Fall Detection System Based On Inertial Motion Sensor

Fall reduces the human activity and can lead to serious physical,psychological and social dysfunction.It is one of the main reasons that results in disability,death,and heavy economic burden for the elderly and patients.In light of high-power,limited range of alarm and inconvenience for carrying in existing fall detection devices,it is of great significance to design a wearable detection device with low power consumption,small size and convenience for carrying,which can improve the medical information intelligence and promptness,and promote the development of a harmonious society.In this thesis,we developed a fall detection system based on mpu6050,using Arduino and Android.We presented the overall framework and performance for the system,with focus on the hardware and software platform of the lower computer used for collecting,transmitting and processing data,and the developing of the upper computer used to locate alarms.This system achieved accurate control of human body's movement data and real-time positioning.The main contribution of this thesis includes four parts as follows:1.In light of the current status of research,the advantages and disadvantages of fall detection device,the demand indices and the overall design of device system was given,and the main innovation of our system was also pointed out.2.We designed the hardware of the device system,which included data sampling module,data transmission module,alarm module and some micro-control module.According to the design requirement of the lower computer,the component selection,circuit schematic and PCB design were completed.At the same time,the circuit board was debugged to achieve the stability and reliability of the system.And we designed the alarm App for the fall detection system,and introduced in detail the design flow of the Baidu map positioning module,the bluetooth serial communication module,and the distress SMS sending module.3.The pattern of the human daily movement was analyzed.We described the selection of data source and sampling frequency,the sampling principle of mpu6050,the preprocessing of data and the extraction of feature quantity in detail.4.The principle of joint Kalman filter and ProtoThread parallel scheduling was discussed in detail.And a parallel threshold algorithm based on data fusion was proposed.The software program of the fall detection system was developed by Arduino IDE,and added some optimizations such as offset compensation to improve the accuracy of the data acquisition system.5.The wearing and the specific operation of the fall detection system in this thesis were described.Through the simulation test of volunteers,the experimental results were compared with the existing algorithms,to verify the stability and reliability of the system.In this thesis,we developed a fall detection system with low cost,low power consumption,high precision and convenience for carrying.It not only reduces the cost of the fall detection equipment and improves the detection precision,but also had important significance for promoting the scientific development of medical equipment,with broad market prospects and good social benefits.