|Home>||Fire||>Alarms and Detection||>EUSAS European Conference details role of Artificial Intelligence in optimising fire safety|
|Home>||Fire||>Evacuation||>EUSAS European Conference details role of Artificial Intelligence in optimising fire safety|
|Home>||Fire||>Passive Fire Protection||>EUSAS European Conference details role of Artificial Intelligence in optimising fire safety|
EUSAS European Conference details role of Artificial Intelligence in optimising fire safety
15 April 2021
THE EUROPEAN Society for Automatic Alarm Systems (EUSAS) recently organised a successful online European Conference focused on the topic ‘Artificial Intelligence in Fire Detection and Security: Without the Hype’. The event showed once again the importance of technological development for a business sector committed to the protection of life.
The conference opened with a discussion on what Artificial Intelligence (AI) really is. The general concepts as well as the history and starting blocks of AI were detailed. Also, the current application fields for AI in ‘Smart Living’ as well as important requirements for the realisation of intelligent ‘Smart Living’ services were presented.
A focal point for the fire industry was the series of presentations on the benefits and opportunities realised by AI for fire detection.
AI and fire safety
In his presentation, Guillermo Rein (Professor of Fire Science at Imperial College London’s Department of Mechanical Engineering) detailed an innovative fire protection system that combines building sensors, computer modelling and AI. Dubbed ‘The Fire Navigator’, this system aims to forecast the movement of a fire inside a large building, providing Fire and Rescue Services with essential information about flames and smoke ahead of time.
‘The Fire Navigator’ bridges the gap between fire safety and Building Information Modelling by making use of the data already produced by high-rise building sensors such as smoke and heat sensors. A fast and simple cellular automated model assimilates sensor data and, via inverse modelling and genetic algorithm techniques, uncovers the fire’s ignition, location, time, flame spread rate and smoke velocity. A test case involving synthetic data was shown for a real iconic building in London.
‘The Fire Navigator’ concept would be particularly suited to the protection of higher risk buildings like high-rise residential blocks and hospitals, or otherwise key infrastructure such as tunnels and power plants.
Paul van der Zanden (general director of Euralarm) elaborated on the connection between AI and the fire industry, adopting an holistic approach by defining AI as “technology used to add value and/or improve the outcome of an existing or new process/system”. The fire industry has a wide scope and covers many aspects. Within Euralarm, fire safety is seen as an ecosystem and, therefore, should be part of the development process.
Assuming that everything is accounted for in the design to prevent a fire from starting, there’s still a chance that a fire incident will happen. A key factor which defines the impact from this incident is time. The timely detection of, and sensitivity to, unnecessary alarms are related. Both factors can be improved by using new technologies including AI.
Can the sector use other future spin-off developments from the AI world for its own positive progression? The introduction of the new sensor technologies available could be one of these spin-offs. Using an example from AI sensor technology development, Paul van der Zanden showed how the future fire detection can be enhanced and moved to the next level.
In his presentation, computer security specialist Ibrahim Daoudi of CNPP presented the vulnerabilities related to the use of AI on security/safety products. There are mainly three categories of vulnerabilities. The first category consists of adversarial attacks where the aim is to generate data sufficiently modified to mislead the model. The second category concerns physical attacks. This is based on adversarial attacks, but applied to real objects. The third category is all about the traditional attacks on information systems leading to the poisoning of the model itself or its training data. All three were discussed and explained.
Using temporal information is crucial for detecting smoke in video sequences. During his delivery, Andreas Wellhausen of Bosch Sicherheitssysteme GmbH outlined work conducted on temporal approaches based on deep learning as applied to video smoke detection. Two methods were outlined.
First, a combination of convolutional neural networks and long short-term memory networks. The second method centres on inflated 3D architecture which consists of 3D convolutions. These are two state-of-the-art approaches to extract spatial and temporal information from video sequences. A new way in which to detect and localise smoke within such sequences was presented. This is known as ‘cell-wise classification’. Furthermore, the advantages of temporal approaches over convolutional neural networks methods, which are commonly used for detection problems in computer vision, were shown.
Training AI on synthetic data
Philip Dietrich of Bosch Sicherheitssysteme GmbH analysed the idea of using synthetic data to train deep learning systems for video-based smoke detection algorithms. Compared to real data, gathering a large-scale database is significantly easier for synthetic data. It was described how deep learning networks can be trained on synthetic videos. The results were compared with real data.
As a means of bridging the domain gap between real and synthetic data, the concept of domain adaptation will be introduced. By forcing networks to extract similar features from real and synthetic data respectively, potential artefacts in synthetic data may not be learned by the network. Experimental results support the hypothesis that domain adaptation improves the generalisation on real data.
Legislation and outlook
While the rapid adoption of AI creates exciting new opportunities for industry and individuals alike, it also poses an important question: ‘Do current laws apply to AI?’ Tadas Tumėnas of Orgalim discussed if and how this new technology should be regulated. He outlined the state of play in Europe in relation to AI, focusing on its definition which should be the essence of the EU legislative framework. Tumėnas also presented the European Commission’s work related to AI.
In the last presentation of the online event, Lance Rütimann (chair of the Fire Section of Euralarm) said that, if the fire safety industry doesn’t take on the task of working with legislators, regulators and standardisation bodies in defining the aforementioned regulatory landscape, then someone else will.
The use of AI to protect lives and assets makes good sense. Understandably, the path ahead is not ultimately clear and there are many, many questions to be asked and answered on this subject. The fact that the work diligently conducted by members of the fire safety industry makes the world a safer place for millions of people is arguably the best motivation to set the focus on a new horizon.