Dynamic wavelet-based pilot allocation algorithm for OFDM-based cognitive radio systems

In a cognitive radio system, the goal is to make better use of the radio electric spectrum, allowing non-licensed users access to those currently unused electromagnetic bands assigned to licensed users (LUs). This can be achieved using OFDM, where the non-licensed users must select the temporarily available subcarriers and turn off those subcarriers used by LUs in order to avoid interference. Hence, only a subset of the subcarriers can be used for data or pilot tone transmission. To this end, some pilot allocation algorithms have been proposed for this dynamic scenario, but they are designed in such away that an equispaced pilot placement is respected (as much as possible) while minimizing the mean squared error of the channel estimate. Nevertheless, this equispaced placement can lead to the use of an increased number of pilots in order to achieve a good channel estimation. In this work, a new pilot allocation algorithm based on wavelet transform is presented. The proposed algorithm uses the discrete wavelet transform to analyze the previous channel state information, taking the knowledge of the available subcarriers into account to provide a suboptimal solution for the pilot positions. This solution leads to a non-equispaced pilot placement, which improves the channel estimation and consequently, the system performance. Likewise, the introduced algorithm allows a reduction of the number of necessary pilots, which aids in increasing the data rate. Finally, simulation results corroborate the effectiveness of the algorithm in dynamic channel scenarios.