A novel task-oriented brain computer interface for human-robot interaction

Since the advent of the technology age, new and innovative ways to communicate with and control machines have been gaining momentum over traditional forms of input. The rise of touch, gesture and speech based control high- light the desire for more natural human com- puter interactions. A Brain-Computer Inter- face (BCI) is an extension of these interaction methods; allowing a user's brain activity to be interpreted as an input into a computer system. In this paper, a successful and novel implemen- tation for controlling a robotic arm using a P- 300 Speller interface is presented, with partic- ular focus on solving high order problems with task speci c commands. The system was val- idated through a packing experiment, where a participant needed to pick and place scattered objects into a container. Initial results show successful completion of the task within four minutes, with a command selection accuracy of approximately 97%.