Document


Title

Adaptive Neuro Fuzzy Inference System (ANFIS) based wildfire risk assessment
Document Type: Journal Article
Author(s): Harkiran Kaur ; Sandeep K. Sood
Publication Year: 2019

Cataloging Information

Keyword(s):
  • ANFIS - adaptive neuro-fuzzy inference system
  • FFVI - Forest Fire Vulnerability Index
  • fire detection
  • fire spread
  • fog computing
  • IoT - Internet of Things
  • K-means clustering
Record Maintained By:
Record Last Modified: July 3, 2019
FRAMES Record Number: 58188

Description

Wildfires are extremely destructive disasters that cause significant loss of lives, forest cover and wildlife. This is due to their uncontrolled, erratic, rapid spread and behaviour. The incidence of wildfires is expected to increase worldwide because of Global Warming. Henceforth, it becomes increasingly important to detect and tackle such fires in their infancy to minimise their adverse effects. IoT technology has shown an exponential growth in recent years. Moreover, deployment of IoT devices to monitor and collect time-critical data is pressing need of hour. This research proposes an effective Fog-IoT centric framework for timely detection of wildfires. The proposed methodology provides an efficient real-time solution to dilute the destruction caused by wildfires. Initially, K-means Clustering is used to detect the wildfire outbreak at fog layer followed by real-time alert generation to the administration and community. Furthermore, cloud layer based Adaptive Neuro Fuzzy Inference System is used for assessing the vulnerability of a forest block to forest fires as well as classifying it into one of the five risk zones based on Forest Fire Vulnerability Index. Implementation results of the proposed framework prove its efficiency in detecting and predicting wildfires. In addition, real-time alert generation further enhances the efficacy of the proposed system.

Online Link(s):
Citation:
Kaur, Harkiran; Sood, Sandeep K. 2019. Adaptive Neuro Fuzzy Inference System (ANFIS) based wildfire risk assessment. Journal of Experimental & Theoretical Artificial Intelligence 31(4):599-619.