Body Posture Detecting And Walking Trends Forecasting Based On FSR And Gyroscopes

In the three main methods of gait or posture detection, the detection method based on theattitude of the inertial sensor is gradually becoming a hot research topic due to its portabilityand less complexity in algorithm. However, the current applications utilizing inertial sensorsare more to detect falling dwon gesture, and the studies of posture detecting motion based oninertial sensores were relatively less. Meanwhile, early in the19th century, study on plantarpressure was done. With the popularity of technology development and medical knowledge,plantar pressure is used in more and more occasions. For instance, distribution of plantarpressure can be used to alert diabetic risk by detecting foot ulcers. And it can also determinethe degree of recovery in the rehabilitation movement. Therefore, the plantar pressuredistribution detection and attitude angle detected by gyroscope combined to detect gesture isof great significance. Development of a wearable device combining plantar pressuredistribution and attitude angle of gyro to detect posture is imminent. This problem is solved bya combination of FSR (force sensitive resistor) and gyroscopes to achieve wearable device andthus determining the different postures.FSR and gyroscopes are used to design and implement a set of wearable device todetermine simple gestures by detecting distribution of plantar pressure and acceleration ofgyroscope. Through doing different actions, plantar pressure distribution and acceleration aredetected of experimenters, which determine some common gestures. According to relevance ofdata postures are predicted. In the implementation process, in order to achieve wearableequipment, IIC communication protocol and Bluetooth wireless communication technologyare used for communication. The system utilizing Arms Cortex-M3core processor, and eachmodules are integrated. FSR pressure sensing components and MPU6050six-axis motionprocessing module are core modules. Due to the useage of wireless communication module,the system uses a battery-powered, and circuit design is of compact and portable. Bycollecting data in12different poses in gesture library of12experimenters, the characteristic ofdata are analyized. Different plantar pressure value on different areas of foot, comparativecorrelation between them and the acceleration are mapped into different postures. In dataprocessing, the Kalman filter is used to obtain better results and reduce the noise in the signal acquisition of posture. DMP technology is used to obtain a more stable attitude angle.This paper presents a comprehensive approach to determine postures utilizing attituderelevance, significant differences of plantar pressure between different regions, as well as theacceleration of the pace of sporty. In this method, the significant differences between the valueof each region are calculated, and the relevance of each action of pressure and attitude areanalyzed. ARMA forecasting model as well as the possibility of subsequent gesture asconstraints are used to predict the attitude. Experimental results show that in a small range ofsamples, the method is simple and feasible. Simple actions can be judged with the actionpredicted by this method.