This paper considers the decision-making problem for a human-driven vehicle crossing a road intersection in the presence of other, potentially errant, drivers. Our approach relies on a novel threat assessment module, which combines an intention predictor based on support vector machines with an efficient threat assessor using rapidly-exploring random trees. This module warns the host driver with the computed threat level and the corresponding best "escape maneuver" through the intersection, if

more »

... he threat is sufficiently large. Through experimental results with small autonomous and human-driven vehicles, we demonstrate that this threat assessment module can be used in real-time to minimize the risk of collision.