Abstract
This paper presents a method for the tolerance of software design faults in a Flight Control System, based on a distributed periodic system in which the processing is performed in nodes. Tasks should be allocated and executed in parallel on different hardware channels with the same copy of programs (no design diversity), although under slightly different conditions (data diversity). A simulation-based fault injection experiment demonstrated that the proposed approach can considerably improve the fault tolerance capabilities of a system as compared with the traditional design. It was observed, however, that injected faults left behind an error and, hence, mechanisms for error detection and error recovery must be included in the system.
Original language | English (US) |
---|---|
Title of host publication | Doktorsavhandlingar vid Chalmers Tekniska Hogskola |
Publisher | Chalmers Tekniska Hogskola |
Edition | 1362 |
State | Published - 1998 |
Externally published | Yes |
ASJC Scopus subject areas
- Engineering(all)