Sensoriamento espectral e detecção de sinais em sistemas MIMO com ruído impulsivo e desvanecimento por múltiplos percursos

Abstract

In mobile systems, information signals must be transmitted with high data throughput and reliability. One of the possible solutions to achieve these desirable requirements is the use of systems with multiple transmitting and/or receiving antennas, featuring a MIMO system (Multiple Input-Multiple Output). However, the most signal processing techniques used in MIMO systems suppose the Gaussianity of the communication channels, decreasing their efficiency in non-Gaussian communication scenarios. In this work, a new signal detection technique in MIMO systems is proposed in scenarios characterized by non-Gaussian noise. The proposed technique is called Maximal Correntropy Detector (MCD), and it can be seen as a generalization of the MLD detector (Maximum Likelihood Detector). The MCD detector is analyzed on Gaussian and non-Gaussian channels, being observed that its performance is superior to classic detectors, without a significant increase in computational complexity. Together, this work also presents a new method of spectrum sensing suitable for detecting signals that show non-linear phase variations over time. The proposed method is based on the use of the theory of angle-time cyclostationary signals, in which transformations are employed in order to mitigate the effect of the nonlinear phase variations. A sensing architecture is evaluated considering a communication scenario with BPSK (Binary Phase Shift Keying) signals. Simulation results indicate a performance gain, in terms of detection rates of primary users, approximately equal to 8 dB, in comparison to sensing by temporal-cyclostationary characteristics.