Predicting Driver Takeover Time in Conditionally Automated Driving
URL: http://arxiv.org/abs/2107.09545v1
archive: archived pipeline: cataloged verified
Abstract
It is extremely important to ensure a safe takeover transition in conditionally automated driving. One of the critical factors that quantifies the safe takeover transition is takeover time. Previous studies identified the effects of many factors on takeover time, such as takeover lead time, non-driving tasks, modalities of the takeover requests (TORs), and scenario urgency. However, there is a lack of research to predict takeover time by considering these factors all at the same time. Toward this end, we used eXtreme Gradient Boosting (XGBoost) to predict the takeover time using a dataset from a meta-analysis study [1]. In addition, we used SHAP (SHapley Additive exPlanation) to analyze and explain the effects of the predictors on takeover time. We identified seven most critical predictors that resulted in the best prediction performance. Their main effects and interaction effects on takeover time were examined. The results showed that the proposed approach provided both good performance and explainability. Our findings have implications on the design of in-vehicle monitoring and alert systems to facilitate the interaction between the drivers and the automated vehicle.
Summary
Explainable machine-learning study (Ayoub, Du, Yang, Zhou) predicting takeover time in conditionally automated (SAE Level 3) driving by aggregating data from a 129-study meta-analysis. eXtreme Gradient Boosting (XGBoost) is used to model takeover time as a function of predictors identified in the literature (takeover lead time, non-driving task type, TOR modality, scenario urgency, etc.), and SHAP (SHapley Additive exPlanations) values are used to globally rank predictor importance and to explain individual predictions. Seven critical predictors were identified that yielded the best prediction performance; main effects and pairwise interactions on takeover time were examined. The authors argue the resulting model provides both good predictive performance and interpretability, with implications for designing in-vehicle monitoring and alert systems that anticipate when a driver will be ready to resume manual control.
Key finding
An XGBoost+SHAP model trained on a 129-study takeover-time meta-analysis identifies seven critical predictors and produces both accurate and interpretable takeover-time predictions for SAE Level 3.
Methodology
secondary_analysis
Sample size: Exp 1: N=10; Exp 2: N=20
Quality score: 5 / 5