A predictive model of discretionary lane change behavior considering human factors in the framework of time parameters

Abstract

Lane changing is regarded as one of the most challenging behaviors of drivers. The lane-changing behaviors are divided into mandatory and discretionary. This study proposes an adaptive neuro-fuzzy model of discretionary lane-changing behavior in real traffic flow. Similar to other behaviors of drivers, lane changing is influenced by human factors, including age, gender, level of driving experience, hastiness, cautiousness, and alertness as well as environmental factors such as road and weather conditions. Identifying and measuring the said factors seem to be difficult or, in some cases, impossible. This study sorts out the lane-changing behavior into moments and two time intervals. In these time intervals, distance and relative speed, affected by the said factors, are accounted for in terms of time parameters and fed as inputs to the proposed predictive model. This is the innovative and distinguishing feature of the present study when compared to other researches. Finally, simulation and comparison based on real data indicate that when time parameters are considered and fed as inputs to model the error between the driver’s behavior and the proposed predictive model is less than when time parameter is not accounted for.

Fatigue level detection using multivariate autoregressive exogenous nonlinear modeling based on driver body pressure distribution

“…The closer the coefficient is to 21 or + 1, the stronger the correlation between the variables. 13,30 Table 1 shows the calculated correlation coefficients between fatigue and each of the 16 areas.…”

“…The structure of the NARX model is based on the nonlinear model of the ARX model, which is used to predict and estimate nonlinear systems, which is shown in Figure 22. The performance functions 13,31 NMSE, MAPE, and R are used for the NARX model shown in relations ( 16):…”

“…11 Other alternative methods have been proposed to diagnose fatigue 12 ; which can be mentioned to analyze driver's behavior in lane changing. 13 Søndergaard et al 14 have studied discomfort in prolonged sitting and curvature of the back, resulting showed changes in COP (pressure center) as shown in Figure 1. Also using linear and nonlinear analysis methods, the relationship between discomfort during prolonged sitting and other variables has been investigated.…”

A novel adaptive lane change method in transient dynamic traffic conditions for highly automated vehicles

“…Some previous studies assumed that lane change condition exist and there is no need to create it, which is considered as a discretionary lane change category. 7–9 On the other hand, the main focus of this paper is on mandatory lane change that means the ego vehicle creates lane change condition, and some of the most important limitations of previous research in this area are discussed here.…”