Detecting Sensorimotor Disorders in Older Adults via Machine Learning and AR/MR technologies.
Abstract: Sensorimotor disorders in older adults affect the interpretation of sensory information and cause difficulties in motor planning and sequencing of movements. In this work, we combine Augmented Reality (AR) and Mix Reality (MR) technologies with a machine-learning technique to detect signs of sensorimotor disorders in older adults. The overall objective is to be able to detect at an early-stage sensorimotor problems and dynamically modify the parameters of the system to enhance user’s motor ability through repetitions.
Recently, the use and development of AR / MR applications have undertaken a leap towards enhancing human interaction in a virtual world because of its effectivity to reduce cognitive load and for multisensory feedback benefits. Furthermore, machine learning is a computational technique that uses analytical models to classify large datasets accurately and intelligently. Our strategy consists of a set of exercises in which the user has to follow a series of physical activities to complete a task. The interface of the program allows the users to grab a virtual object and manipulate it through specific path with certain velocity at different intervals of time. The information retrieved from the program consists of user’s error (in a three-dimensional space) when manipulating the object through the path with different velocities, and number of completed repetitions. Thence, these variables get stored for further analysis via the machine learning model. The number of usages for this type of system can extend to a vast array of areas regarding physical training like; rehabilitation for stroke survivors, Alzheimer and Parkinson patients, and post-flight rehabilitation for astronauts. Ultimately, this research work contributes to the basis for AR/MR applications to detect and assess early stages of physical disorders.
Juan Martinez* and Michael Martinez
Judging Forms – Official judges only