A NOVEL EDGE COMPUTING SYSTEM FOR EARLY FIRE DETECTION IN FOREST

Abstract

Forest fires pose a significant threat to ecosystems, wildlife, and human lives. In this work, we propose an integrated system that addresses multiple aspects of forest fire management. Leveraging historical data and weather conditions, our system predicts the likelihood of forest fires in specific regions. By identifying high-risk areas, we can take proactive measures to prevent fire outbreaks. We deploy static sensor nodes strategically across the forest. These nodes continuously monitor environmental parameters such as temperature, humidity, and wind speed. Based on real-time sensor data, our embedded system classifies forest zones into different categories based on fire intensity. These categories include low-risk, moderate-risk, and high-risk zones. The classification enables targeted resource allocation and efficient firefighting efforts. Our system aims to minimize human intervention in fire-prone areas. By automating early detection and response, we reduce the need for manual patrols. Alerts are sent directly to firefighting teams, allowing them to take immediate action without risking their safety. Forest boundaries are critical for preventing fire spread. However, wildlife often crosses these boundaries, increasing the risk of fire transmission. Our system integrates animal movement data to predict when and where animals are likely to cross forest boundaries. This information informs fire management strategies. Overall, our proposed system combines predictive analytics, sensor technology, and automation to enhance forest fire prevention, response, and ecological conservation.

Specification

Description:In the novel edge computing algorithm (100), six sensor's data are collected and collectively
decide the current status of forest grids. As a single sensor would not be able to predict the
possibility of fire, multiple sensors are used in SNs. Air flow sensor (101), Rain sensor (104),
CO2 sensor (103), flame/smoke sensor (102), temperature sensor (105), wind sensor and
humidity sensor (106) are used in SNs, in which the digital output of each sensor is weighted
by using a definite scaling factor. The weighing factor is assigned according to the precedence
and significance of the sensors contributing to the fire detection based on experimental
analysis. A linear combination of sensor outputs (107) is then compared with the preset
threshold values and accordingly the forest grids are classified into: - safe grid, caution grid,
hot grid and fire grid (108). , Claims:1. A framework of edge computing algorithm in static sensor node (100) comprising
of Air flow rate sensor (101), Smoke Sensor (102), CO2 gas sensor (103), Rain
Sensor (104), Temperature Sensor (105), Humidity sensor (106), Weighted
summation of all sensor inputs (107) and Critical assessment of forest grids and its
classification in to different zones (107)
2. A framework of the static sensor node in the forest grids (200) comprising of
hexagonal grid (201), Acoustic sensing unit with microphones (202), Embedded
microcontroller for processing (203), Ultrasonic proximity sensor (204), Gateway
Unit (205)
3. Monitoring unit (300), which is responsible for taking decisions based on the
algorithm comprising of server unit (301) running with Embedded ML algorithm,
Centralized gateway (302), relay drivers (303), loud speaker (304), Desktop/PC for
User interface (305) , User equipment/Mobile (306) for accessing the web
applications , Siren (307) and Unmanned Aerial Vehicle (308).

Documents