Affordable Nonlinear MIMO Systems


Georgios Psaltopoulos


PhD Thesis, Logos Verlag Berlin, 2011

DOI: 10.3929/ethz-a-006383560

[BibTeX, LaTeX, and HTML Reference]


In this thesis we introduce a new paradigm for low-cost/low power wireless systems: nonlinear MIMO systems. This new type of MIMO systems offers a way to reduce both cost and power consumption of legacy MIMO systems by using nonlinear detection methods, i.e., amplitude and phase detection. Nonlinear MIMO systems have not been considered so far. The nonlinear nature of such systems introduces certain idiosyncrasies: mathematical tractability in nonlinear MIMO systems is not straight-forward anymore and most known results/observations for linear MIMO systems are no more valid. In this inaugural work on nonlinear MIMO systems we explore the fundamental properties and identify the potential of such systems. We develop the theory of nonlinear MIMO systems starting from information theoretic aspects and working towards more realistic systems. First, we show that nonlinear MIMO systems can also exploit the spatial degrees of freedom like MIMO systems, although in a constrained way. This implies that combining multiple antennas with a nonlinear receiver increases the performance with minimal additional cost and power consumption. We then explore ML detection in nonlinear MIMO systems, in conjunction with the achievable receive diversity and spatial multiplexing modulation alphabets. It turns out that nonlinear MIMO systems behave quite different compared to linear MIMO systems. We further develop channel estimation schemes that enable coherent detection in nonlinear MIMO systems, thus completing the basic building blocks required for an operational system. Finally, we demonstrate the validity and potential of nonlinear MIMO systems through a MIMO envelope detection experimental test bed.

This publication is currently not available for download.