The development of AI-based solutions enables cooperative multimodal interaction strategies between several UAVs (fixed wing and multi-rotor) and UGVs and thus an intelligent interaction of the individual systems for an optimised assistance service for emergency forces oriented towards special mission scenarios. The UAV/UGV systems used for this purpose have different performance parameters as well as different sensor technology, geared based on the requirements of the individual scenarios, and enable near-real-time on-board situation analysis. A cooperative task management system manages tasks and goals, prioritises and generates new tasks based on changes in the situation and delegates the autonomous processing of tasks to the UAV/UGV systems. The result of KI-SecAssist is a modular proof-of-concept functional demonstrator, including tests, exercises and evaluation related to the functional, performance and practical suitability for an optimised assistance performance for emergency forces in crisis and disaster scenarios.
Background and Motivation
The increasing number of crisis and disaster situations, as well as dynamically changing security situations, pose enormous challenges to emergency forces, the armed forces and security organisations, and require large human resources as well as complex operational strategies. Essential challenges for efficient crisis and disaster management are the real-time availability of a comprehensive common operational picture (COP) and detailed information on individual objects in order to support time-critical decision-making processes as well as to ensure targeted operational support for the teams on site. Due to great progress in the field of artificial intelligence (AI), sensor development and aerial and ground robots (UAV, UGV), there is a need for targeted R&D to realise a new quality of assistance services to support crisis and disaster management on the basis of cooperative multi-modal solutions.
Objectives and innovation aspects
The aim of KI-SecAssist is to develop autonomous assistance modules based on cooperative task management and various UAV and UGV systems with integrated multi-sensor payload. The development of AI-based solutions enables cooperative multi-modal interaction strategies between several UAVs (fixed wing or multi-rotor) as well as UGVs and thus an intelligent interaction of the individual systems for an optimised assistance service custom tailored for special mission scenarios. The UAV/UGV systems used for this purpose have different performance parameters and different sensors, geared based on the requirements of the individual scenarios. A key objective in KI-SecAssist is the development of a cooperative task management system that manages and prioritises tasks and goals, generates new tasks based on changes in the situation (global goal manager) and delegates autonomous task processing to UAV/UGV systems (activity dispatcher). Other key R&D objectives are AI-based multi-sensor data processing and analysis in order to realise near-real-time object detection (persons, hot spots, etc.), as well as the implementation of multi-sensor data fusion and situational awareness for the dynamic COP. Real-time analysis results, both from UAV and UGV platforms, flow into a common world model and are the basis for dynamic cooperative task management and for decision-making processes of the responsible task forces. Due to the clearly defined operational strategies, a human-in-the-loop process is an integral part of the cooperative task management and is secured by the elaboration of the legal and ethical aspects in cooperation between development partners and supporting organisations.
The result of KI-SecAssist is a modular proof-of-concept functional demonstrator, including tests, exercises and evaluation related to the functional, performance and practical suitability for an optimised assistance service for emergency forces in crisis and disaster scenarios. Another goal in KI-SecAssist is strong international professional networking through the integration of an international expert board and the implementation of international workshops.
AIT Austrian Institute of Technology AIT
Bundesministerium für Landesverteidigung (enduser) BMLV
Berufsfeuerwehr Graz BFWG
Disaster Competence Network Austria (GSK partner) DCNA
Freiwillige Feuerwehr Gumpoldskirchen FGK
IFR - Ing. Richard Feischl IFR
Technische Universität Graz - Institut für Softwaretechnologie IST
MATTRO GmbH MAT
twins GmbH TWI
DI Alexander Almer
JOANNEUM RESEARCH – Institut für Informations- und Kommunikationstechnologien
Steyrergasse 17, 8010 Graz, AUSTRIA
Tel.: +043 316 876 1738
Mobile: +043 699 1876 1738
Fax: +043 316 876 1720
E-Fax: +043 316 876 9 1738