Tan Yaoyuan V, Elliott Michael R, Flannagan Carol A C
Department of Biostatistics, University of Michigan, United States.
Department of Biostatistics, University of Michigan, United States.
Accid Anal Prev. 2017 Sep;106:428-436. doi: 10.1016/j.aap.2017.07.003. Epub 2017 Jul 20.
As connected autonomous vehicles (CAVs) enter the fleet, there will be a long period when these vehicles will have to interact with human drivers. One of the challenges for CAVs is that human drivers do not communicate their decisions well. Fortunately, the kinematic behavior of a human-driven vehicle may be a good predictor of driver intent within a short time frame. We analyzed the kinematic time series data (e.g., speed) for a set of drivers making left turns at intersections to predict whether the driver would stop before executing the turn. We used principal components analysis (PCA) to generate independent dimensions that explain the variation in vehicle speed before a turn. These dimensions remained relatively consistent throughout the maneuver, allowing us to compute independent scores on these dimensions for different time windows throughout the approach to the intersection. We then linked these PCA scores to whether a driver would stop before executing a left turn using the random intercept Bayesian additive regression trees. Five more road and observable vehicle characteristics were included to enhance prediction. Our model achieved an area under the receiver operating characteristic curve (AUC) of 0.84 at 94m away from the center of an intersection and steadily increased to 0.90 by 46m away from the center of an intersection.
随着联网自动驾驶汽车（CAV）进入车队，在很长一段时间内，这些车辆都将不得不与人类驾驶员互动。CAV面临的挑战之一是人类驾驶员不能很好地传达他们的决策。幸运的是，在短时间内，人类驾驶车辆的运动行为可能是驾驶员意图的良好预测指标。我们分析了一组在十字路口左转的驾驶员的运动时间序列数据（例如速度），以预测驾驶员在转弯前是否会停车。我们使用主成分分析（PCA）来生成独立维度，以解释转弯前车辆速度的变化。在整个操作过程中，这些维度保持相对一致，这使我们能够在接近十字路口的不同时间窗口内，计算这些维度上的独立得分。然后，我们使用随机截距贝叶斯加法回归树，将这些PCA得分与驾驶员在执行左转前是否会停车联系起来。我们还纳入了另外五个道路和可观测车辆特征，以提高预测能力。我们的模型在距离十字路口中心94米处的受试者工作特征曲线（AUC）下面积为0.84，并在距离十字路口中心46米处稳步增加到0.90。
