Nav-star: An adaptive signaling model for pedestrian evacuation based on cellular automata using A-star algorithm

Evacuating during emergencies presents significant challenges, particularly in buildings with complex layouts or when exits, hallways, and stairs are hidden from view. In critical situations, traditional static signage systems often prove inadequate, resulting in dangerous delays that can lead to loss of life. Effectively addressing these risks requires robust evacuation systems that optimize routes to the nearest exits in real time. Over the past decade, research on pedestrian evacuation modeling has focused on optimizing exit routes using dynamic signage systems, in which the evacuation strategy is usually precomputed. These systems generally respond only to local events or depend on the environment’s initial state, leaving them unable to account for real-time developments during evacuations, such as sudden blockages or congestion near exits, which are crucial for dynamically adjusting evacuation guidance. To overcome these limitations, we introduce Nav-star, an adaptive signaling model designed for pedestrian evacuation in complex building layouts. Nav-star employs a two-layer architecture. The first layer analyzes building layouts and monitors congestion. It processes this information to adapt IoT signage to guide pedestrians toward optimal exits using a modified A* algorithm that incorporates distance, pedestrian density, and exit capacity constraints. The second layer simulates pedestrian movement in response to adaptive signage, enabling individuals to reach less-congested evacuation routes. Simulation results show Nav-star outperforms traditional static signage systems by nearly 50% under critical evacuation scenarios. The proposed model consistently achieves lower coefficients of variation (CV), demonstrating better load balancing and more even use of exits during high-density evacuations.