Using Learning to Support the Development of Embedded Systems

Project: Research project (funded)Research

Project Details

Layman's description

For many computer-based systems it is important to understand the worst-case behaviour of systems, e.g. how long it takes for a car's braking system takes to activate. This project has developed techniques for this.

Key findings

The project established important machine learning and search-based techniques for Worst-Case Execution Time (WCET). The principal results were machine learning techniques for exactly determining the loop bounds of software and understanding cache behaviours, and how to achieve better confidence when searching for the WCET.
Effective start/end date1/10/0730/09/11


  • EPSRC: £379,941.00