Bus arrival time prediction method based on quantile convolutional network

Abstract

The invention discloses a bus arrival time prediction method based on a quantile convolutional network. Two prediction modes of point prediction and quantile prediction are performed on the arrival time of a bus, richer feature information is extracted from historical data, and future uncertainty is captured in a better way. Meanwhile, the method can solve the problem that historical data of some bus lines are sparse or are unavailable occasionally, and can learn a complex bus running mode by using a small amount of historical bus GPS data and known change characteristics (such as months, weeks, hours, whether to open bus lanes or not) at future moments. And a multi-step prediction result in a future period of time is obtained. In the future, the method can be applied to some specific bus routes with too long bus dispatching intervals and unstable GPS signals.

Description

technical field [0001] The invention relates to the technical field of bus travel time prediction, in particular to a bus arrival time prediction method based on a quantile convolution network. Background technique [0002] Conventional public transport is the oldest urban public transport mode in my country, and it is also the mode that undertakes the largest passenger transport volume. Compared with rail transit such as subways, it has a dense service network and customizable flexibility. [0003] But in recent years, with the rapid development of urban rail transit, the rise of online car-hailing and slow traffic, in most cities across the country, the passenger volume and proportion of public transport undertaken by conventional buses have shown a significant downward trend in recent years. [0004] Strong competition from other modes of transportation is the main external cause of the shrinkage of conventional bus passenger traffic. From the perspective of internal caus...

AI Technical Summary

Problems solved by technology

[0004] Strong competition from other modes of transportation is the main external cause of the shrinkage of conventional bus passenger traffic. From the perspective of internal causes, conventional buses also have problems of insufficient attractiveness: slow travel speed, low punctuality rate of line services, and unpredictable travel time

This algorithm has a good performance in prediction accuracy, but the public transport system is essentially a dynamic random system: the road traffic volume changes with time, and the time distribution of traffic volume is random; residents' demand for rides fluctuates , the arrival of passengers at the bus station is random; the arrival time of the bus at the intersection is random, according to the queuing and arrival and departure characteristics of the bus at the intersection, the delay time of the bus intersection is affected by random factors

Images