Software adaptation for an unmanned undersea vehicle

Research output: Contribution to journalArticle

Authors

  • Avi Pfeffer
  • Curt Wu
  • Gerald Fry
  • Kenny Lu
  • Steve Marotta
  • Mike Reposa
  • Yuan Shi
  • T. K. Satish Kumar
  • Craig A. Knoblock
  • Irfan Muhammad

Colleges, School and Institutes

External organisations

  • Charles River Analytics
  • University of Southern California Information Sciences Institute

Abstract

Most current software systems cannot adapt to changing hardware requirements or changing operational environments. This lack of adaptivity shortens the life of the software and makes it less capable of achieving its mission. We are developing Probabilistic Representation of Intent Commitments to Ensure Software Survival (PRINCESS) to enable software to adapt to both hardware changes and changing environments. We are developing methods to automatically optimize software for new environments, turning non-adaptive code into optimizable adaptive code. We are also developing sensor adapters that enable us to maintain software function despite changing or failing sensors. We are implementing PRINCESS on the navigation system of an unmanned undersea vehicle. An independent evaluation has demonstrated PRINCESS's ability to adapt to degraded sensors, changing environmental conditions, and loss of power.

Details

Original languageEnglish
Pages (from-to)91-96
Number of pages6
JournalIEEE Software
Volume36
Issue number2
Early online date21 Feb 2019
Publication statusE-pub ahead of print - 21 Feb 2019

Keywords

  • remotely operated vehicles, underwater vehicles, current software systems, software adaptation, environmental conditions, unmanned undersea vehicle, Software development management, Probabilistic logic, Navigation, Sensor systems, Optimization