No abstract

No abstract

The development of future driver assistance systems has a strong tendency towards both automation and cooperative intelligent systems (ITS). Two types of ITS can be defined: Car-to-car (C2C) and car-to-infrastructure (C2I). Either way, the driver is equipped with technology to communicate with other equipped drivers and thus ease cooperation. In regard to nowadays relatively low penetration rate of those systems, the question arises of how non-equipped drivers might react to equipped drivers' behavior. In daily traffic, a driver anticipates the behavior of a vehicle in front, e.g. at a traffic light. It may be anticipated that the drivers maintain their speed when approaching a green traffic light. However, the equipped driver, may act differently, e.g. decelerating when the cooperative system communicates that there is no time left to pass the traffic light in time. The goal of this work is evaluating and modelling the effect of those systems on the interaction between differently equipped drivers by using a new data analysis approach. So the impact of cooperative systems e.g. on safety and user acceptance aspects can be estimated reliably.The data was collected with a multi-driver-simulator that can be used to let multiple drivers interact freely with each other in the same simulation scenario. The scenario was designed with two drivers following a computer controlled leading vehicle. This platoon was driving on a ring-shaped road passing several signalized intersections. The leading vehicle was programmed simulating the behavior of a driver equipped with traffic light assistance. Two different use cases were defined and implemented across the intersections: Firstly, stopping at a red traffic light, start during red phase and crossing the stop line simultaneously with the signal switching to green and secondly, approaching a red traffic light that switches t green before the leading vehicle stops. Test runs were executed to collect both experimental and baseline data, where for the latter the computer controlled vehicle simulated non-equipped driving behavior. 64 participants contributed to the dataset within 32 separate trials. For each trial a participant was either in the first (Nr. 2) or the second (Nr. 3) following car. Both baseline and experimental run where completed respectively.Car-following situations are prominent in today's urban traffic. Thus, especially reacting to changes in speed of a leading vehicle is a fundamental task in driving. In 2013, 13.9% of all crashes with physical injuries were caused by unadapted velocities (Statistisches Bundesamt, 2014). So on the other hand for the sake of safety, it is important to estimate a system's impact on car following behavior. On the other hand, an adaption of non-equipped drivers on equipped ones would enhance the efficiency of the whole traffic flow and thus the benefit of the ITS.Car-following analysis inherently requires taking into account the course of the speed trajectory. Most descriptive statistics do not come up to this challenge. Thus more c...