The use of naturalistic driving studies (NDSs) for driver behavior research has skyrocketed over the past two decades. Intersections are a key target for traffic safety, with up to 25-percent of fatalities and 50-percent injuries from traffic crashes in the United States occurring at intersections. NDSs are increasingly being used to assess driver behavior at intersections and devise strategies to improve intersection safety. A common challenge in NDS intersection research is the need for to combine spatial locations of driver-visited intersections with concurrent video clips of driver trajectories at intersections to extract analysis variables. The intersection identification and driver trajectory video clip extraction process are generally complex and repetitive. We developed a novel R package called ndsintxn to streamline this process and automate best practices to minimize computational time, cost, and manual labor. This paper provides details on the methods and illustrative examples used in the ndsintxn R package.

翻译：在过去二十年中,利用自然驾驶研究(NDS)进行驾驶员行为研究的自然驾驶研究(NDS)已大肆发展,交叉路口是交通安全的一个关键目标,美国交通事故造成的死亡人数和受伤率高达25%,在十字路口发生。国家驾驶员越来越多地使用自然驾驶研究(NDS)评估交叉路口的驾驶员行为,并制订改善交叉安全的战略。国家发展司交叉路口研究的一个共同挑战是,需要将驾驶员-来访交叉路口的空间位置与交叉路口司机轨迹的同步视频剪片结合起来,以提取分析变量。交叉识别和驾驶员轨迹视频剪取过程一般是复杂和重复的。我们开发了一个名为ndsintxn的新型R包件,以简化这一过程,并自动化最佳做法,以尽量减少计算时间、成本和人工劳动。本文详细介绍了dsintxn R包使用的方法和示例。