Investigating Human Mobility to Improve Multi-pattern Passenger Prediction Model for Graph Learning

PROJECT TITLE :

Exploring Human Mobility for Multi-pattern Passenger Prediction A Graph Learning Framework

ABSTRACT:

Predicting the flow of traffic is an essential component of an intelligent transportation system and, as such, is fundamental for a variety of applications that are related to traffic. Buses are an essential mode of transportation for urban residents because of their predetermined routes and timetables, which contribute to a certain degree of travel regularity. However, human mobility patterns, particularly the intricate relationships that exist between bus passengers, are obscured to a large extent by this stationary mode of transportation. Even though there are a lot of models that can predict how traffic will flow, the patterns of human mobility have not been investigated very thoroughly. We propose a multi-pattern passenger flow prediction framework called MPGCN that is based on graph convolutional networks in order to address this research gap and learn human mobility knowledge from fixed travel behaviors (GCN). First, using the information from the bus records, we build an original sharing-stop network to model the relationships between different passengers. After that, we make use of GCN to extract features from the graph by learning useful topology information, and we present a deep clustering method as a means of recognizing mobility patterns that are hidden in bus passengers. In addition, in order to make the most of the spatio-temporal information available, we propose the use of GCN2Flow to forecast passenger flow based on the many different mobility patterns. This paper, to the best of our knowledge, is the first work to adopt a multi-pattern approach to predict the bus passenger flow by making use of graph learning. This was done in order to take advantage of the benefits that graph learning offers. We develop a case study for the purpose of route optimization. Extensive testing on a dataset collected from buses operating in the real world reveals that the MPGCN has the potential to be effective in passenger flow prediction and route optimization.

Did you like this research project?

To get this research project Guidelines, Training and Code... Click Here