A novel sensorized shoe system to classify gait severity in children with cerebral palsy

The clinical management of children with Cerebral Palsy (CP) relies upon periodic assessments of changes in the severity of gait deviations in response to clinical interventions. Current clinical practice is limited to sporadic assessments in a clinical environment and hence it is limited in its ability to estimate the impact of CP-related gait deviations in real-life conditions. Frequent home-based quantitative assessments of the severity of gait deviations would be extremely useful in scheduling clinical visits and gathering feedback about the effectiveness of intervention strategies. The use of a wearable system would allow clinicians to gather information about the severity of gait deviations in the home setting. In this paper, we present ActiveGait, a novel sensorized shoe-based system for monitoring gait deviations. The ActiveGait system was used to gather data, under supervised and unsupervised conditions, from a group of 11 children with various levels of CP-related gait deviation severities. We present a methodology to derive severity measures based on features extracted from Center of Pressure (CoP) trajectories. Results show that a Random Forest classifier is able to estimate severity scores based on the Edinburgh Visual Scale with a level of accuracy >;80% adequate for clinical use.