Activity classification in users of ankle foot orthoses

A framework for activity classification using inertial sensors mounted on ankle foot orthoses (AFOs) is presented. A decision tree-nearest neighbor algorithm classifies activities using subject-specific training. Eight volunteer subjects wore modified bilateral AFOs with shank and foot mounted triaxial accelerometers and gyroscopes. The AFOs were fitted with hardware to induce different gait perturbations: free rotation of the ankle, plantarflexion or "equinus" gait, and locked ankle joint. For each condition, the subject performed eight gait activities at varied slopes and standing, sitting, and lying postures. Using video for ground truth, the algorithm had an overall mean sensitivity of 95% using 50% of the data (≈ 140 s) for training and demonstrated upwards of 90% sensitivity with 25% of the data (≈ 70 s) for training. High sensitivities (≥ 87%) and PPV (≥ 90%) were achieved for all annotated gait patterns for all perturbations, excluding stair climbing (63%, 77%) and descending (80%, 78%). Postures were classified with less sensitivity and PPV than gait activities: lying (98%, 93%), standing (80%, 84%) and sitting (64%, 75%). Non-annotated walking (68%) and standing (73%) were classified with less sensitivity than were corresponding annotated events. Our results indicate that AFOs are a suitable sensor platform for future research in activity classification and gait monitoring in AFO users with perturbed gait using limited training data.