ABSTRACT
During path planning, rescue operations in buildings are often hindered due to the lack of critical building information. This paper proposes a real-time navigational platform combining various information sources to mitigate this problem. Various operating agents such as first-responders and UGVs receive specialized assistance by integrating semantic information from digital building models with incident-related real-time data. The proposed methodology includes three steps: (1) Pre-configuration of conditions influencing path choices according to agents‘ capabilities and safety preferences. (2) Representation of the building geometry using a navigation graph. Inclusion of semantic information by weighting the graph with incident-related information, compared with the conditions per agent type and semantic information from the digital model. (3) Development of a knowledge database containing conditions for various types of agents. Evaluation in a real-world scenario revealed that paths calculated by the planning module were faster, shorter, and safer than those determined by first responders.
