基于RFID的城市道路交通流预测方法研究

发布时间：2018-03-29 18:58

本文选题：RFID　切入点：相似性　出处：《吉林大学》2017年硕士论文

【摘要】：RFID(Radio Frequency Identification)是自动识别技术的一种,通过无线射频方式进行非接触双向数据通信,对目标加以识别并获取相关的数据。RFID技术在交通上的应用前景广泛,很多生活小区、工作场所、停车场的出入口都设有RFID检测器,路网上每个关键节点也都会安装相应的RFID检测基站,RFID将会成为越发重要的交通信息采集源。出行者在每日的工作休闲娱乐中会形成一定的出行习惯和规律,通过RFID能够得到每个驾驶员的历史和实时出行信息,从而得到这些规律以便更好的探知交通流运行状况。首先对RFID数据进行了预处理。在RFID数据的传输或存储过程中出现故障会造成数据冗余、错误和缺失,对RFID原始数据进行检测和分析,并清洗数据,可以极大的保证数据质量。在数据清洗完成后,本文更进一步对海量的RFID数据按时间汇集度进行分割并按时间先后顺序进行存储,同时也考虑了交通流在工作日与节假日和交通平高峰时段各自不同的变化规律,分开分类存储,为交通流的分析和预测提供有力的数据保障。其次,基于数据的相似性对交通流进行了预测。传统的预测方法,极少考虑实时的交通流状态和驾驶员的路径选择习惯,造成其预测结果往往偏离交通流的实际变化趋势。基于预处理后的RFID数据,通过计算相似交通流状态下的流量转移率来预测路段交通流量,并进行了实例验证。该方法兼顾实时交通流数据与历史数据,考虑到了驾驶员的历史路径选择概率,与实际情况更相符,且能够适应一定的交通流波动情况,预测精度高,能够满足交通信息服务的要求。基于RFID技术获取驾驶员出行链信息,并利用出行链预测交通流。通过路网中的RFID检测基站就可以得到每辆车的历史出行OD和实时出行OD,将车辆的历史出行OD在Arc GIS上进行叠加可以得到出行链信息,结合实时出行OD就可以通过动态交通分配来预测路网的交通流运行状况,最后利用中观交通仿真软件对模型进行了验证分析,发现在对整个路网的交通流进行预测时,该方法方便简单,结果较为精确。
[Abstract]:RFID(Radio Frequency Identification is a kind of automatic identification technology. It uses wireless radio frequency mode to carry out contactless two-way data communication, identifies the target and acquires the relevant data. The entrance and exit of the parking lot are equipped with RFID detectors. Each key node on the road network will also install a corresponding RFID detection base station, which will become an increasingly important source of traffic information collection. Travelers will form certain travel habits and rules in their daily work, leisure and entertainment. Through RFID, you can get the history and real-time travel information of each driver. In order to obtain these rules in order to better understand the traffic flow status. First, the RFID data is preprocessed. Failure in the transmission or storage of RFID data will lead to data redundancy, errors and defects. The data quality can be greatly guaranteed by detecting and analyzing the RFID raw data and cleaning the data. After the data cleaning is completed, this paper further divides the massive RFID data according to the time collection degree and stores them in order of time. At the same time, it also takes into account the different changes of traffic flow during weekdays, holidays and traffic peak hours, and stores them separately in order to provide strong data protection for traffic flow analysis and prediction. Based on the similarity of the data, the traffic flow is forecasted. The traditional forecasting method seldom considers the real-time traffic flow state and the driver's route choice habit. The prediction results often deviate from the actual trend of traffic flow. Based on the pre-processed RFID data, the traffic flow is predicted by calculating the flow transfer rate in the similar traffic flow state. An example is given to verify that this method takes into account both real time traffic flow data and historical data, and takes into account the probability of driver's historical path selection, which is more consistent with the actual situation, and can adapt to certain traffic flow fluctuation, and has high prediction accuracy. Can meet the requirements of traffic information services. Based on RFID technology to obtain driver travel chain information, Using trip chain to predict traffic flow, the history travel OD and real time travel OD of each vehicle can be obtained by detecting the base station of RFID in the road network, and the travel chain information can be obtained by superposing the historical travel OD of the vehicle on the Arc GIS. Combined with real time travel OD, the traffic flow can be predicted by dynamic traffic assignment. Finally, the model is verified and analyzed by using the mesoscopic traffic simulation software, and it is found that when forecasting the traffic flow of the whole road network, the traffic flow of the whole road network can be predicted. The method is simple and convenient, and the result is accurate.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U491

【相似文献】