Constructing Future Vehicle Digital Twins with Insights from Cloud-Edge Collaboration

Abstract

Intelligent Transportation System (ITS) is embarking on a new development road, in which Digital Twins (DT) provides crucial support for improving the smartness standard of vehicles and realizing a high-level of traffic management in this process. However, the existing solutions face challenges such as limited computing resources, data scarcity and data distribution, which hinder the application prospects of DT for ITS. First, we propose the concept of future-oriented DT aim to address these issues. Specifically, lightweight local DTs are deployed at the different edge servers to provide services for vehicles, and a strong performance global DT is deployed in the cloud server to manage and optimize the local DTs. Then, we introduce a collaborative learning method between two models to construct the DT network architecture and realize the vision of the future-oriented. This method makes use of rich global knowledge and local data, so that the model can dexterously integrate personalization and generalization capabilities. Finally, We use an Incremental Learning algorithm to improve the ability of the model that adapt to new scenarios and changing input data. All experiments are based on real traffic data, and trajectory prediction task as application carrier. Experimental results demonstrate the effectiveness of the proposed method, and the stability and performance of DTs have achieved excellent performance in the construction process.

Topics & Concepts