Decision Tree-Based Maneuver Prediction for Driver Rear-End Risk-Avoidance Behaviors in Cut-In Scenarios

Manjiang Hu,Wenjun Wang,Yuan Liao,Fang Chen,Guofa Li,Bo Cheng

doi:10.1155/2017/7170358

Abstract

Predicting driver rear-end risk-avoidance maneuvers in cut-in scenarios, especially dangerous precrash scenarios, benefits the customization of automatic driving, particularly automatic steering. This paper studies driver rear-end risk-avoidance behaviors in cut-in scenarios on a straight three-lane highway. Data from 24 participants in 1326 valid trials were collected using a motion-based driving simulator. An Eysenck Personality Questionnaire (revised for Chinese participants) was used to obtain the personality traits of the participants. Based on a statistical analysis, the candidate features used in the driver maneuver prediction were determined as a combination of objective risk indicators and driver characteristics. A decision tree-based model was constructed for maneuver prediction in cut-in scenarios. The prediction accuracy of the extracted classification rules was 79.2% for the training data set and 80.3% for the test data set. The most powerful predictive variables were extracted, and their effects on maneuver decisions were analyzed. The results show that driver characteristics strongly influence the prediction of maneuver decisions.

Highlights

Autonomous driving systems have been developed from various discrete automated functions (e.g., Adaptive Cruise Control, ACC)
How does the driver maneuver when another vehicle cuts in front? Do they brake or steer? What are the perceptions and considerations that motivate their decisions? The results of this study provide a reference for decision-making of highly automated driving (HAD) that will contribute to a better user experience
If D F ≤ 17 m, the time to collision between the host vehicle and the lead vehicle in the left lane (LL) (TTC F), A, and N score are involved in the driver maneuver prediction

Summary

Introduction

Autonomous driving systems have been developed from various discrete automated functions (e.g., Adaptive Cruise Control, ACC). ACC combined with braking and steering interventions has been realized by several automakers, such as Daimler and Volvo [2, 3] Such systems assist drivers during emergency braking and/or steering maneuvers based on driver input and surrounding sensor information [4]. Recent studies of automated steering in the context of HAD have focused on hazard assessment for decision-making [8] and dynamic control for collision avoidance [9,10,11]. It has been widely understood by researchers and engineers in intelligent driving research field that stages of autonomous driving are divided into 5 levels. HAD systems refer to the level of limited selfdriving automation, where drivers are required to monitor and occasionally to take over

Methods

Results

Discussion

Conclusion