Car-following Security Situation Estimation Based on Multi-source Information Fusion
Abstract
For intelligent and connected vehicles, analyzing driving scenarios to estimate security situation is the primary premise to ensure vehicles safety,especially in car-following scenario. In this paper, an approach for car-following security situation estimation is proposed based on multi-source information fusion, which uses fuzzy inference system as the estimation framework. By analyzing the driver's diversified driving behaviors in longitudinal direction and dynamic inter-vehicle motion characteristics, the desired acceleration, HWT and TTC-1 are selected to evaluate the security situationunder the current running state. In order to make the security situation evaluation results more accurate and rational, a car-following scenario database is built by PreScan software, and a data-driven analysis method and MOS are employed to determine the membership function of each variable. Finally, through simulation experiment, the RMSE of the approach, proposed in this paper, is 0.0813, compared with MOS, and the security situation evaluation approach is verified.
Keywords
car-following scenario; security situation estimation; multi-source information fusion; data-driven
DOI
10.12783/dteees/iceee2018/27842
10.12783/dteees/iceee2018/27842
Refbacks
- There are currently no refbacks.