Fault tolerance in multisensor networks

Abstract

Replicating sensors is desirable, not only to tolerate sensor failures, but to increase the average accuracy of the ensemble of replicated sensors beyond that obtainable with a single sensor. Such replication is used in a multi-sensor environment or in a distributed-sensor network. Following Marzullo (1990), the authors have modeled a continuous valued sensor as an interval of real numbers containing the physical value of interest. Given n sensors of which at most f can suffer arbitrary failures, this paper presents an efficient O(n/spl middot/log(n)) fault-tolerant algorithm (J/FTA) whose output is reliable (guaranteed to contain the correct value at all times) and is fairly accurate when f<gilb(/sup 1///sub 2/(n+1)). The output of J/FTA can be either a single-interval or a set-of-intervals, depending on the nature of the multi-sensor environment. J/FTA can be used not only to detect all possibly-faulty sensors but to detect all sets (combinations) of possibly-faulty sensors. This paper proves the following results pertaining to the possibly-faulty sensors identified by J/FTA: the number of sets each containing f possibly-faulty sensors is at most (f+1); the number of sets each containing f or fewer faulty sensors is at most (2f+1); and the number of possibly-faulty sensors identified by J/FTA is at most 2f. These results help to: narrow the search to detect faulty sensors; bound the number of intervals needed to construct an accurate and reliable abstract sensor; and identify at least one correct sensor.