Data preparation and visualization of heavy vehicular movement for traffic prediction
Chong, Ji Shen
Date of Issue2017
School of Computer Science and Engineering
The problem with Singapore’s traffic is that Singapore does not have the ability to expand its road network on its limited landmass. Hence in order to maintain smooth traffic, one of the methods is to predict the flow of traffic at any given time and at any given weather. This will enable city planners to identify which roadways are prone to congestions during which periods and how to divert which type of traffic via which route that best enable the least travelling time. This report aims to characterise movements of various types of vehicle from their raw GPS data to filter, pre-process, calculate and visualise the movement of those vehicles in such a manner that will aid in the prediction of their movements. All of the data given is the raw GPS coordinates that has not been screened for error, hence error correction is also needed to prevent erroneous data from affecting the calculations of trips taken by the vehicles. The GPS data gathered are from GPS transponders on-board various vehicle types such has heavy vehicles, delivery vehicles and private buses. The limitation of this collection method is that the GPS transponder might be unable to maintain error-free recording of its GPS location at all time. In order to give the best predictions possible, the trip information of each vehicle types must be as accurate as possible. Hence the python scripts programmed in this report is designed to filter the raw GPS data for errors and correct them to a level that is as low as possible. This report finds that based on the GPS data given, the heavy vehicles have certain preferred route and routine in which guides its movements on the road. This routine is not known to change despite the availability of similar routes that allows the vehicle to reach its destination. However, the size of the vehicle may limit which route it can take. For Private buses, there is no fixed route or routine for most of the bus except one loop service. Hence the road often taken by such buses can be determined to be similar to regular cars and vehicles. This meant that there are a wider range of alternate routes available for these private buses.
DRNTU::Engineering::Computer science and engineering
Final Year Project (FYP)
Nanyang Technological University