Abstract
Predicting city-scale taxi origin-destination (OD) flows takes an important role in understanding passengers' travel demands as well as managing taxi operation and scheduling. But the complex spatial dependencies and temporal dynamics make this problem challenging. In this paper, a hybrid deep neural network prediction model based on convolutional LSTM (ConvLSTM) is proposed. For improving the prediction accuracy, the implicit correlation between travel time and OD flow is explored and they are combined as inputs of the prediction model. Moreover, in order to realize OD flows prediction at the road network level, and solve the problem that grid-based representation method cannot distinguish traffic flow at different heights, such as in multi-layer overpass areas, this paper presents a grid and road nested method to represent ODs. With the time of day partitioned into time slots, OD flows are extracted and predicted in both spatial and temporal domain. In the experiment, real taxi data are used to verify the proposed model and prediction method fully. And the experimental results show that the proposed model can effectively predict city-wide taxi OD flow, and outperforms the typical time sequence models and existing deep neural network models.
Highlights
Due to the limited means of observation, most studies focused on static OD flow estimation, and it was usually assumed that the traffic flow on the road sections was stable during a relatively long observation period (e.g., 1 week, 1 month, or 1 year) [1]
To fully utilize the fine-grained characteristics of taxi trajectory in the spatial-temporal domain, we propose a grid and road nested OD representation method, and utilizes ConvLSTM, Conv2DTranpose, and SeparableConv2D to construct a hybrid deep neural network prediction model, which can be applied to explore the correlation between OD flow and travel time, and estimate the taxi OD flows on the city scale
Prediction modeling is the core part of this study, which will construct a hybrid deep neural network based on ConvLSTM
Summary
To fully utilize the fine-grained characteristics of taxi trajectory in the spatial-temporal domain, we propose a grid and road nested OD representation method, and utilizes ConvLSTM, Conv2DTranpose, and SeparableConv2D to construct a hybrid deep neural network prediction model, which can be applied to explore the correlation between OD flow and travel time, and estimate the taxi OD flows on the city scale. With the increasing enrichment of traffic data, non-parameterized prediction models have begun to play a role in nonlinear, static, and dynamic processes as well as spatial-temporal analysis, becoming the most effective traffic prediction models in the current phase Some of these models can be combined with environmental factors to further improve the accuracy of prediction [35].
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
