Design Of Wrist Fall Detection System Based On LDA And Logistic Regression Algorithm

With the improvement of modern medical care and social material level,the life expectancy of the population is increasing year by year,and the proportion of the elderly is increasing.Followed by the elderly health problems become the focus of increasingly,but the elderly because of the older,the body each function decline,reflect the slow,often prone to accidents.Among them,the most common accidental injury is to fall.In this paper,it is found that the traditional algorithm of fall detection system has the characteristics of low real time and high false alarm rate.Therefore,it is of great significance to design a fall detection system with high accuracy and timely alarm of fall.Fall in this paper,after studying the existing detection technology,designed a set of hardware and software platform terminal wearable fall detection system,the system hardware platform wrist wear in the body,in does not affect the old man under the condition of normal life,through the sensor to detect the state of daily life movement of the body,when detect the elderly fall incidents,send alarm information and location information to the guardian,in an accident situation,make old man can get relief for the first time,limit the damage.The research contents of this paper mainly include the following points:First designed the fall detection system hardware terminal platform,through the analysis of human behavior and falling action,chooses six axis sensors on the sensor acquisition wrist motion acceleration and angular velocity data,the collected data is transmitted to the main control chip data preprocessing,communication positioning using GU620 module send text messages and provide GPS positioning,use lithium batteries on power supply and provide the demand of the power management circuit realizes the step-up and step-down,to provide users with more lasting life.Secondly,in the algorithm,due to the acquisition of feature information by multi-sensor,there will be some redundant information in the extracted feature vector,which will affect the computing time of the system.This article used on eliminating redundancy feature vector linear discriminant analysis,the characteristics of the input dimension,then put forward to identify the fall detection algorithm based on Logisticregression,and optimize the algorithm parameters.Finally,on the software platform,I wrote the smart hand ring service platform App to locate the hardware platform,and built the server to carry out the data transmission between the hardware data and the App.Finally,the whole system is built and tested to verify the effectiveness of the system.In this paper,the wearable fall detection system for technical research,design a wearable fall down test hardware and software platform of testing system,and to verify this algorithm in the system,the experimental results show that this system on the fall detection system has high accuracy and real-time performance,and complete the SMS alarm and communications positioning function.