Project Description
FireCloud is an automated, data-driven, and real-time bushfire monitoring and early warning broadcast system. Our idea is to supplement existing satellite-based fire detection systems with ground-based, IoT remote sensing devices. During bushfire season, it allows the Fire Service and concerned citizens to detect fires and pinpoint its exact location within minutes. It can also broadcast emergency SMS alerts to cellular phones within a certain risk radius of the fires when a certain danger threshold is determined. Off season, it can be used to assess the ecological condition of an area, allowing state and local authorities to concentrate reforestation efforts on the most vulnerable sites.
On the hardware side, FireCloud is a system of LPWAN IoT devices composed of sensor node pairs. The first node is a tower or tree mounted sensor that can collect temperature, humidity, stem water content, and wind data, as well as take images. The second node is a ground sensor that can measure soil moisture. A node pair is connected to each other using Bluetooth Low Energy. Together, the pairs collect data used to assess bushfire risk.
These devices will be deployed in bushfire prone areas such as those identified by NSW’s BushFire Prone Land dataset or Victoria’s Designated Bushfire Prone Area dataset. Being LPWAN IoT, the devices will have longer battery life and thus ideal for rural deployments. They will use Telstra’s cellular IoT technologies, such as the NB-IoT or the Cat-M1, having a SIM which can connect them to the 3G/4G/5G network without having to set up cables.
On the server side, FireCloud has three main components:
FirePath is a predictive modeling tool that uses real-time data gathered by the sensors combined with historical terrain data (such as land classification and elevation) and weather data (such as date of last rainfall) to simulate fire movement using existing rate-of-spread models (such as Spark and Phoenix of NSW or Aurora-Astralis of WA). The results are displayed in a dashboard managed by each region’s Fire Service (show: RFS logo). This data will also be accessible to the public through a mobile app, displaying location-specific risk levels for each user based on the region’s standardized bushfire alert system.
Once a fire is detected, the second component, FireAlarm uses TelstraDev’s Messaging API to automatically broadcast emergency SMS alerts to cellular phones within a certain risk radius of the fires. This empowers citizens and volunteers to make informed decisions about timely evacuation and selection of safe evacuation routes.
The third component, FireSight utilizes the sensors’ cameras to collect images for rapid damage assessment and active recovery monitoring. Using a regression model based on CSIRO’s data on habitat condition scores, the system can assess the ecological condition of an area, allowing state and local authorities to concentrate reforestation efforts on the most vulnerable sites.
Physical Components of a Sensor Pair:
Top Sensor
Sensors: Temperature, Air Humidity, Stem Water Content, Wind (Speed & Direction), Smoke
Camera
Chip
Power (Battery)
Cellular network card
BLE (connect to Ground Sensor)
Ground Sensor
BLE (connect to sensor Top Sensor)
Chip
Power (Battery)
Soil Moisture Sensor
References:
Rogers, R 2020, ‘Presentation during the NSW Frontiers in Sensing Forum’, https://www.nssn.org.au/news/2020/5/11/stream-the-nssn-frontiers-in-sensing-forum-2020.
Matthews, S, Sullivan, A, Gould, J, Hurley, R, Ellis, P & Larmour, J 2012, ‘Field evaluation of two image-based wildland fire detection systems’, Fire Safety Journal, vol. 47, pp. 54–61, viewed 23 April 2020, https://www.bushfirecrc.com/sites/default/files/news/fire-detection-public-report.pdf