Protect emergency services: Potentials of AI in hazardous environments
Whether in rescue missions, fire-fighting operations or inspections in the deep sea, mobile robots that use artificial intelligence (AI) to find their way in unknown situations will be able to effectively support people in activities in dangerous environments in the future. The potential and the concrete benefits of AI in this field of application are shown in a current report of the Learning Systems platform based on two application scenarios. The report was presented today in Karlsruhe. The authors - among them the head of the Fraunhofer IOSB, Prof. Jürgen Beyerer - also name technical and social challenges as well as prerequisites that have to be created for a reliable and economic use of AI in hostile environments.
In the future, mobile self-learning robots will be able to relieve humans of dangerous or health-endangering activities. At the same time, they make operations in difficult terrain more economical or even possible in the first place. From a technical point of view, however, the use of such learning systems in hostile environments still poses a number of challenges. These include autonomous learning in unknown environments. In addition, the cooperation of autonomous robots with humans must be designed.
"The use of artificial intelligence is associated with enormous opportunities for our society. Especially in disaster control, the dismantling of nuclear power plants or in maritime areas, there are great opportunities to effectively support skilled workers with the help of artificial intelligence. This is why the Learning Systems Platform has set up an interdisciplinary working group to discuss how learning systems for hostile environments can be developed and used for the benefit of people," says Professor Holger Hanselka, President of the Karlsruhe Institute of Technology and member of the Platform's Steering Committee on Learning Systems. "IT security will be enormously important, especially for autonomous systems that we use in the event of a crisis. Therefore, KIT's research focuses on protecting not only the external borders of a complex IT system but also each individual part, and in particular contributes its expertise in IT security to the Learning Systems Platform".
In its report, the working group on hostile environments shows on the basis of two application scenarios how artificial intelligence can support disaster control and reconnaissance and maintenance missions in about five years. The application scenario "Fast Help in Rescue Operations" illustrates how AI-based robotic systems can support firefighters on the ground and from the air in the event of a fire in a chemical plant. With the help of multi-sensor technology, the systems are able to quickly generate a detailed situation picture, set up a communication and logistics infrastructure for rescue operations, search for injured persons and identify and contain sources of danger. In the application scenario "Under water autonomously on the move", robotic underwater systems maintain the foundations of an offshore wind turbine. They navigate independently in the deep sea, take over the planned planning steps and, if necessary, request support from divers or remote-controlled systems.
Niche market with special requirements
"The demands on learning systems are particularly high in hostile environments: They have to be intelligent and at the same time robust against extreme conditions, and they have to be able to find their way independently under unpredictable conditions," says Jürgen Beyerer, head of the Hostile Environments working group of the Learning Systems Platform as well as head of the Fraunhofer Institute for Optronics, Systems Engineering, and Image Exploitation IOSB and professor for Interactive Real-Time Systems at KIT. "Until then, AI-based systems can be operated remotely by emergency forces and the collected data can be used for the development of intelligent functions. Gradually, the systems will reach a higher and higher degree of autonomy and can finally improve themselves further by machine learning.
Learning systems for use in hostile environments are still a niche market. Germany is well positioned in the development of these AI systems. In its report, the Working Group on Life-Hostile Environments jointly headed by Jürgen Beyerer (KIT and Fraunhofer IOSB) and Frank Kirchner (Robotics Innovation Center, German Research Center for Artificial Intelligence and University of Bremen) identifies concrete design options to exploit the opportunities of learning systems in life-hostile environments and to serve global markets with the self-learning robots. These range from the establishment of suitable infrastructures, such as comprehensive data pools and reference platforms, to the promotion of innovation, for example through competitions or technology demonstrators, to the creation of standards for industry and research and the flexibilization of the procurement market.