Planar Control of a Quadcopter Using a Zero-Training Brain Machine Interface Platform

Reza A Abiri (1) , Justin A Kilmarx (1) , Mohammad A Raji (1) , Xiaopeng A Zhao (1)

1 The University of Tennessee [Knoxville]

Brain Machine interface (BMI) enables promising applications in neuroprosthesis and neurorehabilitation by controlling robotic devices based on the subject's intentions. In contrast to the earlier techniques using sensorimotor rhythms, the method here intends to directly extract information of imagined body kinematics and thus can significantly reduce training time. We developed a zero-training BMI platform that controls a quadcopter using noninvasively acquired brain signals. Scalp electroencephalogram (EEG) signals of a user's imaginary movements are collected in real-time and translated by a computer to control a quadcopter along a designated path in a two-dimensional space. The results showed a promising performance using the zero-training paradigm that can be utilized in controlling neuroprosthetic limbs and neurorehabilitation devices by patients with BCI illiteracy.

Domaines: Biotechnologie; Interface homme-machine [cs.HC]; Systèmes et contrôle [cs.SY]

Reza A Abiri, Justin A Kilmarx, Mohammad A Raji, Xiaopeng A Zhao. Planar Control of a Quadcopter Using a Zero-Training Brain Machine Interface Platform. Biomedical Engineering Society Annual Meeting (BMES 2016), Oct 2016, Minneapolis, MN, United States. ⟨hal-01574290⟩