Abstract:
The emergence of Autonomous Vehicles (AVs) can significantly alter how people and
goods are transported, as Connected and Autonomous Vehicles (CAVs) offer increased
automation as well as improved connectivity between vehicles and roadside infrastructure.
Although a future where all vehicles on the road are autonomous, which seems to be farfetched, the rapid development of AV technologies worldwide is a clear indication that local
readiness needs to be addressed sooner rather than later.
There are many uncertainties regarding how AVs would operate on the road network
recognising that few studies have investigated how AVs would operate in an urban
environment. This study attempts to provide insight as to how AVs and CAVs perform in
such an environment, thus acting as a form of preparation for the future when AVs are
eventually implemented.
Four levels of autonomous behaviour were evaluated with a microscopic model, developed
in PTV VISSIM, to assess the relative improvements that each of these levels provides to
the road network’s performance at various penetration rates. The performance of the road
network was evaluated by considering various performance indicators, including average
network delays, network travel times, network travel speeds, queue lengths, and the
average number of stops. Specific focus was also placed on the performance of CAVs and
how robust they are to changes in traffic volumes.
AVs, and CAVs, were quite effective in improving the network’s performance across all the
performance indicators. Additionally, CAVs were quite robust when it came to handling
changes in traffic volumes. However, the performance of the AVs were significantly
dependent on their penetration rate. The findings in this study provided valuable insight
into what the future would look like when AVs are implemented. This research provides a
useful, albeit significant, first step in understanding the incremental introduction of CAVs
and AVs in South Africa from a traffic engineering perspective.
