Research On Travel Chain Identification To Evaluate The Collaborative Status Of Urban Transportation

The coordinated development of urban passenger transport is an inevitable trend to meet the multi-level transport demand,and it is also the basis of social sustainable development.The purpose of comprehensive passenger transport collaborative evaluation is to find out the factors that affect the linkage of multiple transportation modes,analyze the factors that restrict the collaborative development,provide data basis for management and planning,so as to balance the traffic supply and the travel demand of residents under the limited resources.The first step to evaluate the collaborative status of urban passenger transport is to grasp the urban comprehensive passenger transport operation status,that is,the characteristics of residents' daily travel.Meanwhile,the study of daily travel characteristics of residents is based on individual travel informationBased on the mobile GPS track data of individual travel,this study develops an individual travel information extraction algorithm for comprehensive passenger collaborative evaluation in two stages respectively.Firstly,the ensemble learning method is used to identify the transfer behavior/transfer points in individual travel.Afterward,a deep learning framework abbreviated as ABLCNN,concatenate the attention mechanism based Recurrent Neural Network and Convolution Neural Network,is proposed to identify four kinds of ground transportation modes:bicycle,walking,car and bus.The results show that the proposed method has high robustness and recognition rate.