Provided by: International Journals of N&N Global Technology(IJNNGT)
In this paper, the authors present and evaluate multi-sensor hardware architecture for monitoring forest fires using Wireless Sensors Network (WSN). In the proposed architecture, each node is equipped with various devices, including, e.g., temperature and carbon monoxide sensors in order to improve accuracy detection. The proposed solution is especially designed to enable low power and higher precision in WSN. The interest of their solution is twofold. First, neural network is used as a powerful and accurate approach which can effectively be applied to fire detection. Second, this approach maintains a high accuracy level despite fluctuations in the sensor values compared to the use of crisp values in traditional algorithms. As a result, this scheme reduces the number of false detection, while still providing accurate event detection.