An investigation into the stability of redundant controller voting schemes

Abstract

Redundant controllers are being designed to provide improved system reliability to meet the needs of various applications. In some implementations of these systems unacceptable system transients have occurred when a single-controller failure occurs. The problem has been treated using case-by-case correction which is not sufficiently general in application to meet the needs of systems which may be user-programmed. Redundant control systems generalized as ideal and nonideal using two specific voting schemes are examined to gain insight into theoretical characterization of their performance under single-point failure modes and the limitations of applicability for the characterizations. State-model computer simulation is used to demonstrate the characterizations. A suggested direction for a general solution to the problem based on the characterizations is described.