Study on Optimization of Takeover Time and Safety Distance of L3 Automatic Driving System

Yao, Bohan · 2025 · Crossref

DOI: 10.54254/2755-2721/2025.27954

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study addresses the critical safety challenges in Level 3 (L3) autonomous driving systems, specifically focusing on optimizing takeover time and dynamic safety distances. L3 systems require human intervention during emergencies, yet traditional safety models rely on fixed parameters that fail to account for real-time driver states such as fatigue or distraction. The research aims to bridge this gap by proposing a machine learning-based framework that dynamically adjusts safety distances based on preceding vehicle speed and real-time driver condition monitoring. The methodology integrates multimodal data acquisition from real-world traffic scenarios, utilizing the Kaggle Czech Vehicle Speed Dataset and on-board sensor data from experimental vehicles. Preceding vehicle speed was predicted using a hybrid Gradient Boosting Machine and Long Short-Term Memory (GBM-LSTM) model. Driver states were monitored via an improved CNN-ResNet50 network, which analyzed facial micro-expressions and behavioral cues to detect fatigue, mobile phone use, and smoking. The study constructed a dynamic safety distance model that couples driver reaction time and caution coefficients with vehicle dynamics. Data preprocessing involved outlier removal using DBSCAN clustering and noise reduction via Kalman filtering. The system’s performance was validated through road tests exceeding 500,000 kilometers. Key findings indicate that 68% of near-collision incidents occur within two seconds, highlighting the urgency of rapid response. The dynamic model demonstrated superior accuracy, with calculation errors controlled within 5%, representing a 30% improvement over traditional models. Specific state-dependent metrics revealed that distracted drivers require a safe distance of 14.5 meters and a takeover time of 2.1 seconds, while fatigued drivers require 16.8 meters and 2.8 seconds. The driver state recognition system achieved 95.2% accuracy, with 98.7% sensitivity for detecting eye-closed states. Furthermore, the implementation of a three-level safety strategy reduced critical risk events by 58% and maintained a nighttime false alarm rate below 0.8%. The significance of this work lies in its demonstration that integrating real-time driver state monitoring with predictive vehicle dynamics significantly enhances L3 system safety. By moving beyond static safety margins, the proposed model provides a robust basis for designing warning strategies and takeover protocols. The study concludes that future advancements should incorporate V2X data and multimodal physiological signals to further refine risk assessment and system robustness in complex driving environments.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-nvidia on 2026-08-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• situational awareness
• following distance

Information type

What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).

• Empirical Findings: behavioral performance data
• Theoretical Contribution: computational model, conceptual framework