To assist in optimizing the accuracy of bus warning system identification to prevent the occurrence of dangerous driving behaviors in urban buses, we extract risk evaluation indexes from multiple sources and propose a vehicle state risk evaluation model. Firstly, the multi-source data such as warnings of driver and vehicle state, vehicle driving characteristics, bus road network and weather information are extracted from the network platform, advanced assistance driving system and driver state monitoring system. Then, we explore the spatial and temporal risk distribution characteristics of vehicle operation state under different weather from time and space based on probability distribution theory and geographic information processing technology, and further investigate the correlation between each warning. Finally, the risk evaluation model of vehicle operation state is established based on the extracted risk evaluation indicators. The findings show that the proposed approach can effectively support the correction of erroneous warnings of the system. This study would provide theoretical and algorithmic references for the optimization of bus warning systems and bus operation management.

CICTP