基于轨迹大数据的城市交通状态识别研究
Research on Urban Traffic State Identification Based on Trajectory Big Data

对城市交通状态的有效识别与评估有助于提高城市交通精细化管理水平和应急响应能力。以轨迹大数据为基础，提出基于多个交通要素的城市交通状态识别模型。首先建立基于路段平均速度、交通流密度、行驶时间指数、拥堵指数和车道占有率的交通状态识别的综合指标体系，并以交通时空大数据为基础，结合主成分分析方法构建综合指标体系中各个指标的客观权重。进一步根据交通状态的模糊性，建立基于核模糊C均值聚类的交通状态识别模型。最后以兰州市出租车GPS数据为基础，对兰州市的城市交通状态进行识别和划分，分析了兰州市的交通状态的时空演化特征。实验结果表明，兰州市交通状态分为5个等级，城市交通整体上呈现拥堵状态，且拥堵路段空间分布有较大差异，集聚特征比较明显。
The effective identification and evaluation of urban traffic state are helpful to improve the fine management level and emergency response capability of urban traffic. Based on trajectory big data, a traffic state recognition model based on multiple traffic elements is proposed. Firstly, a compre-hensive index system for traffic state identification based on average speed of road sections, traffic flow density, travel time index, congestion index and lane occupancy rate is established. Based on trajectory big data, combined with principal component analysis, the objective weight of each index in the comprehensive index system is constructed. Then, according to the fuzziness of traffic state, a traffic state recognition model based on kernel-based fuzzy C-means clustering is established. Fi-nally, the urban traffic state of Lanzhou city is identified and divided based on the taxi GPS data of Lanzhou city, and the temporal and spatial evolution characteristics of the traffic state of Lanzhou city are analyzed. The experimental results show that the traffic state in Lanzhou is divided into five grades. The urban traffic is congested during all working hours. There are great differences in the spatial distribution of congested road sections, and the aggregative characteristics are obvious.