车辆动力学特性和路口交通信号影响下的车辆动态路径规划研究

The dynamic path planning as an effective measure to clear the congested traffic flow becomes a new display technology, how to improve its real-time accuracy is the key problem. This project considering the vehicle dynamics and intersection signals, will study the optimal path planning problem in time varying road network with the least expected running time. The research concentrates on three key scientific problems: the vehicle dynamics analysis and modeling, the travel time short-term prediction method and the optimal path planning algorithm. This project begins with the investigation on rules how the vehicle dynamics, intersection traffic signals and driver factors influence on the road travel time, then the vehicle dynamics model is built up under the vehicle route planning conditions, the road trip time se series is reconstructed in phase space, and the chaos theory is used to establish the short-term link travel time prediction model. Based on above researches, considering the intersection traffic signal control law, known or unknown, the penalty function algorithm and improved labels correction algorithm are proposed respectively, and the normalization of this two algorithms is studied. Finally, the real intersection signal model will set up to experimental verification the proposed planning algorithm, a more realistic vehicle path planning algorithm is expected.

车辆动态路径规划作为疏通拥挤交通流的有效措施而成为研究热点，如何提升其实时性、准确性是关键。本项目考虑车辆动力学特性和路口交通信号的影响，基于最少预期运行时间，对具有信号控制的随机时变路网中的车辆最优路径规划问题进行研究。工作重点围绕路径规划条件下车辆动力学特性分析与建模，路段行程时间短时预测方法和最优路径规划算法这三个关键科学问题，研究车辆动力学特性、路口交通信号和驾驶员因素等对路段短时行程时间影响的内在规律，建立车辆路径规划条件下的车辆动力学模型，对路段行程时间序列进行相空间重构，并应用混沌理论建立路段行程时间短时预测模型，在此基础上，针对路口交通信号控制规律是否已知，分别提出改进的惩罚函数算法和标签修正算法，且对两种算法的归一化进行研究。通过真实路口信号模型，对项目提出的算法进行实验验证并优化，以提高车辆路径规划的实时性、准确性，使车辆路径规划更具现实意义。

本项目针对我国城市交通特点，基于最少预期运行时间对具有信号控制的随机时变路网中单车辆最优路径的运行机理进行了研究，并从实际应用出发，首次综合研究了车辆动力学特性、路口交通信号控制、驾驶员等因素对路段行程时间的影响规律，并将这些影响因素模型进行表述，建立了能体现单车辆路径规划条件下的车辆动力学模型。应用混沌理论构建了一种新的路段行程时间短时预测模型，该模型考虑了车辆动力学特性、路口交通信号等影响因素对路段行程时间序列相空间重构质量的影响，提高了预测的准确度。针对路口交通信号规律是否已知，对信号规律已知的随机时变路网，基于最少预期时间路径策略，提出了一种新的改进惩罚函数算法，考虑了路口交通信号所导致的延时对车辆动态路径规划的影响；对于信号规律不确定，提出一种新的标签修正算法，明确地考虑由于信号控制而导致的附加延迟，而且能够在路口信号控制规律已知的情况下，还原为改进惩罚函数算法。