oordinated signalization has been applied on arterials since 1960s. However, long-term effects of signal coordination on traffic safety have not been fully discussed, especially in terms of the temporal and spatial variations among crashes. With Bayesian Poisson lognormal model, the paper compares three types of spatiotemporal correlations on the coordinated arterials, including the fixed spatial and linear temporal variation, the fixed spatial and linear-quadratic temporal variation, and the time-varying spatial interaction. With 9-year crash data from Michigan, the study demonstrates that (1) different models should be separately applied for the coordinated arterials, e.g. BPLM with the time-varying spatial interaction fits roadway segments, and BPLM with the fixed spatial and linear temporal variation excels at signalized intersections and (2) the temporal and spatial variations are correlated on roadway segments. The findings emphasize the importance of conducting the safety analysis regularly and dynamically monitoring driving behaviors along the roadway segments.

自1960年代以来，协调信号就已应用于动脉。但是，信号协调对交通安全的长期影响尚未得到充分讨论，尤其是在碰撞之间的时间和空间变化方面。利用贝叶斯泊松对数正态模型，在协调的动脉上比较了三种时空相关性，包括固定的空间和线性时间变化，固定的空间和线性二次时间变化以及时变的空间相互作用。根据密歇根州9年的碰撞数据，该研究表明（1）应当对协调的动脉分别应用不同的模型，例如BPLM，时变空间相互作用适合道路段，具有固定的时空线性变化的BPLM在信号交叉口处表现出色，并且（2）在路段上时空变化是相关的。这些发现强调了定期进行安全分析并动态监控沿路段行驶行为的重要性。