Noncoherent OFDM Transmission via Off-the-Grid Joint Channel and Data Estimation

Pilot-aided channel estimation techniques are known to waste the spectral bandwidth. An off-the-grid blind estimator for time-variant orthogonal frequency division multiplexing (OFDM) systems is studied in this letter. In this regard, we propose a blind estimator based on atomic norm minimization (ANM) for OFDM systems. To do so, at the first transmission block, using a lifted ANM (LANM) and simple constraint on l(2) norm of data, we simultaneously estimate the channel and data. For the subsequent blocks, we use a penalized ANM (PANM) to simultaneously track the channel's parameters and detect transmit signals. The proposed problems require an infinite dimensional search, hence are NP-hard. Therefore, we propose two semidefinite programs (SDPs) to implement them. We then derive the total computational complexity of the proposed estimator. The simulation results show the superiority of the proposed estimator to the state-of-the-arts.