In recent times, automotive giants and IT firms have actively pursued autonomous driving technology, even claiming achievements up to level 5 automation. However, commercial technology mainly operates at level 2 or 3, with a transition to level 4. Full autonomy remains premature due to the need for substantial investments and regulatory challenges.
Practical implementation of autonomous driving faces complexities such as coexisting autonomous and non-autonomous vehicles. In this transition period, an effective strategy is essential to maximise benefits of the autonomous technology. In dedicated autonomous spaces, uncertainties of human drivers vanish, allowing even lower-level autonomy (levels 3 or 4) to function independently. Safety and efficiency can be guaranteed in autonomous spaces.
Efforts have been made to create such exclusive autonomous spaces. For instance, Elon Musk's Boring Company invested $52 million in a tunnel for passenger transport in Las Vegas Convention Center using Tesla electric vehicles. Furthermore, the Vegas Loop is planned to consist of 109 kilometers of tunnels with 81 stations, with the potential to serve 90,000 passengers per hour.
In a study by the Seoul Institute (2023), a virtual autonomous driving exclusive network, similar to the Vegas Loop, was envisioned. The study used simulations of a 3x3 autonomous driving network to redefine the capacity of a dedicated autonomous network. In this following research, actual autonomous driving exclusive network designs suitable for Seoul were determined and associated transportation capacity and costs were calculated based on practical guidelines.
Two network designs were proposed: a grid-based network traversing Seoul, offering open or closed routes and a radial network encircling the city center, mainly composed of closed routes. The grid-based network demonstrated an average link capacity of 30,000 vehicles per day, increasing to around 35,000 vehicles per day with a growing demand. Travel time on surface urban expressways was improved by about 5%, reducing a 60-minute trip to 57 minutes. The closed network exhibited traffic capacities of 659,000 passengers per day for the grid-based network and 547,000 passengers per day for the radial network. This study also estimated the cost of constructing a similar network using tunnel boring machines (TBMs) as a benchmark.
This Study presents four policy recommendations: easing constraints on autonomous road network projects, promoting private investments, deploying level 4 autonomous vehicles as shared services in the initial phase, implementing differentiated demand management between private and shared autonomous vehicles and ensuring convenient travel by connecting residential and commercial facilities to the network through small stations and infrastructure development.