Roborueda: Python-based GUI to control a wheelchair and monitor user posture

Neck and/or head movements play an important role in the control of assistive devices such as robotic wheelchairs, considering these systems allow the acquisition of intentionality information and conversion into control commands, which can improve the mobility and independence of people with severe...

Full description

Saved in:
Bibliographic Details
Published inSoftwareX Vol. 24; p. 101555
Main Authors Gonzalez-Cely, Aura Ximena, Blanco-Diaz, Cristian Felipe, Diaz, Camilo A.R., Bastos-Filho, Teodiano Freire
Format Journal Article
LanguageEnglish
Published Elsevier 01.12.2023
Subjects
Online AccessGet full text

Cover

Loading…
More Information
Summary:Neck and/or head movements play an important role in the control of assistive devices such as robotic wheelchairs, considering these systems allow the acquisition of intentionality information and conversion into control commands, which can improve the mobility and independence of people with severe disabilities. This work addresses a gap based on the development of Human–Machine Interfaces (HMIs) by introducing Roborueda, a robust Graphical User Interface (GUI) designed to facilitate the interaction of robotic wheelchair users, with a primary focus on enhancing mobility and preventing future diseases. The main objective of the system is to recognize head and neck movements using strategies based on inertial and Optical Fiber Sensors (OFSs) to generate commands that drive a robotic wheelchair. In addition, the system provides a Posture Recognition System that can help in the preventive treatment of pressure ulcers. The GUI was developed in Python programming language using Wi-Fi communication functionalities, which encourages open software distribution. The system, together with the GUI, was initially evaluated with healthy subjects, where the subsystem in charge of sensing and control obtained an accuracy close to 100% using fuzzy logic techniques, whereas posture recognition using OFS presented accuracy close to 97% using Machine Learning (ML) classifiers, highlighting the k-Nearest Neighbors. The system presented response time of less than 250 ms, which allows observing the feasibility of real-time implementation. Additionally, the subjects showed compliance and familiarity with the GUI, which allowed concluding that the software is user friendly and has potential use in the field of assistive technology and robotic rehabilitation. The implementation of this GUI marks a significant step towards integrating OFS, which contributes to the advancement of assistive technologies. To ensure the effectiveness and practicality of Roborueda, future work will focus on validating the system through extensive testing with wheelchair users, thereby obtaining valuable feedback and making further improvements related to posture control to prevent pressure ulcers and more personalized controllers, such as speed control.
ISSN:2352-7110
2352-7110
DOI:10.1016/j.softx.2023.101555