Login / Signup

Agent-Based Intelligent Interface for Wheelchair Movement Control.

Alberto L BarriusoJavier Pérez-MarcosDiego M Jiménez-BravoGabriel Villarrubia GonzálezJuan F De Paz
Published in: Sensors (Basel, Switzerland) (2018)
People who suffer from any kind of motor difficulty face serious complications to autonomously move in their daily lives. However, a growing number research projects which propose different powered wheelchairs control systems are arising. Despite of the interest of the research community in the area, there is no platform that allows an easy integration of various control methods that make use of heterogeneous sensors and computationally demanding algorithms. In this work, an architecture based on virtual organizations of agents is proposed that makes use of a flexible and scalable communication protocol that allows the deployment of embedded agents in computationally limited devices. In order to validate the proper functioning of the proposed system, it has been integrated into a conventional wheelchair and a set of alternative control interfaces have been developed and deployed, including a portable electroencephalography system, a voice interface or as specifically designed smartphone application. A set of tests were conducted to test both the platform adequacy and the accuracy and ease of use of the proposed control systems yielding positive results that can be useful in further wheelchair interfaces design and implementation.
Keyphrases
  • healthcare
  • machine learning
  • high throughput
  • primary care
  • physical activity
  • quality improvement
  • risk factors
  • deep learning