Transfer Learning Architecture Approach for Smart Transportation System

Document Type

Conference Proceeding

Publication Date

1-1-2022

Abstract

An intelligent and smart transportation system aims at effective transportation and mobility usage in smart cities. In recent years, modern transportation networks have undergone a rapid transformation. This has resulted in a variety of automotive technology advances, including connected vehicles, hybrid vehicles, Hyperloop, self-driving cars and even flying cars, as well as major improvements in global transportation networks. Because of the open existence of smart transportation system as a wireless networking technology, it poses a number of security and privacy challenges. Information and communication technology has long aided transportation productivity and safety in advanced economies. These implementations, on the other hand, have tended to be high-cost, customized infrastructure systems. To address these challenges, a novel machine learning method developed for a transportation system is reused for making it more generic and smart for intelligent carriage. This type of transfer learning enables rapid progress on the task with enhanced results. In this work, together with domain adaptation, a novel weighted average approach is used to build models related to the smart transportation system. A smart system comprising of interconnected sensors along with the gateway devices can lead the way to a more efficient, viable and robust city centers. Finally, in this paper also provides a view of current research in smart transportation system along with future directions.

Publication Title

Communications in Computer and Information Science

First Page Number

162

Last Page Number

181

DOI

10.1007/978-3-031-09469-9_15

This document is currently not available here.

Share

COinS