Hossam Hassanein
Professor and Director, School of Computing, Queen's University
Date/Time:
Title: Vehicular Edge Services
Abstract: The rise of Autonomous Vehicles (AVs) and Connected Vehicles (CVs), coupled with their growing service demands and enhanced computational power, has created a pressing need for innovative edge computing solutions that can both meet their demands and harness their capabilities. By leveraging the underutilized computational resources of such abundant and powerful AVs and CVs, we can offload computationally intensive tasks to these vehicles, enabling parallel processing of partitioned tasks at the extreme edge, thus drastically reducing latency. However, AVs and CVs are user-owned Extreme Edge Devices (EEDs), characterized by a highly dynamic nature, including a dynamic user access behavior. This dynamic nature increases the risk of intermittent availability and the likelihood of continuous changes in the vehicles' available capabilities, leading to uncertainty and discrepancies between perceived and actual performance, ultimately reducing reliability and impacting the Quality of Service (QoS). To address the issue of intermittent availability, we predict the availability of vehicles to adapt to the highly dynamic nature of the vehicular edge computing environment, and we incorporate these predictions into the resource allocation decision. Additionally, to enhance reliability and reduce uncertainty, we develop a comprehensive reputation scoring system to assess the reliability of vehicles based on past performance, and we use such reputation scores to make proactive task replication decisions. Beyond leveraging the computational capabilities of vehicles, we also meet the increasing demands of AVs and CVs by enabling quality-aware offloading of tasks initiated and requested by these vehicles for cooperative perception, thereby enhancing traffic situational awareness. We improve the quality of cooperative perception by minimizing perception redundancy and maximizing the Value of Information (VOI) received by each requesting vehicle. These innovative strategies have the potential to revolutionize vehicular edge networks, leading to more reliable, efficient, and responsive intelligent transportation systems, ultimately enhancing road safety, traffic management, and the overall driving experience.
Josep Laborda
CEO & Managing Partner, Factual-Consulting
Title: How to become a bike-friendly city? Lane Patrol as the international standard for cycling network planning
Abstract: The session will introduce the index developed by FACTUAL and the International Road Federation (IRF), which evaluates the readiness level of road infrastructures for the commercial operation of fully autonomous vehicles. The aim is to collaborate with road infrastructure managers globally to establish the SRI as the first worldwide standard for assessing roads under consistent parameters.
Marc Figuls
Managing Partner, Factual-Consulting
Title: How to become a bike-friendly city? Lane Patrol as the international standard for cycling network planning
Approximately 1.19M road traffic deaths occur annually, with over half of the fatalities involving vulnerable road users, such as cyclists and pedestrians. The CycleRAP methodology is introduced as an evidence-based solution to assess the safety of urban infrastructure, aiding policymakers in prioritizing interventions to improve road safety and the Lane Patrol tool complements this model, enabling automatic image collection, analysis of key parameters using AI techniques from the assessment side, and simulations of potential improvements in the cycling network while becoming an international standard for planning safer cycling networks that has been applied in very different use cases at worldwide level such as Madrid, Skopje, Montevideo and Abu Dhabi.
