关键词: 农村公路, 交叉口, 双变量Probit模型, 事故形态, 事故严重程度

Abstract: In order to reduce crash risks at intersections on rural highways, with environment, highway, facilities, vehicle types and illegal behaviors extracted as independent variables, pattern and severity of crashes taken as dependent variables, a bivariate Probit model was established. Then, intrinsic correlation of these two variables was analyzed and significant influencing factors were identified. The results show that there is a negative correlation between crash pattern and severity, and death rate of single-vehicle accidents is higher. The better road alignment is and the more well-equipped isolation and control facilities are, the lower crash severity is, while the bigger vehicles are, the more serious crash becomes. Besides, traffic offence increases probability of fatal single-vehicle accidents. Safety of intersections on rural highway can by improved by measures such as installing safety facilities, implementing speed-limit management at intersections of highways with significant grade difference, strengthening security control of oversize vehicles and motorcycles, and increasing penalties for driving without license.

Key words: rural highways, intersections, bivariate Probit modal, accident pattern, accident severity