Human body posture recognition system based on wearable sensor

Abstract

The invention discloses a human body posture recognition system based on a wearable sensor, and belongs to the technical field of sensor measurement. Plantar pressure is collected through a pressure sensor module, filtering, denoising and smoothing are conducted through a signal processing module, and a sensor signal is wirelessly sent to an upper computer through a wireless transmission module; the upper computer adopts a hidden Markov algorithm and a random forest fusion algorithm to recognize postures; the system comprises a pair of flexible pressure sensor arrays, and the sensor arrays are worn on the soles of the feet and collect pressure signals of the soles of the feet; the signal processing module is completed by an upper computer, the wireless transmission module adopts a Bluetooth module, and the upper computer carries out real-time identification after training the model; the system improves the recognition rate, is high in adaptability, and is suitable for the fields of medical treatment, human-computer interaction, exercise training and the like.

Description

Technical Field

[0001] The present invention relates to the field of sensor measurement technology, and in particular to a human posture recognition system based on a flexible plantar pressure sensor array. Background Art

[0002] In today's era of rapid technological advancement, human posture recognition technology plays a vital role in numerous fields. In sports science, analyzing athletes' posture can improve training efficiency and performance. In augmented reality (AR) and virtual reality (VR), precise human posture tracking provides a more immersive experience. In health and medical systems, posture recognition helps monitor patient recovery and prevent falls. In human-computer interaction, posture recognition can simplify operating interfaces and enhance user experience. Gaming and entertainment systems use posture recognition to provide players with a more natural way to interact.

[0003] While computer vision-based human gesture recognition technology offers numerous advantages, such as contactless operation and a wide range of application scenarios, it still faces challenges in practical applications. These challenges include sensitivity to environmental conditions such as illumination and occlusion, as well as hardware reliance on computationally intensive algorithms such as deep learning. Furthermore, while traditional gesture recognition methods, such as optical sensors and inertial measurement units (IMUs), offer an alternative solution, they also have limitations, such as dependence on lighting and angles, and the inconvenience of being wearable.

[0004] Therefore, it is particularly important to develop a new human posture recognition technology that is more adaptable, can work stably in different environments, and is cost-effective and easy to promote and apply.

[0005] The emergence of flexible pressure sensors has brought new breakthroughs in this field. These sensors offer exceptional flexibility and wearability, adapting to the body's curves and providing continuous and precise pressure data. This data can be used to accurately infer human posture and movement. Compared to traditional visual and inertial sensors, flexible pressure sensors are not restricted by lighting conditions or viewing angles and offer significant advantages in wearability and comfort. Furthermore, they typically consume less energy, making them suitable for extended use.

[0006] In summary, flexible pressure sensors have enormous potential for application in the field of human gesture recognition. By further studying and optimizing the performance and integration methods of these sensors, we can develop more efficient, reliable, and user-friendly human gesture recognition systems, thereby promoting technological advancement and expanding applications in related fields. Summary of the Invention

[0007] In response to the shortcomings of the existing technology, the present invention provides a human posture recognition system based on a flexible plantar pressure sensor array. The system uses the flexible plantar pressure sensor array to capture plantar pressure information of different human postures, and processes the signal through an improved hidden Markov algorithm and random forest fusion algorithm to realize human posture recognition, solving the problems of accuracy and comfort existing in the background technology.

[0008] In order to achieve the above-mentioned purpose, the present invention specifically adopts the following technical solutions: A human posture recognition system based on wearable sensors is characterized by comprising a sensor module, a wireless transmission module and a host computer, wherein the host computer comprises a signal processing module and a recognition module; The sensor module collects plantar pressure signals of both feet; The wireless transmission module wirelessly transmits the plantar pressure signals collected by the two sensors to subsequent modules for further processing; The host computer receives data from the wireless transmission module, processes the data and extracts features, and finally uses an improved hidden Markov algorithm and a random forest fusion algorithm to recognize human body posture.

[0009] Furthermore, the sensor module includes a flexible pressure sensor array.

