Authors: Tao Geng, John Q. Gan, Huosheng Hu
Addresses: Department of Computing & Electronic Systems, University of Essex, Colchester, CO4 3SQ, UK. ' Department of Computing & Electronic Systems, University of Essex, Colchester, CO4 3SQ, UK. ' Department of Computing & Electronic Systems, University of Essex, Colchester, CO4 3SQ, UK
Abstract: This paper presents the design and online experiments of a self-paced online brain-computer interface (BCI) for controlling a simulated robot in an indoor environment. Three one-vs-rest linear discriminant analysis (LDA) classifiers are combined to control the switching between automatic control (AC) and subject control (SC) modes. The hierarchical structure of the controller allows the most reliable class (mental task) in a specific subject to play a dominant role in the robot control. A group of simple rules triggered by local sensor signals are designed for safety and obstacle avoidance in the AC mode. Due to the intuitive nature of the controller and the small number of AC rules, a subject has much flexibility and full control of the robot. Online experiments have shown that subjects successfully control the robot to circumnavigate obstacles and reach some specified targets in separate rooms by motor imagery of their hands and feet.
Keywords: brain-computer interface; BCI; brain actuated robot control; electroencephalography; EEG; mobile robots; robot simulation; indoor environments; linear discriminant analysis; LDA; obstacle avoidance; collision avoidance; motor imagery.
International Journal of Advanced Mechatronic Systems, 2010 Vol.2 No.1/2, pp.28 - 35
Published online: 10 Jan 2010 *Full-text access for editors Access for subscribers Purchase this article Comment on this article