| Summary: | 碩士 === 國立中央大學 === 通訊工程研究所 === 100 === Channel estimation is an important issue for successfully implementing an orthogonal frequency division multiplexing (OFDM) system. Pilot-aided channel estimation is the most common approach, and the least square (LS), maximum likelihood (ML), or minimum mean square error (MMSE) criteria are often applied to estimate the multipath channel gains. It has been evident from the literature that the uniformly distributed pilot tones which satisfy the Nyquist criterion is the best way to attain the minimum square error for channel estimation.
For the next-generation wireless communication, the wireless channels could behave like a sparse multipath channel in which merely contains a few significant paths with a long delay spread owing to high data rate transmission and large cell coverage. Therefore, it is required to embed a great amount of pilot tones over subcarriers to get an accurate channel estimate at the sacrifice of spectral efficiency. Otherwise, a significant performance degradation can be observed for the conventional channel estimation methods in sparse multipath channels. In this thesis, we apply the compressive sensing (CS) technique to estimate and track the sparse multipath channels in wireless mobile environments. We first formulate a Bayesian ℓ1-norm optimization problem by maximizing the posteriori probability of the channel parameters. The associated ℓ1-norm optimization problem, which involves the features of the channel sparsity and channel variation, is solved by the convex optimization technique. Besides, an efficient iterative subgradient algorithm is derived to attain the optimal solution, and an adaptive step size is proposed to effectively reduce the iteration number. Compared with the conventional discrete Fourier transform (DFT)-based channel estimation, simulation results show that the proposed channel estimation scheme can greatly improve the channel estimation accuracy in sparse multipath channels.
|
|---|
