Reducing the impact of vibration-caused artifacts in a brain-computer interface using gyroscope data

We implemented an artifact prediction method for a saline-pad wireless electroencephalograph equipped with two-axis gyroscope used as a basic brain-computer interface (BCI). The BCI unit serves two purposes in the scope of the project Innotruck. Firstly, it enables remote control of vehicles and other systems over a limited set of trained mental activity. Secondly, it is a source of data for the passive analysis of the operator's mental fitness, which is further integrated into the driver assistance systems. The latter aspect has been the focus of our work. Saline-pad electrodes used in consumer grade electronics are prone to errors stemming from vibrations and sudden head movements. The implemented approach successfully preconditions the signal processing pipeline to take such artifacts into account and reduces the later processing overhead.