TRUE lab utilizes various kinds of physiological data to analyze driver’s fatigue or stress. Measuring physiological signals can show overall patterns that are not intrinsic to the data. These trends can sometimes hinder the data analysis and must be removed. Consider two electrocardiogram (ECG) signals with different trends. ECG signals are sensitive to disturbances such as power source interference. In order to solve this problem, we can use curve fitting method. Curve fitting is the process of constructing a curve, or mathematical function, that has the best fit to a series of data points, possibly subject to constrains. Curve fitting can involve either interpolation, where an exact fit to the data is required, or smoothing. By using ‘polyfit’ and ‘polyval’ function in Matlab, the trend can be removed. Polyfit computes a least squares polynomial for a given set of data. Polyfit generates the coefficients of the polynomial, which can be used to model a curve to fit the data. Polyval evaluates a polynomial for a given set of x values. Therefore, polyval generates a curve to fit the data based on the coefficients found using polyfit.
Based on EEG stress index extracted from actual EV (electric vehicle) driving experiment, TRUE lab observed how operational and physiological data are related to EV driving stress. The plots below show the relationship between EEG stress index and remaining driving range (RDR), velocity, altitude and gazing rate of each 1/100 progress of the experiment. The subjects didn’t show common trend but some of them had significant aspects which can be the evidences of well-known suspicions. In the upper set of plots, driver’s stress increases as remaining driving range decreases below around 40 km. And other operational data doesn’t show any significant trend with stress. Therefore, we think this subject felt ‘range anxiety’ for driving EV. On the other hand, the other set of plots implies that subject felt more stress when she drove uphill road at high altitude. TRUE lab will apply a certain technique to analyze those non-linear relationships quantitatively.
TRUE Lab analyzes various traffic accident data, and we have utilized 5 year Seoul traffic accident data in order to find out the vulnerability of senior occupants in various seat positions and restraint use with multinomial logit model. The general approach to multinomial data is to nominate one of the response categories as a baseline, calculate odds ratio for all other categories relative to the baseline. In this analysis, we picked the non-severe injury as a baseline, and calculated odds ratio for severe injury for each explored variable. This case, the multinomial logit model converges to a binary logit model since binary response (severe and non-severe) was used. We examined how much the senior occupants are at risk regarding various vehicle seat positions and use of seatbelts. The graph below shows that seniors are more likely to have severe injury than mid-age group in all seat positions, especially in the front passenger seat and the rear seat. When seatbelt is off, the ORs increased, but it decreased sharply with a large magnitude when seatbelt is on.
Time series models were developed to examine the long term trends with the effect of time. This model is often affected by intervention events such as a new regulation/policy, holidays, promotions, and natural disasters. On April 1, 2004, KTX (Korea Train eXpress), the first HSR (High-Speed Rail) in Korea, was introduced to Gyeongbu Line. The introduction of the KTX service led to a change in the number of passenger for Gyeongbu Line. Previous studies have analyzed the pre and post-event changes of the intervening events by simple statistics or intervention ARIMA analysis. True lab analyzed the effect of intervention event on the number of passengers using the Gyeongbu line based on the time series outlier detection which can overcome limitations in the previous studies. Time series outlier detection can analyze the time, effect type and size of an intervention event without the assumption of the time and effect type of the intervention event. The data was collected from the Korea Transport Database (KTDB) and included to the years 2003 to 2014. As a result of the analysis, the size of the influence type in the same intervention event was different for the major city routes, and the intervention event which could not be found by previous study method was found.
To address the anxiety about battery capacity and driving range from driving electric vehicles, TRUE lab suggested an empirical index calculated by the ratio of remaining driving amount and state of charge. For example, if a driver need to ride more than expected driving range, he or she may feel anxiety.
EV drivers’ EEG data was collected from field operational test and TRUE lab extracted EEG stress index values of each subject from the data which represent EV drivers real stress level during driving. The plot below shows relationship between our suggested range anxiety index and extracted EEG stress index from one subject. Most of subjects had statistically positive relationship and this implies that suggested index may reflect EV drivers’ range anxiety.
It is known that some people who just began driving an electric vehicle feel certain fear from insufficient driving range or state of charge. This is called 'Range Anxiety' and often said to be a barrier to buying electric vehicles.
TRUE lab is analyzing driver experiment data to find out the existence and appearance condition of EV Range Anxiety scientifically. Many kinds of data are considered and 'eye fixation' is also one of the data we focus on. Based on a hypothesis that drivers must look at the parts of driving range and state of charge in instrumental panel more when they feel more anxiety of range, we check how long and how often driver subjects see the panel.
Looking some relationship between the eye fixation and state of charge, more human factors like EEG, ECG are being considered together. Please look forward to the next post!