European Commission logo
English English
CORDIS - EU research results

Reducing the impact of environmental disasters: AI-based assistant for fronline emergency management

Periodic Reporting for period 2 - IRIS (Reducing the impact of environmental disasters: AI-based assistant for fronline emergency management)

Reporting period: 2021-12-01 to 2023-06-30

Increasingly complex natural & man-made disasters are overwhelming Emergency Services capacity to respond.

a) Too much data causing saturation in the field: Emergency Services are deploying IT technologies in their operations (drones, advanced GIS, wearables...) to face these threats.
b) Inefficient coordination: emergency response of wildfires & natural disasters requires many teams and emergency services to work together. The lack of a tool to share real-time information makes the coordination complicated, jeopardizing the whole operation.
c) Lack of scalable tools: other threats are raising in parallel, like terrorism or the increasing complexity of overpopulated cities. There are no transversal tools to support Emergency Services in the field.

IRIS is the first Digital Assistant based on Artificial Intelligence for Emergency Services. It’s a web tool that integrates incident related data from different sources, both static (standard operational guidelines, emergency plans, maps, emergency risks and plans, historical data) and dynamic (drones, cameras, GPS, IoT sensors), processes and analyses the data to deliver specific advice to commanders, and enables fluid interaction amongst incident response team members, through a simple-to-use interface.
• Monitoring and control of the project in all its dimensions through weekly and monthly meetings with the leaders of the different areas: technical, economic, business.
• Generation of documentation for EIC stakeholders (due diligence documents, investor decks, project deliverables, monthly updates, etc). Monthly updates since June 2021 have been provided to both EIC Fund representatives and EIC Project Officer (see Annexes).
• Coordination with external stakeholders for development activities (i2CAT), penetration tests during field trials and other required professional services for the development of IRIS product (ISO certification & product multimedia materials).
• Assembly of a dedicated QA (Quality Assessment) team within the technical area.
• Incorporation of a specialised IT support service (Oxon Tech) to provide 24/7 attention to UK clients and pilots with higher activity.
• Successful completion of AI developments from i2CAT achieving the construction of two different prediction models with a fair degree of accuracy.

A short video presentation has been produced as wrap up of the project:

The delivered complete system design includes:
i) Set of scenarios where IRIS must provide value
ii) List of desired features/services
iii) Followed process to collect requirements and establish development priorities
iv) Product version pipeline with expected timelines
v) Solution architecture.

The current version of IRIS 1.0 prototype keeps evolving and adapting new functionalities, such as:
(i) AI feature at the Decision Log, which allows voice records to effectively auto-fill messages in one go
(ii) the redesigning of the Operational Field of View (OFV) to follow a more intuitive navigation and interaction with end-users
(iii) the creation of a Forms Dashboard, designed to gather many different types of documents to be reported at a single and easily approachable place
(iv) the conceptualization of a Casualty Management functionality which logs and gives absolute control on critical information about the casualties in an incident
(v) full makeover to IRIS Tactics mobile app, a real-time connection with the fire ground that keeps alive the communication between personnel in the field and commanders, which is now even more robust and informative than in previous releases.
The aforementioned technical items and many more are thoroughly described in the updated versions of the following documents with its corresponding detailed annexes: “A. IRIS Technical Summary”, “B. IRIS Functionalities Summary” and “C. Unblur IRIS AI Summary”.

In our initial proposal, we had estimated 25 pilot sites for comprehensive field trials. However, as the project progressed, we strategically chose to focus our efforts on a select number of sites, with higher requirements that were most likely to become potential clients for our innovative firefighting solutions.
By the end of the project, we successfully achieved in-depth field trials at 6 pilot sites. These selected sites held significant importance as they included esteemed organisations such as the London Fire Brigade, the largest fire service in the UK, West Midlands Fire Service, and the Fire Service College, a prominent educational authority for firefighting in the country.

Unblur has been very active with dissemination and communication activities during RP Final (see details in Deliverable 5.1).
Through conferences, forums, webinars, and targeted advertising in digital media and magazines, Unblur has successfully amplified brand visibility and captured the attention of key stakeholders.

Our innovation IRIS is a state-of-the-art human-like reasoning system designed to assist first responders in synthesizing high-level data while at the scene of an emergency. IRIS learns, analyses, reasons, predicts, collaborates, and provides data fusion to provide direction for first responders on the scene.
Integrating machine learning with a foundational expertise in physical science, data science, and leading authorities in emergency management and risk management, we use layers of complex data, drawing relationships between numerous data points to map the impact of disasters in a way never before possible.
Our powerful models process large amounts of unstructured data at different levels of resolution and quality, creating actionable insights that help humans make better, more informed decisions. Data sources include real-time incident data, private and public collaborative datasets, extensive built and natural environment datasets.


Incident Commanders must be able to make decisions in a very short period of time and usually under very adverse circumstances. Therefore, it is essential that any IT solution they have to use in the front line has minimal friction to use.
IRIS includes unique interaction features, such as: i) voice-enabled commands to operate IRIS; ii) automated incident guidance to show to the commander what's going on when he arrives to the incident; iii) real-time recommendations for emergency resource allocation based on historical data and advanced machine learning.


Within this period, the Unblur team has gained wide experience not just on how Public Emergency Services operate, but also about how this technology applies in new segments such as Critical Infrastructure, Industrial Safety, Private Security and Safety in Mass Events. The current technology of IRIS serves as a base to develop further modules to solve the specific challenges of these new segments. Some examples are integrating 3rd party simulations on hazardous materials explosions and plumes, integrating computer vision to detect threats (e.g. suspect vehicles, mass dynamics), or integration of alternative geolocation devices such as new generation radios.