EEG-based brain-computer interface for human-robot collaboration

Li, Yao

Permalink

https://hdl.handle.net/2142/109363 Copy

Description

Title

EEG-based brain-computer interface for human-robot collaboration

Author(s)

Li, Yao

Issue Date

2020-11-19

Director of Research (if dissertation) or Advisor (if thesis)

Kesavadas, Thenkurussi

Doctoral Committee Chair(s)

Kesavadas, Thenkurussi

Committee Member(s)

• Reis, Henrique M
• Sreenivas, Ramavarapu S
• Chowdhary, Girish

Department of Study

Industrial&Enterprise Sys Eng

Discipline

Systems & Entrepreneurial Engr

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Date of Ingest

2021-03-05T21:36:58Z

Keyword(s)

• BCI
• EEG
• SSVEP
• Industrial robot

Abstract

One of the expectations for the next-generation industrial robots is to work collaboratively with humans. Collaborative robots must be able to communicate with human collaborators intelligently and seamlessly. However, industrial robots in prevalence are not good at understanding human intentions and decisions. We propose to develop human-robot interactions based on Brain-Computer Interfaces (BCIs) transferring human cognition to robots directly. By collecting and encoding brain activities with BCIs, human can actively send commands to robots in thought or passively let robots monitor mental activities. We conduct two major experiments, i.e. BCI for welding robot and BCI for defective part picking robot, through which human operators can actively communicate with robots and work collaboratively on manufacturing tasks. The BCI for welding robot allows operators to select weld beads and command the robot to weld in thought. In the picking robot study, the robot picks defective part from a conveyor based on the decisions made when operators examining the qualities visually. Besides, to build faster and more accurate BCIs, we propose a Conv-CA model, which combines convolutional neural network (CNN) and canonical correlation analysis (CCA) to improve the performance of the state-of-art steady-state visually evoked potential (SSVEP) algorithm. We also conduct a study for passive BCI communication, i.e. the robot detects the circumstance when operators feel unsafe in the human-robot collaboration. When a fear response is detected, the robot can stop immediately to protect human safety.

Graduation Semester

2020-12

Type of Resource

Thesis

Permalink

http://hdl.handle.net/2142/109363

Copyright and License Information

Copyright 2020 Yao Li

Owning Collections