Project number: NCST-20-08

Link to final report: https://escholarship.org/uc/item/7x58z00c

With rapid population growth and urban development, traffic congestion has become an inescapable issue, especially in large cities. Many congestion reduction strategies have been proposed in the past, ranging from roadway extension, transportation systems management and operations, to demand management. In particular, as a demand management approach, congestion pricing and incentive offering schemes have been used as reinforcements for traffic control. Despite extensive research on congestion pricing mechanisms, almost all studies focus on traditional mobility systems and little has been done for future mobility services. With recent technological advancements, the shape of mobility services is drastically changing. Traditionally, the driver is the car owner and is the ultimate decision-maker on the origin, destination, routing, and the time of travel. In contrast, future mobility systems consist of different organizations and companies that completely (or partially) influence the behavior of individual human (or AI-based) drivers. Such organizations include car-sharing services (e.g. Zipcar, Turo), ride-hailing services (e.g. Uber, Lyft), crowdsourcing delivery systems (e.g. Amazon Flex, Instacart, Doordash), navigation applications (e.g. Google maps and Waze), and even companies producing autonomous cars with built-in navigation systems (e.g. Tesla), to name just a few.

In this research, we study and develop mechanisms for offering incentives to organizations and companies to change the behavior of individual drivers in their organization (or individuals using their organization's services). Such mechanisms can be more effective than traditional individual-level incentive offering mechanisms since each organization can control a large pool of individual drivers; thus moving the traffic flow toward the optimal "system-level" performance. In addition, such an "organization-level" incentive offering enjoys more flexibility than the individual "driver-level" incentive mechanisms. For examples, as we will discuss in this proposal, the time of travel or the choice of routing can be influenced significantly by providing incentives to organizations rather than to individual drivers. Even when congestion in certain areas is inevitable, incentives can be offered to organizations to use vehicles with less of a carbon footprint (such as electric cars) in congested areas to reduce the total carbon emission of the system. Our proposed approach is real-time and relies on historical data as well as demand estimates provided by organizations to continually predict traffic flow of the network; and provides incentives to organizations to reduce their carbon footprint by changing the behavior of individual drivers in their organization.

Finally, we will evaluate the performance of our method using data from the Los Angeles area as well as the models we develop during this research. The Los Angeles region is ideally suited for being the validation area as one of the most congested cities in the US. Additionally, researchers at USC have developed the Archived Data Management System (ADMS) that collects, archives, and integrates a variety of transportation datasets from Los Angeles, Orange, San Bernardino, Riverside, and Ventura Counties. ADMS includes access to real-time traffic datasets from i) 9500 highway and arterial loop detectors providing data approximately every 1 minute, and ii) 2500 bus and train GPS location (AVL) data operating throughout Los Angeles County. We will use this data to evaluate the performance of our proposed algorithm. As a byproduct of this research, we will also study the effectiveness of this "organization-level" method compared to individual-level incentive offering mechanisms.