A multi-modal physiological indicator fusion estimates damage treadmill

Abstract

This invention relates to a treadmill for estimating injury using multimodal physiological index fusion, comprising a multimodal monitoring unit. The multimodal monitoring unit includes a foot pressure sensor array embedded below the running platform assembly, a visual capture system mounted on the support frame, a physiological signal monitoring module, and a data processing module. This invention offers advantages such as high-precision multidimensional force sensing, multimodal fusion monitoring, in-situ testing in multiple environments, medical-grade diagnostic assistance, structural optimization and energy saving, and intelligent early warning and safety features.

Description

Technical Field

[0001] This invention relates to the field of medical rehabilitation training technology, and in particular to a treadmill for estimating injury by fusing multimodal physiological indicators. Background Technology

[0002] Current conventional methods for testing lower limb motor function mainly involve doctors directly observing the patient's gait or using tools such as stopwatches, measuring tapes, and ink to record and assess lower limb movement and footprints. This method requires a high level of professional knowledge and experience from doctors and suffers from strong subjectivity and difficulty in data quantification. Furthermore, existing three-dimensional gait analysis systems are mostly fixed laboratory equipment, which is expensive and difficult to widely apply in community or home rehabilitation settings.

[0003] A search revealed that existing patent CN113869260A discloses a motion capture-based fitness assistance system that uses motion capture technology on a treadmill to correct running posture. However, this solution struggles to capture the pressure distribution and multidimensional torque information in the subtle areas of the sole, lacking in-depth detection of the directionality and distribution characteristics of the force. Furthermore, it lacks a real-time injury warning mechanism during patient movement and fails to effectively integrate physiological indicators such as heart rate with gait data. If patients experience gait abnormalities, exhaustion, or excessive fatigue during training, they are prone to falls or secondary injuries, posing a significant safety hazard. Summary of the Invention

[0004] The purpose of this invention is to provide a treadmill that uses multimodal physiological index fusion to estimate injury, in order to solve the above-mentioned problems.

[0005] The technical solution of this invention is: a treadmill for estimating injury by fusing multimodal physiological indicators, comprising a frame assembly, a drive system, and a running platform assembly, and further comprising: A multimodal monitoring unit, comprising a plantar pressure sensor array embedded below the treadmill assembly, a visual capture system mounted on the support frame, a physiological signal monitoring module, and a data processing module; A control terminal is communicatively connected to the data processing module and is used to display data and receive user commands. The data processing module has a built-in preset gait analysis algorithm, which is used to perform in-situ gait detection and pre-diagnosis and fatigue assessment based on the data collected by the plantar pressure sensor array, the visual capture system and the physiological signal monitoring module. Specifically, the following steps are included: S1. Data Preprocessing: Based on the characteristics of the pressure sensor data and heart rate spectrum, perform data cleaning, preprocessing and standardization. S2. Feature Engineering: Use Pearson correlation coefficient method and principal component analysis to extract key gait features, such as gait frequency, ground contact time, and contact force distribution cloud map; S3. Model Construction and Mapping: A database is established based on the runner's physical characteristics to construct a mapping relationship between "runner's condition - lower limb dynamics - lower limb fatigue". The mapping relationship is established using a multiple linear regression model or a support vector machine classification model. The criterion for judging lower limb fatigue is based on a weighted comprehensive score of heart rate variability index and gait symmetry index. When the comprehensive score is lower than a preset threshold, it is judged as a fatigue state. S4. Report Output: The report includes plantar pressure heatmap data, pressure center trajectory data, joint angle change curve data, joint torque peak data, muscle activation estimation data, and heart rate-gait-fatigue mapping analysis conclusions.

[0006] The surface of the treadmill assembly is provided with a modular textured panel; the modular textured panel is detachably mounted on the surface of the treadmill belt.

[0007] The plantar pressure sensor array consists of a thin-film pressure sensor array and a multi-dimensional force sensor module.

[0008] The physiological signal monitoring module uses an optical heart rate sensor based on photoplethysmography; the visual capture system includes a depth camera or an infrared motion capture camera to collect data signals of left and right knee joint angles, left and right hip joint angles, and stride length.

[0009] The pressure center trajectory data is calculated using a pressure value weighted centroid algorithm collected by a pressure sensor array; the joint angle change curve data is calculated using an inverse kinematics algorithm based on the feature point coordinates collected by a visual capture system; the joint torque peak data is calculated using a Newton-Euler inverse dynamics algorithm combined with a human biomechanical model; and the muscle activation estimation data is estimated based on a biomechanical model of the relationship between joint torque and muscle lever arm.

