Noncooperative Optimization of Space-Time Signals in Ad hoc Networks

Recent results on noncooperative space-time waveform optimization for Ad hoc networks are first reviewed. The optimization criterion is sum power minimization under a QoS constraint (decoupled capacity.) Noncooperative iterative minimum mean-square (IMMSE) algorithms are advocated here as a pragmatic solution to the QoS constrained optimization problem. Connections between IMMSE, time-reversal (TR), taxation concepts in game theory, and Lagrangian optimization are discussed. The ST waveform noncooperative algorithm is then extended to ST eigencoding. The idea of interference taxation leads to a water-filling algorithm based on generalized eigenvalues. A matrix IMMSE-TR update is proposed for ST eigencoding, for which a subset of fixed points correspond to a noncooperative optimum.