Genetic Algorithm for Effective Fall Detection with Wrist Wearable Device
Abhilash Unnikrishnan1, Abraham Sudharson Ponraj2
1Abhilash Unnikrishnan, Department of Electronics Engineering, VIT, Chennai (Tamil Nadu), India.
2Abraham Sudharson Ponraj, Department of Electronics Engineering, VIT, Chennai (Tamil Nadu), India.
Manuscript received on 14 December 2019 | Revised Manuscript received on 22 December 2019 | Manuscript Published on 31 December 2019 | PP: 159-164 | Volume-9 Issue-1S3 December 2019 | Retrieval Number: A10321291S319/19©BEIESP | DOI: 10.35940/ijeat.A1032.1291S319
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Falls have always been a major cause of injury related deaths among the old aged population in our country. It causes mental trauma and severe fractures to the bones and spine which impacts their quality of life. Therefore a proper fall prediction and alert system along with a timely rapid response could enable us to tackle such serious fall events and decrease the fatality. Various sensors and embedded controllers are used in conjunction with various machine learning classifiers to help us predict and optimize the falls effectively. This work presents a wrist wearable device using MPU-6050 sensor and raspberry-pi controller with help of machine learn algorithm which help us to predict the falls. Five different supervised learning algorithms and one unsupervised algorithm was implemented and evaluated on the basis of their accuracy, sensitivity and specificity. Out of all these classifiers, the decision tree with an accuracy of 85% was implemented in the system which classified the fall from the real time non-fall data sets. Further the performance of system was increased using genetic algorithm which gave better classification results unlike the normal decision tree classifier. Once the falls are predicted we can give a real-time response which can be an added feature to this system.
Keywords: Decision Tree, Fall Detection, Genetic Algorithm, Machine Learning.
Scope of the Article: Parallel and Distributed Algorithms