Gait-Ground Reaction Force Sensors Selection Based on ROC Curve Evaluation

Abstract

Classification of normal gait from pathological gait as then can be used as indicator of falling among subjects requires the correct choice of sensor location in the insole. Such a flexi force-sensor can be used underneath foot to measure vertical ground reaction force. To start with, the most relevant information (parameters) that can characterize the recorded signals are extracted from the vertical ground reaction force signals. Then Receiver Operating Characteristic curve is used to evaluate the features upon 8 sensors underneath each foot located at different locations. To confirm results obtained, features are passed upon a chosen classifier, in this paper K-nearest neighbors algorithm is chosen. Results show that the sensor located at the inner arch of the sole of the foot (i.e. at the mid foot) holds the most relevant information needed for better classification compared to other sensors.

Vertical ground reaction force marker for Parkinson’s disease

“…Alkhatib et al . achieved 83% classification accuracy by proposing ANN for classifying normal and pathological gait, ignoring SVM [ 29 ]. Zhang et al .…”

“…As PD patients tend to put less pressure during placing the heel strike and toe off than control subjects [ 29 ], maximum VGRF at heel strike and toe off for each gait cycle was computed. The mean and standard deviation of the VGRF overall gait cycles were taken as features for classification.…”

“…However, this model utilizes information from one sensor, discarding potentially useful information from other force sensors integrated into the shoe. Gait variability of PD patients based on VGRF has been investigated in other studies [ 25 – 29 ], in which a significantly increased variability in duration of gait tasks, such as swing time and stride time, has been observed in PD patients. This work accounts for such variability by including the swing time and stride time variability obtained from the VGRF data.…”

“…Classification of normal and abnormal gait using the k-Nearest Neighbor (kNN) classifier was conducted by Alkhatib et al . [ 29 ] on the same datasets used by Zhang [ 28 ]. In this research, VGRF data from one sensor, instead of an array of sensors, was used to calculate features, and a classification accuracy of 83% was achieved.…”

Análise no Desempenho de Algoritmos de Aprendizagem Supervisionada na Classificação da Marcha em Parkinsonianos

“…Visto que as técnicas de seleção de atributos já foram aplicadas em algoritmos de aprendizagem supervisionada na classificação de dados e mostraram um significativo desempenho na precisão (superior a 90%) no diagnóstico de portadores de DP. Support Vector Machine (SVM) [Pant and Krishnan 2014], k-NN [Alkhatib et al 2015] e Redes Neurais Artificiais (RNA) [Lee and Lim 2012], já foram utilizados em pesquisas com pacientes parkinsonianos aproveitando dados da marcha disponibilizados na base Physionet. Contudo, nenhum dos trabalhos publicados aplica mais de um algoritmo de AM manipulando atributos da base Physionet.…”