Distributed Adaptive Trace Ratio Optimization in Wireless Sensor Networks

The trace ratio optimization (TRO) problem consists of finding an orthonormal basis for the discriminative subspace that maximizes two operators on covariance matrices corresponding to distinctive classes or signal components. TRO is encountered in various processing problems such as dimensionality reduction, enhancement, and analysis. In this paper, we propose a distributed adaptive algorithm solving context wireless sensor networks (WSNs), where involved correspond (unknown) spatial correlation signals across nodes network. We first focus fully-connected every node can communicate with each other, but only compressed observations be shared reduce communication cost. After showing convergence, modify operate WSNs more general topologies. Simulation results are provided validate complement theoretical results.