Investigating the feasibility of a BCI-driven robot-based writing agent for handicapped individuals

27th International Conference on CADCAM, Robotics and Factories of the Future 2014

工业技术;机械制造

Syan, Chanan S.^1 ; Harnarinesingh, Randy E. S.^1 ; Beharry, Rishi^1

Department of Mechanical and Manufacturing Engineering, University of the West Indies, St. Augustine, Trinidad and Tobago^1

关键词: 3 degrees of freedom;    Assistive technology;    Brain computer interfaces (BCIs);    Classification accuracy;    Disabled individuals;    Local area networks (LAN);    Private information;    Wheelchair controllers;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/65/1/012020/pdf
DOI  :  10.1088/1757-899X/65/1/012020

学科分类：工业工程学

来源: IOP

【 摘 要 】

Brain-Computer Interfaces (BCIs) predominantly employ output actuators such as virtual keyboards and wheelchair controllers to enable handicapped individuals to interact and communicate with their environment. However, BCI-based assistive technologies are limited in their application. There is minimal research geared towards granting disabled individuals the ability to communicate using written words. This is a drawback because involving a human attendant in writing tasks can entail a breach of personal privacy where the task entails sensitive and private information such as banking matters. BCI-driven robot-based writing however can provide a safeguard for user privacy where it is required. This study investigated the feasibility of a BCI-driven writing agent using the 3 degree-of- freedom Phantom Omnibot. A full alphanumerical English character set was developed and validated using a teach pendant program in MATLAB. The Omnibot was subsequently interfaced to a P300-based BCI. Three subjects utilised the BCI in the online context to communicate words to the writing robot over a Local Area Network (LAN). The average online letter-wise classification accuracy was 91.43%. The writing agent legibly constructed the communicated letters with minor errors in trajectory execution. The developed system therefore provided a feasible platform for BCI-based writing.