[0010] The control terminal is fixedly connected to a touch screen; the system has a preset safety threshold. When the monitored lower limb movement or physiological state exceeds the preset safety threshold, the system will issue an early warning by flashing the touch screen and using a speaker.

[0011] The control terminal is fixedly connected to handrails at both ends, and the handrails are connected to the support column through a lifting mechanism.

[0012] A solar panel is rotatably connected to the support frame via a rotating shaft. The solar panel is electrically connected to the energy storage battery inside the electromechanical box via a charging controller.

[0013] The working process of a treadmill is as follows: Step S1: Information entry; Step S2: Scene selection; The user selects the desired treadmill incline and replaces the modular texture panel; Step S3: Movement and Data Acquisition; The user begins running training, and the drive system moves the treadmill components; the plantar pressure sensor array collects pressure distribution and multi-dimensional torque data in real time, the visual capture system simultaneously collects video streams of lower limb joint movement, and the PPG sensor simultaneously collects heart rate data. Step S4: Real-time analysis and early warning; The data processing module uses the built-in gait analysis algorithm to solve the data in real time and perform heart rate-gait-fatigue mapping analysis; If abnormal gait, force exceeding the safety threshold, or excessive fatigue is detected, the control terminal will immediately issue an audible and visual warning; Step S5: Report Generation; After training, a report is generated that includes plantar pressure heatmap data, pressure center trajectory data, joint angle change curve data, joint torque peak data, muscle activation estimation data, and heart rate-gait-fatigue mapping analysis conclusions, which can be exported for doctors' diagnostic reference.

[0014] The beneficial effects of this invention are as follows: (1) High-precision multidimensional force sensing: The integrated thin-film pressure sensor array based on elastomer and strain gauge can capture X, Y and Z forces and torques. Compared with traditional single-point sensors, it greatly improves the comprehensiveness and accuracy of data collection and deeply detects the directionality and distribution characteristics of force.

[0015] (2) Multimodal fusion monitoring: Combining strain sensing technology, optical imaging principle and PPG physiological monitoring, multi-dimensional data such as plantar pressure, joint angle, torque and heart rate are collected simultaneously to construct a "heart rate-gait-fatigue" mapping relationship, and the data is more comprehensive and accurate.

[0016] (3) Multi-environment in-situ testing: The slope of the treadmill (flat / uphill) is adjusted by modular texture panel design to meet the testing needs of human beings in various life and movement scenarios, and to reveal the in-situ testing mechanism of the patient's lower limb movement status in multiple environments.

[0017] (4) Medical-grade diagnostic assistance: Construct disease early warning and assessment algorithms to transform doctors' subjective judgments into objective scientific data, which are applicable to the assessment of gait disorders in patients with stroke, Parkinson's disease, etc.

[0018] (5) Structural optimization and energy saving: The use of buffer pads to reduce shock absorption improves the accuracy of the sensor and protects the internal components; a solar power supply system can be selected to reduce the energy consumption of the equipment and make it green and environmentally friendly.

[0019] (6) Intelligent early warning and safety: Real-time monitoring of movement status, timely sound and light warning when the data exceeds the safety threshold, and can be linked to handrails or emergency stop to ensure patient training safety and prevent secondary injury. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the external structure of the treadmill of the present invention.

[0021] Figure 2 This is a schematic diagram of the structure of the treadmill assembly and pressure sensor array of the present invention.

[0022] Figure 3 This is a schematic diagram of the cross-sectional structure of the support plate.

[0023] Figure 4 This is an example interface for the lower limb dynamic analysis report generated by this invention.

[0024] In the picture: 1-Touchscreen display, 2-Handrail, 3-Running belt, 4-Lifting mechanism, 5-Vision capture system, 6-Data acquisition chip, 7-Signal integration module, 8-X / Y / Z axis data acquisition unit, 9-Force distribution acquisition card, 10-Multi-dimensional force sensor module, 11-Rotating shaft, 12-Thin film pressure sensor array, 13-Foot pressure sensor array, 14-Support plate, 15-Protective shell, 16-Buffer pad, 17-Clamping groove, 18-Modular textured panel. Detailed Implementation

[0025] 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.

