On Urban Vehicular Ad-hoc Networks and Parking Systems

  • Pietro E Carnelli

Student thesis: Doctoral ThesisEngineering Doctorate (EngD)

Abstract

The nexus between wireless communication and transportation systems is investigated for urban areas. Both near and longer term scenarios are considered. In the near term, we research potential methods for reducing inner-city vehicle congestion through use of wireless enabled vehicle-parking occupancy detection and information dissemination. Subsequently, we developed a low-energy, smart-phone based vehicle-parking activity detection system. Our proposed system achieved high parking activity detection accuracies using low-energy sensors found in typical smart-phones.
With regards to longer term urban transport and wireless systems, we investigated the feasibility of using a connected fleet of (potentially autonomous) vehicles, also referred to as a VANET (Vehicular Ad-hoc NETwork), to provide temporary network connectivity to a city-wide (delay tolerant) sensor network.
Initially we investigated how such a system would perform using real-world taxi
trace datasets that were publically available online for Rome (Italy) and San
Francisco (USA). We combined the (filtered, map-matched and ‘folded’) taxi traces with a road network and Line-of-Sight (LoS) model (based on Open-
StreetMap road and building footprint datasets) for our VANET simulations.
Simulation results show that increasing our connected fleet size relative to the
number of sensors reduces end-to-end delay. However, a performance plateau
was observed after increasing taxi fleet sizes beyond a certain ratio of vehicles
to sensors.
Finally, we investigated methods of improving the performance of our citywide
VANET system. We developed an agent-based VANET simulation that
allowed for strategic re-routing of connected autonomous vehicles, as they performed their passenger trips, in order to increase sensor message mean packet delivery ratio values. Results show that even a minor increase in typical passenger journey lengths (circa 500m or approximately a tenth of the median passenger journey length) increased the final mean sensor packet delivery ratio values.
Date of Award23 Jun 2020
Original languageEnglish
Awarding Institution
  • The University of Bristol
SupervisorR E Wilson (Supervisor) & George Oikonomou (Supervisor)

Cite this

'