Traffic Sign Recognition Evaluation for Senior Adults Using EEG Signals.

Elderly people are not likely to recognize road signs due to low cognitive ability and presbyopia. In our study, three shapes of traffic symbols (circles, squares, and triangles) which are most commonly used in road driving were used to evaluate the elderly drivers' recognition. When traffic signs are randomly shown in HUD (head-up display), subjects compare them with the symbol displayed outside of the vehicle. In this test, we conducted a Go/Nogo test and determined the differences in ERP (event-related potential) data between correct and incorrect answers of EEG signals. As a result, the wrong answer rate for the elderly was 1.5 times higher than for the youths. All generation groups had a delay of 20-30 ms of P300 with incorrect answers. In order to achieve clearer differentiation, ERP data were modeled with unsupervised machine learning and supervised deep learning. The young group's correct/incorrect data were classified well using unsupervised machine learning with no pre-processing, but the elderly group's data were not. On the other hand, the elderly group's data were classified with a high accuracy of 75% using supervised deep learning with simple signal processing. Our results can be used as a basis for the implementation of a personalized safe driving system for the elderly.

An Evaluation Method of Safe Driving for Senior Adults Using ECG Signals.

The elderly are more susceptible to stress than younger people. In particular, heart palpitations are one of the causes of heart failure, which can lead to serious accidents. To prevent heart palpitations, we have devised the Safe Driving Intensity (SDI) and Cardiac Reaction Time (CRT) as new methods of estimating the correlations between effects on the driver's heart and the movement of a vehicle. In SDI measurement, recommended acceleration value of vehicle for safe driving is inferred from the suggested correlation algorithm using machine learning. A higher SDI value than other people means less pressure on the heart. CRT is an estimated value of the occurring time of heart palpitations caused by stressful driving. In particular, it is proved by SDI that elderly subjects tend to overestimate their driving abilities in personal assessment questionnaires. Furthermore, we validated our SDI using other general statistical methods. When comparing the results using a t-test, we obtained reliable results for the equivalent variance. Our results can be used as a basis for evaluating elderly people's driving ability, as well as allowing for the implementation of a personalized safe driving system for the elderly.

Filter Bank Regularized Common Spatial Pattern Ensemble for Small Sample Motor Imagery Classification.

For the last few years, many feature extraction methods have been proposed based on biological signals. Among these, the brain signals have the advantage that they can be obtained, even by people with peripheral nervous system damage. Motor imagery electroencephalograms (EEG) are inexpensive to measure, offer a high temporal resolution, and are intuitive. Therefore, these have received a significant amount of attention in various fields, including signal processing, cognitive science, and medicine. The common spatial pattern (CSP) algorithm is a useful method for feature extraction from motor imagery EEG. However, performance degradation occurs in a small-sample setting (SSS), because the CSP depends on sample-based covariance. Since the active frequency range is different for each subject, it is also inconvenient to set the frequency range to be different every time. In this paper, we propose the feature extraction method based on a filter bank to solve these problems. The proposed method consists of five steps. First, motor imagery EEG is divided by a using filter bank. Second, the regularized CSP (R-CSP) is applied to the divided EEG. Third, we select the features according to mutual information based on the individual feature algorithm. Fourth, parameter sets are selected for the ensemble. Finally, we classify using ensemble based on features. The brain-computer interface competition III data set IVa is used to evaluate the performance of the proposed method. The proposed method improves the mean classification accuracy by 12.34%, 11.57%, 9%, 4.95%, and 4.47% compared with CSP, SR-CSP, R-CSP, filter bank CSP (FBCSP), and SR-FBCSP. Compared with the filter bank R-CSP ( , ), which is a parameter selection version of the proposed method, the classification accuracy is improved by 3.49%. In particular, the proposed method shows a large improvement in performance in the SSS.

Improving the Accuracy and Training Speed of Motor Imagery Brain-Computer Interfaces Using Wavelet-Based Combined Feature Vectors and Gaussian Mixture Model-Supervectors.

In this paper, we propose a set of wavelet-based combined feature vectors and a Gaussian mixture model (GMM)-supervector to enhance training speed and classification accuracy in motor imagery brain-computer interfaces. The proposed method is configured as follows: first, wavelet transforms are applied to extract the feature vectors for identification of motor imagery electroencephalography (EEG) and principal component analyses are used to reduce the dimensionality of the feature vectors and linearly combine them. Subsequently, the GMM universal background model is trained by the expectation-maximization (EM) algorithm to purify the training data and reduce its size. Finally, a purified and reduced GMM-supervector is used to train the support vector machine classifier. The performance of the proposed method was evaluated for three different motor imagery datasets in terms of accuracy, kappa, mutual information, and computation time, and compared with the state-of-the-art algorithms. The results from the study indicate that the proposed method achieves high accuracy with a small amount of training data compared with the state-of-the-art algorithms in motor imagery EEG classification.

An R-peak detection method that uses an SVD filter and a search back system.

In this paper, we present a method for detecting the R-peak of an ECG signal by using an singular value decomposition (SVD) filter and a search back system. The ECG signal was detected in two phases: the pre-processing phase and the decision phase. The pre-processing phase consisted of the stages for the SVD filter, Butterworth High Pass Filter (HPF), moving average (MA), and squaring, whereas the decision phase consisted of a single stage that detected the R-peak. In the pre-processing phase, the SVD filter removed noise while the Butterworth HPF eliminated baseline wander. The MA removed the remaining noise of the signal that had gone through the SVD filter to make the signal smooth, and squaring played a role in strengthening the signal. In the decision phase, the threshold was used to set the interval before detecting the R-peak. When the latest R-R interval (RRI), suggested by Hamilton et al., was greater than 150% of the previous RRI, the method of detecting the R-peak in such an interval was modified to be 150% or greater than the smallest interval of the two most latest RRIs. When the modified search back system was used, the error rate of the peak detection decreased to 0.29%, compared to 1.34% when the modified search back system was not used. Consequently, the sensitivity was 99.47%, the positive predictivity was 99.47%, and the detection error was 1.05%. Furthermore, the quality of the signal in data with a substantial amount of noise was improved, and thus, the R-peak was detected effectively.