|
Abstract:
|
Lane changes by drivers are significant safety hazards, causing as many as of 20% of accidents. They are also major contributors to congestion, especially in freeway “weaving” sites at which entering, exiting and through vehicles are all present. We assess the capability of modern, connected-vehicle-based and infrastructure-based datasets to illuminate where, when and how lane changes occur, using a variety of statistical methods and visualizations. Issues range from identifying lane changes (with the attendant tradeoff between false positives and false negatives) to distinguishing between mandatory and discretionary lane changes to proposing and evaluating potential interventions. Data are drawn from sources in California and North Carolina.
|
