While curbside management may seem a small part of the transportation system, it is actually a crucial component. Fascinatingly, it is a huge bottleneck for cities in terms of congestion, construction permit generation, business district vitality, livability, and looking forward, integration of automated vehicles. As we modify the transportation landscape in urban settings to accommodate automated vehicles and infrastructure, it is important to understand how peripheral services such as curbside management will be impacted. Current challenges that cities face include limited visibility into the consumption of parking resources, legacy equipment and regulatory constraints, and human behavior (e.g., on-street parking users rarely plan ahead, and instead drive to a location and circle outward). Motivated by practical considerations in the management of curb usage (with a particular focus on on-street parking) in cities, this talk will introduce some very interesting and challenging problems at the intersection of learning and control. The perspective taken will center on the development solutions that work within the constraints cities have (e.g., monetary budget, regulations/policy, and slow adoption of technology), while also creating solutions that are flexible in the sense that as new technology is introduced, the solutions still remain relevant. With this perspective in mind, I will briefly discuss some current promising approaches including active learning for estimation of occupancy and pricing/incentives in limited feedback, non-stationary environments. I will conclude the talk with a discussion on how highlighted the challenges and opportunities for curbside management are impacted by the future transportation system where automation will play a more significant role.
Copyright. All Rights Reserved.