IEEE Transactions on Aerospace and Electronic Systems · 2018 · 12 citations · 26 references
ComputingComputer ScienceElectrical EngineeringDistributed SystemsOptimizationSensor FusionDecision-level FusionDigital Signal ProcessingDistributed SensingSignal Detection TheoryBayesian ProbabilityDistributed DetectionFusion RulesGenetic Algorithm OptimizationIndependent ReportingInteger Nonlinear ProgrammingSensor Data FusionSoft-decision Decoding
Prior works on distributed detection with correlated Gaussian observations primarily dealt with single-bit reports to the fusion center and suboptimal fusion rules. Considering multiple bits and independent reporting channel errors, we formulate the optimal detection problem as a nonlinear integer programming problem for which the genetic algorithm is applied to find a suboptimal solution. The Bayesian probability of error performance of our suboptimal solution, first, for one-bit, is better than the performance of some previously proposed schemes and, second, for multiple bits, approaches that of the centralized detection scheme.
26