[0026] like Figure 1 The treadmill shown is used for estimating injury through the fusion of multimodal physiological indicators. It includes a frame assembly, a drive system, a running belt assembly, and a multimodal monitoring unit. The frame assembly includes a support plate and a support frame, with a control terminal fixedly connected to the support frame. The drive system drives the running belt assembly, and the surface of the running belt 3 is provided with a modular textured panel 18. The modular textured panel 18 can be detachably installed on the surface of the running belt 3 through an overlapping layered structure or a Velcro structure to construct various test scenarios such as flat, sloping, and rugged surfaces.

[0027] like Figure 2As shown, the multimodal monitoring unit includes a foot pressure sensor array 13 embedded below the treadmill assembly. To improve sensor accuracy and protect the equipment, a clamping groove is provided on the inner side of the support plate, and the treadmill assembly is fixedly connected to a buffer pad within the clamping groove. The buffer pad uses a high-damping rubber or spring shock-absorbing structure to isolate mechanical vibrations generated by the drive system and prevent interference with sensor signals. A protective shell is fixedly connected between the support plates, covering the bottom of the running belt 3 and the tail of the support plates to prevent foreign objects from entering.

[0028] The working principle and configuration of the multimodal monitoring unit system are as follows: 1. Data Acquisition and Signal Transmission Configuration The plantar pressure sensor array 13 is composed of a thin-film pressure sensor array 12 and a multi-dimensional force sensor module 10. The thin-film pressure sensor array 12 is laid beneath the surface of the running platform assembly to collect planar distribution information of plantar pressure and generate a plantar pressure heat map; the multi-dimensional force sensor module 10 is used to measure the overall force value and magnitude. It employs a specially designed elastomer structure internally, such as... Figure 3 As shown, the elastic body, made of alloy steel and with a cantilever beam structure, is located around the module. When subjected to externally applied forces, the elastic body undergoes microscopic deformation. This structure has self-adjusting capabilities, effectively reducing nonlinear errors during the detection process. High-precision strain gauges are mounted on the surface of the elastic body to acquire X, Y, and Z force components and torque components. The thin-film pressure sensor array 12 and the multi-dimensional force sensor module 10 are fused. The thin-film pressure sensor 12 is used to detect the plantar pressure distribution map, and the multi-dimensional force sensor module 10 is used to measure the value and magnitude of the force, thereby improving spatial resolution and the accuracy of force measurement.

[0029] Signal transmission link: (1) Front-end acquisition: The information collected by the strain gauges is filtered and classified according to the X, Y and Z directions.

[0030] (2) Directional splitting and digitization: The signals are initially processed by the X-axis / Y-axis / Z-axis data acquisition units 8 respectively. The data acquisition unit 8 integrates a data acquisition chip 6, which is responsible for converting the analog signal into a digital signal and performing preliminary filtering.

[0031] (3) Signal integration: The digitized data is transmitted to the signal integration module 7, while the original information is aggregated and transmitted to the force distribution acquisition card 9. The signal integration module 7 uses an FPGA or a high-performance MCU to realize the synchronization and time-scale alignment of multi-channel data.

[0032] (4) Terminal output: The signal integration module and the force distribution acquisition card 9 send the processed data to the terminal board of the multi-dimensional force sensor module 10 respectively, and complete the digitization and uploading of the data to the data processing module.

[0033] Physiological signal acquisition: The physiological signal monitoring module uses an optical heart rate sensor based on photoplethysmography (PPG). When LED light is reflected from the skin tissue to the photosensitive sensor, the light absorption by arterial blood flow differs from that of muscle, bone, and other tissues, and the changes in light attenuation are converted into heart rate and heart rate variability data.

[0034] Kinematic data input: Position information signals from the visual capture system 5. The visual capture system 5 includes a depth camera or an infrared motion capture camera, mounted on top of the support frame, and is used to collect data signals of the left and right knee joint angles, left and right hip joint angles, and stride length.

[0035] 2. Signal Processing and Output Configuration Data preprocessing: Based on the characteristics of pressure sensor data and heart rate spectrum, data cleaning, preprocessing and standardization are performed to remove power frequency interference and motion artifacts.

[0036] Feature engineering: Pearson correlation coefficient method and principal component analysis (PCA) are used to extract key gait features, such as gait frequency, ground contact time, and contact force distribution cloud map.

