[.c-pullout-quote-little-wrapper][.c-pullout-quote]“There’s nothing like this, certainly in Australia." NSW Rural Fire Service Assistant Commissioner Ben Millington - October 2023[.c-pullout-quote][.c-pullout-quote-little-wrapper]
[.c-link-quote-wrapper][.c-link-quote]NSW fires: RFS AI tool Athena to help save homes, lives (smh.com.au)[.c-link-quote][.c-link-quote-wrapper]
[.c-richtext-text-wrapper][.c-richtext-text]In response to this pressing need, RFS took a significant step forward by collaborating with Kablamo to deploy a state-of-the-art fire management solution known as Athena. This custom implementation, built on top of Kablamo's existing bushfire intelligence platform, Firestory, set out to fill the critical gaps and windows of risk that the RFS previously faced.[.c-richtext-text][.c-richtext-text] Designed as a cloud-native platform, Athena was built to not only harness extensive bushfire data, but also convert it into meaningful intelligence that guides fire management efforts. It does this by adapting itself to the unique workflows, systems, decision-makers, communication platforms and preferences of each fire-fighting agency. This advanced level of customisation, tailor-made to meet the very specific needs of the RFS, is a direct result of Kablamo's commitment to working closely with their team and clients to garner the most optimal technological outcome.[.c-richtext-text][.c-richtext-text]Athena is built with a serverless architecture, leveraging AWS Lambda, RDS Aurora, S3, Fargate, Step Functions, and Event Bridge. Fire prediction engines in use are provided by CSIRO (Spark) and Fire Prediction Services (Phoenix RapidFire). All of the intelligence gathering from the ground and air is provided to Athena through ESRI GIS services. AWS Rekognition is used to perform real-time image recognition from social media.[.c-richtext-text][.c-richtext-text-wrapper]
[.c-richtext-title-80][.c-richtext-title]Real-Time Data Fuels Real-Time Prevention[.c-richtext-title][.c-richtext-title-80]
[.c-richtext-text-wrapper][.c-richtext-text]The results of this pioneering collaboration between RFS and Kablamo have been significant for Australia's firefighting efforts. The newly introduced architecture, interface and machine learning capabilities have redefined how incident data is captured and, more importantly, how it is learned from. This has given RFS a renewed sense of confidence, enabling them to tap into non-traditional data sources, including social media, to obtain verifiable intelligence.[.c-richtext-text][.c-richtext-text] One of the most noteworthy features of Athena is its ability to deliver timely notifications and alerts to RFS incident management teams. Critical objects such as education facilities, health facilities, essential infrastructure and human settlements are now closely monitored, ensuring that they receive timely warnings when they are predicted to be impacted by fire spread. This consolidation of information into a single, user-friendly interface marks a significant departure from the laborious, manual workflows that previously required the observation of various maps and disparate data sets across multiple tools and platforms.[.c-richtext-text][.c-richtext-text-wrapper]
[.c-richtext-quote-wrapper][.c-richtext-quote]Athena “takes a fire and can instantly look at what that fire could do within the next two hours and how many properties might be affected.”NSW Rural Fire Service Commissioner Rob Rogers[.c-richtext-quote][.c-richtext-quote-wrapper]
[.c-richtext-text]Athena’s intelligence system consists of modules, each of which has brought an array of powerful functionalities to the RFS: [.c-richtext-text]
- Multi-Model Ensemble Fire Predictions: This module includes the world’s first operational side-by-side cloud implementation of Phoenix RapidFire and CSIRO Spark to perform automated fire ignition ensemble predictions and established fire ensemble predictions. Included in the module is a Jobs Board to enable the human assessment and validation of predictions, as well as the ability to upload manual predictions created by human analysts. These features significantly enhance the RFS's ability to predict and manage fire incidents.
- Risk Visualisation: This module offers a lightweight 3D tiled spatial map, multi-incident predictions, incident prioritisation, potential impact alerts, incident and activity logs and an impressive array of over 30 spatial layers. These tools enable incident management teams to visualise risk amongst various terrains so firefighters can prioritise responses effectively. Each activity log also produces rich data sets to train future AI models and improve their capabilities in the fire space.
- Social Intel: By incorporating geotagged social media into incident maps and providing incident social streams that send automatic alerts to the NSW RFS regarding bushfires mentioned on social media instead of being reported via emergency hotlines, this module taps into the power of social media to enhance situational awareness and response capabilities.
- Dynamic Coverage: This module keeps track of and alerts users when there are shortages of fire trucks or aircraft in each district or brigade area across the state. By considering available, tasked, dispatched, and offline resources, the system offers a dynamic state map display and sends push notifications when significant coverage gaps are identified.
- Aviation Safety: After the tragic crash of Bomber 134 during the 2019/20 Fire Season, resulting in the loss of three aircrew members when their C-130 lost control while laying retardant lines in the Snowy Mountains, the RFS made a commitment to develop an Electronic Flight Risk Assessment Tool. This tool considers current and forecasted weather and environmental conditions on the fireground and assesses them against the capabilities of commonly used aircraft in firefighting missions. The tool provides a red, amber, green assessment of whether it’s safe to deploy aircraft to the fireground. It also tracks aircraft tasking across the state and offers automated updates to decision-makers if conditions change.
[.c-richtext-title-80][.c-richtext-title]Paving the way for a brighter, safer future[.c-richtext-title][.c-richtext-title-80]
[.c-richtext-text-wrapper][.c-richtext-text]The collaboration between RFS and Kablamo to expand Firestory’s capabilities, ultimately leading to the creation of Athena, has enabled a new era of firefighting in Australia. By harnessing the power of advanced technology, plus data and machine learning, firefighters are better equipped than ever to protect lives and properties in the face of bushfires.[.c-richtext-text][.c-richtext-text][.c-richtext-text]Since Athena went live, it has tracked over 15,000 fire incidents. The platform also boasts a 100% uptime despite the Kablamo team performing continuous seamless deployments to the production environment.[.c-richtext-text][.c-richtext-text]Kablamo’s innovative approach to intelligence and data management is not only streamlining operations for the NSW RFS but is also poised to set new standards in bushfire and wildfire management across the world. The 'Black Summer' was a harsh teacher, and the lessons learned have paved the way for a brighter, safer future — one where advanced technology is a vital ally in the fight against bushfires.[.c-richtext-text][.c-richtext-text-wrapper]
