| Summary: | 碩士 === 國立嘉義大學 === 資訊管理學系研究所 === 105 === Abstract
In recent years, applications on the wireless sensor networks (WSNs) become more and more popular because of the rapid development of the WSNs. The crowd guidance services for emergency evacuation based on WSNs receive more and more attention. Previous studies on emergency evacuation have got some certain achievements, but most of the studies didn’t consider the possibility of the congestion situation. To solve the congestion problem, the capacity of the corridor, the distances to the exits and the places where emergency events happen are considered in this research.
This paper proposed an emergency evacuation algorithm adaptive to emergency events for an underground parking garage scenario to disperse the crowd fast and safely. In our proposed approach, sensor nodes will calculate the initial dangerous level based on the scenario and re-calculate the dangerous level with the amount of victims. If one of distance value and victims’ amount value is too big, it may dominate the calculation of the dangerous level. To prevent this, the proposed algorithm uses the percentage of distance and percentage of victims’ amount instead of original one. After calculating the dangerous level, sensor nodes will guide victims to the optimal exit and change the dangerous level immediately to prevent the congestion. Besides, the emergency events usually make the specific place blocked. The propose approach will guide victims to a safe place soon and shut down the area immediately to prevent guiding victims to a blocked area. The preliminary results show the proposed method occur less congestions and evacuate faster than Load-balancing Emergency Guiding System (LEGS), Multiple Streaming Crowd Guidance (MSCG), Emergency Support System plus (ESS+), HEX Inc.’s algorithm, and Dijkstra’s algorithm when there is lots of people in the scenario. The proposed algorithm are safer even if there is no congestion in the scenario.
Keywords: Guiding Navigation, Emergency services, Wireless Sensor Networks, Crowd Guidance Services
|
|---|
