Adaptive pilot pattern for OFDM systems

In OFDM communication systems over fading channels, link adaptation based on the available channel state information at the transmitter (i.e., power allocation, adaptive modulation and coding rate) increases the spectral efficiency and reliability of the link. In this work, we explore the possibility to extend the set of transmission parameters to be adaptively selected to the pilot arrangement (i.e., to the collocation of pilot subcarriers on the time-frequency grid). Based on the prediction of the channel estimation error at the receiver, the transmitter can minimize the number of pilot subcarriers that guarantees a sufficiently reliable channel estimate according to Quality of Service requirements. Prediction of the channel estimation error is performed by computing the error covariance matrix of Kalman filters. The adaptive pilot arrangement problem is formulated and a "greedy" solution is proposed. Simulations show the effectiveness of the algorithm with respect to periodic re-training.