Analysis of vehicle interactions on interstate highways: Discrete choice and linear systems approaches
The research in this dissertation developed statistical and linear system models to predict driving behavior according to driver attributes, vehicle characteristics, car-following dynamics and/or driving conditions. The scope of this research was limited to the interaction of two vehicles traveling in the same direction on interstate highways. The statistical models proposed in this research investigated the contributions of the different vehicle types on the manner and likelihood of collision. Discrete choice models were used to estimate the probability of the types of collision (rear-end, angle and sideswipe) as functions of driver attributes, striking and struck vehicle types, pre-crash driving actions and traffic and environmental conditions. The National Automotive Sampling System-General Estimates System (NASS-GES) crash data set from 2005 to 2009 was used to develop and validate the model. This research demonstrated that different types of vehicles have different coefficients in the utility functions of the discrete choice models. Using linear system analysis, the states of a pair of vehicles over time were predicted based on vehicle-following dynamics and vehicle characteristics. The internal characteristics of the system composed of driver and type of vehicle were represented in a state-space equation. Assuming relative speed and space headway as the states of the system, the optimal safe distance between the vehicles was predicted using a gain feedback matrix and a state estimator. Using Kalman Filter, the best estimation of future states was obtained to control a "safe vehicle-following" based on the actions of the leader and the follower. Vehicle trajectories from the Next Generation SIMulation (NGSIM) database were used to calibrate and validate the state-space prediction model. This research identified the differences in the characteristic matrices of the different types of vehicle due to their dynamic properties. This research recommended that different types of vehicle based on their dynamic properties should be differentiated to quantify safe and optimal interaction between cars and trucks.
Statistics|Civil engineering|Electrical engineering
Romo, Alicia, "Analysis of vehicle interactions on interstate highways: Discrete choice and linear systems approaches" (2013). ETD Collection for University of Texas, El Paso. AAI3597250.