HomePublicationsPhD Thesis

Synchronous Intelligent Intersections for Sustainable Urban Mobility
Ref: CISTER-TR-231002       Publication Date: 2023

Abstract:
According to the United Nations urbanization report 2018, 55% of the world's population lives in urban areas and is expected to reach 68% by 2050. Current urban transportation services (except during COVID-19) are already strained, causing fossil fuel dependency, environmental pollution, and associated human health. For instance, in the United States of America, urban dwellers lost 8.7 billion hours and almost 3.5 billion gallons of fuel due to traffic congestion in 2019 alone. In the case of the European Union, member states lost an estimated 110 billion euros annually for similar reasons. As per the European Environment Agency, 80% of European urban dwellers are exposed to dangerous vehicular emissions, particularly at intersections, bus stops, and other points of traffic confluence. This exposure can cause heart disease, cancer, respiratory disease, and, in the worst case, death. Therefore, in the worst-case scenario, the urban future would bring disastrous consequences, including transportation and mobility, human health, and environmental challenges. Hence, sustainable transportation is a major concern, requiring affordable energy-efficient transportation and emitting low to zero air pollutants, including alternate energy sources like electricity.
Intelligent transportation systems (ITS) incorporating communication, electric and autonomous technologies provide new opportunities for sustainable urban mobility (SUM). In the USA, the National Highway Traffic Safety Administration (NHTSA) anticipates that autonomous vehicles (AVs) would reduce nearly 50 minutes of average commuter delay daily. However, autonomous-only urban transportation is not expected before 2045, and transitioning from human-driven internal combustion engine (ICE) vehicles to communicating and electric AVs will take long. Therefore, ITS-based solutions of the near future must support mixed traffic of AVs and human-driven vehicles (HVs).
Motivated by these facts, we identify the vital role that road intersections play in urban transportation, where growing trends in queue length, waiting delays, and associated adverse effects can be observed. Existing intersection management (IM) strategies permit vehicles to access intersections sequentially from one road at a time or parallelly from opposite road lanes while blocking the traffic of other roads and lanes. This behavior imposes unnecessary waiting delays at intersections. Thus, this thesis introduces a reactive synchronous framework for maximizing vehicle intersection access from all non-conflicting roads based on their arrival at the intersection entrance instead of reserving lengthy time slots or imposing strong differentiation between HVs and AVs.
To study the practicality and performance of the synchronous framework, we resort to the Simulation of Urban MObility (SUMO) microscopic traffic simulator throughout this thesis. The synchronous framework is then applied to both isolated intersections (single-lane and multi-lane) and networks of multi-lane intersections. Two left-lane (dedicated and shared) configurations for multi-lane intersections are considered. The performance metrics, such as the throughput, average travel time loss, and associated fuel consumption, are measured for low-speed urban conditions. The simulation results of the tested scenarios show that synchronizing vehicles intersection access improves throughput between 3 to 30%, reduces travel time loss between 56 to 129%, and minimizes associated fuel consumption between 18.2 to 67.4% against the next best baseline approaches.
We also evaluate the operational efficiency of the IM systems in what concerns the worst-case reaction to given traffic scenarios described statistically. For this purpose, we borrowed a service metric from real-time systems concepts namely the Worst-Case Response Time (WCRT). The WCRT evaluates the maximum time a vehicle may experience since it enters an input road (i.e., origin) until it leaves the last intersection in its path before heading to the respective destination. The commonly available parameters such as the geographical settings of the road networks (road lanes length and intersection space within), traffic-related information (average speed, maximum queue length, and capacity), and IM-specific parameters (green phase time and total cycle time) are utilized to formulate the WCRT. The analytical WCRT values are validated with simulation studies. On average, the computed WCRT values in diverse traffic conditions were up to 20% and 18% higher than the observed WCRT values at isolated single-lane and multi-lane intersections, respectively. In the case of networks of intersections, these values were up to 86.7% (64.5%), 49.5% (34.2%), and 47.6% (30.25%) higher for crossing single, two, and three dedicated (and shared) left lane intersections, respectively. This means the pessimism of estimated WCRT values gradually reduces for increasing number of crossed intersections and it also reduces from dedicated to shared left lane intersections.
Finally, we study the transportation sustainability of introducing growing penetration rates of AVs, either considering propulsion systems with gasoline ICE or with electric batteries (BEVs). Energy savings and emissions reduction are considered as performance metrics. The results show that our \textit{synchronous framework} is more sustainable than the benchmark approaches in all the tested scenarios, speeds, and intersections, mainly when human-driven ICE vehicles are mixed with BEVs. In the networks of intersections case, the overall energy savings when increasing the AVs penetration from 10% to 90% are up to 43% (gasoline) and 1.45% (electricity), and emissions reduction is up to 42.6% (CO_2), 48.2% (NOx), 64.9% (PMx), 79.5% (HC), and 89.45% (CO). In the isolated intersections cases, these improvements are even more significant.
Overall, our research shows that the proposed synchronous framework has a high potential to improve the performance of intersections in urban scenarios during the transition period of human-driven to autonomous-only vehicles.

Radha Reddy

Document:

View
PDF Download
PDF