A passive estimator of functional degradation in power mobility device users

This paper documents the development of a passive technique for assessing a power mobility device user´s driving proficiency during everyday driving activities outside formal assessment conditions by therapists. This is approached by first building a model by means of an Artificial Neural Network to infer longer-term destinations for discretized bouts of travel, and subsequently drawing cues indicative of decline in driving proficiency for the duration of point-to-point navigation rather than relying on instantaneously calculated metrics. This resultant quantity, which we refer to as `functional degradation´, can then provide therapists with additional information concerning user health or serve as a leveraging parameter in combinatory shared-control mobility frameworks. Experiments conducted by able-bodied users subject to simulated noise scaled to varying degrees of functional degradation reveal a quantitative correlation between these longer-term proficiency metrics and the magnitude of degradation experienced; a promising outcome that sets the scene for a larger-scale clinical trial.