Equalização adaptativa utilizando seleção de dados em transceptores em bloco com redundância reduzida

Abstract

systems, mostly due to their welldefined structure and blockwise encoding. Among the main challenges encountered by mobile applications, there is an inherent interblock interference, due to superpositions of delayed signal copies, which is commonly eliminated with the addition of redundancy between adjacent data blocks. In addition to that, channel equalization is also usually employed, in order to further mitigate channel interferences. However, the amount of redundancy may be overestimated, which opens an opportunity for reduced-redundancy superfast transceivers, whose features include high spectral efficiency and low computational cost. Although the superfast approach aims at achieving low complexity, equalizer-coefficient updates are still very complex tasks due to channel variations, and most designs do not employ methodologies for computational-effort reduction. The present work addresses this problem and proposes a new design strategy for block-based transceivers, which provides semiblind equalization with data-selective update, besides the possibility of a generalized approach, based on the fast Fourier transform and diagonal matrices. Simulation results show that our approach updates less than 60% of the equalizer coefficients duringsupervised and blind period and maintain a competitive throughput for single-carrier and multicarrier transmissions.