Avaliação da marcha humana utilizando palmilhas sensorizadas e algoritmo de aprendizado de máquina

Abstract

Human gait analysis can provide an excellent source for identifying and predicting pathologies and injuries. In this respect, instrumented insoles also have a great potential for extracting gait information. However, there are technical and commercial difficulties that can limit the diffusion of this technology. The insoles available on the market have a high cost and closed software. On the other hand, the low-cost academic prototypes do not present enough information about the design parameters, manufacturing techniques, and guidelines for developing. In addition, data processing is highly complex and requires a high degree of user knowledge. The present study proposes a proof-of-concept of a system based on vertical ground reaction force (vGRF) acquisition with a sensorized insole that uses a machine learning algorithm to identify different patterns of vGRF and extract biomechanical characteristics that can help during clinical evaluation. For this, a low-cost instrumented insole was developed, with customized sensors that was validated using a double-belt instrumented treadmill (Bertec, 1000 Hz, USA) as the “gold standard”. A new calibration methodology was developed, which increased by 12% the correlation with the force plate in relation to the usual calibration method. The study had the participation of 32 volunteers (18 men and 14 women). Each volunteer walked on the instrumented treadmill while wearing an experimental resistive sensorized insole. The acquired data are processed using algorithms based on machine learning responsible to to identify different patterns of vGRF and extract biomechanical characteristics that can help during clinical evaluation. The data was clustered by an Immunological Algorithm (IA) based on vGRF during gait. These clusters underwent a data mining process using the Classification and Regression Tree algorithm (CART), where the main characteristics of each group were extracted, and some rules for gait classification were created. As a result, the system proposed was able to collect and process the biomechanical behavior of gait. After the application of IA and CART algorithms, six groups were found. The characteristics of each of these groups were extracted and verified the capability of the system to collect and process the biomechanical behavior of gait, offering verification points that can help focus during a clinical evaluation.