Safest-path planning approach for indoor fire evacuation

Many evacuation approaches have been proposed to assist people in escaping from fires. Traditional approaches focus on finding the nearest exit and planning the shortest path. However, they involve disadvantages as they are inflexible and time-consuming to adapt to the dynamic spread of fires. Advanced approaches exploit real-time data provided by a sensor network to dynamically determine evacuation exits and paths. Their disadvantages include local optimal solutions and excessive dependence on the sensor network, which may be destroyed in fires. In contrast to previous approaches, in this study, we adopt fire prediction to determine the safest path. In addition, a path planning algorithm is proposed to utilize the prediction data in parallel. In contrast to previous approaches, in this study, we adopt fire prediction to determine the safest path. In addition, a path planning algorithm is proposed to utilize the prediction data in parallel. In the simulations, we adopted the fire trends simulated by the fire dynamic simulation tool, PyroSim. Common safety procedures are flexibly applied to our approach by transforming and modeling. The effects of a crowded area are disregarded. The survival rate and computation time are considered critical performance measurements. The experimental results indicate that our approach is 10 times faster than other approaches and exhibits a survival rate more than 10% higher than prior methods.