[论文解读] Using Eye-tracking Data to Predict Situation Awareness in Real Time during Takeover Transitions in Conditionally Automated Driving

Feng Zhou, X. Jessie Yang+1 more|arXiv (Cornell University)|Mar 27, 2021

Human-Automation Interaction and Safety被引用 8

一句话总结

本研究提出了一种基于眼动追踪数据与LightGBM机器学习算法的条件自动化驾驶中实时情境意识（SA）预测模型，并通过SHAP提升可解释性。仅使用眼动追踪特征，该模型与真实情境意识的相关性达到0.719，表现出优异的预测性能，并识别出影响接管转换期间情境意识的关键视觉注意力因素。

ABSTRACT

Situation awareness (SA) is critical to improving takeover performance during the transition period from automated driving to manual driving. Although many studies measured SA during or after the driving task, few studies have attempted to predict SA in real time in automated driving. In this work, we propose to predict SA during the takeover transition period in conditionally automated driving using eye-tracking and self-reported data. First, a tree ensemble machine learning model, named LightGBM (Light Gradient Boosting Machine), was used to predict SA. Second, in order to understand what factors influenced SA and how, SHAP (SHapley Additive exPlanations) values of individual predictor variables in the LightGBM model were calculated. These SHAP values explained the prediction model by identifying the most important factors and their effects on SA, which further improved the model performance of LightGBM through feature selection. We standardized SA between 0 and 1 by aggregating three performance measures (i.e., placement, distance, and speed estimation of vehicles with regard to the ego-vehicle) of SA in recreating simulated driving scenarios, after 33 participants viewed 32 videos with six lengths between 1 and 20 s. Using only eye-tracking data, our proposed model outperformed other selected machine learning models, having a root-mean-squared error (RMSE) of 0.121, a mean absolute error (MAE) of 0.096, and a 0.719 correlation coefficient between the predicted SA and the ground truth. The code is available at https://github.com/refengchou/Situation-awareness-prediction. Our proposed model provided important implications on how to monitor and predict SA in real time in automated driving using eye-tracking data.

研究动机与目标

• 开发一种在条件自动化驾驶接管转换期间实时、非侵入性的情境意识（SA）预测方法。
• 克服传统SA测量工具（如SAGAT、SART）因基于探测或事后评估而无法适用于实时应用的局限性。
• 利用眼动追踪数据作为视觉注意力与SA的代理指标，鉴于其与动态驾驶任务中感知与理解的强关联性。
• 通过SHAP值提升模型可解释性，识别最具影响力的眼动追踪特征及其对SA预测的正向或负向影响。
• 通过实时监测驾驶员SA并利用无感生理数据，实现在自动化驾驶系统中的主动干预。

提出的方法

• 从33名参与者处收集了其观看32段模拟驾驶视频时的眼动追踪数据，视频时长为1–20秒，每段视频展示最多六辆具有动态位置与速度的车辆。
• 通过整合三项性能指标（车辆位置、距离与速度估计准确性）将情境意识（SA）标准化为0至1之间的数值。
• 应用LightGBM（一种基于树的集成梯度提升模型）仅基于眼动追踪特征（如注视持续时间、瞳孔直径、扫视幅度等）预测SA。
• 使用SHAP（SHapley Additive exPlanations）解释LightGBM模型，识别最重要的预测变量并量化其对每项SA预测的独立贡献。
• 基于SHAP值进行特征选择，以提升模型性能并减少过拟合。
• 通过均方根误差（RMSE，0.121）、平均绝对误差（MAE，0.096）以及预测值与真实值之间相关系数（0.719）验证模型性能。

实验结果

研究问题

• RQ1 仅使用眼动追踪数据是否能够实现在条件自动化驾驶接管转换期间的实时情境意识预测？
• RQ2 哪些特定的眼动追踪特征（如注视持续时间、瞳孔直径、扫视幅度）对情境意识最具预测力？
• RQ3 各个眼动追踪变量如何影响SA预测，其影响方向是正向还是负向？
• RQ4 SHAP值是否能有效解释树基机器学习模型在驾驶员SA背景下的预测结果，从而实现可解释性并提供可操作的洞察？
• RQ5 当模型仅基于眼动追踪数据训练，不依赖额外生理或行为输入时，是否仍能保持优异性能？

主要发现

• 仅使用眼动追踪数据训练的LightGBM模型，预测值与真实情境意识之间的相关系数达到0.719，RMSE为0.121，MAE为0.096。
• 注视持续时间与SA呈正相关，表明对场景关键要素的长时间注视有助于提升SA，尽管相关性中等（r < 0.5）。
• 瞳孔直径测量值（pupilMean与pupilStd）是重要预测变量，其中pupilMean的影响强于pupilStd，但未观察到显著的线性关系。
• 扫视幅度（sAmpMean）与SA呈负相关，而sAmpStd（幅度的变异性）与SA呈正相关，表明更不规则的扫描模式与更高的SA相关。
• SHAP分析显示，在某一实例中，对SA预测最具影响力的前三个特征为“road = 26”、“difficulty = 29”和“backMirror = 7”；在另一实例中则为“backMirror = 7”、“road = 26”和“fStd = 491.7”，表明特征重要性具有情境依赖性。
• 通过SHAP实现的模型可解释性使我们能够识别出特定视觉注意力模式（如注视后视镜或道路）如何提升或降低预测的SA，为人机系统设计提供了可操作的洞察。