Drowsiness Detection Algorithm

There are four major types of continuous rhythmic sinusoidal EEG activity.

They are recognized as alpha (8-12Hz), beta (above 12Hz), delta (below 4Hz) and theta (4-8Hz) and are listed in Table 5-1:

Table 5- 1 Characteristics of EEG bands

Types Band range Description

γ

(gamma)

30Hz~ Gamma rhythms may be involved in

higher mental activity, including perception, problem solving, fear, and consciousness.

β (beta)

13 – 30Hz Beta with low amplitude beta with multiple and varying frequencies is often associated with active, busy or anxious thinking and active concentration. Rhythmic beta with a dominant set of frequencies is associated with various pathologies and drug effects, especially benzodiazepines.

α (alpha)

8 – 12 Hz Alpha is characteristic of a relaxed, alert state of consciousness. For alpha rhythms to arise, usually the eyes need to be closed. Alpha attenuates with drowsiness and open eyes, and typically come from the occipital (visual) cortex.

An alpha-like normal variant called mu is sometimes seen over the motor cortex (central scalp) and attenuates with movement, or rather with the intention to move.

θ 4 - 8 Hz Theta is associated with drowsiness,

(theta) childhood, adolescence and young adulthood. This EEG frequency can sometimes be produced by hyperventilation. Theta waves can be seen during hypnagogic states such as trances, hypnosis, deep day dreams, lucid dreaming and light sleep and the preconscious state just upon waking, and just before falling asleep.

δ (delta)

~ 4 Hz Delta is often associated with the very young and certain encephalopathies and underlying lesions. It is seen in stage 3 and 4 sleep.

As the characteristic of EEG activity described above, there is an important phenomenon found by the team of brain research center (BRC), NCTU (National Chao Tung University) while recording EEG from forehead. That is, if a person is mild drowsiness, the alpha wave will tend to be superior in EEG activity, and its power will increase time after time. After that, if the person tends to fall asleep, the theta wave will tend to be superior in EEG activity, and the power of alpha will decrease while the theta will still increase time after time. With the phenomenon, the algorithm can be mapped in to a table listed below.

Table 5- 2 Criterion of Drowsiness

State of EEG band State of consciousness

alpha ↑ theta↑ Mild drowsiness

alpha ↓ theta↑ Deep drowsiness alpha ↑ theta↓ Conscious

alpha ↓ theta↓ Conscious For the criterion of drowsiness is based on the changes of EEG activity in the frequency domain. We use the short-time Fourier transform (STFT), or alternatively short-term Fourier transform, which is a Fourier-related transform used to determine the sinusoidal frequency and phase content of local sections of a signal as it changes over time. Here we have tested the short-time Fourier transform on the embedded processor we use. It has produced good results while processing both lower frequency sinusoid (theta band from 4 to 8 Hz) and higher frequency sinusoid (alpha band from 8 to 12 Hz) of the frequency band we desired(4 to 12 Hz). As shown in Fig.5-4, the power of 4Hz and 5Hz sinusoid showed in Fig.5-5 result in peak of 4Hz and 5Hz.

Also, in Fig.5-6, the power of 10Hz and 12Hz sinusoid showed in Fig.5-7 result in peak of 10Hz and 12Hz.

Fig.5- 4 The original signal of 4Hz & 5Hz sinusoid

Fig.5- 5 The power spectrum of 4Hz & 5Hz sinusoid

Fig.5- 6 The original signal of 10Hz & 12Hz sinusoid

Fig.5- 7 The power spectrum of 10Hz & 12Hz sinusoid

The flowchart of our algorithm is described in Fig.5-8. First, the 4-channel EEG data is re-sampled to sampling rate 64Hz. After gathering 512 points of EEG data, the data was fed into ICA. After that, some estimated components are rejected with the criterion of standard deviation described before. We select the last 192 points of ICA component from rest components and use FFT to estimate the power of EEG data.

Finally with the online drowsiness detection algorithm, we recognized the drowsiness index. The detail of the FFT procedure is shown in Fig.5-9. The 192-point EEG data of each channel is overlapped with 64-point update for each sec. Then the 192 points are divided into 32 points sub-windows with overlap 24 points. Then The first 16 points and the last 16 points of 64-point FFT were padded with zero.

Fig.5- 8 The flowchart of our algorithm

Fig.5- 9 The procedure of FFT

After we take 1~13 Hz of the decibel value of FFT and find the trend of alpha and theta band power, we can estimate drowsiness index with a 20 seconds time window. The flow of from FFT to estimation of drowsiness index is shown in Fig.5-10. The moving window size we use is 20sec. Since each time we update 2 seconds. We calculate the whole slope of ten 2-sec power when each time we update

the alpha and theta power value. Choosing the moving window of 20 seconds is the middle-of-the-road policy here because whether the window is too big or too small isn’t good for drowsiness index estimation. If the window is too big, the change of drowsiness index will be too slow. On the contrary, if the window is too small, the change of drowsiness index will be too soon.

Fig.5- 10 The flow of drowsiness estimation algorithm