[0037] Model Construction and Mapping: A database was established based on runners' physical characteristics to construct a mapping relationship between "runner status - lower limb dynamics - lower limb fatigue." This mapping relationship was established using a multiple linear regression model or a support vector machine (SVM) classification model. The criterion for judging lower limb fatigue was based on a weighted comprehensive score of heart rate variability (HRV) indicators (such as SDNN, RMSSD) and gait symmetry index, expressed by the formula: Fatigue Index = α × (HRV normalized value) + β × (gait variability coefficient). A fatigue state was determined when the comprehensive score was below a preset threshold. The study explored the characteristics and patterns of heart rate variability and gait, achieving a correlation analysis between foot kinematics and biomechanical information.

[0038] Report Output: The main control unit is configured to output a lower limb dynamic analysis report signal. For example... Figure 4As shown, the report signal includes plantar pressure heatmap data, pressure center trajectory data, joint angle change curve data, joint torque peak data, muscle activation estimation data, and heart rate-gait-fatigue mapping analysis conclusions. Specifically, the pressure center trajectory data is calculated using a pressure value weighted centroid algorithm collected by a pressure sensor array; the joint angle change curve data is calculated using an inverse kinematics algorithm based on feature point coordinates collected by a visual capture system; the joint torque peak data is calculated using a Newton-Euler inverse dynamics algorithm combined with a human biomechanical model; and the muscle activation estimation data is estimated based on a biomechanical model of the relationship between joint torque and muscle lever arm.

[0039] Storage and Export: The main control unit is equipped with a storage interface, which supports exporting report signals to medical diagnostic auxiliary file formats (such as PDF or DICOM compatible formats).

[0040] 3. Control terminal and early warning hardware configuration The control terminal is fixedly connected to a touch screen 1 for displaying real-time monitoring data and disease warning information. The system presets safety thresholds (such as the upper limit of peak joint torque, the upper limit of gait asymmetry, and the heart rate overload threshold). When the monitored lower limb movement or physiological state exceeds the preset safety threshold, the system issues a warning through flashing (visual) touch screen 1 and a speaker (sound), prompting the user to stop exercising or adjust their posture.

[0041] 4. Accessibility Configuration Handrail adjustment: Handrails 2 are fixedly connected to both ends of the control terminal. The handrails 2 are connected to the support column through the lifting mechanism 4. The lifting mechanism 4 is a hydraulic support rod or an electric push rod, which is used to adjust the height of the handrail to accommodate users of different heights and rehabilitation patients, and to provide support and protection.

[0042] Energy Management: A solar panel is rotatably connected to the support frame via a pivot 11. The solar panel is electrically connected to the energy storage battery inside the electromechanical box via a charging controller, providing auxiliary power to the multi-modal monitoring unit and control terminal to achieve energy-saving operation. The pivot 11 supports angle adjustment to optimize the sunlight reception angle of the solar panel.

[0043] 5. System Workflow The system workflow in this embodiment is as follows: Step S1: Information Entry. Users input or import their basic personal information, medical history, and imaging data through the control terminal.

[0044] Step S2: Scene Selection and Calibration. The user selects the desired treadmill incline (flat / uphill) and replaces the corresponding modular texture panel. After the system starts, sensor zero-point calibration and vision system calibration are performed.

[0045] Step S3: Movement and Data Acquisition. The user begins running training, and the drive system moves the treadmill components. The plantar pressure sensor array 13 acquires pressure distribution and multi-dimensional torque data in real time, the visual capture system 5 simultaneously acquires video streams of lower limb joint movements, and the PPG sensor simultaneously acquires heart rate data.

[0046] Step S4: Real-time Analysis and Early Warning. The data processing module uses a built-in gait analysis algorithm to process the data in real time and perform heart rate-gait-fatigue mapping analysis. If gait abnormalities, stress exceeding safety thresholds, or excessive fatigue are detected, the control terminal immediately issues an audible and visual warning and automatically reduces the treadmill speed if necessary.

[0047] Step S5: Report Generation. After training, the system generates a report containing plantar pressure heatmap data, pressure center trajectory data, joint angle change curve data, joint torque peak data, muscle activation estimation data, and heart rate-gait-fatigue mapping analysis, which can be exported for doctors' diagnostic reference.

[0048] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and not restrictive.

