A new methodology for multi-step-ahead forecasting was proposed herein which combined the wavelet transform(WT), artificial neural network(ANN) and forecasting strategies based on the changing characteristics of available parking spaces(APS). First, several APS time series were decomposed and reconstituted by the wavelet transform. Then, using an artificial neural network, the following five strategies for multi-step-ahead time series forecasting were used to forecast the reconstructed time series: recursive strategy, direct strategy, multi-input multi-output(MIMO) strategy, DIRMO strategy(a combination of the direct and MIMO strategies), and newly proposed recursive multi-input multi-output(RECMO) strategy which is a combination of the recursive and MIMO strategies. Finally, integrating the predicted results with the reconstructed time series produced the final forecasted available parking spaces. Three findings appear to be consistently supported by the experimental results. First, applying the wavelet transform to multi-step ahead available parking spaces forecasting can effectively improve the forecasting accuracy. Second, the forecasting resulted from the DIRMO and RECMO strategies is more accurate than that of the other strategies. Finally, the RECMO strategy requires less model training time than the DIRMO strategy and consumes the least amount of training time among five forecasting strategies.
The impacts of four different car-following types onrear-end crash risks at a freeway weaving section wereevaluated using trajectory data, in which Type 1 represents carfollowing car, Type 2 represents car following truck, Type 3represents truck following car and Type 4 represents truckfollowing truck. The time to collision (TIC) was introducedas the surrogate safety measure to determine the rear-end crashrisks. Then, the trajectory data at a freeway weaving sectionwas used for the case-controlled analysis. Three logisticregression models were developed with different TICthresholds to quantify the impacts of different car-followingtypes. The explanatory factors were alSO analyzed toinvestigate possible reasons for the results of logisticregressions. Results show that the rear-end crash risk of Type3 is 3, 167 times higher than that of Type 1 when the TICthreshold is 2 s. However, the odds ratios of Type 2 and Type4 are both smaller than 1, which indicates a safer condition.The analysis of explanatory factors also shows that Type 3 hasthe largest speed differences and the smallest net gaps. This isconsistent with vehicle operation features at a weaving sectionand is also the reason for the larger rear-end crash risks. Theresults of this study reflect the mechanism of rear-end crashrisks of different car-following types at the freeway weavingsection.
为获取居民公交出行的换乘信息,设计了一套基于多分类支持向量机(multi-class support vector machine)的公交换乘识别方法.通过融合GPS数据和公交IC卡数据获取训练样本,利用多分类支持向量机进行样本训练,选取最佳训练样本量,并采用网格搜索法结合粒子优化算法对模型参数进行标定,以获取最优SVM分类模型.测试结果显示模型分类精度可达90%.以佛山市公交车GPS数据和IC卡数据对算法进行验证,并获取公交换乘量、公交换乘比例等基本换乘数据.结果表明:算法可在少样本条件下完成公交换乘识别,且分类识别精度高,尤其适用于公交线网复杂的大城市公交换乘识别,有助于在公交前期规划时进行线路布设和枢纽选址.