"Treadmill Data"

"Motion symmetry forms an important aspect in rehabilitation, fitness monitoring, and the detection of neurological disorders. In the simplest case, objective assessment of movement can be performed through the analysis of three-axial accelerometric data. This paper presents a methodology for data acquisition using a single accelerometer and subsequent computational analysis based on frequency components detection. The proposed methodology is applied to the estimation of gait symmetry using data collected on the treadmill, replacing analysis of the normal overground gait. The study focuses on the analysis of motion data relevant to the detection of stability disorders associated with neurological problems, providing additional information for biomedical specialists. Symmetry estimation is performed by analyzing the relative energy in selected frequency regions corresponding to odd and even steps, and the left and right step times. Artificial intelligence techniques and neural networks  provided the best performance in distinguishing left and right steps features, achieving the classification  accuracy of walking symmetry 78.8\\% for the walking speed of 6km\/h. The results demonstrate the potential of computational intelligence for the detection of specific neurological disorders in clinical practice. In addition, the proposed general tools may also find applications in engineering and robotics."