An L1-minimization based algorithm to measure the redundancy of state estimators in large sensor systems

Linear models have been successfully used to establish the connections between sensor measurements and system states in sensor networks. Finding the degree of redundancy for structured linear systems is proven to be NP-hard. Previously bound-and-decompose, 0–1 mixed integer programming and hybrid algorithms embedding 0–1 mixed integer feasibility checking within a bound-and-decompose framework have all been proposed and compared in the literature. In this paper, we exploit the computational efficiency of linear programs to present a novel heuristic algorithm which solves a series of l1-norm minimization problems in a specific framework to find extremely good solutions to this problem in remarkably small runtime.