Nonlinear MIMO Systems

A way to reduce both complexity and power consumption of MIMO systems is using nonlinear detection methods, i.e. amplitude and phase detection. These are known as low-cost approaches to receiver design, since they can be implemented with much simpler and more energy-efficient circuitry. In fact, phase detectors are not prone to system nonlinearities, since amplitude carries no information, while amplitude detection can be simply implemented using an envelope detector, thus avoiding the power consuming down-conversion. Conventional MIMO systems studied so far strictly employ linear receivers, meaning that an I/Q demodulator yields a complex baseband received signal. Linear MIMO systems have been extensively studied and are well understood. Their linearity property is attractive since it enables mathematical tracking. Our work considers the combination of nonlinear receivers and MIMO systems. The nonlinear nature of such systems introduces certain idiosyncrasies, apart from the aforementioned advantages. Mathematical tractability is not straight-forward anymore, while most known results/observations for linear MIMO systems are no more valid. Hence, nonlinear MIMO systems offer a new unexplored field for research, which at the same time is very relevant to future needs. This technology is especially lucrative for systems that require high numbers of very low-cost and low-power devices. Sensor networks are a typical example of application. The very low cost and long lifetime of sensor nodes is very critical in such systems. Another, example is RFID tagging systems. As technology advances, it is possible to build more complicated tags, while at the same time the demand for higher data rates increases. Furthermore, as carrier frequencies continue to increase up to and beyond 60 GHz, practical implementation of receivers becomes very cumbersome, while nonlinear reception offers a means to circumvent such problems. In general, nonlinear MIMO systems are relevant in a future environment where wireless access is enabled by an ubiquitous network of sensors and nodes.

Our work on nonlinear MIMO systems has led to the implementation of the sMILE test bed. This is the first MIMO test bed that employs envelope detectors at the receiver.

Publications

• Nonlinear MIMO: Affordable MIMO Technology for Wireless Sensor Networks
  G. K. Psaltopoulos and A. Wittneben, IEEE Transactions on Wireless Communications, Vol. 9, Iss. 2, pp. 824-832, Feb. 2010.
• Channel Estimation for Very Low Power MIMO Envelope Detectors
  G. K. Psaltopoulos and A. Wittneben, IEEE International Conference on Communications, ICC 2010, Cape Town, May 2010.
• Diversity and Spatial Multiplexing of MIMO Amplitude Detection Receivers
  G. K. Psaltopoulos and A. Wittneben, IEEE International Symposium on Personal, Indoor and Mobile Radio Communications, PIMRC 2009, Tokyo, September 2009
• Achievable Rates of Nonlinear MIMO Systems with Noisy Channel State Information
  G. K. Psaltopoulos and A. Wittneben, IEEE International Symposium on Information Theory, ISIT 2008, Toronto, July 2008
• On Achievable Rates of MIMO Systems with Nonlinear Receivers
  G. K. Psaltopoulos, F. Trösch and A. Wittneben, IEEE International Symposium on Information Theory, ISIT 2007, Nice, June 2007