CBANet: A Compact Attention-Based CNN-BiLSTM Network for Aggressive Driving Event Detection

Researchers introduced CBANet, an attention-based CNN-BiLSTM model for aggressive driving event detection from vehicle sensor data.

The approach combines engineered vehicle dynamics features with imbalance-handling methods such as SMOTE and focal loss.

On a new naturalistic driving dataset, it outperformed standard deep learning baselines, especially for rare minority-class events.

Improved detection of aggressive driving could support safer, more reliable in-vehicle monitoring in real-world conditions.