基于交通冲突技术和神经网络模型的高速公路出入口匝道安全性研究

发布时间：2018-05-06 12:13

本文选题：高速公路 + 匝道　；参考：《重庆交通大学》2015年硕士论文

【摘要】：高速公路出入口匝道是高速公路的重要组成部分,是交通事故频发地区,但由于其自身的特殊性以及诸多因素的影响,导致匝道发生交通事故后不仅会造成人员造成伤亡,还会严重影响高速公路的行车畅通。本文运用交通冲突技术和神经网络算法,综合考虑行驶车辆、驾驶员和高速公路线形等因素对出入口匝道的影响,开展高速公路出入口的匝道安全研究。首先,通过引入交通冲突技术研究高速公路出入口匝道安全性,对交通冲突技术的特点、交通冲突的产生过程和交通冲突技术的适应范围进行分析。对出入口匝道交通流特性进行了分析,对不同交通流状况的特点、运行机理和安全特征进行分析。然后,对影响高速公路出入口匝道的安全性的因素进行了分析。主要包括出入口匝道交通流的安全性、驾驶行为的安全性和几何道路线形以及限速对高速公路出入口匝道安全性的影响,为之后的安全预测模型建立提供指标依据。最后,对人工神经元、人工神经网络及特点进行分析。分析BP网络的特点,构成和学习算法,并选取BP网络建立了基于BP神经网络的高速公路出入口安全性预测模型。通过使用人工调查为主录像调查为辅的调查方法获取京津高速牛堡屯收费站出入口匝道的交通数据,并使用基于BP神经网络的高速公路出入口匝道安全性预测模型对交通冲突进行预测。将交通冲突数的预测结果与实际交通冲突数进行比较发现,预测结果能较好的反映高速公路入口匝道的交通冲突的变化规律,基于BP神经网络的高速公路出入口匝道安全性预测模型对高速公路出入口匝道安全性研究具有一定的意义。
[Abstract]:Freeway ramp is an important part of expressway, it is an area with frequent traffic accidents. However, due to its particularity and the influence of many factors, the ramp will not only cause casualties after the traffic accident. It will also seriously affect the smooth running of the highway. In this paper, traffic conflict technology and neural network algorithm are used to study the ramp safety of freeway exit and exit by considering the influence of vehicle, driver and highway alignment on the ramp. Firstly, by introducing traffic conflict technology to study the safety of freeway exit and exit ramp, the characteristics of traffic conflict technology, the process of traffic conflict and the adaptation range of traffic conflict technology are analyzed. The traffic flow characteristics of the ramp are analyzed, and the characteristics, operation mechanism and safety characteristics of different traffic flow conditions are analyzed. Then, the factors influencing the safety of freeway ramp are analyzed. It mainly includes the safety of traffic flow, the safety of driving behavior and geometric road alignment, and the influence of speed limit on the safety of freeway ramp, which provides the index basis for the establishment of safety prediction model. Finally, the artificial neuron, artificial neural network and its characteristics are analyzed. The characteristics, composition and learning algorithm of BP neural network are analyzed, and BP neural network is selected to establish the safety prediction model of expressway entrance and exit based on BP neural network. The traffic data of the entrance and exit ramps of the Beijing-Tianjin High-speed Niupoutun Toll Station were obtained by using the method of manual investigation as the main video survey auxiliary. The traffic conflict prediction model based on BP neural network is used to predict the safety of freeway ramp. By comparing the traffic conflict number with the actual traffic conflict number, it is found that the forecast result can better reflect the change rule of traffic conflict on the freeway ramp. The prediction model of freeway ramp safety based on BP neural network has a certain significance for the study of freeway ramp safety.
【学位授予单位】：重庆交通大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U491

【相似文献】