[0010] 2. A flexible plantar pressure sensor array is attached to the soles of both feet to detect pressure changes. The plantar pressure sensor array module uses a 6x4 flexible pressure sensor array patch, which contains 18 sensor array patch units. The flexible array pressure sensor uses a piezoresistive sensing mechanism. The resistance value of the sensitive unit changes with the contact force. The object to be measured is the resistance change of each sensitive point array. Furthermore, the wireless transmission module is composed of a Bluetooth module; The wireless transmission module uses a Bluetooth transmission module and connects it to the sensor acquisition module. Each flexible sensor is equipped with a Bluetooth sending module to wirelessly transmit the collected plantar pressure data to the host computer. The host computer uses a CH340USB to serial port module as a Bluetooth receiving module with a baud rate of 115200.

[0011] Furthermore, the host computer uses an improved hidden Markov algorithm and random forest fusion algorithm to recognize posture; The host computer filters and preprocesses the received data, performs PCA dimensionality reduction, and then calculates the plantar pressure center trajectory of both feet as one of the features. At the same time, the data is input into the hidden Markov model to calculate the maximum likelihood of the hidden state as the second feature. Finally, these two features are spliced ​​and input into the random forest for supervised learning. After obtaining the trained model, the model is loaded for real-time recognition.

[0012] Compared with the prior art, the present invention provides a human posture recognition system based on a flexible plantar pressure sensor array, which has the following beneficial effects: The present invention calculates the HMM likelihood of the plantar pressure center and different postures as features, wherein the spatial dynamic representation is: the COP trajectory directly reflects the balance control strategy of the human body on the support surface, including the geometric characteristics of the pressure distribution (such as trajectory length, elliptical area, front-to-back / left-to-right displacement) and temporal dynamic characteristics (such as speed and acceleration); the HMM modeling action phase changes (such as the support phase and swing phase in the gait cycle) through the implicit state sequence. The likelihood number it outputs represents the probability that the current observation sequence (pressure signal) belongs to a certain type of posture, the temporal dependence of the implicit action and the state transition law. The HMM reduces the influence of individual differences (such as stride size) through probabilistic modeling, and the random forest selects universal features through feature importance. The two work together to enhance the model's adaptability to unseen users.

[0013] The present invention achieves robustness and efficiency in human posture recognition by fusing COP trajectory with HMM likelihood features and using random forest for classification, and requires fewer samples than deep learning. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 This is a schematic structural diagram of a flexible pressure sensor array according to an embodiment of the present invention; Figure 2 This is a structural diagram of a sensor module according to an embodiment of the present invention; Figure 3 The relationship between pressure and resistance of the flexible sensor according to an embodiment of the present invention; Figure 4 A flowchart of the system according to an embodiment of the present invention; Figure 5 This is a waveform diagram of data collected from 18 channels on the sole of the foot according to an embodiment of the present invention; Figure 6 A COP trajectory distribution diagram calculated according to an embodiment of the present invention. DETAILED DESCRIPTION

[0015] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0016] Example like Figure 1-6As shown, an embodiment of the present invention proposes a human posture recognition system based on wearable sensors, which mainly consists of the following parts: a flexible pressure sensor array, a signal processing module, a wireless transmission module and a host computer.

[0017] 1. Sensor module (1) Flexible pressure sensor array: The flexible pressure sensor array is a highly sensitive sensing system composed of multiple small sensor units designed to be easily worn on the sole of a shoe. Each sensor has extremely high sensitivity, accurately capturing a series of subtle pressure changes generated by the human body in motion or at rest, and converting them into electrical signals for subsequent signal analysis and processing.

[0018] The sensor array uses an advanced manufacturing process, precision printing technology, through which nanoscale force-sensitive materials and conductive silver paste are precisely printed onto a soft polymer film substrate. After printing, these materials undergo a thermal curing process to ensure they are firmly bonded to the substrate and maintain long-term stability and reliability.

