Assessing Sensorimotor Problems Via Bayesian Theory and Hidden Markov Models

Abstract

This work introduces a novel framework that combines Bayesian Statistics for motor control with a probabilistic graphical model to estimate sensorimotor problems. This problem is relevant because as we age, our motor skills tend to decay. A person with this type of problems finds difficult to perform even simple tasks such as walking, cooking, and driving. They become challenging activities due to the alterations to the motor control, which might lead to accidents or injuries. Therefore, the continuous assessment of the sensorimotor functions of a person could help in identifying potential problems at an early stage. This framework aims to provide a substantial estimation of the presence, or absence, of a sensorimotor problem over time. Our strategy is based on the integration of three main components, i.e., data collection during the execution of basic activities via mixed reality, estimation of coordination under uncertainty via Bayesian statistics, and the probability estimation of a sensorimotor problem at different instances of time via hidden Markov model (HMM).