Reinforcement learning as a pre-diagnostic tool for TCP/IP protocols on in-car networks

Sanghun Yun, Jahyun Kim, Hyogon Kim

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

In-car networks such as Automotive Ethernet will enable new services in smart vehicles, but they are an untrodden territory for many network protocols that should support the applications. For instance, TCP has been newly incorporated as an integral part of AUTOSAR, the standard software framework for electronic control units (ECUs). However, the in- vehicle network environment can be starkly different from the Internet where TCP has been optimized. The disparity between the two environments warrants re-evaluation of the Internet protocols such as TCP in the forecasted in-car network environments. In this paper, we propose a novel approach where we employ reinforcement learning as a pre-diagnostic tool to predict potential problems and to present possible remedies.

Original languageEnglish
Title of host publication2019 IEEE 90th Vehicular Technology Conference, VTC 2019 Fall - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781728112206
DOIs
Publication statusPublished - 2019 Sept
Event90th IEEE Vehicular Technology Conference, VTC 2019 Fall - Honolulu, United States
Duration: 2019 Sept 222019 Sept 25

Publication series

NameIEEE Vehicular Technology Conference
Volume2019-September
ISSN (Print)1550-2252

Conference

Conference90th IEEE Vehicular Technology Conference, VTC 2019 Fall
Country/TerritoryUnited States
CityHonolulu
Period19/9/2219/9/25

Bibliographical note

Publisher Copyright:
© 2019 IEEE.

Keywords

  • AUTOSAR
  • Optimal policy
  • Reinforcement learning
  • TCP
  • Throughput

ASJC Scopus subject areas

  • Computer Science Applications
  • Electrical and Electronic Engineering
  • Applied Mathematics

Fingerprint

Dive into the research topics of 'Reinforcement learning as a pre-diagnostic tool for TCP/IP protocols on in-car networks'. Together they form a unique fingerprint.

Cite this