一种基于序贯Monte Carlo算法的交通事件重构方法

A Sequential Monte Carlo Method-based Event Reconstruction Method for Traffic Event Simulation

为了能准确地重构出当前道路场景中的交通流事件,需要合适的模型与方法以及能够代表交通流状态的实时数据。基于交通流非线性非高斯的特点,提出了一种基于序贯Monte Carlo方法的交通流堵塞事件重构模型。提出的模型能够不断的同化真实道路上实时的传感器数据使仿真中的交通流状态与真实路况不断接近。通过分析仿真数据推测出当前真实道路上的堵塞事件及其相关属性,并据此在仿真环境中模拟堵塞,进而实现对真实道路上堵塞事件的重构。理论研究和实验结果表明该模型能够根据重构结果评估当前的道路状况,合理推测引起拥堵的位置与堵塞范围。

In order to reconstruct the traffic incidents on the road accurately, appropriate models and methods as well as real-time data that can represent the traffic flow states are necessary. According to the nonlinear and non-Gaussian character istics of the traffic flow, we proposed a SMC based traffic flow congestion event reconstruction framework. The simulation ＇s states can get close to the real scene continuously along with the data assimilation model to assimilate the real-time sensor data constantly. The congestion event in real scene can be estimated based on the simulation data. Thus, we can simulate the congestion in different particles and finally reconstruct the congestion event. Theoretical research and experimental results demonstrate the framework can evaluate the current roads＇ states based on the reconstruction results, the range and the start position of the congestion can be determined as well.