Self-Learning Systems in Hostile-to-Life Environments (Report)
Whether in rescue operations, maintenance work in the deep sea or the exploration of terrain that is difficult to access: AI-based mobile robots or assistance systems can effectively support specialists in dangerous environments in the future. They reduce the risk for the personnel deployed, shorten the reaction time for time-critical events and take over tasks in which humans cannot react appropriately. Learning systems of this kind thus promise high social benefits and hold great potential for industry and research. From a technical point of view, however, there are still several challenges to be overcome for their use in hostile environments.
In this executive summary, the working group Hostile-to-Life Environment of Plattform Lernende Systeme outlines the benefits of AI-based systems in areas that are difficult to access, using concrete application scenarios. It analyses specific requirements and challenges for their use and shows design options for further research, the development of infrastructures, the promotion of innovations as well as necessary framework conditions.