Performance analysis of different fall detecting algorithms with different combinations of sensors.

Abstract

Fall is one of the major reasons for the death of elderly people. Fall detection systems with different sensors based on different algorithms are now quite well admired. In this paper we analyzed the performance of different algorithms that can be used to detect fall. We used four different types of machine learning algorithms for this project. At first, we have created our own data with accelerometer and gyroscope separately and simultaneously. Then we used this data on each algorithm and found the accuracy rate. After that we added Magnetometer and compared the new result with the previous results and the threshold difference among these algorithms. Our final result is which algorithm has the highest rate to detect fall comparing all the sensors individually and all together and we found SVM algorithm with using accelerometer and gyroscope together gives the highest accuracy of about 97%.