Design of human lower limb motion data acquisition system based on multi-sensor

In order to protect the health of older people, a wearable lower extremity motion data acquisition system was designed. With STM32F103C8T6 as the core, the system uses an inertial attitude sensor to collect lower limb motion attitude information. In view of the large difference in plantar pressure data of different fall types, the distributed flexible film plantar pressure sensor is used to collect plantar pressure information as an auxiliary discrimination means, which can accurately distinguish the fall behavior from other behaviors, and carry out data transmission through the wireless module. The device is small in size, light in weight, low in power consumption, and wearable, which is easy to apply in the field of health monitoring, and can also be used as a reference for the implementation of corresponding protection strategies for active lower limb prosthetics.


Introduction
With the development and progress of society, the number of older adults is becoming increasingly significant.It is predicted that by 2053, the population of older adults in China will exceed 480 million.At the same time, the number of patients with hemiplegia and disability caused by cardiovascular and cerebrovascular diseases such as stroke is increasing, which makes society pay more attention to the prevention and treatment of movement disorders in persons aged 65 or older.Studies have shown that the human central nervous system can reorganize its structure and function after the lower body is damaged, and through practice, the physical condition can be significantly improved, in addition to the initial treatment, the later rehabilitation is of great significance.
In the rehabilitation process of lower limb patients, the collection of lower limb movement data is particularly important, and the collection and analysis of lower limb movement data is a common means to evaluate the rehabilitation effect of patients.Therefore, the monitoring of lower limb movement is particularly important.
In terms of lower limb data collection, the human body data collection methods developed at home and abroad can be divided into contact collection methods and non-contact collection methods according to whether contact with people is required.Among them, the typical contact acquisition method is to fix the sensor in a specific position of the human body, and directly record the data of the human body in the process of movement; The non-contact acquisition method relies on the strong computing power of modern computers to solve the data of the human body by analyzing the continuous image features.There is also a data acquisition method between the two, the motion capture system based on marked points.This method does not need to fix the sensor on the human body but only needs to paste small marked points on the specific position of the human body.The motion state of the marked points is captured by the camera to restore the movement data of the human body.As far as the current research is concerned, all three methods have their own advantages and disadvantages.The fall monitoring system based on vision is not conducive to the protection of privacy, the equipment is expensive and easy to monitor dead corners, and cannot always monitor the health of older people.The fall detection system based on the environment is highly interfered with by environmental factors.For example, the fall detection system based on sound waves has a great difference in detection effect due to different building materials and is easily affected by noise [1] .Therefore, in view of the shortcomings of the above research, this paper designed a wearable human lower limb motion data acquisition system with small volume, lightweight, and low power consumption, which can effectively detect the Angle data of the hip joint, knee joint, and ankle joint.Compared with the existing human lower limb motion data, the effectiveness of this system was verified [2].Based on Angle detection, plantar pressure information is collected.Due to tripping and slipping, the plantar pressure data of the two fall types are significantly different, which can be used as auxiliary data to provide effective support for subsequent judgment of fall types.Joint Angle information and plantar pressure information can support each other, providing effective data support for rehabilitation planning of patients with lower limb movement disorders [3][4] , and providing reference for better implementation of corresponding protection strategies in the future.

Overall system design
The design block diagram of the multi-sensor human lower limb motion data acquisition system is shown in Figure 1.In this paper, STM32F103C8T6 is used as the underlying data processing unit to design a multisensor human lower limb motion data acquisition system.RX-ES42-18 distributed flexible film plantarpressure sensor, N100 inertial attitude sensor, and G01-S wireless module are used for data transmission.The MCU packets the collected data and sends it to the host computer through the wireless transmission module.Considering the convenience of the acquisition system, an 18650 lithium battery is used to power the whole system, and the booster module is used to boost the output voltage of the lithium battery to 5 V to power the whole system.The lithium battery has a light structure and can be charged multiple times, which is very suitable for the power supply of the motion data acquisition system in this paper.The inertial sensor and the development board are fixed by Velcro.Since the leg data acquisition system uses the inertial measurement unit as the sensor, the inertial measurement unit is very sensitive to any space movement.In order to minimize the data noise caused by the installation method, the inertial measurement module is fixed by Velcro in this paper.It can offset the jitter caused by some hardware parts such as wearing gaps, so that the output data waveform of the inertial measurement unit is smooth, and the difficulty of later data processing is reduced.The system transmits the collected data to the host computer for storage through wireless transmission [5][6] .The multi-sensor human lower limb motion data acquisition system is shown in Figure 2.