Claims

1. A treadmill for estimating injury by fusing multimodal physiological indicators, comprising a frame assembly, a drive system, and a running platform assembly, characterized in that, Also includes: A multimodal monitoring unit, comprising a plantar pressure sensor array embedded below the treadmill assembly, a visual capture system mounted on the support frame, a physiological signal monitoring module, and a data processing module; A control terminal is communicatively connected to the data processing module and is used to display data and receive user commands. The data processing module has a built-in preset gait analysis algorithm, which is used to perform in-situ gait detection and pre-diagnosis and fatigue assessment based on the data collected by the plantar pressure sensor array, the visual capture system and the physiological signal monitoring module. Specifically, the following steps are included: S1. Data Preprocessing: Based on the characteristics of the pressure sensor data and heart rate spectrum, perform data cleaning, preprocessing and standardization. S2. Feature Engineering: Use Pearson correlation coefficient method and principal component analysis to extract key gait features, such as gait frequency, ground contact time, and contact force distribution cloud map; S3. Model Construction and Mapping: A database is established based on runners' physical characteristics to construct a mapping relationship between "runner status - lower limb dynamics - lower limb fatigue". The mapping relationship is established using a multiple linear regression model or a support vector machine classification model. The criterion for judging lower limb fatigue is based on a weighted comprehensive score of heart rate variability index and gait symmetry index. When the comprehensive score is lower than a preset threshold, it is judged as a fatigue state. S4. Report Output: The report includes plantar pressure heatmap data, pressure center trajectory data, joint angle change curve data, joint torque peak data, muscle activation estimation data, and heart rate-gait-fatigue mapping analysis conclusions.

2. The treadmill for estimating injury through multimodal physiological index fusion according to claim 1, characterized in that, The surface of the treadmill assembly is provided with a modular textured panel (18); the modular textured panel (18) is detachably mounted on the surface of the running belt (3).

3. The treadmill for estimating injury through multimodal physiological index fusion according to claim 1, characterized in that, The plantar pressure sensor array (13) consists of a thin-film pressure sensor array (12) and a multi-dimensional force sensor module (10).

4. The treadmill for estimating injury through multimodal physiological index fusion according to claim 1, characterized in that, The physiological signal monitoring module uses an optical heart rate sensor based on photoplethysmography (PPG); the visual capture system includes a depth camera or an infrared motion capture camera to collect data signals of left and right knee joint angles, left and right hip joint angles, and stride length.

5. The treadmill for estimating injury through multimodal physiological index fusion according to claim 1, characterized in that, The pressure center trajectory data is calculated using a pressure value weighted centroid algorithm collected by a pressure sensor array; the joint angle change curve data is calculated using an inverse kinematics algorithm based on the feature point coordinates collected by a visual capture system. Peak joint torque data were calculated using the Newton-Euler inverse dynamics algorithm combined with a human biomechanical model; muscle activation estimation data were estimated based on a biomechanical model of the relationship between joint torque and muscle lever arm.

6. The treadmill for estimating injury through multimodal physiological index fusion according to claim 1, characterized in that, The control terminal is fixedly connected to a touch screen (1); the system presets a safety threshold. When the monitored lower limb movement state or physiological state exceeds the preset safety threshold, the system will issue a warning by flashing the touch screen and using a speaker.

7. The treadmill for estimating injury through multimodal physiological index fusion according to claim 1, characterized in that, The control terminal is fixedly connected to handrails (2) at both ends, and the handrails (2) are connected to the support column through the lifting mechanism (4).

8. The treadmill for estimating injury through multimodal physiological index fusion according to claim 1, characterized in that, A solar panel is rotatably connected to the support frame via a rotating shaft (11), and the solar panel is electrically connected to the energy storage battery inside the electromechanical box via a charging controller.

9. The treadmill for estimating injury through multimodal physiological index fusion according to claim 1, characterized in that, The working process of a treadmill is as follows: Step S1: Information entry; Step S2: Scene selection; The user selects the desired treadmill incline and replaces the modular texture panel (18). Step S3: Exercise and data acquisition; The user starts running training, and the drive system drives the treadmill components to move; The plantar pressure sensor array (13) collects pressure distribution and multi-dimensional torque data in real time, the visual capture system (5) collects the lower limb joint movement video stream in real time, and the PPG sensor collects heart rate data in real time. Step S4: Real-time analysis and early warning; The data processing module uses the built-in gait analysis algorithm to solve the data in real time and perform heart rate-gait-fatigue mapping analysis; If abnormal gait, force exceeding the safety threshold, or excessive fatigue is detected, the control terminal will immediately issue an audible and visual warning; Step S5: Report generation; After training, a report is generated that includes plantar pressure heatmap data, pressure center trajectory data, joint angle change curve data, joint torque peak data, muscle activation estimation data, and heart rate-gait-fatigue mapping analysis conclusions, which can be exported for doctors' diagnostic reference.

Images