Single-Channel Real-Time Drowsiness Detection Based on Electroencephalography.

The need of a reliable drowsiness detection system is arising today, as drowsiness is considered as a major cause for accidents as much as alcohol. In this paper, we propose a real-time drowsiness detection algorithm based on a single-channel electroencephalography (EEG) for wearable devices without demanding computing and power resources. The proposed algorithm adopts a cumulative counter to extract important features from 8 different frequency bands: delta (1-3 Hz), theta ($\not\subset-7$ Hz), low-alpha (8-9 Hz), high-alpha (10-12 Hz), low-beta (13-17 Hz), high-beta (18-30 Hz), low-gamma (31-40 Hz), and high-gamma (41-50 Hz). These features are then processed by a support vector machine (SVM) to distinguish between drowsy and awake states. Our preliminary results demonstrate that the proposed algorithm is capable of detecting drowsiness with superior accuracy (83.36%) over the conventional method (70.62%).

Low-power hardware implementation of movement decoding for brain computer interface with reduced-resolution discrete cosine transform.

This paper describes a low-power hardware implementation for movement decoding of brain computer interface. Our proposed hardware design is facilitated by two novel ideas: (i) an efficient feature extraction method based on reduced-resolution discrete cosine transform (DCT), and (ii) a new hardware architecture of dual look-up table to perform discrete cosine transform without explicit multiplication. The proposed hardware implementation has been validated for movement decoding of electrocorticography (ECoG) signal by using a Xilinx FPGA Zynq-7000 board. It achieves more than 56× energy reduction over a reference design using band-pass filters for feature extraction.