Fault-Tolerant Design of Barrier Coverage for Periodically-Repairable Wireless Sensor Networks

Abstract

Wireless sensor networks are sometimes deployed in circumstances where failed sensors may not be able to be repaired or replaced immediately after failure, e.g. a remote, harsh environment. By deploying redundant sensors, such networks will be more likely to tolerate individual sensor failures without experiencing an intolerable degradation in performance. This paper analyzes the redundancy required in a barrier-covering wireless sensor network where repairs to broken sensors can occur at periodic intervals, called “epochs.” Exact, tractable methods are derived for computing the total gap in coverage, the number of gaps in coverage, and the largest gap in coverage in an epoch. Together, the distributions of these metrics allow network designers to select how much redundancy is required to maintain their desired performance with a specified probability. Numerical simulations then demonstrate the accuracy of the methods, and show how network designers can explore the tradeoff between deploying larger numbers of cheap, poor-quality sensors and deploying smaller numbers of expensive, robust sensors.