Skip to Main Content
These notes develop an approach for characterizing uncertainty in sequences of observable variables. Unlike traditional analyses, observations are not treated as samples from associated random variables. Instead, they are simply assumed to be quantities that satisfy sets of loose constraints. Their exact values remain uncertain and are immaterial to the derived results. When required, additional structural assumptions are introduced by associating these sequences with the state-to-state transitions of conventional finite state machines.
Date of Conference: 17-20 Sept. 2012