An automated direction setting algorithm for a smart exit sign

Abstract The aim of this study is to develop an automated direction setting algorithm (ADSA) for a smart exit sign system. The challenges of the ADSA are to translate various space types into nodes on a graph, and to provide directions to the shortest safe egress to any evacuees at any point in a building using publicly available exit signs. We have developed a classification for space types and an ADSA that can exclude unsafe paths and calculate the shortest safe path from unknown points in a building to the closest exit points. We do this by reversing and simplifying the start and end nodes of the typical Dijkstra's algorithm by adding a virtual node. The validity of the ADSA was tested through scenario-based simulations of test cases with various conditions and performance comparison with existing algorithms.