刘源

摘  要：提出了一种基于V2V通信的车辆前方道路实时路况的预测方法，该方法可在不增加通信负载的前提下，快速预测道路拥堵情况。首先，根据V2V通信的特点，建立广播信息的数据类型列表，并按规则清洗数据;然后，提出采用瞬时速度和5min内平均速度作为评价道路拥堵情况的指标，并给出等级划分标准;接着，利用车头方向和GPS坐标变化率筛选出车辆即将通行道路上车辆的广播信息作为分析数据;最后，利用模糊C值聚类算法（FCM）对样本聚类，将所有聚类中心坐标进行加权，得到该路段车的瞬时速度和平均速度值。通过实例分析，该方法得到的道路预测结果与实际情况相符，是一种有效的实时路况计算方法。

关键词：V2V通信;实时路况预测;综合加权;FCM

中图分类号：F252    文献标识码：A

Abstract： A real-time road condition prediction method based on V2V communication is proposed， which can quickly predict the road congestion without increasing the communication load. Firstly， the data type list of broadcast information is established according to the characteristics of V2V communication， and the data is cleaned in accordance with some rules. Then， the instantaneous speed and the average speed within 5 minutes are used as the indicators to evaluate the road congestion， and the grading standards are given. After that， the broadcast informantion coming from vehicles on approaching section should be sreened by the locomotive direction and the change rate of GPS coordinates， so as to are used to screen the broadcast information which come from vehicles on the road to be passed as analysis data. Finally， the fuzzy C-Means algorithm is used to cluster the samples. All the coordinates of cluster centers are weighted to obtain the instantaneous speed and average speed in this section. Through analysis of an example， the road prediction results obtained by this method are consistent with the actual situation. This suggests that it is an effective method to predict real-time road condition.

Key words： V2V communications; predict of real-time traffic condition; comprehensive weighted; FCM

物流作为国家发展的重要经济命脉，利用先进技术提高物流车辆的全面感知能力、智能交互能力，是提升运输效率、节约能源和改善环境的重要途径。当物流车辆进入城市路网时，由于对前方路况缺乏预判，常常遇到道路拥堵问题，频繁的启停不仅会劣化车辆的燃油经济性，而且容易引发交通事故。造成这种现象的原因是由于物流车辆无法感知周围车辆信息，不能形成实时路况判断。

对于道路短时间内的路况预测，主要分为基于模型的预测算法和数据驱动的预测算法[1]。基于模型的路况预测代表性算法主要有卡尔曼滤波模型[2]、基于时间序列模型[3]。这类模型在处理动态、非线性数据上存在缺陷。……