DE-GAM: A dual-encoder graph-attention mixture-of-experts framework for post-crash traffic speed forecasting during freeway all-lane-closure incidents

Contribution

Models post-crash traffic speed during freeway all-lane-closure incidents with a dual-encoder graph-attention mixture-of-experts framework.

The article addresses the operational consequences of crashes through post-crash speed forecasting. Its focus on all-lane-closure incidents connects safety analysis with traffic recovery and incident management.