Authors:  A. N. Terekhova∗ and A. M. Gershteyna

JSA-Vol. 4 (2025),

1 St. Petersburg University, Saint Petersburg, 199034 Russia.

* Correspondence: a.terekhov@spbu.ru

Received: 1 March 2025; Accepted: 5 April 2025; Published: 9 April 2025.

Abstract: Road traffic accidents represent a persistent challenge for urban mobility systems, resulting in significant human and economic losses worldwide. Recent safety-aware routing studies have demonstrated that avoiding statistically dangerous road segments can substantially reduce accident exposure, albeit at the cost of moderately increased travel distances. However, most existing approaches rely on static accident risk estimation derived from historical data, thereby neglecting the temporal variability inherent in traffic safety patterns. This paper proposes a dynamic and time-aware risk-based routing framework that extends penalty-based safety routing by incorporating temporal accident profiles and adaptive edge penalties. Road segments are assigned time-dependent risk scores computed from historical accident distributions across multiple temporal windows. These scores are integrated into a modified shortest-path algorithm to enable context-aware route selection. Extensive experiments conducted on an urban road network demonstrate that the proposed approach achieves additional reductions in relative accident risk beyond static safety-aware routing while maintaining acceptable increases in route length and topological complexity. The proposed framework contributes toward the development of intelligent, safety-oriented navigation systems suitable for real-world deployment in smart cities.

Keywords: Road safety, dynamic routing, time-aware navigation, intelligent transportation systems, graph algorithms