 subjects to collect gait data in multiple scenarios; the experimental tasks included: walking on flat ground, walking on a treadmill at different speeds, going up and down stairs, walking on slopes, walking freely indoors and outdoors, and simulated falls under protective conditions;

[0064] After preprocessing and feature extraction, the collected data is divided into training and validation sets. Machine learning algorithms (such as random forest, support vector machine or neural network) are used to train the model, with the weighted sum of cross-entropy loss and mean squared error loss as the objective function. The parameters are updated iteratively through the Adam optimizer. The model is evaluated on the validation set for accuracy, recall and F1 score until the performance meets the set threshold and then it is deployed.

[0065] To ensure comfort and ease of use, the system hardware is detachably packaged, with the main body integrated into the shoe tongue. The sensors and the main control board are connected by flexible circuitry to avoid interfering with normal walking. The overall design takes into account waterproofing, shockproofing, and durability, making it suitable for long-term daily wear.

[0066] This invention combines multi-sensor fusion, edge intelligent processing, and cloud-based deep analysis to achieve high-precision, low-power, and wearable gait health detection and fall warning. The system has an adaptive sampling and communication mechanism, which extends battery life while ensuring the real-time performance and reliability of monitoring. It is suitable for daily health monitoring and safety protection for the elderly, rehabilitation patients, and other groups.

[0067] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A data-driven intelligent gait health monitoring and positioning system, characterized in that, include: The sensor unit includes a three-point flexible thin-film pressure sensor array, an IMU inertial measurement unit, and a GPS positioning module, all mounted on the insole. The control panel is installed in the insole of the left and right shoes respectively, and is used to collect sensor data of the corresponding feet. The main control board, which communicates with the slave control board, is used to receive and fuse data from multiple sensors and perform gait analysis and state recognition. The communication module is used for data interaction with external terminals or cloud platforms; The main control board has a built-in first-level recognition model, which is used to judge the user's motion status in real time and control the sampling frequency and data transmission strategy.

2. The data-driven intelligent gait health detection and positioning system according to claim 1, characterized in that, The three-point flexible thin-film pressure sensor array is arranged on the inner forefoot, outer forefoot, and heel areas of the insole to collect temporal data of plantar pressure distribution during the gait cycle.

3. The data-driven intelligent gait health detection and positioning system according to claim 1, characterized in that, The slave control board uses an STM32G4 series microcontroller, and the master control board uses an STM32H7 series microcontroller. The two communicate via serial port to upload data and send commands.

4. The data-driven intelligent gait health detection and positioning system according to claim 1, characterized in that, The main control board is also used to perform the following steps: S1. Preprocess the received sensor data and segment the gait cycle; S2. Extract time-domain and frequency-domain features and fuse multi-source information such as pressure, attitude, and acceleration; S3. Based on the first-level recognition model, determine whether the user is stationary, walking, running, or suspected of falling; S4. Enter low-power mode when stationary, and increase sampling rate and trigger data upload when in motion or suspected of falling.

5. The data-driven intelligent gait health detection and positioning system according to claim 1, characterized in that, It also includes a cloud server for running the secondary recognition model, receiving multi-step continuous data uploaded by the main control board, performing gait pattern subdivision and health assessment, and generating fall warning information.

6. The data-driven intelligent gait health detection and positioning system according to claim 5, characterized in that, The secondary recognition model is constructed in the following way: S1. Collect plantar pressure, IMU and positioning data of multiple subjects in different sports scenarios as training set; S2. Label the data, including gait type labels and fall event labels; S3. Use machine learning algorithms to train the model and optimize the recognition accuracy and recall on the validation set.

7. The data-driven intelligent gait health detection and positioning system according to claim 1, characterized in that, The system is integrated into the shoe tongue and uses a detachable package. The sensor unit and the main control board are connected by a flexible circuit to ensure wearing comfort and system reliability.

8. The data-driven intelligent gait health detection and positioning system according to claim 1, characterized in that, The communication module supports multiple communication methods such as 4G / 5G, WiFi, or Bluetooth, and is used to send early warning information, location coordinates, and health data to the bound mobile terminal or monitoring platform.

9. The data-driven intelligent gait health detection and positioning system according to claim 1, characterized in that, The system also includes a fall warning rule base, which uses historical data and real-time feature matching to enable early identification and warning of fall precursors.

10. The data-driven intelligent gait health detection and positioning system according to claim 1, characterized in that, The IMU (Inertial Measurement Unit) includes a three-axis accelerometer, a three-axis gyroscope, and a three-axis magnetometer, used to collect the user's attitude angle, angular velocity, and acceleration information.

Images