Deterministic Pilot Design for Sparse Channel Estimation in MISO/Multi-User OFDM Systems

We study the pilot design problem for sparse channel estimation in OFDM systems where multiple channels are estimated at a single antenna receiver. Such design is applicable to downlink of massive-MIMO systems and also to scenarios where multiple users transmit to a base station at the same carrier frequency. In our design, we deviate from the conventional orthogonal pilot arrangements by assigning the same pilot subcarriers to all transmitters. In the proposed setting, the achieved improvement in spectral efficiency (by reducing pilot overhead) may come at the expense of a more challenging channel estimation block at the receiver. To address this challenge and distinguish between different signals that are arriving at the receiver at the same subcarrier, we propose to select pilot subcarriers through minimizing the coherence of the associated Fourier submatrix, as well as properly assigning different pilot values (complex numbers) to each individual transmitter. We demonstrate that if the channels are sparse enough in time domain, there are simple sparse recovery techniques to simultaneously estimate all the channels, although all transmitters share the same pilot subcarriers. Simulation results demonstrate that the proposed design outperforms existing methods in terms of both mean-square channel estimation error and bit error rate.