[0019] In terms of physical properties, these flexible pressure sensors exhibit a significant piezoresistive effect. When the sensor is subjected to external pressure, its resistance decreases. This change follows a power function relationship, indicating a nonlinear but predictable relationship between resistance and applied pressure. More importantly, the relationship between the inverse of resistance (i.e., conductivity) and pressure is nearly linear, which is particularly critical for the processing and interpretation of sensor signals because it simplifies the mathematical models and algorithms for inferring the raw pressure value from the electrical signal.

[0020] Overall, this flexible pressure sensor array not only provides real-time monitoring capabilities of human posture changes, but also demonstrates high flexibility and precision in design and manufacturing, making it an ideal choice for developing a new generation of human-computer interaction interfaces and health monitoring equipment.

[0021] 2. Wireless transmission module: The wireless transmission module is responsible for transmitting the processed signals from the signal processing module to the host computer's receiver. The system uses the BLE02 (ATK-MW579) Bluetooth module as the wireless transmission module. Its operating voltage range is between 3.3V and 5V, it supports Bluetooth version 4.2, and operates at a 2.4GHz ISM frequency. The module is a master-slave system and supports multiple operating modes, including a low-power mode, data transmission, and one-to-many data broadcasting. The Bluetooth module is connected to the flexible sensors on both feet. The host computer uses a CH340 USB-to-serial port module as the Bluetooth receiver, with a baud rate of 115200.

[0022] 4. Host computer: The host computer includes data processing and hidden Markov and random forest fusion algorithms.

[0023] The Hidden Markov Model (HMM) is an algorithm that uses observable sequences to infer implicit sequences. This algorithm is capable of predicting nonlinear time series. Using the transition probability matrix and initial probability matrix of each observable sequence, the probability of the state at the next moment can be inferred. The Dynamic Time Warping (DTW) algorithm module is used to perform time series matching and recognition on the feature information output by the signal processing module. The Hidden Markov Model (HMM) is a powerful time series signal processing technology. The pressure signal from the plantar surface is also a time series signal. By treating pressure as an observable sequence and posture as an implicit sequence, the algorithm can predict different human postures.

[0024] The purpose of integrating the random forest algorithm is to reduce the uncertainty of the output probability of the HMM algorithm. The output of the HMM is input into the random forest model as a feature, and finally the human body posture can be obtained through decision making.

[0025] Based on this, the Hidden Markov Model (HMM) and the Random Forest (Random Forest) fusion algorithm are integrated. First, the raw data is smoothed and cleaned, and the distribution of the plantar pressure trajectory is calculated as a feature. The processed data is then subjected to PCA dimensionality reduction before being fed into a Hidden Markov Model (HMM). Finally, the resulting optimal output sequence is used as a feature input into a Random Forest (Random Forest) model for final classification. This compensates for the instability of the HMM in posture prediction, improving both recognition rate and speed.

[0026] After the host computer has trained the model, the trained model can be used directly by inputting the data. The test set and training set can be divided in advance to output the human body posture recognition.

[0027] A human posture recognition system based on wearable sensors uses a flexible pressure sensor array to measure the pressure change data of the tester's foot at a certain moment, and recognizes and outputs the tester's movements through a hidden Markov and random forest fusion algorithm. The system workflow is as follows: Figure 4 shown.

[0028] Since each unit of the flexible pressure sensor array outputs the pressure change of each unit in the form of an electrical signal, the electrical signal will change to varying degrees during an acquisition cycle. Since the acquisition is performed in an array manner, the sensor array of one foot will collect different data changes of 18 units. The pressure of each unit on the sole of the foot caused by different actions appears in order, so the present invention proposes a method for calculating the distribution of the center of plantar pressure to assist in identification.

[0029] The center of pressure is the point where the resultant vertical reaction forces (pressures) at all points of contact between an object and a supporting surface (such as the ground or a platform) act. It characterizes the spatial concentration of pressure distribution and is a core parameter used in analyzing balance, stability, and movement patterns in biomechanics, robotics, and engineering.

[0030] The trajectory refers to the path that the CoP moves on the support surface over time, reflecting the pressure changes during dynamic processes (such as the migration of plantar pressure during walking). The CoP trajectory of the present invention is achieved by a flexible pressure sensor array on the sole of the foot, and the steps are as follows: 1. Sensor calibration: (1) Determine the position coordinates of each sensor ( , ).

