Search

Taxi trip-sharing is of great significance to improve the overall traffic efficiency of the city. The calculation of the matching possibility between any two taxi trips in a given time range is large,so it is difficult to meet the application needs of the whole urban space . This paper proposed a fast filtering strategy based on time, distance and direction constraints. It applied the strategy to the trip-sharing matching method which was based on graph. The results verified by the New York city taxi data set, show that the proposed filtering strategy can greatly reduce the matching time without affecting the effect of trip-sharing( the matching time of a single trip is reduced by about 82%). Based on the GPS trajectory data of 8 531 taxis in Qingdao on March 15th (Wednesday) and March 18th (Saturday) during 2017, this paper made an empirical analysis. It shows that when the de- lay threshold is set at 5 min, the number of shared trips in the city on workdays can reach 45% of the total trips, the total travel time can be saved close to 5 000 hours, and the total distance can be saved 80 000 km. The results of Saturday and Wednesday are very close. The number of trips during the morning rush hour from 6 :00 to 9 :00 is relatively small and the spatial distribution is relatively scattered. Therefore, the proportion of shared trips, the total saved time and saved distance are the least com- pared to other period. The benefit of trip-sharing during this period is the smallest. [ABSTRACT FROM AUTHOR]

Reliable traffic flow prediction is of great significance in traffic management and public safety. However, this is also a challenging task because it is affected by spatial dependencies, temporal dependencies and some additional factors (weather and emergencies, etc . ) . Most existing works can only consider part of the attributes of traffic data, resulting in insufficient modeling and unsatisfactory prediction performance. This paper proposed a novel end-to-end deep learning model, called spatio-temporal attention ConvLSTM ( ST-AttConvLSTM ), for traffic flow prediction. ST-AttConvLSTM was divided into three branches for modeling. For each branch, the residual neural network was used to extract local spatial features, and external factors were also combined with them. Then, it employed two components consisting of ConvlLSTM and attention model to discover the potential relationship of traffic flow, and capture the correlations of data in both spatial and temporal dimensions. It used two real trips data sets in Beijing and New York to evaluate the proposed method . The experimental results show that this method achieves higher prediction accuracy than well-known baselines. [ABSTRACT FROM AUTHOR]