International Journal of Physical Medicine & Rehabilitation
Open Access

Department of Health and Human Performance, University of Houston, Texas, USA

Publications

• Research Article   
  Digital Twin-Based Controls in Gait Analysis: A Machine Learning Approach
  Author(s): Paul M Trusov, Dacia Martinez Diaz and Charles S Layne*
  
  Human gait refers to the way people walk, which can vary widely between individuals due to factors such as body structure, age and health conditions. Traditional gait analysis often compares the walking patterns of individuals with and without medical conditions, which may incorrectly attribute natural gait variations to these conditions, thus introducing biases in diagnosing and understanding gait abnormalities. This study proposes a novel approach using machine learning to create synthetic control subjects that emulate a “healthy twin” for affected individuals. This method allows for a more refined comparison by accounting for individual-specific gait characteristics, thereby isolating abnormalities more accurately attributable to the medical condition. The method utilizes a Long Short-Term Memory (LSTM) model to analyze gait waveform data. A gait waveform is a graphical.. View More»
  DOI: 10.35248/2329-9096.24. S27.003

  Abstract HTML PDF