城市快速路交通流故障数据修复方法研究

发布时间：2019-06-27 14:39

【摘要】：摘要：交通流数据是交通状态辨识、交通管理及控制等交通领域研究及工作开展的基础。随着交通信息采集系统的发展,海量交通流数据不断涌现,但由于检测器自身故障、传输网络故障及环境因素等的影响,采集到的交通流数据难免会出现各种质量问题(不完整、错误、噪音等)。有效地对交通流故障数据(包括缺失数据和异常数据)进行识别和修复,使其能够真实地反映交通运行状态,才能为后续各项研究的顺利开展提供完整的数据支持和基础保障。本论文从检测器采集到的交通流数据出发,在对故障数据进行有效识别及分析的基础上研究对故障数据进行修复的方法。首先对数据进行时空特性分析,确定用来进行故障数据修复的时间特征参数和空间特征参数；然后进一步分析了交通故障数据识别和修复的方法,提出了基于平滑估计阈值的故障数据识别方法和基于统计相关分析的故障数据修复方法,并进行了实验验证；最后基于“分解—组合”思想,以基于最小二乘支持向量机的数据修复模型作为局部修复模型,以基于最大熵的交通状态概率分布估计模型作为自适应权重模型,提出一种自适应权重的两阶段故障数据修复组合模型,并结合北京市微波检测器的实际交通流量数据进行实验验证。实验结果表明,该方法能够较大程度上减少交通系统随机因素的干扰,具有较高的修复精度。
[Abstract]:Abstract: traffic flow data is the basis of traffic state identification, traffic management and control. With the development of traffic information collection system, massive traffic flow data continue to emerge, but due to the fault of detector itself, transmission network fault and environmental factors, the collected traffic flow data will inevitably have various quality problems (incomplete, error, noise, etc.). Effectively identify and repair the traffic flow fault data (including missing data and abnormal data), so that it can truly reflect the traffic operation state, in order to provide complete data support and basic support for the smooth development of the follow-up research. Based on the traffic flow data collected by the detector, this paper studies the method of repairing the fault data on the basis of the effective identification and analysis of the fault data. Firstly, the temporal and spatial characteristic parameters used to repair the fault data are determined, and then the methods of traffic fault data identification and repair are further analyzed, and the fault data recognition method based on smoothing estimation threshold and the fault data repair method based on statistical correlation analysis are proposed and verified by experiments. Finally, based on the idea of "decomposition-combination", taking the data repair model based on least square support vector machine as the local repair model and the traffic state probability distribution estimation model based on maximum entropy as the adaptive weight model, a two-stage fault data repair combination model with adaptive weight is proposed, and the experimental verification is carried out with the actual traffic flow data of Beijing microwave detector. The experimental results show that the method can greatly reduce the interference of random factors in traffic system and has high repair accuracy.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U491.112

【参考文献】