Main controller design
The main control module of the system uses STM32F103C8T6 as the bottom data processing unit of the lower limb motion data acquisition system, which is a microprocessor developed by microelectronics and is responsible for the data analysis and control of each module.STM32 MCU transmits data through the communication interface between itself and the sensor, communication module, and control output module, controls the corresponding sensor to collect data, and realizes the data sending and receiving through the wireless communication module.If the single-chip microcomputer STM32 fails, we will press the reset key to initialize.The minimum system of a singlechip microcomputer includes a main control unit, clock circuit, power management circuit, JTAG circuit, reset circuit, etc.The internal crystal oscillator or external clock circuit can be used as STM32F103C8T6 clock.SYSCLK clock source can be HSI, HSE, or PLL [7][8][9] .

Design of inertial measurement unit
An attitude sensor is a sensor that can measure the orientation and position of an object in threedimensional space.By measuring the Angle, acceleration, angular velocity, and other physical quantities of the object, it can precisely control and locate the trajectory of the object.The commonly used motion attitude sensors are motion accelerometers, rotating gyroscopes, magnetometers, and so on.In view of the need for wearable devices, after the evaluation of several IMUs, this paper finally decided to use the N100 inertial state sensor of China WHEELTEC company to collect the movement data of human legs.The N100 has a nine-axis position sensor with a data output frequency of 400 Hz, including a three-axis gyroscope, a three-axis accelerometer, and a three-axis magnetometer.The builtin Kalman fusion filter algorithm can obtain accurate information such as three-axis angular velocity, three-axis geomagnetic field, and three-axis Angle through accurate adjustment of the algorithm.This paper chooses to use the IMU as the leg motion data acquisition sensor [10][11] .

Plantar pressure collection design
According to the previous research, this paper selects an RX-ES42-18 flexible array pressure sensor to collect plantar pressure signals.This sensor is a plantar pressure sensor with European standard 42 code, its total length is 260 mm, and it adopts 6 rows * 4 columns array design, with 18 independent sensing units, and each sensing point is carefully designed.Detect pressure changes with minimal interference, low power consumption, and can be worn for a long time to collect plantar pressure data.Therefore, this paper adopts a flexible array pressure sensor to measure plantar pressure [12] .

Data transmission part design
Considering that the acquisition system should not cause interference and inconvenience to the movement of the wearer, the system does not use a wired connection to connect the computer but uses a wireless module for the transmission of motion data.Motion data transmission requires fast speed and high delay.Most of the wireless modules that meet such requirements are large and not easy to wear and use.Therefore, to find a wireless mode that is small enough in size and stable enough in performance, this paper selects the G01-S wireless communication module produced by China Chipset Micro Corporation.The communication frequency of the module is 2.4 GHz, the data throughput is large, and the delay is very low.In the case of a short distance, it almost reaches the level of wired transmission, which meets the data transmission requirements of the acquisition system in this paper.Moreover, the module is a patch module, and the size is very small, which can greatly reduce the difficulty of hardware design [13] .
The hardware circuit design of the multi-sensor human lower limb motion data acquisition system based on the above modules is shown in Figure 3.

System software design
The software program of the fall detection controller based on STM32 mainly includes communication with the debugging computer, fall detection control algorithm, information collection, data processing, and transmission.Since the CAN driver chip is not added to the N100 inertial navigation module, an external CAN driver chip is adopted.After the main program is initialized, CAN communication is configured in the upper computer, and CAN communication is started.The N100 inertial navigation module sends two kinds of data by default, namely MSGAHRS and MSG IMU, whose instruction categories are different.The data length is also different, and the data needs to be received by classification.After receiving the CAN message, the message is parsed, the monitored data is computed, and transmitted to the host computer through the communication interface.In order to facilitate subsequent program transplantation and generality of program writing, C language was chosen for programming development, and KEIL MDK software was used as the development environment.The debugging computer is used as the upper computer, and the debugging software is designed based on Python, which is mainly used to monitor and display the information collected by the sensor [14] .

Experimental paradigm
In this study, three healthy adult volunteers aged 22-23 years, height 170-175 cm, weight 65-80 kg, all of whom had not undergone knee and ankle surgery, volunteered to participate in this experiment.As mentioned above, the sensors were worn to test the Angle data of left and right hip joints, knee joints, and ankle joints of three subjects respectively.The Angle data of left and right hip joints, knee joints, and ankle joints of subject 1 are shown in Figure 4 and Figure 5.

Conclusion
This paper designs a lower limb motion data acquisition system, which is based on an STM32 single chip microcomputer as the main control chip, collects sole pressure through thin film pressure insole, inertial measurement unit collects human lower limb motion information, and transmits data through G01-S wireless module.The raw information collected mainly includes the angular velocity and acceleration information of the hip joint, knee joint, and ankle joint.Compared with the limited collection method in the laboratory, the subject's range of activity can be unrestricted.The system can be used in the field of health monitoring, and can also provide an effective reference for the implementation of protective strategies for active lower limb prosthetics.

Figure 1 .
Figure 1.Structure diagram of the detection system.

Figure 2 .
Figure 2. Multi-sensor human lower extremity motion data acquisition system.

Figure 3 .
Figure 3. Hardware circuit of human lower limb motion data acquisition system.