Identification of driving behavior in continuous diverging sections of expressway system interchange based on CNN-BiLSTM
View abstract on PubMed
Summary
This summary is machine-generated.Researchers developed a driving behavior spectrum (DBS) to analyze complex expressway interchanges. This system accurately identifies driving behaviors and their transitions, offering insights for safer road design.
Area Of Science
- Traffic Engineering
- Artificial Intelligence in Transportation
- Human Factors in Driving
Background
- Expressway interchanges present complex driving environments with significant safety risks.
- Understanding driving behavior and transitions is crucial for mitigating these risks.
Purpose Of The Study
- To investigate driving behavior and transition patterns in continuous diverging expressway sections.
- To develop a novel method for analyzing and classifying driving behaviors in complex traffic scenarios.
Main Methods
- A simulated driving experiment was conducted to collect driving behavior data.
- A driving behavior spectrum (DBS) was constructed for continuous diverging sections.
- A CNN-BiLSTM model was used to identify specific driving behaviors (straight driving, lane changing, deceleration, turning).
- A Hidden Markov Model (HMM) quantified transitions between driving states.
Main Results
- The DBS effectively captures temporal changes in driving behavior.
- The CNN-BiLSTM model achieved 98% average accuracy in identifying four typical driving behaviors.
- Lane changes were observed to occur at specific distances before diverging points (121m and 78m).
- The HMM successfully elucidated driving state transition patterns.
Conclusions
- The developed models provide a robust framework for analyzing driving behavior in complex expressway environments.
- Findings offer valuable data for identifying hazardous zones and optimizing interchange design for improved safety.
- This research contributes to the advancement of intelligent transportation systems and road safety engineering.