Research and Design of a Fall Detection System Based on Multi-axis Sensor

This paper proposed a falling detection system based on MPU6500 senor, this system uses the method of combining hardware and mobile phone APP, the hardware bases on STM8 chip, using MPU6500 gathering human three-axis attitude angle and three-axis acceleration motion, after Kalman filtering for Euler angle and four-element method are used to calculate the human body posture angle. If the discrimination exceeds the set threshold value, turn into the discriminant of based on the characteristics acceleration vector of the SVM. The discrimination can be filtered non-falling state, achieve the purpose of accurate identification and alarm. Software supports data retrieval and improvement. Experimental results verify that the algorithm accuracy reaching 97% and have high accuracy and stability.