The application of commercial floating car data for speed-based traffic state evaluation in the Sub-Saharan African context

Date
2023-03
Journal Title
Journal ISSN
Volume Title
Publisher
Stellenbosch : Stellenbosch University
Abstract
ENGLISH ABSTRACT: Floating car data (FCD) are traffic data that are passively reported from within the traffic stream by GPSenabled probe devices commonly carried in vehicles, including smartphones, on-board navigation devices, and vehicle tracking systems. Commercial FCD are collected, aggregated, stored, and sold by third-party traffic data entities, providing network-wide speed, travel time, and origin-destination data. Commercial FCD eliminate the need for traffic sensors and communications networks, while simultaneously reducing the data analysis demands typical of Big Data because commercial FCD are characteristically made available in structured and readily usable formats. Commercial FCD are poised to become a primary source of traffic data in low- and middle-income countries where traditional, sensor-based traffic data are only sparsely collected over a vast road network, thereby leapfrogging the extensive traffic data collection systems established in high-income countries. Presently, commercial FCD are not widely used or researched in Sub-Saharan Africa. This dissertation is, therefore, well timed in its aim to assess the correct application of commercial FCD in the Sub-Saharan African context. This dissertation identified that the sample providing commercial FCD results in an inherent bias (called sample bias) because the reporting devices are purchased according to socioeconomic status, particularly in Sub-Saharan Africa, where average income, smartphone penetration and technology uptake is low. This dissertation represents the first time that the term sample bias has been described and systematically investigated as a characteristic of commercial FCD. The research found that commercial FCD are impacted by sample bias according to the socioeconomic, demographic, and geographic distribution of the sample that reports FCD in Sub-Saharan Africa. Sample bias was also proven to impact the accuracy of FCD speeds to the extent that speed accuracy differed between cities in South Africa and between commercial FCD sources. Mathematical models for the correct application of FCD in Sub-Saharan Africa that can be applied irrespective of sample bias, were then considered. Sample bias can be excluded in congestion measurement using ratio indices. FCD were found to enable accurate and comparative congestion measurement, both of recurrent and non-recurrent congestion, using variations of the Speed Reduction Index. Finally, three unique use-case studies, specifically applicable to Sub-Saharan Africa, were carried out to demonstrate the usefulness of FCD to evaluate the impact of roadworks projects, identify potholes along rural routes, and map changes in traffic patterns during the COVID-19 pandemic. The primary contribution of this research will be to steer Sub-Saharan Africa towards applicable use-cases of commercial FCD for transport planning purposes. This dissertation should allow commercial FCD to be applied with confidence for the correct use-cases in Sub-Saharan Africa. This research has provided a guideline for the evaluation of potential sample bias and the impact thereof, demonstrated the huge potential for FCD-based traffic pattern monitoring in a Sub-Saharan African context – for both urban and rural applications – and suggested future research for continued Sub-Saharan-specific FCD application. This research is critical to guide commercial FCD towards a significant role in providing primary traffic data over the extensive road network of our continent.
AFRIKAANS OPSOMMING: Swewende voertuigdata (FCD) is verkeersdata wat passief gerapporteer word van binne die verkeerstroom deur GPS-geaktiveerde toestelle reeds in voertuie, insluitend slimfone, navigasietoestelle en voertuigopsporingstelsels. Kommersiële FCD word versamel, saamgevoeg, gestoor en verkoop deur derdepartyverkeersdata- entiteite, wat spoed-, reistyd-, en oorsprong-bestemmingdata verskaf. Kommersiële FCD skakel die behoefte aan verkeersensors en kommunikasienetwerke uit, en verminder data-ontledingsvereistes verwant aan “Big Data” omdat kommersiële FCD tipies in gestruktureerde formate beskikbaar gestel word. Kommersiële FCD is gereed om die primêre bron van verkeersdata in lae- en middelinkomstelande te word, waar tradisionele, sensor-gebaseerde verkeersdata nie langs die meerderheid van die padnetwerk gerapporteer is nie. Daardeur word die uitgebreide verkeersdata-insamelingstelsels wat in hoë-inkomstelande gevestig is, oorgeslaan. Kommersiële FCD is nog nie wyd benut of nagevors in Sub-Sahara Afrika nie. Hierdie verhandeling is dus gepas betyds om die toepassing van kommersiële FCD in Sub-Sahara Afrika te assesseer. Hierdie verhandeling het geïdentifiseer dat die steekproef wat kommersiële FCD verskaf, 'n inherente begunstiging insluit (genoem steekproefbegunstiging) omdat die rapporteringstoestelle volgens sosioekonomiese status aangekoop word; veral in Sub-Sahara Afrika, waar gemiddelde inkomste, slimfoon penetrasie en tegnologie-opname laag is. Hierdie verhandeling verteenwoordig die eerste keer dat die term steekproefbegunstiging as 'n kenmerk van kommersiële FCD beskryf en sistematies ondersoek is. Dié navorsing het bewys dat kommersiële FCD beslis beïnvloed word deur steekproefbegunstiging volgens die sosio-ekonomiese, demografiese, en geografiese verspreiding van die FCD steekproef in Sub-Sahara Afrika. Daar is ook bewys dat steekproefbegunstiging die akkuraatheid van FCD-spoed beïnvloed tot die mate dat spoed-akkuraatheid tussen stede in Suid-Afrika, asook tussen bronne van kommersiële FCD, verskil. Wiskundige modelle vir die korrekte toepassing van FCD in Sub- Sahara Afrika wat toegepas kan word ongeag steekproefbegunstiging, is daarna oorweeg. teekproefbegunstiging in oponthoudmeting kan deur verhoudingsindekse elimineer word. Dit is bevind dat FCD akkurate en vergelykende oponthoud-meting moontlik maak van beide herhalende en nie-herhalende oponthoud deur variasies van die Spoedverminderingsindeks te gebruik. Laastens is drie unieke gebruiksgevallestudies, spesifiek tot Sub-Sahara Afrika, wat die bruikbaarheid van FCD bewys uitgevoer: evaluasie van die impak van padwerkprojekte, identifisering van slaggate, asook evaluasie van veranderinge in verkeerspatrone tydens COVID-19. Die primêre bydrae van hierdie navorsing sal wees om Sub-Sahara Afrika te lei na die toepaslike gebruiksgevalle van kommersiële FCD vir vervoerbeplanningsdoeleindes. Hierdie verhandeling laat dat kommersiële FCD met vertroue toegepas kan word vir die korrekte gebruiksgevalle in Sub-Sahara Afrika. Hierdie navorsing het 'n riglyn verskaf vir die evaluering van potensiële steekproefbegunstiging, het die potensiaal vir FCD-gebaseerde verkeersmonitering in 'n Sub-Sahara Afrika-konteks, vir beide stedelike en landelike toepassings getoon, en toekomstige navorsing voorgestel vir verdere FCD-toepassing in Sub-Sahara Afrika. Hierdie navorsing is van kritieke belang om kommersiële FCD te lei tot 'n beduidende rol in die verskaffing van primêre verkeersdata oor die uitgebreide padnetwerk van ons vasteland.
Description
Thesis (PhD)--Stellenbosch University, 2023.
Keywords
Citation