[0031] (2) Calibrate the force-electric signal relationship to ensure , precise measurement.

[0032] 2. Data acquisition: High-frequency sampling (usually 100-1000Hz) to obtain the transient vertical force of each sensor (t).

[0033] 3. Real-time calculation: (1) Calculation of resultant force: = , (2) Calculate the weighted position sum: (t)= (t)= 4. Instantaneous output CoP: (t)= , (t)= Application areas: The human posture recognition system of this invention has extremely broad application prospects. In the medical field, the system can monitor changes in a patient's posture, such as gait monitoring and rehabilitation training posture, providing doctors with valuable information to help them more comprehensively understand the patient's condition and recovery status. In the rehabilitation field, it can be used to evaluate the effectiveness of patients' rehabilitation training and provide personalized training recommendations to improve the effectiveness and quality of rehabilitation training.

[0034] In the field of human-computer interaction, this system can be applied to scenarios such as smart homes and smart offices, enabling smarter and more convenient human-computer interaction. For example, by recognizing the user's posture, a smart home system can automatically adjust environmental parameters such as lighting and temperature to provide a more comfortable living experience. Smart office equipment can automatically adjust the screen angle and height based on the user's posture, improving work efficiency and comfort.

[0035] In addition, this human posture recognition system also has broad application potential in sports training, virtual reality / augmented reality, security monitoring and other fields, bringing more convenience and safety to people's lives and work.

[0036] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art will be able to modify the technical solutions described in the aforementioned embodiments or substitute equivalents for some of the technical features. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A human posture recognition system based on wearable sensors, characterized by: It includes a sensor module, a wireless transmission module and a host computer, wherein the host computer includes a signal processing module and an identification module; The sensor module collects plantar pressure signals of both feet; The wireless transmission module wirelessly transmits the plantar pressure signals collected by the two sensors to subsequent modules for further processing; The host computer receives data from the wireless transmission module, processes the data and extracts features, and finally uses a hidden Markov algorithm and a random forest fusion algorithm to recognize human posture.

2. The wearable sensor-based human gesture recognition system according to claim 1, characterized in that: The sensor module includes a flexible pressure sensor array.

3. The wearable sensor-based human gesture recognition system according to claim 1, characterized in that: The flexible plantar pressure sensor array is fixed to the two soles of the human body to detect pressure changes on the soles of the feet.

4. The wearable sensor-based human gesture recognition system according to claim 3, characterized in that: The plantar pressure sensor array module uses a 6*4 flexible pressure sensor array patch, which contains 18 sensor array patch units. The flexible array pressure sensor adopts a piezoresistive sensitive mechanism. The resistance value of the sensitive unit changes with the change of contact force. The object to be measured is the resistance change of each sensitive point array.

5. The wearable sensor-based human gesture recognition system according to claim 1, characterized in that: The wireless transmission module is composed of a Bluetooth module.

6. The wearable sensor-based human gesture recognition system according to claim 5, characterized in that: The Bluetooth transmission module is connected to the sensor acquisition module. Each foot is equipped with a Bluetooth sending module to wirelessly transmit the collected plantar pressure data to the host computer. The host computer uses a CH340USB to serial port module as a Bluetooth receiving module with a baud rate of 115200.

7. The wearable sensor-based human gesture recognition system according to claim 1, characterized in that: The upper computer uses an improved hidden Markov algorithm and random forest fusion algorithm to recognize posture.

8. The wearable sensor-based human gesture recognition system according to claim 7, characterized in that: The host computer filters and preprocesses the received data, performs PCA dimensionality reduction, and then calculates the plantar pressure center trajectory of both feet as one of the features. At the same time, the data is input into the hidden Markov model to calculate the maximum likelihood of the hidden state as the second feature. Finally, these two features are spliced ​​and input into the random forest for supervised learning. After obtaining the trained model, the model is loaded for real-time recognition.

Images