From Small-Scale to Full-Scale: Assessing the Potential for Transferability of Experimental Results in Small-Scale CAV Testbeds

Connected and Autonomous Vehicles (CAVs) promise to enhance traffic safety and efficiency, yet full-size system testing remains resource-intensive and costly. This paper explores small-scale CAV testbeds as a cost-effective and low-maintenance alternative and systematically compares them with their full-size counterparts, explicitly focusing on the ability to replicate the real-world behavior. This paper separates the CAV domain into subcomponents: vehicle, communication, automation, and environment. We systematically assess each subcomponent, with some components easily classified as representable based on existing literature, while others require evaluation due to a lack of relevant discussion in the literature. We aim to provide a understanding of the representability of CAV subcomponents in small-scale testbeds and highlights their strengths and limitations in testing various aspects of CAV technology. Our results indicate that small-scale testbeds excel in evaluating software components related to motion planning, communication, and other algorithms. However, limitations arise in the interaction with the surrounding environment and in some parts may fall short of what traditional computer simulations are capable of.