Abstract
Achievable rates of wireless communication systems with pilot-based channel estimation are investigated for the case of time-selective fading. Novel analytical expressions for the maximum achievable rates of such systems are derived in terms of the system signal-to-noise ratio (SNR), fading rate and estimation scheme deployed. The frame size is optimized jointly based on the SNR and the fading rate. The maximum rate achieving coding scheme is suggested and shown to be a modified version of the classical water-filling algorithm that accounts for imperfect channel state information (CSI) at the transmitter. The impact of the estimation scheme and the angular spread of the received signal on the quality of estimation and achievable rates is evaluated. A number of numerical simulations are provided to illustrate the dependence of the optimal block length and achievable rates on SNR, fading rate, estimation scheme and angular spread of the channel.