Developing Demand-Responsive Mobility Services Based on Autonomous Driving Technology: A Vision for the Future of Public Transportation
Developing Demand-Responsive Mobility Services Based on Autonomous Driving Technology: A Vision for the Future of Public Transportation
▲ Research Specialist Jang Ji-yong, Department of Highway & Transportation Research, KICT
Prologue
The city of Seoul began its operation of autonomous buses in Cheonggyecheon in November 2022, which was followed by the launch of late-night autonomous buses running between Hapjeong Station and Dongdaemun Station in December 2023. Both of these were public transportation services provided along limited, fixed routes, with a driver's seat and a driver on board, yet are examples of commercialized public transportation services leveraging autonomous driving technology at the local government level. In the past, services such as Hyundai's "Shucle," "Zero Shuttle" in Pangyo, and "Majung" in Siheung in Gyeonggi Province have been trialed, though these were more akin to pilot operations. It seems that the autonomous driving technology we are approaching may first be experienced by most of us through public transportation.
Public transportation, a service relied on by many for mobility within a city, provides greater convenience to the public as its service area expands. However, issues such as manpower and budget constraints impose limits on expanding the service area beyond a certain level. As one alternative in the public transportation sector, Demand Responsive Transit (DRT) services have been expanded to improve the quality and utility of public transportation services (Korea Research Institute of Transportation Industries, 2024). However, even DRT-based services cannot be completely free from operational manpower and financial constraints. For this reason, there has been active research into combining autonomous driving technology with the demand-responsive public transportation services attempted by Seoul and other local governments as an alternative that can overcome the inherent limitations of conventional public transportation. Advanced autonomous driving technology does not require drivers, which means that it can potentially be used to overcome some of the current limitations of public transportation, at least in terms of operational manpower and related financial constraints.
Since April 2021, the Korea Institute of Civil Engineering and Building Technology (KICT) has been conducting a national research and development project called “Development of Real-Time Demand-Responsive Autonomous Public Transportation Mobility Service Technology” (Principal Researcher: Moon Byung-sup, Senior Research Fellow) to develop a public transportation mobility service utilizing autonomous driving technology. The goal is to develop a demand-responsive autonomous public transportation service that expands the service concept of existing public transportation, including DRT. This paper introduces what differentiates this service from existing ones and why it is called a "Vision for the Future of Public Transportation.”
Definition of Demand-Responsive Autonomous Mobility Services
This service is a demand-responsive public transportation mobility service based on autonomous driving technology. It aims to provide a first-and-last-mile service using Level 4 autonomous vehicles as defined by the Society of Automotive Engineers (SAE), transporting passengers to their desired destinations without fixed routes (Figure 1). To enable a safe public transportation service, a small vehicle equipped with a Level 4 autonomous driving system is being developed for demand-responsive service. What distinguishes this system from previous similar demand-responsive services is its ability to learn and remember individual users' travel patterns. Using this learned information, it generates optimal dynamic routes considering real-time changes in road and traffic conditions, and transports passengers accordingly. To provide this service, a 9-seater small vehicle is being made, allowing ride-sharing within pre-allocated routes and travel time allowances. The features of learning individual travel patterns to predict usage demand and preferred routes and proposing these to users, along with the capability for ride-sharing in an autonomous bus, clearly differentiate this service from previous offerings, making it a new vision for the future of public transportation.
Configuration and Functions of Demand-Responsive Autonomous Mobility Services
To provide a safe and comfortable demand-responsive public transportation service using small buses equipped with Level 4 autonomous driving systems, a central system responsible for service operation and control is required. Additionally, as this is a public transportation service based on autonomous driving technology, an evaluation system is required to assess the service’s public availability and operational efficiency. In addition to the autonomous small bus, central system, and evaluation system, facilities for vehicle storage and charging are needed. The system configuration for providing a demand-responsive autonomous public transportation mobility service is shown in Figure 2.
The core functions for providing demand-responsive autonomous mobility services are included in the central system, vehicles, and user mobile app (Figure 3). First, a user mobile app is required to provide a public transportation service based on a driveress Level 4 autonomous system. The mobile app has functions for service requests, user authentication, billing, and checking reservation and operation information. The central system is responsible for the core functions that enable demand-responsive services. This involves algorithms that analyze passengers' travel history to predict call demand and pre-allocate the required number of vehicles to service areas. It also includes algorithms for selecting the nearest virtual stop to the user's call point. Additionally, the system generates optimal dynamic routes from origin to destination, reflecting real-time road and traffic conditions, and updates routes with minimal detour time when ride-sharing requests are made. The vehicle itself is equipped with an autonomous driving system, an in-vehicle terminal for user authentication, and a human-machine interface for interaction between onboard safety personnel and the autonomous driving system.
The central system and vehicles exchange Travel Information Messages (TIM), Waypoint messages, and Probe Vehicle Data (PVD) in real time to provide services. Here, PVD is a message that contains the vehicle status information, including the driving trajectory of an autonomous small bus. The Waypoint message is a core message for implementing driverless autonomous public transportation services. It contains global path information representing the vehicle's route of movement and essentially includes the coordinates of nodes the vehicle passes through and the Estimated Time of Arrival (ETA) between nodes.
Efforts to Develop Future Public Transportation Services
Level 4 autonomous driving implies a "Mind-off" state, wherein the human driver is not required to be aware of the surroundings, make driving decisions or control the vehicle. Since public transportation services that apply driverless autonomous driving technology cater to a large number of users, the development of the service itself is important, but it is equally crucial to develop thorough verification technologies. Looking at previous research related to Autonomous Mobility-on-Demand (AMoD) services utilizing autonomous driving technology, most studies have only performed performance checks of the developed systems (Zhang et al., 2016; Barbier et al., 2019). To ensure passenger safety and successful establishment as a public transportation service, I am developing new service verification techniques by incorporating traffic engineering theories into the unavoidable verification technology development (Jang et al., 2023). Despite being public transportation, this world-first service concept learns individual travel patterns to predict usage demand and preferred routes in advance, and proposes them to users. It is an autonomous public transportation service that allows ride-sharing while following dynamic routes without fixed lines. Along with the development of autonomous public transportation service verification technology that considers public safety, these advancements are expected to lead a new future of public transportation that we will soon experience.
―――――――――――――――――
References
• Korea Research Institute of Transportation Industries (2024) Bus Transportation, Vol. 81, pp. 24-37.
• Barbier, M., Renzaglia, A., Quilbeuf, J., Rummelhard, L., Paigwar, A., Laugier, C., Legay, A., Ibanez-Guzman, J., and Simonin, O. (June 2019), Validation of Perception and Decision-Making Systems for Autonomous Driving via Statistical Model Checking. 2019 IEEE Intelligent Vehicles Symposium (IV), Paris, France, pp. 252-259.
• Zhang R., Rossi, F., and Pavone, M. (May 2016) Model Predictive Control of Autonomous Mobility on Demand Systems. 2016 IEEE International Conference on Robotics and Automation (ICRA), Stockholm, Sweden, pp.1382-1389.
• Jang, J., Moon, B., and Ha, J. (2023) Development of Performance Verification Methodology for Level 4 Autonomous Driving Technology-based Demand-Responsive Mobility System. International Journal of Highway Engineering, 25(6), pp. 357-367.
