Soft computation technique based fire evacuation system

Occurrence of fire is an unpredictable activity needing very high attention. Early response for fire may lead to less damage of life and property. In this paper a work is proposed for evacuating the people to a safe point once fire accident occurs. The objective of the proposed work is to process the data acquired from smoke and temperature sensor to indicate the direction for movement of people. Computational technique is used to collect data from sensors and display safest path for people to evacuate from the building. Artificial Neural Network (ANN) is trained using particle swarm optimization to produce evacuation information. To test working of proposed technique it is subjected to several test cases, results obtained proves successful implementation of proposed work.