机构地区:[1]中交第一公路勘察设计研究院有限公司,西安市710065 [2]西安航空学院车辆工程学院,西安市710077 [3]长安大学运输工程学院,西安市710018
出 处:《公路》2023年第11期182-191,共10页Highway
基 金:国家重点研发计划项目,项目编号2017YFC0803900;国家重点研发计划项目,项目编号2022YFC3002600;陕西省自然科学基础研究计划项目(一般项目(青年)),项目编号2022JQ-540。
摘 要:为研究高速公路互通立交分/合流区行驶速度特性,采用基于无人机视频的实时交通参数获取方法,提取试验路段全样本高精度时序速度数据,对比分析互通立交分流区和合流区车道级速度分布差异;研究不同车型速度特性;针对时序数据的长期依赖问题,构建基于时序Transformer的车道级行驶速度短时预测模型,运用平均绝对误差(MAE)和相对误差指标(MAPE)对比分析不同车道的预测精度。分析结果表明:分/合流区施划虚实线两侧客货车道切换过程中,速度变化幅度最大,且2个车道速度离散度最高;分/合流区速度分布基本符合内侧车道实际运行车速比外侧车道高的特性,但合流影响区加速车道较相邻主线车道速度统计值高;主线最高限速值与各车道85%分位车速平均差值为11.77km/h;两种车型速度分布总体呈现双峰特征,且小型车速度分布更为离散;最终所构建的模型预测准确率可达到98.35%,平均绝对误差为0.996km/h。上述结果表明:在工程设计、安全设施布置优化等方面需考虑客货车道的速度顺适过渡以及加速车道和相邻主线车道的速度协调性关系;主线限速标准对于互通分/合流区并不完全适用,驾驶人车速控制主要受行车环境和驾驶需求影响;小型车较大型车安全性差;本文所建立的时序Transformer模型适应于高速公路互通立交分/合流区所有车道短时速度预测,且模型效果较好,可为后续长时交通流参数预测奠定基础。In order to study the driving speed characteristics of interchange diverging and merging areas of expressway,the real-time traffic parameter acquisition method based on UAV video is adopted to extract full sample high-precision time-sequence speed of the test section.The speed distribution difference of lane level in diverging and merging areas of interchange is compared and analyzed.The speed characteristics of different vehicle types are studied.Aiming at the long-term dependence of time series data,a short-term prediction model of lane-level driving speed based on time-sequence Transformer is built,and the prediction accuracy of different lanes is compared and analyzed by mean absolute error(MAE)and relative error index(MAPE).The results show that the speed variation of passenger and cargo lanes on both sides of dashed solid line in diverging and merging zones is the greatest and the speed dispersion of both lanes is the highest.The speed distribution in diverging and merging zones basically conform to the characteristics that the actual driving speed of inner lane was higher than that of outer lane,but the speed statistics of acceleration lane in merging zone were higher than those of adjacent main lane.The average difference between the maximum speed limit of the main line and 85%of the vehicle speed of each lane is 11.77km/h.The speed distributions of the two models are generally bimodal,and the speed distributions of small vehicles are more discrete.The prediction accuracy of the model could reach 98.35%and the average absolute error is 0.996km/h.The above results indicate that the proper speed transition of passenger and cargo lanes and the speed coordination relationship between acceleration lanes and adjacent main lanes should be considered in engineering design and layout optimization of safety facilities.Main line speed limit standard is not fully applicable to interflow diverging and merging areas.Driver speed control is mainly affected by driving environment and driving requirements.Safety of smalle
关 键 词:交通安全 速度预测 Transformer模型 高速公路互通立交 车速控制
分 类 号:U491.112[交通运输工程—交通运输规划与管理]
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...
正在载入数据...