Digital representations that replicate 3D physical objects at low latency and high accuracy, known as ?live 3D digital twins,? are the next advance in cyber-physics systems. In the transportation domain where vehicles are highly instrumented with 3D sensors that, e.g., include LiDAR and stereo cameras, their data will be continuously streamed and fused into a digital representation that provides accurate real-time situational awareness for the driver and the transportation infrastructure itself, promoting greater safety and efficiencies. This project develops and implements techniques, known as pipelines, to help build and compose digital twin models that are adaptive with guarantees on performance, accuracy, and dependability. <br/><br/>This project develops a framework that enables live 3D digital twins? pipelines to adapt to underlying network and compute resources and ensure low latency and high accuracy across a range of operating conditions. This project develops common abstractions (operators) with which we can compactly build and represent these pipelines. Doing so not only enables the capability to holistically reason about them, but also speeds up the development of pipelines and makes them less error prone. In addition, this project develops a control plane that uses a distributed resource monitor to understand the underlying network and compute conditions.<br/><br/>This research on live 3D digital twins will lead to safer transportation systems as applications such as autonomous driving and drone delivery increasingly become part of future transportation infrastructures. The project prepares students to be part of the future transportation research and development workforce through curriculum that integrates education with this research.<br/><br/>This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.