Prediction based indoor fire escaping routing with wireless sensor network

Fire hazard causes lots of economic loss and personal injuries every year. Many ways are proposed to help people escape quickly from dangerous region. As one key step for fire escaping, the fire escaping system detects fire and dynamically provides escaping route to help people escape from fire scene. With the advanced technique, Wireless Sensor Network (WSN), the fire escaping system is developed to be more promising for fire escaping than before. Most existing fire escaping systems ignore or simplify the dynamics of fire hazard. Thus people’s safety is not guaranteed with fire spreading and growing. This paper designs a new fire spread model based on confidential data created by the powerful simulation system: Fire Dynamics Simulator (FDS). Based on the model, this paper predicts the Available Egress Duration (AED) of all locations in the building. Considering both the length and AED of each escaping route, this paper designs a faSt firE Escaping algorithm (SEE). To evaluate the performance of our approach, this paper conducts experiments on a real WSN platform with TelosB nodes. Experiment results confirm that the fire spread model in this paper can achieve high prediction accuracy. SEE outperforms the existing prediction based approaches by utilizing more AED, so that people can escape with higher probability.