Cognitive Fire Alert Framework for Realizing Fire in Unfriendly Environment using Wireless Sensor Network

Nowadays, Wireless Sensor Networks (WSNs) are extensively distributed in environmental monitoring, operational health monitoring, and industrial monitoring. It is an inexpensive application. It contains small, less energy-consuming, lower-cost devices combined with restricted computation, observation, and radio transmission capacities. People nowadays use fire alarms with sensors to detect fires. Fire sensors are unreliable. With the fast growth of WSNs, people can now remotely watch their surroundings to ascertain the current status. A standard fire alarm system is only able to detect one of the environmental parameters at a time. It is possible that there is a fire if the detection value is greater than the threshold; this would indicate that there is a fire. Copper, insulated, or cables are used to make the connections between detectors and alarms in a traditional fire alarm system. Due to the presence of a large number of elements at the site of the fire, the smoke alarm system will be unable to make an accurate determination as to the location of the flames. This article proposes a cognitive fire alert framework (CFAF) for realizing fire in unfriendly environments using WSN. It has a phase for wireless transmission, a phase for cognitive fire recognition, and a phase for the Random Forest Method (RFM). This approach utilizes the RFM to classify the fire by data classification. The results of the experiments show that the suggested technique minimizes the number of false alarms by evaluating the delay value by utilizing data samples and comparing the temperature, humidity, pressure, and dust level.