|Published online: June 5, 2015||$US5.00|
Since falls are the leading cause of injury and injurious death among older adults, the development of techniques for prediction and prevention of falls are of great value to the aging society. The purpose of this study is to examine a single descriptor of walking dynamics, derived from an inverted-pendulum model, that is capable of distinguishing older adults with a history of falls from those without. Thus, we used an inverted pendulum approach to quantify the error between the motion trajectory of a person’s center of mass (COM) calculated by the inverted pendulum model and that measured by a 3D motion capture system. As a pilot study, we recruited and tested two groups of participants. One group has 15 participants who fell in the last 12 months and thus is called the fallers group and the other has 16 participants of similar ages who did not fall in the same time period. The test results reveal that the errors between the model-calculated COM and the measured COM for the two groups are significantly different. Three machine learning (ML) methods were implemented to help classify the test data. The results of the study demonstrate that the proposed distinguisher can separate the individuals with a fall history from those without a fall history and, therefore, it may be potentially used to predict the risk of falls of an older adult.
|Keywords:||Falls-risk Prediction, Inverted Pendulum Model, Machine Learning|
Research Assistant, Department of Mechanical and Aerospace Engineering, New Mexico State University, Las Cruces, New Mexico, USA
Professor, Mechanical Engineering, New Mexico State University, Las Cruces, NM, USA
Assistant Professor, Department of Human Performance, New Mexico State University, Las Cruces, NM, USA
Professor and Academic Had, Department of Human Performance, New Mexico State University, Las Cruces, NM, USA