Research On Vehicle Speed Control Model Based On Vigilance Detection Using EEG

In the recent years, with the increasing traffic accident number caused by drowsy driving, the driver-centered vehicle safety problem has become one of the most important road traffic safety issues in China and even the world. Therefore, the research of vigilance detection based vehicle active safety technology has become a hot topic in vehicle field and the vigilance detection device will gradually became one of a part of Advanced Driver Assistance Systems (ADAS). In this paper, we focus on the research of driver-based vehicle active safety technology and present a novel vehicle speed control model based on vigilance detection using EEG and low-rank matrix decomposition. The studies included in this paper are summarized as follows:First, an 8-channel wireless wearable BCI system is designed based on the relationship of electroencephalogram (EEG) and drowsy to collect driver’s EEG signal. The high quality electrode is placed on the drowsy related cerebral areas. After that we design experiments to collect the driver’s EEG signal when driver is drowsy or alert. Moreover, a 64-channel EEG collection device provided by Brain Products also is used for EEG collection experiment under the same experiment condition.Second, we will adopt wavelet filter and de-noising algorithm to remove the interference of EEG based on the deep knowledge about EEG disturbing factors. After that, EEG logarithmic power spectrum density will calculate using FFT to model the vigilance detection model. We will discuss the effect of data size on feature extraction.Third, Low-rank matrix decomposition algorithm is adopted in dictionary learning and vigilance classification. The original EEG signal is segmented by second and then shaped into a matrix to form the original dictionary. Then, the Low-rank matrix decomposition algorithm is presented to calculate a new dictionary with stronger representation ability.At last, the vigilance detection based vehicle speed control model is present to decide when and how will the speed control (deceleration, braking) should be made. Moreover, a vehicle safety deceleration/braking based on vehicle following minimum distance model, is proposed to calculate a safety accelerometer which the vehicle should adopt in the course of speed control to avoid rear-end collision. In the last of this paper, the vehicle dynamical model is established to verify the above method. Then, analyze the speed and accelerometer curve when driver is alert and drowsy. The final performance evaluation demonstrates the validity of the proposed vehicle speed control model.