Orthogonal frequency division multiplexing (OFDM) is a multicarrier transmission technique that promises high spectral efficiency and system flexibility. This work investigates the link level performance of wireless OFDM systems with antenna arrays under real-world conditions like e.g. imperfect channel estimation, nonlinear transmitter power amplifiers (PAs) and fading channels.
First, robust channel estimation schemes for linear single-input multiple-output (SIMO) and multiple-input multiple-output (MIMO) system models are investigated. For both indoor environments with almost time-invariant (TI) channels and outdoor environments with high mobility and time-variant (TV) channels, the receiver design is based on a spatial diversity model to increase the robustness of the resulting demodulation for unknown channel model order w.r.t. direction-based channel models. In the TI case, by exploiting the fact that in most cases the number of subcarriers exceeds the number of relevant propagation paths in the diversity model, we design different subspace based channel estimation schemes with optimum training sequences for fully loaded/non-fully loaded SIMO and MIMO OFDM systems based on a least squares approach. For TV channels, we propose a Kalman filter based channel tracking with decision feedback in the time domain (TD) and in the frequency domain (FD). Both the TD and FD schemes show identical tracking performance, while the TD scheme requires a lower computational complexity than the FD scheme.
In the second part of the work, we adopt the Bussgang theorem to describe the PA nonlinearity by a linear model assuming Gaussianity of the underlying multicarrier signal. For perfect channel state information, a signal-to-noise measure being relevant for the performance of SIMO and MISO systems with different signaling formats is derived. Furthermore, the average bit-error rate is calculated as a function of the input back-off and signal-to-noise ratio of systems with Rayleigh fading. Simulation results are used to confirm the theoretical findings and to quantify the impact of the nonlinearity on the channel estimation and bit-error rate performance in an IEEE 802.11a system.

Keywords: OFDM, antenna arrays, channel estimation, power amplifier nonlinearity, spatial diversity, bit-error rates, Rayleigh fading.