A Dataflow-based Mobile Brain Reading System on Chip with Supervised Online Calibration - For Usage without Acquisition of Training Data

Abstract

Brain activity is more and more used for innovative applications like Brain Computer Interfaces (BCIs). However, in order to be able to use the brain activity, the related psychophysiological data has to be processed and analyzed with sophisticated signal processing and machine learning methods. Usually these methods have to be calibrated with subject-specific data before they can be used. Since future systems that implement these methods need to be portable to be applied more flexible tight constraints regarding size, power consumption and computing time have to be met. Field Programmable Gate Arrays (FPGAs) are a promising solution, which are able to meet all the constraints at the same time. Here, we present an FPGA-based mobile system for signal processing and classification. In addition to other systems, it is able to be calibrated and adapt at runtime, which makes the acquisition of training data unnecessary.